(Quantitative Perspectives On Behavioral Economics and Finance) James Ming Chen (Auth.) - Finance and The Behavioral Prospect - Risk, Exuberance, and Abnormal Markets-Palgrave Macmillan (2016)

FINANCE AND
THE BEHAVIORAL
PROSPECT
Risk, Exuberance, and Abnormal Markets
JAMES MING CHEN

Quantitative Perspectives on Behavioral
Economics and Finance
Series Editor
James MingChen
College of Law
Michigan State University
East Lansing,Michigan, USA
Aims of the Series
The economic enterprise has firmly established itself as one of evaluating
human responses to scarcity not as a rigidly rational game of optimiza-
tion, but as a holistic behavioral phenomenon. The full spectrum of social
sciences that inform economics, ranging from game theory to evolution-
ary psychology, has revealed the extent to which economic decisions and
their consequences hinge on psychological, social, cognitive, and emo-
tional factors beyond the reach of classical and neoclassical approaches
to economics. Bounded rational decisions generate prices, returns, and
resource allocation decisions that no purely rational approach to optimiza-
tion would predict, let alone prescribe.
Behavioral considerations hold the key to longstanding problems in
economics and finance. Market imperfections such as bubbles and crashes,
herd behavior, and the equity premium puzzle represent merely a few
of the phenomena whose principal causes arise from the comprehensi-
ble mysteries of human perception and behavior. Within the heterodox,
broad-ranging fields of behavioral economics, a distinct branch of behav-
ioral finance has arisen.
Finance has established itself as a distinct branch of economics by apply-
ing the full arsenal of mathematical learning on questions of risk manage-
ment. Mathematical finance has become so specialized that its practitioners
often divide themselves into distinct subfields. Whereas the P branch of
mathematical finance seeks to model the future by managing portfolios
through multivariate statistics, the Q world attempts to extrapolate the
present and guide risk-neutral management through the use of partial dif-
ferential equations to compute the proper price of derivatives.
The emerging field of behavioral finance, worthy of designation by the
Greek letter psi (), has identified deep psychological limitations on the
claims of the more traditional P and Q branches of mathematical finance.
From Markowitzs original exercises in mean-variance optimization to the
Black-Scholes pricing model, the foundations of mathematical finance
rest on a seductively beautiful Gaussian edifice of symmetrical models and
crisp quantitative modeling. When these models fail, the results are often
catastrophic.
The branch of behavioral finance, along with other postmodern
critiques of traditional financial wisdom, can guide theorists and practi-
tioners alike toward a more complete understanding of the behavior of
capital markets. It will no longer suffice to extrapolate prices and forecast
market trends without validating these techniques according to the full
range of economic theories and empirical data. Superior modeling and
data-gathering have made it not only possible, but also imperative to har-
monize mathematical finance with other branches of economics.
Likewise, if behavioral finance wishes to fulfill its promise of transcend-
ing mere critique and providing a more comprehensive account of finan-
cial markets, behavioralists must engage the full mathematical apparatus
known in all other branches of finance. In a world that simultaneously lauds
Eugene Famas efficiency hypotheses and heeds Robert Shillers warnings
against irrational exuberance, progress lies in Lars Peter Hansens com-
mitment to quantitative rigor. Theory and empiricism, one and indivisible,
now and forever.
More information about this series at

http://www.springer.com/series/14524
James MingChen
Finance and the

Behavioral Prospect
Risk, Exuberance, and Abnormal Markets
James MingChen
College of Law
Michigan State University
East Lansing, Michigan, USA
Quantitative Perspectives on Behavioral Economics and Finance

ISBN 978-3-319-32710-5 ISBN 978-3-319-32711-2 (eBook)
DOI 10.1007/978-3-319-32711-2
Library of Congress Control Number: 2016950218
The Editor(s) (if applicable) and The Author(s) 2016

The author(s) has/have asserted their right(s) to be identified as the author(s) of this work
in accordance with the Copyright, Designs and Patents Act 1988.
This work is subject to copyright. All rights are solely and exclusively licensed by the
Publisher, whether the whole or part of the material is concerned, specifically the rights of
translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on
microfilms or in any other physical way, and transmission or information storage and retrieval,
electronic adaptation, computer software, or by similar or dissimilar methodology now
known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are
exempt from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information
in this book are believed to be true and accurate at the date of publication. Neither the
publisher nor the authors or the editors give a warranty, express or implied, with respect to
the material contained herein or for any errors or omissions that may have been made.
Cover illustration TongRo Images / Alamy Stock Photo
Printed on acid-free paper
This Palgrave Macmillan imprint is published by Springer Nature

The registered company is Macmillan Publishers Ltd. London
To Heather Elaine Worland Chen, with all my love
ACKNOWLEDGMENTS
This book incorporates ideas from papers I have presented at the University
of Cincinnati, Florida State University, Georgetown University, Michigan
State University, the University of Pennsylvania, the University of Virginia,
and the Faculty of Economics of the University of Zagreb (Ekonomski
Fakultet, Sveuilite u Zagrebu). The International Atlantic Economic
Society and the ACRN Oxford Academic Research Network have pro-
vided multiple platforms for the work underlying this book. Along the
way, I have benefited from scholarly and professional interactions with
Anna Agrapetidou, Abdel Razzaq Al Rababaa, Moisa Altar, Christopher
J.Brummer, Irene Maria Buso, Adam Candeub, Seth J.Chandler, Felix
B. Chang, Tendai Charasika, Csar Crousillat, David Dixon, Robert
Dubois, John F. Duffy, Daniel A. Farber, Christopher C. French,
Santanu K. Ganguli, Tomislav Gelo, Periklis Gogas, Gil Grantmore,
Andy Greenberg, Losbichler Heimo, Hemantha Herath, Jesper Lyng
Jensen, Jagoda Kaszowska, Daniel Martin Katz, Yuri Katz, Imre Kondor,
Carolina Laureti, Cordell Lawrence Jr., Cordell Lawrence Sr., Matthew
Lee, Othmar Lehner, Heimo Losbichler, Gerry Mahar, Milivoj Markovi,
L. Thorne McCarty, Steven C. Michael, Ludmila Mitkova, Jos Mara
Montero Lorenzo, Kevin Lynch, Laura Muro, Vivian Okere, Merav
Ozair, Elizabeth Porter, Mobeen Ur Rehman, Carol Royal, Bob Schmidt,
Jeffrey A.Sexton, Galen Sher, Ted Sichelman, Jurica imurina, Nika Sokol
imurina, Robert Sonora, Lisa Grow Sun, Elvira Takli, Peter Urbani,
Robert R.M.Verchick, Benjamin Walther, Karen Wendt, Gal Zahavi, and
Johanna F. Ziegel. Christian Diego Alcocer Argello of Michigan State
Universitys Department of Economics provided very capable research
ix
x ACKNOWLEDGMENTS
assistance. I am also grateful for contributions by several students at

Michigan States College of Law: Angela Caulley, Yuan Jiang, Morgan
Pitz, Emily Strickler, Paul M. Vogel, and Michael Joseph Yassay. The
research services of the Michigan State University Law Library and admin-
istrative support by Marie Gordon were indispensable. Special thanks to
Heather Elaine Worland Chen.
CONTENTS
1 The Structure ofaBehavioral Revolution 1
2 Mental Accounting, Emotional

Hierarchies,andBehavioralHeuristics 29
3 Higher-Moment Capital Asset Pricing andIts

BehavioralImplications 57
4 Tracking theLow-Volatility Anomaly Across

Behavioral Space 73
5 The Intertemporal Capital Asset Pricing Model:

HedgingInvestment Risk Across Time 93
6 Risk Aversion 111
7 The Equity Risk Premium andtheEquity

Premium Puzzle 137
8 Prospect Theory 181
xi
xii CONTENTS
9 Specific Applications ofProspect Theory to

Behavioral Finance 213
10 Beyond Hope andFear: Behavioral Portfolio Theory 247
11 Behavioral Gaps Between Hypothetical Investment

ReturnsandActual Investor Returns 283
12 Irrational Exuberance: Momentum Crashes and

SpeculativeBubbles 301
13 The Monster andtheSleeping Queen 323
Index 327
CHAPTER 1
The Structure ofaBehavioral Revolution
1.1 ABNORMAL MARKETS, IRRATIONAL INVESTORS

This book represents one of the first two volumes in the series, Quantitative
Perspectives on Behavioral Economics and Finance. Its companion vol-
ume, Postmodern Portfolio Theory: Navigating Abnormal Markets and
Investor Behavior, addresses leading departures from the putative efficiency
of financial markets.1 Intense pressure on the conventional capital asset
pricing model gave rise to theoretical innovations such as Eugene Fama
and Kenneth Frenchs three-factor model. Postmodern Portfolio Theory
traces this story through the four statistical moments of the distribution of
financial returns: mean, variance, skewness, and kurtosis.
This book conducts a fuller exploration of behavioral phenomena in
finance, such as the low-volatility anomaly, the equity premium puzzle, and
momentum in stock returns. Mental accounting, persistent gaps between
hypothetical investment return and actual investor return, and alternatives
to modern portfolio theory and the conventional capital asset model con-
tribute to the development of behavioral approaches to portfolio design
and risk management. Gaps in perception and behavioral departures from
rational decision-making appear to spur momentum, even irrational exu-
berance and speculative bubbles. Ultimately, this book hopes to explain
emotion-laden deviations from the strict rationality traditionally associated
with mathematical finance. Together with Postmodern Portfolio Theory, this
The Editor(s) (if applicable) and The Author(s) 2016 1

J.M. Chen, Finance and the Behavioral Prospect, Quantitative
Perspectives on Behavioral Economics and Finance,
DOI10.1007/978-3-319-32711-2_1
2 J.M. CHEN
book synthesizes observations on abnormal markets and irrational inves-

tors into a coherent behavioral account of financial risk management.
Chapter 1 traces the rise of the behavioral revolution in portfolio theory
and, more generally, in mathematical finance. Like any other story in the
history and philosophy of science, the transformation of portfolio theory
begins with the identification of anomalies. Only after identifying anoma-
lies and challenging an established paradigm can dissenters lay a credible
claim to a competing intellectual movement.
Behavioral accounting, arising from the irresistible human urge to keep
emotional score, undermines modern portfolio theorys rational prem-
ises. The separation theorem and the two mutual fund theorem counsel
investors to consolidate all assets into a single portfolio along the effi-
cient frontier. But individual and even institutional investors consistently
reject that sort of normative guidance. Assets and portfolios are imbued
with affect, and positive and negative emotions warp investment deci-
sions. Chapter 2 explores two seemingly divergent but ultimately similar
manifestations of emotion in economics: Maslowian portfolio theory and
behavioral environmental economics.
Chapter 3 introduces some of the most important mathematical tools in
behavioral finance. After summarizing the conventional capital asset pricing
model, Chapter 3 presents a higher-moment approach to capital asset pric-
ing as an outgrowth of the Taylor series expansion of logarithmic returns.
Finance proceeds from the assumption that risk and return are posi-
tively correlated. If investors are generally risk averse, they will presumably
demand a higher return in exchange for buying assets whose prices exhibit
higher variance. Departures from this relationship between risk and return
undermine this theoretical foundation of finance. In reality, some of the
highest returns are available on the stocks exhibiting the lowest levels of
volatility. Along with its analogue in accounting, Bowmans paradox, the
low-volatility anomaly poses a serious challenge to the conventional finan-
cial narrative. Chapter 4 tracks the low-volatility anomaly in behavioral
space by examining beta on either side of mean returns and analyzing the
separate volatility and correlation components of beta. Chapter 5 intro-
duces the intertemporal capital asset pricing model and the prospect of
explaining the low-volatility anomaly according to time as well as space.
Chapter 6 outlines a quantitative approach to risk aversion. It speci-
fies the ArrowPratt measures of absolute and relative aversion, as well
as the famously tractable model of hyperbolic absolute risk aversion, as
a prelude to examining a more behaviorally sensitive account of human
responses to risk. Two paradoxes, Allaiss paradox and the St. Petersburg
THE STRUCTURE OFABEHAVIORAL REVOLUTION 3
paradox, suggest that conventional accounts of risk aversion do not

provide a comprehensive explanation of economic behavior in the face of
risk or uncertainty.
Risk aversion provides at least a partial explanation for the historic pre-
mium that equities have commanded over lower-risk investments such as
bonds. Though accounts vary, the equity risk premium rests in the neigh-
borhood of 36% per year. The magnitude of this premium, however,
poses a formidable (and arguably, still unresolved) theoretical challenge
to conventional asset pricing models. Building on Chapter 6s measures
of risk aversion, Chapter 7 explores both the equity risk premium and the
econometric puzzle to which that premium has given rise. The equity
premium puzzle is to behavioral finance as the low-volatility anomaly is to
modern portfolio theory and the conventional capital asset pricing model:
Without resolving contradictions of this magnitude, many of the theoreti-
cal suppositions of mathematical and behavioral finance will be squarely
contradicted by the behavior of real markets and real investors.
The final chapters of this book present two leading accounts of behav-
ioral finance, prospect theory and SP/A (security-potential/aspiration)
theory. Those chapters also offer thoughts on speculative bubbles in
finance. Chapter 8 introduces prospect theory, arguably the most promi-
nent manifestation of behavioral economics in finance. The theorys four-
fold pattern provides the most widely accepted account of risk-averse as
well as risk-seeking behavior. Prospect theory explains the financial impact
of fear and greed. Humans depart from purely rational utility in three
ways. First, humans heed reference points. Second, humans hate losing
more than they like winning. Third, humans grow less sensitive to eco-
nomic changes as gains or losses increase. A fourfold pattern provides a
comprehensive account of risk-averse as well as risk-seeking behavior.
Chapter 9 applies prospect theory to a set of related problems and puz-
zles in finance. Affirmative risk-seeking, something not readily accommo-
dated by expected utility theory, is predicted by prospect theory. Skewness
preference manifests itself across a large number of financial settings in
the form of investor demand for instruments that couple low expected
returns with high potential jackpots. The two-way mispricing of initial
public offeringsunderpriced in the short run to issuers detriment, over-
priced in the long run at the expense of investorsprovides an especially
vivid illustration of this preference for lottery-like instruments. In addi-
tion, prospect theory supports a distinct body of proposed solutions to
Bowmans paradox and the equity premium puzzle.
4 J.M. CHEN
Chapter 10 presents a competing account of behavioral finance. SP/A

theory describes the competing forces of security, potential, and aspiration
within financial decision-making. SP/A theory transforms the dynamics of
hope and fear into a behavioral account of portfolio theory. Intriguingly,
behavioral portfolio theory extends the safety-first principle that inspired
the very first departures from the perfectly symmetrical and rational sup-
positions of modern portfolio theory. Behavioral portfolio also bears a
deep resemblance to value-at-risk analysis, albeit as a method for evaluat-
ing extreme positive outcomes.
Over time, investor behavior consistent with the predictions of pros-
pect theory and SP/A theory has had a profound impact on financial mar-
kets. Cycles of fear and greed have systematically eroded investor returns.
Behavioral finance can measure those gaps in investor performance, rela-
tive to simple buy-and-hold strategies, through a new statistic, . Chapter
11 presents a method for calculating as the gap between hypothetical
investment returns and actual investor returns. A market where ordinary
investorsand many active fund managerssystematically underperform
buy-and-hold strategies is a market that exhibits both heterogeneity and
significant limits to arbitrage. Chapter 12 accordingly explores the behav-
ioral origins of momentum as well as the rise of speculative bubbles.
1.2 ANOMALIES, FAST ANDSLOW

How mathematical finance came to abandon its original, strictly ratio-
nal and utilitarian suppositions and to adopt a sophisticated awareness of
investor behavior is by no means a unique story in science. Indeed, the
development of contemporary financial theory follows the usual progres-
sion of scientific progress. Normal science does not aim at novelties of
fact or theory and, when successful, finds none.2 But when fundamen-
tal novelties of fact and theory arise, [d]iscovery commences with the
awareness of anomaly, i.e., with the recognition that nature has somehow
violated the paradigm-induced expectations that govern normal science.3
Once an awareness of anomaly ha[s] lasted so long and penetrated so
deep as to plunge a scientific discipline into a state of growing crisis,
a succeeding period of pronounced professional insecurity over the
persistent failure of the puzzles of normal science prompts a fruitful
search for new rules.4
The quest for scientific understanding assumes even greater urgency in
finance, a field devoted to elaborating uncertainty, both in theory and
[in] empirical implementation.5 The starting point for every financial
model is the uncertainty facing investors, and the substance of every financial
model involves the impact of uncertainty on the behavior of investors and,
ultimately, on market prices.6 The interplay between theory and empiri-
cal work is a dialectic in which [t]heorists develop models with testable
predictions and empiricists document puzzles, or stylized facts that
fail to fit established theories and thereby stimulate[] the development of
new theories.7 What makes finance in general and asset pricing in particular
such fantastic instances of the scientific process is that the random shocks
that propel knowledge forward happen also to be the subject matter to
which these branches of economic theory devote themselves.8
The presence of efficiency-defying anomalies such as market swings
in the absence of new information and prolonged deviations from under-
lying asset values invites challenges to the efficient markets hypothesis.9
Not all departures from market efficiency carry the same cognitive weight,
however. Chapters 4 and 7 of Postmodern Portfolio Theory distinguished
between the volatility and correlation components of beta partly on the
basis of differences in the way investors perceive, evaluate, and respond to
those quantifiable aspects of financial markets. Even in the shadow of high-
frequency trading,10 contemporary markets exhibit meaningful differences
in the rate at which they absorb different types of information. Differences
in processing speed distinguish two basic models of human reasoning: a
speedy, intuitive mode prone to cognitive bias and mistakes in judgment,
and a slower, more rational mode that counterbalances humans innate
heuristics with comprehensive evaluation of evidence.
Behavioral finance reflects the interplay between the fast heuristics of
human behavior and the slow processing of rational evidence.11 Adopting
labels proposed in the psychological literature,12 Daniel Kahneman has
assigned the names System 1 and System 2, respectively, to these fast and
slow modes of thought.13 System 1 operates automatically and quickly, with
little or no effort and no sense of voluntary control.14 Intuitive, fast think-
ing, such as the automatic[] and effortless[] recognition of anger in a
human face, requires no work.15 It just happen[s].16 By contrast, System
2 allocates attention to effortful mental activities, often those associ-
ated with the subjective experience of agency, choice, and concentration.17
Solving even a simple multiplication problem such as 17 24 demands
slow thinking. Ponder, even for a second, whether the right answer to
that problem is 568 or 408.18 Slow thinking slogs through a sequence of
steps requiring deliberate, effortful, and orderly mental work.19
Complex financial computations presumably belong to the domain of
System 2.20 The common connection among the highly diverse operations
6 J.M. CHEN
of System 2 is their need for attention.21 Drawing attention away from a

task assigned System 2 will disrupt or even defeat this process of slow think-
ing.22 Successful discharge of System 2 responsibilities demands the commit-
ment of constrained mental resources: [Y]ou dispose of a limited budget of
attention that you can allocate to activities, and if you try to go beyond your
budget, you will fail.23 In particular, activities that impose high demands
on System 2 require self-control, an exercise that is depleting and unpleas-
ant.24 System 2s dependence on mental energy is more than a mere meta-
phor;25 it literally commandeers and demands blood glucose.26
It is traditionally assumed that the slow rationality of System 2, albeit
imperfectly, curbs the fast heuristics and emotional excesses of System
1.27 The most complex financial calculations are assigned to System 2, the
place where the conscious, reasoning self of neoclassical economics and
mathematical finance carefully marshals its beliefs, makes choices, and
decides what to think and what to do.28 But this compounds the usual
error of giving too much credit to rationality at the expense of instinct.
Although System 2 believes itself to be where the action is, the automatic
System 1 is the [real] hero of human cognition.29 The mind at work may
assign even surprisingly complex patterns of ideas to the automatic
operations and the freewheeling impulses and associations of System
1.30 For a chess master, finding a strong move constitutes an automatic
activit[y] attributed to System 1.31
Moreover, the persistence of superstition and magical thinking, even
among educated and emotionally stable adults,32 suggests that System 2
may consist of two distinct processes: a moderately slow mechanism for
detecting cognitive errors committed by System 1, and an even slower
mechanism for correcting those errors.33 Instances where humans detect
their mistakes but choose not to correct them arguably represent an
entirely distinct response to uncertainty: acquiescence.34
1.3 SELL INMAY ANDGOAWAY?

It therefore behooves us to distinguish anomalies that might offer insight
into investor behavior from those that do little beyond identifying quanti-
tative curiosities, at least within the limits of existing technology and psy-
chological learning. Fluctuations in security prices according to the time
of year or even the day of the week undermine confidence that rationality
rather than human frailty rules the market.35
To earn genuine respect, however, technical considerations must sup-
ply practical investment advice, or at least inform financial decisions.
Should a rational investor really sell in May and go away?36 Maybe, or

maybe not.37 Although no less an authority than Eugene Fama has found
a January effect, whereby stock returns, especially on small stocks,
on are average higher in January,38 no anomaly carries economic signifi-
cance, let alone undermines the efficient markets hypothesis, unless it is
strong enough to outperform a buy-and-hold strategy on a risk adjusted
basis.39 To the extent that these calendar anomalies ever held sway, the
passage of time appears to have dissipated them, or at least substantially
attenuated their power.40 Presumably, savvy trading has exhausted any
arbitrage opportunity presented by the identification of the anomaly.41
If calendar anomalies are to offer any insight into investor behavior,
those anomalies must arise from factors affecting emotion and judg-
ment. We do know that humans respond to news and environmental
stimuli. Reading sad rather than happy newspaper articles, for instance,
predisposes people to raise their estimates of the risk of various causes
of death and their levels of concern over those sources of mortality.42
Although the existence of the effect and its extent are contested,43
some studies have found that good weather positively influences stock
returns.44 Even geomagnetic storms are alleged to affect financial deci-
sions.45 Some of these effects, if indeed they exist, are almost surely
attributable to the market impact of putatively random events with
emotional content, ranging from the trivial (sports events)46 to the
tragic (aviation disasters).47
By contrast, seasonal changes in climate may outweigh the potential
of ephemeral events such as the weather to exert a powerful and sys-
tematic influence on returns.48 For everything there is a season, said
the Preacher, and a time for every matter under heaven.49 Investors
evidently agree: Mutual funds flows around the world reveal an inves-
tor preference for safer funds in the fall and riskier funds in the spring.50
These preferences hold in Australia as well as in Canada and the USA.51
Because spring and fall are reversed on either side of the equator, these
three developed Anglophone markets demonstrate that risk-seeking
among investors rises with seasonal temperature, and not according to
fixed calendar dates.
1.4 LAW ONTHEMARKET

I now present a markedly distinct illustration of the failure of markets to
satisfy the strict assumptions of modern portfolio theorythat returns be
normally distributed and new information be assimilated in frictionless
8 J.M. CHEN
fashion into security prices.52 In tacit homage to the fraud on the mar-
ket doctrine in federal securities law, we may call this phenomenon law
on the market.53 The Supreme Court of the USA routinely decides cases
involving publicly traded parties, or at least significant legal issues with
potential impact on security prices. Applying standard event study meth-
odology,54 one survey of Supreme Court decisions from October Term
1999 through October Term 2013 (which ended in June 2014) found
79 decisions associated with abnormal returns on 118 securities.55 Those
79 decisions represented 5.5% of the Courts docket during the relevant
time span.56 Share price changes in 118 securities in direct response to a
Supreme Court decision reached an estimated total of $140 billion.57
For our purposes, the crucial finding of this survey was the rate at which
new information from a Supreme Court decision diffused through the
securities market. In the algorithmically driven, high-frequency trading
environment of contemporary markets,58 security prices often move within
fractions of a second in response to central bank announcements,59 sur-
veys of consumer sentiment,60 and other financial news.61 High-frequency
trades typically move in the direction of permanent price changes, which
presumably reflect future efficient price moves, and in the opposite
direction of transitory pricing errors.62 Although high-frequency traders
impose adverse selection costs on other investors, they play a beneficial
role in price efficiency and supply liquidity in stressful times such as the
most volatile days and around macroeconomic news announcements.63
By contrast, the full diffusion of Supreme Court decisions through
financial markets may take hours, even an entire trading day.64 In spec-
tacular instances, the market affirmatively misinterprets a Supreme Court
decision and, at least initially, sends the prices of affected securities in the
wrong direction. In the 2012 case of National Federation of Independent
Business v. Sebelius,65 apparent misreporting on the actual nature of the
closely watched, hotly controversial Obamacare decision66 sparked very
high volatility in the stock prices of health insurance companies such as
Aetna (AET), Humana (HUM), and Anthem/WellPoint (WLP).67
Even more dramatically, the 2013 decision in Association for Molecular
Pathology v. Myriad Genetics Inc.68 accounted for a 10% abnormal increase
in the stock price of Myriad Genetics (MYGN) in the first hour of trading
after the 10a.m. announcement of the decision, which was reversed into
a 10% abnormal decrease during the final two hours of the trading day.69
Over two trading days, the Supreme Courts decision accounted for 20%
negative abnormal returns in the price of MYGN.70
Admittedly, the legal reasoning in the Sebelius and Myriad decisions

was highly complex. Expert legal analysts, let alone capital markets, had
evidently failed to anticipate that the high court could somehow reject
the governments characterization of the Affordable Care Act as regula-
tion of interstate commerce, but nevertheless uphold health care reform
as an exercise of Congresss powers over taxation. Nor did the relevant
legal or financial actors appear to anticipate that the Court would invali-
date Myriads patent claim to DNA mutations in the BRCA1 and BRCA2
genes (which are associated with a heightened risk of breast and ovar-
ian cancer), but manage to uphold Myriads patent in complementary
DNA extracted from the same genetic material. The important implica-
tion for finance is that economically significant information from Supreme
Court decisions diffuses throughout markets over the course of minutes,
hours, or even entire trading days, a veritable eternity in the age of high-
frequency trading.
The following heat map, covering two trading days after the announce-
ment of 79 financially significant, law on the market decisions by the
Supreme Court, shows considerable amounts of blue, green, and yellow
to the leftcolors indicating less than full assimilation of new information
by the securities market (Fig. 1.1).71
The behavioral implications of law on the market, if any, lie in the time
lags between the arrival of new information and the assimilation of that
information by a putatively efficient market. Barriers to the diffusion of
financially significant information from the Supreme Court, to say nothing
of less salient legal tribunals, appear to rise from legal complexity and the
nuance involved in interpreting the real economic impact of certain legal
decisions. I do not mean to suggest that law is immune to computational
analysis; algorithmic analysis promises new weaponry, for instance, against
tax evasion.72 But there remain meaningful differences in the speed with
which certain types of information are digested and diffused throughout
the market.
In finance, as in other domains, the quality of decision-making is a func-
tion of time pressure and information load.73 Among individual investors,
the speed and convenience accompanying the transition from phone-based
to online trading platforms led to more trading, more speculation, and
lower profits.74 Nearly instantaneous machine processing of announce-
ments concerning macroeconomic variables and consumer confidence
fuels high-frequency trading strategies that move billions in market capi-
talization within seconds.75 Security-specific news from court decisions, at
10 J.M. CHEN
S 6/25/2014
SBGI 6/25/2014
FOXA 6/25/2014
CBS 6/25/2014
HAL 6/23/2014
BHI 6/23/2014
XOP 6/23/2014
XLE 6/23/2014
XES 6/23/2014
C 6/23/2014
SANM 6/9/2014
CTS 6/9/2014
LLNW 6/2/2014
XLK 1/14/2014
XTL 12/10/2013
VOX 12/10/2013
F 12/2/2013
MA 6/20/2013
XPH 6/17/2013
ACT 6/17/2013
MYGN 6/13/2013
XLP 5/13/2013
XOP 4/17/2013
XLE 4/17/2013
PSO 3/19/2013
WLP 6/28/2012
MGLN 6/28/2012
HUM 6/28/2012
HNT 6/28/2012
HCA 6/28/2012
CI 6/28/2012
AET 6/28/2012
XLE 6/21/2012
KMI 6/21/2012
XHE 6/18/2012
STN 6/18/2012
XLV 4/17/2012
UBS 3/26/2012
DB 3/26/2012
CS 3/26/2012
GT 6/27/2011
XLY 6/20/2011
XES 6/6/2011
XLY 5/31/2011
JAH 5/31/2011
XLF 5/16/2011
XTL 4/27/2011
XPH 3/29/2011
XTL 3/1/2011
VOX 3/1/2011
GM 2/23/2011
WEX 1/24/2011
AXP 1/24/2011
XLI 4/21/2010
CDE 6/22/2009
CNA 6/18/2009
PFE 3/4/2009
DD 1/26/2009
MO 12/15/2008
XLY 6/23/2008
S 6/23/2008
XLV 2/20/2008
VOX 4/30/2007
TFX 4/30/2007
VOX 4/17/2007
GLBC 4/17/2007
MO 2/20/2007
EBAY 5/15/2006
MER 3/21/2006
ITW 3/1/2006
YRCW 6/20/2005
MRK 6/13/2005
XLP 5/16/2005
XLI 12/13/2004
XOM 6/24/2004
HES 6/24/2004
CVX 6/24/2004
UNH 6/21/2004
XLI 5/3/2004
V 4/21/2004
XLF 12/2/2003
XLY 6/26/2003
TRV 6/23/2003
XLP 6/9/2003
XLV 5/19/2003
PFE 5/19/2003
BMY 5/19/2003
AZN 5/19/2003
PHS 4/7/2003
XLI 3/10/2003
UNP 3/10/2003
NSC 3/10/2003
WLB 1/15/2003
XLY 5/20/2002
T 5/20/2002
T 5/13/2002
XLE 1/9/2002
GWO 1/8/2002
XLP 12/10/2001
XLI 12/10/2001
VGR 6/28/2001
XLI 6/4/2001
CC 3/21/2001
HMC 2/27/2001
GM 2/27/2001
XLV 2/21/2001
XLE 12/4/2000
XLF 6/12/2000
PLA 5/22/2000
KSU 4/17/2000
JNJ 4/3/2000
BMY 4/3/2000
ABT 4/3/2000
XOM 3/6/2000
CVX 3/6/2000
9:30 10:00 Day 1 Close Day 2 Close
0.5 0.0 0.5 1.0 1.5
Fig. 1.1 Cumulative abnormal returns from Supreme Courts decisions, 1999
through 2014, as a function of time
least under existing technology, still requires additional evaluation, both

legal and financial, before informed trading can take place. Given an extra
second, minute, or hour, would markets dare/Disturb the universe?76
In a minute, after all, there is time/For decisions and revisions which
a minute will reverse.77 The time it takes informed traders to act on law
on the market marks the temporal boundaries of the domain within which
instinct and calculationaction and reaction within the behavioral uni-
verse of financeinfluence security prices.
From calendar anomalies to law on the market, we have now found
our focus within our quest for behavioral departures from strict rational-
ity in the evaluation of financial risk. Cracks in the edifice of the efficient
markets hypothesis are the openings from which a more comprehen-
sive and more accurate account of behavioral finance will emerge. Not
every anomaly has enough economic significance to provide material for
a workable trading strategy. And not every trading strategy arises from
cognitive biases and behavioral heuristics. These limitations safely con-
sign slogans such as sell in May and go away to the domain of popular
financial journalism. By contrast, the financial impact of Supreme Court
decisions tantalizingly suggests a trading strategy whose temporal window
of opportunity may be orders of magnitude wider than that of the usual
high-frequency trading algorithm.78
1.5 RAW EMOTION

Theories of behavioral finance become necessary only in the presence of
uninformed investors and noise traders.79 A market composed solely of
information traders is a market where price efficiency and the CAPM
hold, where [r]isk premia are determined solely by beta and distribution
of returns on the market portfolio, and where option prices80 and the
term structure of bonds81 follow mathematically beautiful models reflect-
ing comparably rational assumptions about those corners of the financial
marketplace.82 The actions of noise traders weaken the relation between
security returns and beta, but they create a positive conditional corre-
lation between abnormal returns and beta.83 As behavioral anomalies
exert steady and forceful pressure upon the twin paradigms of price
efficiency and the CAPM, a correspondingly compelling need arises for a
behavioral theory of capital asset prices and the volume of trade.84
At an even broader level of generality, behavioral limits undermine
the assumption of rationality that permeates not merely modern portfo-
lio theory, but all of neoclassical economics.85 Real consumers and real
12 J.M. CHEN
investors simply do not behave like the stylized actors of neoclassical eco-
nomics rational expectations hypothesis.86 We will never really under-
stand important economic events unless we confront the fact that their
causes are largely mental in nature.87 Risk, the prime mover in finance, is
experienced and understood in emotional terms.88 And the primary forces
that appeal to emotion take verbal, visual, and narrative form: [M]uch of
the human thinking that results in action is not quantitative, but instead
takes the form of storytelling and justification.89
And storytelling is gossip, the steady deliverer of secrets, , the carrier
of speculation and suspicion.90 To live without gossip is to forfeit the
perilous cost of being born human:91
Gossip is theology translated into experience. In it we hear great stories of

conversion as well as stories of failure. When we gossip we are also
praying, not only for them but for ourselves.92
[T]here is one story in the world, and only one .93 (Or perhaps as
many as seven, as we shall soon see.) Every individual,94 every organiza-
tion,95 every country,96 every religion97 lays claim to some form of unique-
ness: Cultures of all kinds stress uniqueness and claim to be superior or
to offer the one true faith.98 But claims to organizational uniqueness ulti-
mately reduce to no more than seven stories, which are not unique at all,
but universal.99 The simultaneous recognition of these stories in diverse
domains resembles the scientific phenomenon of multiple discovery,100
which arguably portrays the typical way by which science advances.101
Social organizations, including businesses, thus repeat the human sto-
rytelling experience, which consists of seven basic plots.102 Of particular
interest to finance is the archetypical tale of rags to riches.103 Implicit in the
tale of successful rise from obscurity, poverty and misery to a state of great
splendor and happiness104 is the dark version of the Rags to Riches
plot, in which failed or pyrrhically victorious protagonists reach their
self-destructive [destiny] by precisely the same rules which govern the
attainment of material and spiritual satisfaction.105 So perhaps, there is only
one story in the world after all, and we revel in telling it again and again.106
Unsurprisingly, the impact of language, down to the very words we use,
depends on its connection to the physical senses.107 When making financial
decisions, investors weigh[] a story, which has no quantitative dimension,
against the observed quantity of financial wealth that they have available
for consumption.108 Over time, the most cognitively appealing narratives
congeal into conventional wisdom,109 which investors, financial advisors,
commentators, and politicians can all exploit, with varying degrees of per-
sonal success and societal impact. [P]opular narratives, particularly human
interest stories, are fundamental drivers of motivation.110
At a minimum, the mixed, and changing, feelings that accompany
the admittedly arbitrary social convention of starting a new year may
drive a large number of calendar effects, of the sort described in 1.2.111
[P]eople love almanacs and have an irrational belief in them because
they purport to find order in random events, like weather or the seasons.112
More ambitiously, all storytellershistorians, novelists, psychoanalysts, or
even behavioral economistsstrive to feel[] what [they] believe[] others
are feeling and to attribute[] what [they] remember[] to others as [they]
observe[] or describe[] their subjects in order to achieve verisimilitude
or, in the narrative craft at its finest, wonderfully vivid mimesis.113
Although the earliest critiques of modern portfolio theory emphasized
downside risk,114 fear of loss is hardly the only behavioral pitfall in finance.
Arguably greater danger lurks in the allure of speculation, fueled by unre-
alistic expectations of runaway gains (a special concern in ground floor
investment opportunities perceived as offering lottery-like returns, such as
venture capital, business development companies, initial public offerings,
crowdfunding, and even penny stocks).115 Not even John Maynard Keynes
was immune.116 While managing an investment fund at Cambridge,
Keynes allegedly engaged in speculative investment practices and reached
cowboy proportions of risk,117 in the sense that he reached 80% of the
maximum levels justified under the Kelly criterion for determining the
optimal size of a series of bets.118
The performance of real markets provides no basis for the traditional,
sanguine assumption that all actors are rational and aim to maximize their
own welfare in a universally uniform, objective way. Brain-damaged peo-
ple, unconstrained by fear or memory, actually beat unimpaired people at
an investment game, because those players performances were hobbled
by the memory of losses from previous rounds.119 (Of course, stacking
the deck the other wayquite literallyso that players with damage to
their prefrontal lobes could not remember dangerous aspects of the game
would tilt the contest back in favor of unimpaired players.)120 If anything,
overconfidence in personal business acumen explains a meaningful mea-
sure of price movements in capital markets.121
At worst, the affect heuristic122 encourages economic agents to evaluate
the magnitude of risk and expected loss according to raw likes and dis-
likes.123 In more conventional financial settings, after all, the addition or
14 J.M. CHEN
removal of dot-com from a company name resulted in abnormally posi-

tive returns, depending on whether the name change took place before or
after the technology bust of the early 2000s.124 In cruder terms, human
decision-making is thrall to the laws of sympathetic magicthe emo-
tional laws dictating that contact with disgusting objects constitutes
permanent contamination (food touching a cockroach is repulsive) and
that visual similarity constitutes qualitative equivalence (food resembling a
cockroach is also repulsive).125
1.6 TRADE LIKE AGIRL

To confound matters even more, add sex to the equation. There is a pro-
nounced sex-based difference in investor overconfidence, and it lopsid-
edly impairs men.126 Men are especially prone to overconfidence in areas
stereotypically perceived as domains where men excel.127 And overconfi-
dence, simply put, is hazardous to your wealth.128 Sensation-seeking is also
more pronounced in men than in women, though it does decline with age
in both sexes.129 Sheer seniority pays a maturity dividend in the form of
better diversification and less destructively aggressive trading.130 Boys will
be boys, but if they are lucky, some boys do grow up and become men.
Despite the mellowing effects of age, sex-based differences still stand
out. The greater an investors distance from women, the worse the results
from trading: Single men traded less sensibly than married men, and
married men traded less sensibly than single women: the less the female
presence, the less rational the approach to trading in the markets.131 From
these insights arises the hilarious (but not inaccurate) title, Warren Buffett
Invests Like a Girl: And Why You Should, Too.132 The simple version of the
lesson might be summarized thus: Muliebrity good, masculinity bad.
The basic ingredients of behavioral degradation of financial returns
appear to be quite simple: fear and greed.133 A study of the impact of
psychological factors on trading performance revealed that intense,
emotional reactions to monetary gains and losses, on both the positive
and the negative side, significantly impaired performance among day-trad-
ers.134 The same study offers some reassurance through its failure to dis-
cover any specific personality profile associated with success in trading.135
In that case, trading skills may not be innate, and proper guidance can
enable traders with different personality types to succeed.
Somewhat less reassuring is the suggestion that fear and greed track
the sexual boundary between men and women. The literature on sex dif-
ferences in cooperation offers elaborate theoretical grounds for assuming
that men respond more strongly to greed and that women respond more
strongly to fear.136 But much of that literature rests on the prisoners
dilemma, a game known to contain elements of both fear and greed.137
The quest for empirical evidence to verify or falsify the greedy men/fear-
ful women hypothesis now hinges on the design of experiments that are
better able to isolate sex-specific responses to fear and to greed.138
Using a more sophisticated two-parameter specification of prospect the-
orys weighting function,139 one study has found that women are indeed
more pessimistic than men with respect to their evaluation of potential
gains.140 Relative to men, women are less sensitive to changes in probabil-
ity.141 Women also tend to lower their estimates that they will secure large
gains.142 These findings align research on men and womens distinctive
approaches to financial decision making with traditional scholarship on
gender-specific differences in response to non-financial risks, especially
physical or life-threatening risks.143
Specifically:
On average, female probability weighting functions differ from male ones

in a specific way. Womens curves are more curved [in all circumstances].
Moreover, in the domain of gains, women tend to underestimate larger
probabilities much more strongly than do men. This gender difference is
particularly pronounced when the decisions are framed in investment terms
rather than in abstract terms. And indeed, we find women to be more risk
averse than men when facing investment choices.144
Most intriguingly of all, especially in light of other research into mens

propensity to overestimate their financial prowess, the context-specific
pessimism expressed by women appears to arise from sex-specific differ-
ences in overconfidence and financial market knowledge.145 As their
objectively measured knowledge of finance increases, men become
relatively more risk averse.146 By contrast, as women know more about
finance, they become relative more risk prone.147 The net effect of these
differences is that the appetite for financial risk is greatest in most foolishly
overconfident men and the most cautiously sophisticated women.148
Closer examination of investment choices by women and by men
demonstrate[s] that the effects of gender on investment decisions are
more complicated than [some] research has suggested.149 Cursory sur-
veys of participant-directed retirement accounts initially suggested, in line
with a financially dimorphic view of the sexes, that women were more con-
servative investors than men and more likely than men to populate their
16 J.M. CHEN
accounts with bonds than with stocks.150 This apparent difference, how-
ever, evaporates upon consideration of marital status: [T]hough women
and men do not differ, married women are more likely than single women
to choose mostly bonds.151 Just as [s]urprisingly, perhaps, neither
education nor age seems to affect allocation decisions.152 Filtering away
these factors leaves something not likely to be coded within defined con-
tribution retirement accounts: the presence of children and intertemporal
(indeed, intergenerational) financial considerations spanning not only dif-
ferent periods of an investors life, but also the lives of her heirs.
More recent research evaluating investor behavior within defined con-
tribution retirement plans shows that women are 14% likelier than men
to participate in a workplace savings plan.153 Once enrolled, women at all
income levels save at higher rates than their male counterparts.154 Though
women and men take similar levels of portfolio risk (partly because the use
of target-date plans as a default subjects a large number of participants,
female or male, to identical levels of risk), men have greater marginal pro-
pensity to trade and to hold more aggressive equity allocations.155 Because
they earn more, men hold higher balances in absolute terms.156 Controlling
for income, however, women save more and have higher balances.157
Studies emphasizing sex-based differences in financial behavior encoun-
ter the limitations that constrain all efforts to describe behavioral hetero-
geneity by identifying demographic distinctions such as sex or age and
assuming, somewhat ironically, that all investors fitting that characteristic
behave alike.158 It is quite evident that trading like a girl involves far
more than a binary, state-switching algorithm devised by a rogue sociobi-
ologist. Even as it sorts humanity into distinct male and female bins, the
opening passage in the classic American novel, Their Eyes Were Watching
God, offers a subtler, more persuasive view:
Ships at a distance have every mans wish on board. For some they come
in with the tide. For others they sail forever on the horizon, never out of
sight, never landing until the Watcher turns his eyes away in resignation, his
dreams mocked to death by Time. That is the life of men.
Now, women forget all those things they want to remember, and remem-
ber everything they dont want to forget. The dream is the truth. Then they
act and do things accordingly.159
An appropriately humble and nuanced approach to mathematical

finance therefore acknowledges the presence of anomalies that strictly
neoclassical economic models such as modern portfolio theory cannot
explain. The descriptive abnormalities of financial markets cannot be

separated from irrational investor behavior. Indeed, those abnormalities
may arise from investors own departures from rational conduct. Even
the most disciplined traders may err in their discovery and evaluation of
price signals. All humans, especially nave traders prone to react on the
basis of noise, are prone to cognitive biases. In an effort to bring comfort-
ing but ultimately illusory order to chaos, we look for patterns that are
not there, and persuade ourselves against all evidence that we have found
them.160 This is to say nothing of the possibility of whimsy and caprice.
Our contemporary selves can scarcely forecast our own future preferences,
let alone give proper evaluation to markets today.
We therefore continue this tour of behavioral finance with an examina-
tion of mental accounting, emotional hierarchies, and behavioral heuristics.
NOTES
1. James Ming Chen, Postmodern Portfolio Theory: Navigating Abnormal
Markets and Investor Behavior (2016).
2. See Thomas S.Kuhn, The Structure of Scientific Revolutions 52 (2d ed.
enlarged, 1970).
3. Id. at 5253.
4. See id. at 6667.
5. John Y.Campbell, Andrew W.Lo & A.Craig MacKinlay, the Econometrics
of Financial Markets 3 (1997).
6. Id.
7. John Y. Campbell, Asset Pricing at the Millennium, 55 J. Fin. 1515
1567, 1515 (2000).
8. See id.
9. Halliburton Co. v. Erica P. John Fund, Inc., 134 S. Ct. 2398, 2421
(2014) (Thomas, J., concurring in the judgment) (quoting Donald
C. Langevoort, Taming the Animal Spirits of the Stock Markets: A
Behavioral Approach to Securities Regulation, 97 Nw. U. L. Rev. 135
188, 141 (2002)).
10. See, e.g., Jonathan Brogaard, Terrence Hendershott & Ryan Riordan,
High-Frequency Trading and Price Discovery, 27 Rev. Fin. Stud. 2267
2306 (2014); Terrence Hendershott, Charles M. Jones & Albert
J. Menkveld, Does Algorithmic Trading Improve Liquidity?, 66 J. Fin.
133 (2011); Albert J.Menkveld, High Frequency Trading and the New
Market Makers, 16J.Fin. Mkts. 712740 (2013); Martin L.Scholtus,
Dick J.C. Van Dijk & Bart Frijns, Speed, Algorithmic Trading, and
Market Quality Around Macroeconomic News Announcements,
38J.Banking & Fin. 89105 (2014).
18 J.M. CHEN
11. See generally Daniel Kahneman, Thinking, Fast and Slow 19105 (2011);
Chen, Postmodern Portfolio Theory, supra note 1, 4.4, at 4449 (sum-
marizing Kahnemans dichotomy between System 1 and System 2
thinking).
12. See Keith E. Stanovich & Richard F. West, Individual Differences in
Reasoning: Implications for the Rationality Debate, 23 Behav. & Brain
Scis. 645665 (2000); see also In Two Minds: Dual Processes and Beyond
(Jonathan St. B.T. Evans & Keith Frankish eds., 2009) (recognizing a
similar divide within the dual nature of human thought); Jonathan St.
B. T. Evans, Dual-Processing Accounts of Reasoning, Judgment, and
Social Cognition, 59 Ann. Rev. Psych. 255278 (2008) (same).
13. See Kahneman, Thinking, Fast and Slow, supra note 11, at 2021, 450.
14. Id. at 20 (emphasis in original).
15. Id. at 1920.
16. Id. at 20.
18. Id. at 20. The right answer is 1724=408.
19. Id. at 20.
20. Id. at 22.
21. Id.
22. See id.
23. Id. at 23.
24. Id. at 42. See generally Martin S.Hagger, Chantelle Wood, Chris Stiff &
Nikos L.D.Chatzisarantis, Ego Depletion and the Strength Model of Self-
Control: A Meta-Analysis, 136 Psych. Bull. 495525 (2010).
25. Kahneman, Thinking, Fast and Slow, supra note 11, at 43.
26. See generally Matthew T.Gailliot & Roy F.Baumeister, The Physiology of
Willpower: Linking Blood Glucose to Self-Control, 11 Personality & Soc.
Psych. Rev. 303327 (2007); Matthew T. Gailliot, Roy F. Baumeister,
C. Nathan DeWall, John K. Maner, E. Ashby Plant, Dianne M. Tice,
Lauren E.Brewer & Brandon J.Schmeichel, Self-Control Relies on Glucose
as a Limited Energy Source: Willpower Is More Than a Metaphor,
92J.Personality & Soc. Psych. 325336 (2007); cf. Roy F.Baumeister,
W.Scott Simpson, Stephen J.Ware & Daniel S.Weber, The Glucose Model
of Mediation: Physiological Bases of Willpower as Important Explanations for
Common Mediation Behavior, 15 Pepperdine Dispute Resolution L.J.
377413 (2015).
27. See generally, e.g., Daniel Kahneman & Shane Frederick, Representativeness
Revisited: Attribute Substitution in Intuitive Judgment, in Heuristics and
Biases: The Psychology of Intuitive Judgment 4981 (Thomas Gilovich,
Dale W.Griffin & Daniel Kahneman eds., 2002); Daniel Kahneman &
Shane Frederick, A Model of Heuristic Judgment, in The Cambridge
Handbook of Thinking and Reasoning 267293 (Keith J. Holyoak &
Robert G.Morrison eds., 2005).
28. Kahneman, Thinking, Fast and Slow, supra note 11, at 21.
29. Id.
30. Id.
31. Id.
32. See, e.g., Jane L.Risen & Thomas Gilovich, Why People Are Reluctant to
Tempt Fate, 95J.Personality & Soc. Psych. 293307 (2008).
33. See Jane L. Risen, Believing What We Do Not Believe: Acquiescence to
Superstitious Beliefs and Other Powerful Intuitions, Psych. Rev. (Oct. 19,
2015) (available online at http://psycnet.apa.org/index.cfm?fa=buy.opt
ionToBuy&id=2015-47838-001).
34. See id.
35. See, e.g., M.Hakan Berument & Nukhet Dogan, Stock Market Return
and Volatility: Day-of-the-Week Effect, 36 J. Econ. & Fin. 282302
(2011); Kenneth R. French, Stock Returns and the Weekend Effect,
8J.Fin. Econ. 5569 (1980).
36. See Sven Bouman & Ben Jacobsen, The Halloween Indicator, Sell in May
and Go Away: Another Puzzle, 92 Am. Econ. Rev. 16181635 (2002);
Sandro C.Andrade, Vidhi Chhaochharia & Michael E.Fuerst, Sell in
May and Go Away Just Wont Go Away, 69:4 Fin. Analysts J. 94105
(July/Aug. 2013).
37. See Edwin D. Maberly & Raylene M. Pierce, Stock Market Efficiency
Withstands Another Challenge: Solving the Sell in May/Buy After
Halloween Puzzle, 1 Econ. J.Watch 2946 (2004).
38. Eugene F. Fama, Efficient Capital Markets II, 46 J. Fin. 15751617,
15861587 (1991) (noting, in addition, that much of the higher
January return on small stocks comes on the last trading day in December
and the first 5 trading days in January); see also Werner F.M. de Bondt
& Richard H. Thaler, Further Evidence on Investor Overreaction and
Stock Market Seasonality, 42J.Fin. 557581 (1987); Donald B.Keim,
Size-Related Anomalies and Stock Return Seasonality: Further Empirical
Evidence, 12J.Fin. Econ. 1332 (1983); Richard Roll, Vas Ist Das? The
Turn-of-the-Year Effect and the Return Premia of Small Firms,
9:2 J. Portfolio Mgmt. 1828 (Winter 1983); Michael S. Rozeff &
William R. Kinney Jr., Capital Market Seasonality: The Case of Stock
Returns, 3 J. Fin. Econ. 379402 (1976); Seha M. Tinic & Richard
R.West, Risk and Return: January Versus the Rest of the Year, 13J.Fin.
Econ. 561574 (1984); cf. James N.Bodurtha, Jr. & Nelson C.Mark,
Testing the CAPM with Time-Varying Risks and Returns, 46 J. Fin.
14851505 (1991) (devising a conditional, time-varying capital asset
pricing model to test the January effect).
39. Maberly & Pierce, supra note 37, at 30; see also id. (quoting Burton Malkiel
and Richard Roll for the proposition that calendar time anomalies, to
serve as evidence of market inefficiency, should present an exploitable
opportunity). See generally Michael C.Jensen, Some Anomalous Evidence
20 J.M. CHEN
Regarding Market Efficiency, 6 J. Fin. Econ. 95101 (1978); Mark

Rubinstein, Rational Markets: Yes or No? The Affirmative Case, 57:3 Fin.
Analysts J. 1529 (May/June 2001).
40. G.William Schwert, Anomalies and Market Efficiency, in 1 Handbook of
the Economics of Finance 939972, 945 (George M. Constantinides,
Milton Harris & Ren M.Stulz eds., 2003).
41. See Haim Levy, The Capital Asset Pricing Model in the 21st Century:
Analytical, Empirical, and Behavioral Perspectives 4 n.11 (2012).
42. See Eric J. Johnson & Amos Tversky, Affect, Generalization, and the
Perception of Risk, 45J.Personality & Soc. Psych. 2031, 2324 (1983).
43. See Mark A.Trombley, Stock Prices and Wall Street Weather: Additional
Evidence, 36 Q.J.Bus. & Econ. 1121 (1997).
44. See Edward M.J.Saunders, Jr., Stock Prices and Wall Street Weather, 83
Am. Econ. Rev. 13371345 (1993); David Hirshleifer & Tyler Shumway,
Good Day, Sunshine, 58J.Fin. 10091032 (2003) (extending Saunderss
work to 26 countries).
45. See Anna Krivelyova & Cesare Robotti, Playing the Field: Geomagnetic
Storms and the Stock Market, Federal Reserve Bank of Atlanta Working
Paper 2003-5b (Oct. 2013) (available at http://www.frbatlanta.org/
research/publications/wp/2003/03_05.aspx).
46. See Alex Edmans, Diego Garcia & yvind Norli, Sports Sentiment and
Stock Returns, 62J.Fin. 19671998 (2007).
47. See Guy Kaplanski & Haim Levy, Sentiment and Stock Prices: The Case of
Aviation Disasters, 95J.Fin. Econ. 174201 (2009).
48. See Mark J.Kamstra, Lisa A.Kramer & Maurice D.Levi, Winter Blues: A
SAD Stock Market Cycle, 93 Am. Econ. Rev. 324343 (2003). SAD
refers to seasonal affective disorder. See generally Justin Osborn, Jacqueline
Raetz & Amanda Kost, Seasonal Affective Disorder, Grief Reaction, and
Adjustment Disorder, 98 Med. Clinics N.Am. 10651077 (2014). But see
Megan K. Traffanstedt, Shela Mehta & Steven G. LoBello, Major
Depression with Seasonal Variation: Is It a Valid Construct?, Clin. Psych.
Sci. (Jan. 19, 2016) (available at http://cpx.sagepub.com/content/early
/2016/01/18/2167702615615867) (concluding that seasonal varia-
tion in depression diagnoses, though strongly rooted in folk psychol-
ogy, is unrelated to latitude, season, or sunlight and ultimately is not
supported by objective data).
49. Ecclesiastes 3:1 (Revised Standard Version).
50. See Mark J.Kamstra, Lisa A.Kramer, Maurice D.Levi & Russ Wermers,
Seasonal Asset Allocation: Evidence from Mutual Fund Flows (August 12,
2014) (available at http://ssrn.com/abstract=1907904).
51. See id.
52. See generally Chen, Postmodern Portfolio Theory, supra note 1, 2.7, at
1315.
53. See Basic Inc. v. Levinson, 485 U.S. 224, 241247 (1988); Halliburton
Co. v. Erica P.John Fund, Inc., 134 S.Ct. 2398, 24072413 (2014).
54. See generally A.Craig MacKinlay, Event Studies in Economics and Finance,
35 J. Econ. Lit. 1339 (1997); cf. Sanjai Bhagat & Roberta Romano,
Event Studies and the Law, Part I: Technique and Corporate Litigation, 4
Am. L. & Econ. Rev. 141168 (2002).
55. See Daniel Martin Katz, Michael J. Bommarito II, Tyler Soellinger &
James Ming Chen, Law on the Market? Evaluating the Securities Impact
of Supreme Court Decisions 5, 22 (Aug. 24, 2015) (available at http://
ssrn.com/abstract=2649726).
56. See id. at 22.
57. See id. at 3.
58. See, e.g., Terrence Hendershott, Charles M.Jones & Albert J.Menkveld,
Does Algorithmic Trading Improve Liquidity?, 66 J. Fin. 133 (2011);
Albert J.Menkveld, High Frequency Trading and the New Market Makers,
16J.Fin. Mkts. 712740 (2013); Martin L.Scholtus, Dick J.C.Van Dijk
& Bart Frijns, Speed, Algorithmic Trading, and Market Quality Around
Macroeconomic News Announcements, 38 J. Banking & Fin. 89105
(2014).
59. Compare Benjamin S.Bernanke & Kenneth N.Kuttner, What Explains
the Stock Markets Reaction to Federal Reserve Policy?, 60J.Fin. 1221
1257 (2005) (Federal Reserve System of the USA) with David-Jan.
Jansen & Jakob de Haan, Were Verbal Efforts to Support the Euro Effective?
A High-Frequency Analysis of ECB Statements, 23 Eur. J.Pol. Econ. 245
259 (2007) (European Central Bank).
60. See Shumi Akhtar, Robert Faff, Barry R. Oliver & Avanidhar
Subrahmanyam, Stock Salience and the Asymmetric Market Effect of
Consumer Sentiment News, 36J.Banking & Fin. 32893301 (2012).
61. See Axel Gro-Klussmann & Nikolaus Hautsch, When Machines Read the
News: Using Automated Text Analytics to Quantify High Frequency News-
Implied Market Reactions, 18J.Empirical Fin. 321340 (2011); Robert
F.Schumaker & Hsinchun Chen, A Quantitative Stock Prediction System
Based on Financial News, 45 Info. Processing & Mgmt. 571583 (2009).
62. Jonathan Brogaard, Terrence Hendershott & Ryan Riordan, High-
Frequency Trading and Price Discovery, 27 Rev. Fin. Stud. 22672306,
2268 (2014).
63. Id.
64. See Katz, Bommarito, Soellinger & Chen, supra note 55, at 2021.
65. 132 S.Ct. 2566 (2012).
66. See Josh Blackman, Unprecedented: The Constitutional Challenge to
Obamacare 237252 (2013).
22 J.M. CHEN
68. See 133 S.Ct. 2107 (2013).

70. See id.
71. See id. at 21 (figure 6).
72. See Erik Hemberg, Jacob Rosen, Geoff Warner, Sanith Wijesinghe &
Una-May OReilly, Tax Non-Compliance Detection Using Co-Evolution of
Tax Evasion Risk and Audit Likelihood, ICAIL 15: Proceedings of the
15th International Conference on Artificial Intelligence and Law 7988
(Association for Computing Machinery, 2015).
73. See, e.g., Minhi Hahn, Robert Lawson & Young Gyu Lee, The Effects of
Time Pressure and Information Load on Decision Quality, 9 Psych. &
Marketing 365378 (1992); Jonathan C.Pettibone, Testing the Effect of
Time Pressure on Asymmetric Dominance and Compromise Decoys in
Choice, 7 Judgment & Decision Making 513523 (2012).
74. See Terrance Odean & Brad M.Barber, Online Investors: Do the Slow Die
First?, 15 Rev. Fin. Stud. 455487 (2002).
75. On the impact of consumer confidence on asset prices and subsequent
returns, see generally Michael Lemmon & Evgenia Portniaguina,
Consumer Confidence and Asset Prices: Some Empirical Evidence, 19 Rev.
Fin. Stud. 14991529 (2006).
76. T.S.Eliot, The Love Song of J.Alfred Prufrock, in Collected Poems, 1909
1962, at 37, 4 (1991) (1st ed. 1963).
77. Id.
79. See Hersh Shefrin & Meir Statman, Behavioral Capital Asset Pricing
Theory, 29J.Fin. & Quant. Analysis 323349, 323 (1994). See generally
Fischer Black, Noise, 41J.Fin. 529543 (1986). For a persuasive dem-
onstration of the so-called noisy markets hypothesis, see Jeremy J.Siegel,
The Noisy Markets Hypothesis, Wall St. J., June 14, 2006, at A14, is a
mathematically flawed effort to undermine the efficiency of capitaliza-
tion-weighted equity market portfolios, see Andr F. Perold,
Fundamentally Flawed Indexing, 63:6 Fin. Analysts J. 3137 (Nov./
Dec. 2007); cf. Jason Hsu, Cap-Weighted Portfolios Are Sub-Optimal
Portfolios, 4:3J.Inv. Mgmt. 4453 (3d quarter 2006); Jack Treynor, Why
Market-Valuation-Indifferent Indexing Works, 61:5 Fin. Analysts J.
6569 (Sept./Oct. 2005).
80. See generally Fischer Black & Myron S.Scholes, The Pricing of Options
and Corporate Liabilities, 81 J. Pol. Econ. 637654 (1973); Robert
C.Merton, The Theory of Rational Option Pricing, 4 Bell J.Econ. 141
183 (1973).
81. See generally John C.Cox, Jonathan E.Ingersoll, Jr. & Stephen A.Ross,
A Theory of the Term Structure of Interest Rates, 53 Econometrica 385
408 (1985); Stephen J. Brown & Philip H. Dybvig, The Empirical
Implications of the Cox, Ingersoll, Ross Theory of the Term Structure of

Interest Rates, 41J.Fin. 617630 (1986); Roger H.Brown & Stephen
M. Schaefer, The Term Structure of Real Interest Rates and the Cox,
Ingersoll, and Ross Model, 35J.Fin. Econ. 342 (1994).
82. Shefrin & Statman, Behavioral Capital Asset Pricing Theory, supra note
79, at 323.
83. Id. at 346. See generally Allan W. Kleidon, Anomalies in Financial
Economics: Blueprint for Change?, 59J.Bus. 469499 (1986).
79, at 323.
85. See, e.g., Gary S. Becker, An Economic Approach to Human Behavior
(1976); Richard A.Posner, Economic Analysis of Law (7th ed. 2007).
86. See John A. Muth, Rational Expectations and the Theory of Price
Movements, 29 Econometrica 315335 (1961). See generally Rational
Expectations and Economic Policy (Stanley Fischer ed., 1980); Thomas
J. Sargent, Rational Expectations and Inflation (1986); Steven
M.Sheffrin, Rational Expectations (2d ed. 1996).
87. George A. Akerlof & Robert J. Shiller, Animal Spirits: How Human
Psychology Drives the Economy, and Why That Matters for Global
Capitalism 1 (2009).
88. See, e.g., George F. Loewenstein, Emotions in Economic Theory and
Economic Behavior, 65 Am. Econ. Rev. 426432 (2000); George
Loewenstein, Elke U. Weber, Chris K. Hsee & Ned Welch, Risk as
Feelings, 127 Psych. Bull. 267286 (2001).
89. Robert J.Shiller, Irrational Exuberance 168 (3d ed. 2015) (emphases in
original).
90. Cynthia Ozick, The Novels Evil Tongue, N.Y.Times, Dec. 20, 2015, at
BR1 (available at http://nyti.ms/227GDFj).
91. Id.
92. Kathleen Norris, Dakota: A Spiritual Geography 76 (2001; 1st ed. 1993).
93. John Steinbeck, East of Eden 413 (1st ed. 1952).
94. See C.R. Snhyder & Howard L. Fromkin, Uniqueness: The Human
Pursuit of Difference (1980).
95. See Philip Selznick, Leadership in Administration: A Sociological
Interpretation 139 (1957) (defining an organizations distinctive com-
petence as work other organizations cannot perform); Charles C.Snow
& Lawrence G. Hrebiniak, Strategy, Distinctive Competence, and
Organizational Performance, 25 Admin. Sci. Q. 317336 (1980).
96. See Clifford Geertz, The Interpretation of Cultures (1973); Josep
R.Llobera, Foundations of National Identity: From Catalonia to Europe
101105 (2005) (discussing Ernest Gellner, Nations and Nationalism
(1983); Ernest Gellner, Nationalism (1997)).
24 J.M. CHEN
97. See Craig Ott & Harold A.Netland, Globalizing Theology: Belief and
Practice in a Era of World Christianity (2006).
98. Joanne Martin, Martha S.Feldman, Mary Jo Hatch & Sim B.Sitkin, The
Uniqueness Paradox in Organizational Stories, 2 Admin. Sci. Q. 438
453, 438 (1983).
99. See id.
100. See generally, e.g., David Lamb & Susan M.Easton, Multiple Discovery:
The Pattern of Scientific Progress (1984); Robert K.Merton, Resistance
to the Systematic Study of Multiple Discoveries in Science, 4 Eur. J.Sociol.
237282 (1963), reprinted in Robert K. Merton, The Sociology of
Science: Theoretical and Empirical Investigations 371382 (1973).
101. See Robert K.Merton, Singletons and Multiples in Scientific Discovery: a
Chapter in the Sociology of Science, 105 Proc. Am. Phil. Socy 470486
(1961), reprinted in Merton, The Sociology of Science, supra note 100,
at 343370.
102. See generally Christopher Booker, The Seven Basic Plots: Why We Tell
Stories (2004).
103. See generally id. at 5168.
104. Id. at 65.
105. Id. at 66, 68.
106. See id. at 4 (identifying five identifiable stages in all stories, from the
initial mood of anticipation, through a dream stage when all seems to be
going unbelievably well, to the frustration stage when things begin to
go mysteriously wrong, to the nightmare stage where everything goes
horrendously wrong, and ultimately to a climactic resolution).
107. See Ezgi Akpinar & Jonah Berger, Drivers of Cultural Success: The Case of
Sensory Metaphors, 109J.Personality & Soc. Psych. 2034 (2015) (sug-
gesting that linguistic phrases that relate to physical senses in metaphoric
ways, such as a cold person, have enjoyed greater cultural success).
108. Shiller, Irrational Exuberance, supra note 89, at 168.
109. See John Kenneth Galbraith, The Affluent Society 617 (2002) (1st ed.
1958).
110. Robert J.Shiller, How Stories Drive the Stock Market, N.Y.Times, Jan. 24,
2016, at BU5 (available at http://nyti.ms/1OCxI7F). See generally
Jerome Bruner, Actual Minds, Possible Worlds (1987) (propounding a
theory of narrative psychology, in which the imaginative creation of pos-
sible worlds propels a large number of human endeavors).
111. Shiller, How Stories Drive the Stock Market, supra note 110.
112. James B. Stewart, For Stock Markets, January Is a Cloudy Crystal Ball,
N.Y.Times, Jan. 8, 2016, at B1 (available at http://nyti.ms/1VMMw6g)
(quoting Nassim Nicholas Taleb).
113. Ramsay MacMullen, Feelings in History: Ancient and Modern 8 (2012;
1st ed. 2003).
114. See generally Chen, Postmodern Portfolio Theory, supra note 1, 5.1, at
5960.
115. See infra 9.1 and 9.2, at 215224.
116. See Rachel E.S. Ziemba & William T. Ziemba, Scenarios for Risk
Management and Global Investment Strategies 2931 (2007) (Good
and Bad Properties of the Kelly Criterion); Leonard C. Maclean,
Edward O.Thorp & William T.Ziemba, Long-Term Capital Growth: The
Good and Bad Properties of the Kelly and Fractional Kelly Capital Growth
Criteria, 10 Quant. Fin. 681687 (2010). The speculative spirit of the
Jazz Age is captured by Philip Carret, The Art of Speculation (1930).
117. See J.H.Chua & R.S.Woodward, The Investment Wizardry of J.M.Keynes,
39:3 Fin. Analysts J. 3537 (MayJune 1983); cf. J.E.Woods, On Keynes
as an Investor, 37 Camb. J. Econ. 423442 (2013) (tracing Keyness
eventual abandonment of speculation in favor of an approach that
would be recognizable today as value investing).
118. See generally J.L. Kelly, A New Interpretation of Information Rate, 35
Bell Sys. Tech. J. 917926 (1956); Edward O.Thorp, The Kelly Criterion
in Blackjack, Sports Betting, and the Stock Market, in 1 Handbook of
Asset and Liability Management: Theory and Methodology 385428
(Stavros A. Zenios & William T. Ziemba eds., 2006); E.O. Thorp,
Optimal Gambling Systems for Favorable Games, 37 Rev. Intl Stat. Inst.
273293 (1969); Vasily Nekrasov, Kelly Criterion for Multivariate
Portfolios: A Model-Free Approach (Sept. 30, 2014) (available at http://
ssrn.com/abstract=2259133). f*, the fraction of a bankroll that a gam-
bler should wager on a single bet, may be expressed as the ratio of
expected net proceeds from a winning bet to the nominal value of the net
winnings from a winning bet. Formally:
bp - q p ( b + 1) - 1
f* = =
b b
where p is the probability of winning, q=1p is the probability of losing,

and b expresses the odds-based payout on a winning bet, as expressed in
the form of b to 1 odds. http://en.wikipedia.org/wiki/Kelly_criterion
119. See Baba Shiv, Goerge Loewenstein, Antoine Bechara, Hanna Damasio &
Antonio R. Damasio, Investment Behavior and the Negative Side of
Emotion, 16 Psych. Sci. 435439 (2005).
120. See Antoine Bechara, Hanna Damasio, Daniel Trand & Antonio
R.Damasio, Deciding Advantageously Before Knowing the Advantageous
Strategy, 275 Science 12931295 (1997). For an older but entertaining
account of the psychological effects of prefrontal damage and other
insults to the brain, see Antonio Damasio, Descartes Error, Emotion,
Reason, and the Human Brain 5282 (1994).
26 J.M. CHEN
121. See Kent Daniel, David Hirshleifer & Avanidhar Subrahmanyan,

Overconfidence, Arbitrage, and Equilibrium Asset Pricing, 56 J. Fin.
921965 (2001); Harrison Hong, Jos Scheinkman & Wei Xiong, Asset
Float and Speculative Bubbles, 61J.Fin. 10731117 (2006); Lin Peng &
Wei Xiong, Investor Attention, Overreaction, and Category Learning,
80J.Fin. Econ. 563602 (2006). On the psychology of overconfidence,
see generally Don A. Moore & Paul J. Healy, The Trouble with
Overconfidence, 115 Psych. Rev. 502517 (2008).
122. See generally Melissa L. Finucane, Ali Alhakami, Paul Slovic & Stephen
M. Johnson, The Affect Heuristic in Judgments of Risks and Benefits,
13J.Behav. Decision Making 117 (2000); Paul Slovic, Melissa Finucane,
Ellen Peters & Donald G.Macgregor, The Affect Heuristic, in Heuristics
and Biases: The Psychology of Intuitive Judgment 397420 (Thomas
Gilovich, Dale Griffin & Daniel Kahneman eds., 2002); Paul Slovic, Ellen
Peters, Melissa L.Finucane & Donald G.MacGregor, Affect, Risk, and
Decision Making, 24 Health Psych. S35-S40 (2005); Meir Statman,
Kenneth L.Fisher & Deniz Anginer, Affect in a Behavioral Asset-Pricing
Model, 64:2 Fin. Analysts J. 2029 (March/April 2008); R.B. Zajonc,
Feeling and Thinking: Preferences Need No Inferences, 35 Am. Psychologist
151175 (1980).
123. See Justin Pidot, Deconstructing Disaster, 2013 BYU L.Rev. 213257,
242.
124. Compare Michael J.Cooper, Orlin Dimitrov & P.Raghavendra Rau, A
Rose.com by Any Other Name, 56 J. Fin. 23712388 (2001) (finding
positive value in the adoption of a dot-com name during the technology
boom) with Michael J.Cooper, Ajay Khorana, Igor Osobov, Ajay Patel &
P. Raghavendra Rau, Managerial Actions in Response to a Market
Downturn: Valuation Effects in the Dot.com Decline, 11 J. Corp. Fin.
319335 (2005) (finding positive value in the removal of dot-com from
corporate names after the technology crash).
125. See Paul Rozin, Linda Millman & Carol Nemeroff, Operation of the Laws
of Sympathetic Magic in Disgust and Other Domains, 50J.Personality &
Soc. Psych. 703712 (1986).
126. See Brad M. Barber & Terrance Odean, Boys Will Be Boys: Gender,
Overconfidence, and Common Stock Investment, 116 Q.J.Econ. 261292
(2001).
127. See Kay Deaux & Elizabeth Farris, Attributing Causes for Ones Own
Performance: The Effects of Sex, Norms, and Outcome, 11J.Research in
Personality 5972 (1977); Kay Deaux & Elizabeth Farris, Complexity,
Extremity, and Affect in Male and Female Judgments, 43J.Personality
379389 (1975).
128. See, e.g., Daniel Kahneman & Dan Lovallo, Timid Choices and Bold
Forecasts: A Cognitive Perspective on Risk Taking, 39 Mgmt. Sci. 1731
(1993); Stuart Oskamp, Overconfidence in Case-Study Judgments,
29 J. Consulting Psych. 261265 (1965); cf. William B. Locander &
Peter W. Hermann, The Effect of Self-Confidence and Anxiety on
Information-Seeking in Consumer Risk Reduction, 16 J. Marketing
Research 268274 (1979).
129. See David W.Fulker, Sybil B.G.Eysenck & Marvin Zuckerman, A Genetic
and Environmental Analysis of Sensation Seeking, 14 J. Research in
Personality 261281 (1980).
130. See Daniel Dorn & Paul Sengmueller, Trading as Entertainment?, 55
Mgmt. Sci. 591603 (2009); William N. Goetzmann & Alok Kumar,
Equity Portfolio Diversification, 12 Rev. Fin. 433463 (2008).
131. Michael Lewis, Boomerang: Travels in the New Third World 37 (2011).
132. See LouAnn Lofton, Warren Buffett Invests Like a Girl: And Why You
Should, Too (2011).
133. See generally Hersh Shefrin, Beyond Greed and Fear: Understanding
Behavioral Finance and the Psychology of Investing (2000).
134. See Andrew W.Lo, Dimitry V.Repin & Brett N.Steenbarger. Fear and
Greed in Financial Markets: A Clinical Study of Day-Traders, 95 Am.
Econ. Rev. 352359, 357 (2005).
135. See id.
136. See Brent Simpson, Sex, Fear, and Greed: A Social Dilemma Analysis of
Gender and Cooperation, 82 Soc. Forces 3552, 36 (2003).
137. See id. at 3738.
138. See id. at 4748; Ko Kuwabara, Nothing to Fear But Fear Itself: Fear of
Fear, Fear of Greed and Gender Effects in Two-Person Asymmetric Social
Dilemmas, 84 Soc. Forces 12571272 (2005).
139. See infra 10.3. Prospect theory is the subject of Chapter. 8.
140. See Helga Fehr-Duda, Maneule de Gennaro & Renate Schubert, Gender,
Financial Risk, and Probability Weights, 60 Theory & Decision 283
313, 299 (2006).
141. See id.
142. See id.
143. Id. at 283; cf., e.g., James P.Byrnes, David C.Miller & William D.Schafer,
Gender Differences in Risk Taking: A Meta-Analysis, 125 Psych. Bull.
367383 (1999).
144. Fehr-Duda, de Gennaro & Schubert, supra note 140, at 285; see also id.
at 304305. For insight into systematic, cohort-specific differences in
probability weighting functions, see generally Herbert Walther, Normal-
Randomness Expected Utility, Time Preferences, and Emotional Distortions,
52J.Econ. Behav. & Org. 253266 (2003).
28 J.M. CHEN
145. Fehr-Duda, de Gennaro & Schubert, supra note 140, at 305.

146. Id. at 305306.
147. Id. at 306.
148. See id.
149. Annika E.Sundn & Brian J.Surette, Gender Differences in the Allocation
of Assets in Retirement Savings Plans, 88 Am. Econ. Rev. 207211, 209
(1998).
150. See Richard P.Hinz, David D.McCarthy & John A.Turner, Are Women
Conservative Investors? Gender Differences in Participant Directed Pension
Investments, in Positioning Pensions for the Twenty-First Century
91103 (Michael. S. Gordon, Olivia S. Mitchell & Marc M. Twinney
eds., 1997). See generally James P. Byrnes, David C. Miller & William
D. Schaefer, Gender Differences in Risk Taking: A Meta-Analysis, 125
Psych. Bull. 367383 (1999); Catherine C.Eckel & Philip J.Grossman,
Sex Differences and Statistical Stereotyping in Attitudes Toward Financial
Risk, 23 Evol. & Human Behav. 281295 (2002).
151. See Sundn & Surette, supra note 149, at 209.
152. Id.
153. See Jean A.Young, Women Versus Men in DC Plans 2 (Oct. 2015) (avail-
able at https://institutional.vanguard.com/iam/pdf/GENDRESP.pdf).
154. See id.
155. See id. at 67.
156. See id. at 8.
157. See id. at 9.
158. Robert A.Nagy & Robert W.Obenberger, Factors Influencing Individual
Investor Behavior, 50:4 Fin. Analysts J. 6368 (July/Aug. 1994); accord,
e.g., Athar Iqbal & Sania Usmani, Factors Influencing Individual Investor
Behavior (The Case of the Karachi Stock Exchange), 3 S.Asian J.Mgmt.
Scis. 1516 (2009); Sania Usmani, Factors Influencing Individual Investor
Behaviour in Karachi, 2 Intl J.Asian Soc. Sci. 10331047 (2012). See
generally James Farrell, Demographic and Socioeconomic Factors of Investors,
in Investor Behavior: The Psychology of Financial Planning and Investing
117134 (H.Kent Baker & Victor Ricciardi eds., 2014).
159. Zora Neale Hurston, Their Eyes Were Watching God 1 (1937).
160. Cf. Joshua Benjamin Miller & Adam Sanjurjo, Surprised by the Gamblers
and Hot Hand Fallacies? A Truth in the Law of Small Numbers, IGIER
Working Paper No. 552 (Sept. 15, 2015) (available at http://ssrn.com/
abstract=2627354) (finding a subtle but substantial bias in a standard
measure of the conditional dependence of present outcomes on streaks of
past outcomes in sequential data, whose correction would reverse the
findings of prominent studies of the gamblers fallacy, the hot-hand fal-
lacy, and other instances of the so-called law of small numbers).
CHAPTER 2
Mental Accounting, Emotional Hierarchies,

andBehavioral Heuristics
2.1 KEEPING EMOTIONAL SCORE

If there is one fixed star in the firmament of economic science, it is the
principle that sunk costs are just that, sunk.1 Vorbei ist vorbei; reden wir
nicht mehr davon.2 Or in the words of plain English taught to generations
of American college students: One of the most important lessons of eco-
nomics is that you should look at the marginal costs and marginal benefits
of decisions and ignore past or sunk costs.3
If only we could follow such economically impeccable advice. The
endowment effect, a bedrock element of humans innate heuristics for
evaluating risk,4 leads agents to overvalue preexisting wealth and to take
economically unwarranted account of sunk costs. Consider, for example,
the formation of public policies regarding natural disasters, extreme events
involving physical violence far exceeding that of most financial setbacks.
Behavioral biases routinely confound disaster management. Governments
systematically mismanage the choice between ex ante investments in
disaster preparedness and ex post expenditures on disaster relief. By one
estimate, each dollar in disaster preparedness is worth roughly $15 in
mitigated future damage.5 Even though an ounce of prevention is almost
literally worth a pound of cure, individuals and governments systemati-
cally underinvest in disaster preparedness ex ante and overinvest in disaster
relief ex post.6 Along every spatial, temporal, and behavioral dimension, the

DOI10.1007/978-3-319-32711-2_2
30 J.M. CHEN
political economy of disaster assistance dictates perverse outcomes,7 which

we may ruefully but truthfully describe as accidents waiting to happen,8
or catastrophic responses to catastrophic risks.9
Aspire as we might to guide by the light of reason,10 virtually every
domain of human decision-making is refracted through the lens of emo-
tion. Human beings cannot resist the temptation to keep emotional score.11
Some reference points are so powerful that even sustained efforts are
framed around achieving some sort of arbitrary benchmark. Consider, for
example, finishing times in marathons. Despite their superhuman athletic
talents, marathon runners emotional preferences are thoroughly human.
So overpowering is the allure of finishing on the hour or the half-hour
that marathon runners exert extraordinary effort to attain these round-
number benchmarks.12 Notably, human judgment warps even under the
influence of random numbers selected by a wheel of fortune marked 1 to
100numbers that should have no emotional significance.13 With round-
number benchmarks in an endeavor as demanding and personal as mara-
thon running, athletes and other humans dont stand a chance.
Similarly, arbitrary benchmarks mark the many paths in the race of
life. The vaunted $1 billion threshold separating unicorns from other
venture-backed companies represents one of the most vivid of these arbi-
trary benchmarks.14 More generally, prior financial outcomes exert pow-
erful influence over humans risky choices.15 Financial experiences guide,
even dictate, future savings and investment decisions.16 The ability (if any)
to learn from investment mistakes17 requires knowledge of past financial
outcomes.18 Mental accounting takes stock of the numerous manifesta-
tions of the human inability to forget economic events.19 Specifically,
mental accounting refers to the implicit methods individuals use to code
financial outcomes, whether in investments, gambles, or other economic
transactions.20
Mental accounting helps simplify the most cognitively taxing chores
in finance. For instance, it requires considerable effort to process covari-
ance, correlation, and other properties of joint probability distributions.21
When confronted with complex choices and cognitively challenging tasks,
humans often respond by disregarding shared components and focusing
on distinguishing characteristics.22 Reliance on emotions increases with
the complexity of information and with stress.23
In one experiment, subjects asked to construct portfolios of three secu-
rities simply ignored covariances among the securities.24 Even institutional
investors take mental shortcuts. Despite the potential loss in efficiency
MENTAL ACCOUNTING, EMOTIONAL HIERARCHIES, ANDBEHAVIORAL... 31
from failing to integrate their assets in a single portfolio, institutional

investors often optimize their international holdings in piecemeal fashion.
They maintain separate portfolios, keeping global securities within one
layer and currencies within an overlay whose management is assigned
to an outside specialist.25 At a stipulated level of volatility at 15%, the suc-
cessively less efficient strategies of joint, partial, and separate optimization
of security and currency accounts, respectively, provide an estimated 172,
139, and 72 basis points in diversification benefits every year.26
Even more spectacular mental shortcuts abound in finance. For exam-
ple, many investors engage in strictly nave asset allocation: they design
their portfolios by allocating exactly 1/n of total wealth to each of n
instruments or asset classes.27 Moreover, the so-called rank effect pre-
dicts that investors are more likely to sell the extreme winning and the
extreme losing positions in their portfolios, without regard to the holding
period of those positions, the absolute level of positive or negative returns,
or any firm-specific information.28 Focusing narrowly on the worst and the
best positions is tantamount to ignoring the rest of the portfolio.
This chapter will explore two variations on the theme of mental account-
ing. The first of those themes takes a look at the inner world of Maslowian
portfolio theory, a way of understanding financial decision-making in
harmony with each individuals hierarchy of physiological, social, and
emotional needs. This chapter then reverses field and explores behavioral
economics in a far more physically expansive domain. The making of envi-
ronmental policy is likewise a species of risk management, one where the
vectors of physical uncertainty and emotional reaction differ from those of
finance in degree more than in kind.29 Physics is far from alone among sci-
ences contributing to financial evaluations of abnormal markets.30 Among
branches of economics, environmental economics provides an especially
rich source of insights into the impact of emotion, cognitive bias, and
behavioral heuristics on risk assessment and management.
2.2 MASLOWIAN PORTFOLIO THEORY

In its strictest, traditional formulation, modern portfolio theory assumes
purely rational investors who trade solely on information and have no
needs apart from those that are best served by holding a mean-variance
optimized portfolio along the efficient frontier.31 As a potential vector
for irrational and therefore irrelevant decision-making criteria, mental
accounting lies beyond contemplation.
32 J.M. CHEN
This purely rational approach to investing and risk management befits

hypothetical decision-making agents, not real human beings.32 The coun-
tervailing approach of Maslowian portfolio theory33 aligns assetallocation
and portfolio design with the hierarchy of human needs articulated by
the legendary psychologist, Abraham Maslow,34 from simple physiology
to love and social esteem. The resulting infusion of behavioral realism
gives portfolio theory its own potential shot at greatness: an opportunity
to transcend its limits as a strictly descriptive theory and to become a pow-
erfully accurate predictive theory, perhaps even an emotionally persuasive
prescriptive theory.
Maslow arranged human needs as layers of a pyramid. The base of that
pyramid comprises basic physiological needs (such as food, hydration, and
a stable body temperature) and safety, in the sense of freedom from imme-
diate threats to health or safety.35 At higher levels of the pyramid, Maslow
placed love (in the sense of caring for family and offspring) and esteem,
which he defined as achievement, reputation, prestige, and social stand-
ing.36 The value of layering to a behavioral understanding of finance lies in
the recognition that human needs are not all addressed simultaneously,
but rather in a hierarchical progression from survival and safety to love
and esteem.37 Notably, money has a twofold role: In addition to its
instrumental value as a means of exchange for goods that confer esteem,
money itself can be a source of esteem and recognition.38 Individual
attitudes toward economic risk take account of the psychological impact
of prospective changes in wealth, including the impact of financial gain or
loss on an individuals self-regard and social standing.39
Even more significantly, Maslowian portfolio theory recognizes that
investors, contrary to the separation theorem and two mutual fund theo-
rem of modern portfolio theory,40 mentally arrange their wealth accord-
ing to specific, consumption-oriented goals.41 Whereas modern portfolio
theory and mean-variance optimization are silent about ultimate port-
folio consumption goals, such goals are central to Maslowian portfolio
theory and other behaviorally sensitive approaches to finance.42 From an
unapologetically instrumental perspective, individual investors want their
portfolios to satisfy goals such as a secure retirement, college education for
the children, or being rich enough to hop on a cruise ship whenever they
please.43 Institutional investors likewise wish to pay[] promised benefits
and to add[] new benefits.44
Critically, a single investor can have multiple risk profiles, not simply
across different Maslowian levels of need, but also for investment goals
within one need level.45 Quite realistically, an investor might be very

averse to risk with respect to retirement, much less averse with respect
to childrens education (since there are scholarships and subsidized loan for
college but not for retirement), and willing to take any risk, even be risk
seeking, with [investments] devoted to getting rich.46 Investment strate-
gies accordingly follow the priorities of life; overall wealth is divided
into multiple sub-portfolios, each linked to a separate goal within its
own need level.47 Changes in [n]eeds as life unfolds dictate dif-
ferent financial investment strategies.48 Over the course of the investors
lifetime, overall well-being remains portfolio theorys Holy Grail.49
Investors allocating assets according to Maslowian principles do not
consider their portfolio as a whole.50 Risk for each layer of human needs
and for each goal within each layer is separately measured by the prob-
ability of failing to reach the threshold level of return needed to attain a
particular goal.51 Identifying the probability of not reaching the [return]
threshold of each particular goal necessarily departs from more conven-
tional modes of optimization that define risk as the standard deviation
of the return of the overall portfolio.52
Human beings are far better prepared to define their consumption-
oriented goals than they are to express the risk-aversion coefficients
demanded of them by conventional portfolio optimization techniques.53
If investment goals are understood as current preferences for future
expenditures, it becomes readily apparent that such future liabilities .
are usually fuzzy and imprecise both in magnitude and timing.54 For
[a] client who is concerned about not reaching a target return, it hardly
serves an advisor to discuss[] variance.55 Humans are likewise more able
to define specific goals (such as retirement, education, or life-changing
wealth) than to quantify their ambitions in abstract, aggregate terms.56
Ideally, sensitivity to the investors hierarchy of needs, and to the possibil-
ity that needs and preferences might change over time, should provide[]
structure and discipline for emotion-free investment decision making,
with the specific benefit of reduc[ing] panic selling during market down-
turns and euphoric buying following exuberant times.57
Maslowian portfolio theory does depart from Abraham Maslows orig-
inal psychological vision in an important respect. Contemporary appli-
cations of Maslow to behavioral finance seek to anchor[] the hierarchy
of human motives more firmly in the bedrock of modern evolution-
ary theory.58 One ambitious effort, in particular, supplies an updated
and revised hierarchy of human motives according to insights from
34 J.M. CHEN
evolutionary biology, anthropology, and psychology, with the conscious

purpose of elevating the quaint visual artifact of Maslows pyramid of
layered human needs and motives to its deserved status as one of the
most cognitively contagious ideas in the behavioral sciences.59
Financially significant consequences flow from the rearrangement of
Maslows hierarchy according to the evolutionary function, developmen-
tal sequencing, and cognitive priority of motivations such as physiologi-
cal survival, safety, love, and esteem.60 These modifications of Maslows
original hierarchy recognize that humans have multiple motivational
and learning systems, which operate using different rules and, indeed,
involve architecturally distinct areas of the brain.61 As Maslow himself
acknowledged, [W]e could never understand fully the need for love no
matter how much we might know about the hunger drive.62
Moreover, neo-Maslowian wisdom recognizes that certain motivations
negate each other. No behavioral inclination is likely to operate without
costs, and fear reactions are a good example.63 The logic underlying this
intuition bridges evolutionary biology and behavioral finance. Although
fearful avoidance minimizes losses to predators or poisonous animals
or hostile strangers, that same instinct also eliminates risky situations
that, if confronted, could yield payoffs.64 The cognitive architecture
of fearful avoidance has evolutionary origins. When the risk of physi-
cal damage is highly costly, threat-avoidance systems are likely to be set
like smoke alarms, favoring false positive alarms rather than false negative
complacency, since an individual who unnecessarily flee[s] is likelier to
survive than one who mistakenly remain[s] in danger[].65
Maslowian psychology is likely to make a profound contribution to
behavioral finance toward the lower layers of the hierarchy of needs, or at
least when the threat of loss looms larger than the prospect of gain. When
humans feel threatened, they are less inclined to take riskswhether that
meant, in ancient times, venturing farther away from caves, or gambling
on the unknown in modern markets, politics, or social settings.66
At the other end of the emotional spectrum, subjective perceptions of
power, in the form of greater optimism and a firmer sense of control, tend
to prompt risk-seeking behavior.67 Introducing threats with the potential
to remove that power or sense of control, however, reverses the generally
risk-promoting effect of power. When the dominance hierarchy becomes
unstable, powerful and power-motivated individuals will retreat into a
more conservative, risk-averse mode of decision-making.68 In the presence
of ultimate threats, not just to welfare or power but also to life, mortality
salience erodes generosity toward others, requiring active work to pro-

mote the victory of any cultural norm favoring generosity over competing
norms (and personal predispositions) favoring greed.69
2.3 SHOTS AT GREATNESS: REHABILITATING SELF-

ACTUALIZATION INNEO-MASLOWIAN THOUGHT
ASATRADING STRATEGY
Neo-Maslowian psychology does reject outright what was perhaps the

most distinctive element of Abraham Maslows theory of human moti-
vation. Maslow placed self-actualization at the top of his pyramid of
needs. A musician must make music, he wrote, an artist must paint, a
poet must write, if he is to be ultimately at peace with himself.70 Even if
the United States Army has abandoned the slogan, Maslowian psychology
exhorts: Be all that you can be.71
But in elevating the self, Maslow erred in disconnecting the desire to ful-
fill ones own unique potential from the biological foundations of human
motivation.72 Self-actualization fails on strictly sociological grounds, since
any self-inflating tendencies not calibrated to others respect could have
maladaptive consequences for success in social groups.73 What the social
psychologist calls self-actualization may prove to be nothing more impres-
sive than financial overconfidence or, given a sufficiently large departure
from social acceptance, unfiltered narcissism. Contemporary psychological
sources informing Maslowian portfolio theory have therefore remov[ed]
self-actualization from a revised version of Maslows pyramid, as a con-
cession to the reality that the privileged position once accorded to self-
actualization cannot be compelled [or] justified by the functional logic
of human evolutionary biology.74 Even Maslow himself, in later elabora-
tions of his own theory, felt compelled to add altruism, spirituality, and a
sense of something outside or larger than the self to the top layer of his
pyramid of needs and motivations.75
Despite its demotion within neo-Maslowian thought, self-actualization
in the sense of a personal shot at greatness persists in behavioral finance.
Indeed, it is not too much to suggest that the longing for shots at great-
ness pervades behavioral finance. As we will see in 9.1 and 9.2, the
willingness to gamble for over-the-top returns explains a wide variety of
phenomena, from the simple failure to diversify portfolios to systematic
skewness preference and a seemingly universal affinity for lotteries and
financial instruments promising lottery-like returns.
36 J.M. CHEN
True to the admonition that no departure from rational market effi-

ciency is significant, letalone meaningful as a window on human behavior,
unless it can be framed as a trading strategy,76 perhaps we can rehabilitate
the traditional Maslowian focus on self-actualization in purely pecuniary
terms. The managerial literature on employee motivation devotes ample
attention to the deep desire of workers for meaning on the job, well beyond
what Maslow would have considered the bare essentials.77 Custodians at
an academic hospital regarded their interactions with patients, their fami-
lies, and medical staff as the most rewarding aspects of their jobseven
though their formal job descriptions excluded any mention of other peo-
ple and, of course, provided no financial compensation for these putatively
voluntary activities.78
Whether businesses bear any social responsibility beyond maximiz-
ing shareholder profits within the bounds of the law79 is the domain of
the vast literature on corporate social responsibility (CSR).80 There are
sources finding annual abnormal returns as high as 8.7% from an invest-
ment strategy based on buying stock in companies with high CSR rat-
ings and selling stock in companies with low ratings.81 Admittedly, other
sources find no evidence of significant differences in risk-adjusted returns
between ethically conscious and conventional mutual funds.82 Still other
sources find that ethical funds not only fail to outperform their conven-
tional counterparts, but also underperform the benchmark established by
the so-called four-factor FamaFrenchCarhart model, which emphasizes
value, size, and momentum in addition to some measurement of system-
atic risk.83 Findings of no better than mixed success characterize studies
focusing on more narrowly defined approaches to CSR, such as Islamic
equity funds and banks honoring Quranic prohibitions on the payment
and receipt of interest (among other financial limitations).84 The absence
of clear financial signals from CSR-motivated management and investing
probably stems from the complexity involved in measuring ethical busi-
ness behavior85 and from investors variable responses to different activities
in different industrial sectors.86
One strain in the CSR literature does appear to guide financial perfor-
mance. In the financial sector, where ethical ratings are provided by the
Ethisphere Institute, highly rated firms offer higher returns and lower vol-
atility than counterpart firms failing to secure Ethisphere approval.87 These
differences ostensibly arise because positive perceptions among inves-
tors improve investment flow and access to capital in ethically operated
companies.88 Similarly, another study finds a strong negative correlation
between volatility and high third-party ratings for environmental, social,

and governance-related (ESG) corporate practices.89 The low-volatility
anomaly should enable active fund managers to delete the lowest ESG-
rated firms at virtually no cost in screening and with a positive impact on
average and maximum portfolio return and on the probability distribu-
tion of returns.90 These observations offer insight into the low-volatility
anomaly and Bowmans paradox, related departures from the expected
relationship between risk and return that will draw closer examination in
4.1 and 9.3.
Lowering volatility is one way of enhancing risk-adjusted returns. The
other strategy requires finding abnormally high returns. Because the quest
for alpha, or above-market returns, may be as quixotic as a search for
El Dorado or the Fountain of Youth,91 paying heed to each firms treat-
ment of its own employees may bear sweeter fruit. Beyond confirming
Maslowian psychologys emphasis on social interaction and acceptance as
an essential bridge to self-actualization and a shot at greatness, insights
into the value of emotionally engaged employees have notable implica-
tions for finance. The American semiconductor industry of the 1970s and
1980s, legendary for its technological prowess and its contribution to the
countrys economic hegemony, owed much of its success to its handling
of human resources.92 Corporate culture, best defined as the cluster of
relationships centered on the individual employee, affects job satisfac-
tion at least as much as the position, title, and salary commanded by an
employee.93 At its best, corporate culture captures the motivation tri-
fecta of autonomy, mastery, and purpose.94
Despite the usual supposition that corporate strategies largely cannot
be purchased or sold,95 differences in corporate engagement of employ-
ees do affect the valuation of firms. Indeed, this factor is so significant
that it may be possible to craft a trading strategy on the basis of corporate
policies on human resources and employee compensation.96 A study of
136 nonfinancial companies that issued initial public offerings in 1988,97
a highly diverse cohort ranging from biotechnology to food service
retailing and var[ying] widely in size,98 found dramatic differences in the
probability that these companies would survive at least five years after their
initial public offerings (IPOs):
[T]he difference in survival probability for firms one standard deviation

above and one standard deviation below the mean on the human resource
value scale was almost 20 percent. The difference in survival depending on
38 J.M. CHEN
where the firm scored on the rewards scale [i.e., nonmanagerial employees
access to stock options and profit-sharing] was even more dramatic, with a
difference in five-year survival probability of 42 percent between firms in the
upper and lower tails of the distribution.99
The results are so compelling that prudent investors may want to use
these results in evaluating new companies.100 The same wisdom may be
generalizable to ongoing enterprises: You get the distinct impression that
if youre trying to decide where to make an investment, the best place to
look is those annual lists of the 100 best places to work.101 Hierarchical
theories of needs thus connect the economics of human capital,102 by way
of the exit-voice-loyalty-neglect model of organizational behavior,103 to
mathematical and behavioral finance.
2.4 BEHAVIORAL ENVIRONMENTAL ECONOMICS

Maslowian portfolio theory modulates mental accounting in finance accord-
ing to the physiological, social, and emotional needs of individuals. It rep-
resents one of the most explicit applications of the behavioral sciences to
finance, a branch of economics that has historically preferred strictly formal
models assuming perfectly rational actors. But behavioral psychology is hardly
the only branch of the social sciences that contests conventional approaches
to economics and accounting. There are different ways of accounting, and
different actors will resort to some of these methods, even if the government
or some abstract ledger treats only one system of accounting as correct.104
Environmental economics levels an analogous criticism at the use of
conventional econometrics as a baseline for evaluating environmental
policy.105 Conventional measures of social welfare such as gross domestic
product (GDP) allegedly give little or no weight to ecosystem services.106
In place of GDP, or least alongside it, ecological economists have devised a
wide variety of measures intended to capture elements of human and eco-
logical welfare that carry no weight in standard national income and prod-
uct accounts.107 These measures include the genuine progress indicator108
and the human development index.109 The adoption of gross national hap-
piness by the Himalayan kingdom of Bhutan connects the quest for proper
economic measurement with human emotion, a link that environmental
economics seeks to complete.110
The realm of environmental protection and conservation, where beauty
and mystery seized us at the beginning and continues to inspire the
human intellect that masters [it] all,111 serves as an apt reminder that
neither finance nor any other branch of economics can be severed from
aesthetic judgment and human emotion.112 Homebuyers do not just see
a house; they see a handsome house, an ugly house, or a pretentious
house.113 The recognition that even the briefest introduction to a sub-
jectmere exposurealters emotional reactions to novel stimuli114 opens
the door to complete economic consideration of the affect heuristic.115
Instantaneous, automatic feelings associated with [such] stimulus words
[as] treasure or hate116 motivate entire branches of business and econom-
ics associated with advertising, marketing, and publicity.117 We should like-
wise expect the affect heuristic to influence evaluations of risk.
Emotions drive the price of the assets that are putatively invisible to
conventional pricing models. Art,118 collectibles,119 anything beautiful.120
Ye gods, real estate.121 No less than prices for wine,122 stock market prices
reflect investor sentiment as well as rational factors.123 What does come
at a price is relaxing the supposition, embedded in standard financial
theory, that affect plays no role in the pricing of financial assets.124
Human emotion and behavior do affect financial markets and portfolios,
and the corresponding recognition that affect plays a role in pricing
models of financial assets gives rise to the development of a behavioral
asset-pricing model.125
Vigilance against downside risk animates the most temperamentally
(if not politically) conservative principle in environmental law and safety
regulation. As a counterweight to conventional cost-benefit analysis, the
precautionary principle discourages risk-taking that may hurt the public at
large, or an especially vulnerable segment of it.126 The need to accumu-
late and safeguard wealth for immediate consumption, directed at survival
or safety, is likelier to consume a deeper portion of a poor familys total
wealth.127 This sensitivity to unforeseen, even unforeseeable, risk and to
wealth effects finds a welcome home in the normative toolkit of environ-
mental economics.
On the other hand, the risk of excessive social consumption in dis-
regard for environmental disruption and other long-term consequences
does appear to reach its apex during periods of nominal economic growth.
Environmental hazards that are viewed as familiar, commonplace,
everyday risks are often underestimated, especially by local residents
who rely (mistakenly) on the perceived collective judgment of others
around them who have seemingly concluded that their community is safe
enough.128 Too often, public responses to environmental risk fall under
40 J.M. CHEN
the sway of the gamblers fallacy.129 Consequently, governments often

overinvest in reconstruction in the wake of disaster based on a mistaken
assumption that a period of repose and relative safety will follow.130 At
the same time, many individuals ignore ecosystem services and other posi-
tive environmental externalitiesfor example, flourishing forests, healthy
wetlands, thriving honeybee populations, and a stable climateuntil the
loss of such societal benefits becomes the face of disaster.131
These mistakes in environmental judgment arise from what behavioral
economists call the availability heuristic.132 The salience of data, rather
than a scientifically sober account of its significance, ultimately drives
decision-making.133 The interconnected nature of human society quickly
compounds the power of narratives whose plausibility and vividness
exceed their validity into availability cascades that overwhelm proper,
critical evaluation.134 Availability cascades have created an entire disas-
ter mythology in which catastrophic events prompt looting, violence,
and general lawlessness.135 Availability cascades bedevil other domains of
public health and environmental policy, such as acceptance of the safety of
vaccines136 and of the science demonstrating the anthropogenic nature of
climate change.137
Informational cascades are even more devastating when they validate
ideas that the audience is already predisposed to favor. Behavioral econom-
ics calls this problem confirmation bias.138 As if these problems were not
demoralizing enough, increased levels of literacy, numeracy, and scientific
sophistication do not change minds or hearts about pressing issues of risk
management. Instead, they merely entrench all parties even more deeply
in their predispositions and biases.139 Cultural cognition theory,140 though
by no means immune to criticism,141 suggests that the public will turn its
attention to deep threats such as climate change only when legal and polit-
ical actors learn how to communicate in ways that resonate with deeply
held societal values.142 Active vigilance supplies the best safeguard against
erect[ing] our prejudices into legal principles; in exercis[ing] high
power entrusted to public officials, we must let our minds be bold.143
2.5 FABLES OFTHERECONSTRUCTION

This narrative bridge between mathematical finance and environmental
economics evokes the fable of the ant and the grasshopper. Both versions
of this morally ambiguous fable inform the connection between these
branches of economics. Aesops more traditional version of the tale144
suggests that we should treat behavioral departures from strict, techno-

cratic rationality as the moral equivalent of the grasshoppers behavior,
singing throughout the summer rather than gathering food. A parallel
source of ancient wisdom counsels, Go to the ant, O sluggard; consider
her ways, and be wise. Without having any chief, officer or ruler, she
prepares her food in summer, and gathers her sustenance in harvest.145
Approaching the fable of the ant and the grasshopper as a didactic allegory
suggests that failures in disaster law, public health, and climate change
policy demonstrate human disregard of the world whose unfathomable
complexity [and] sublime beauty gave rise to the human thirst for new
ideas in the first place.146
But there is a different interpretation of this fable, one that counsels a
more cautious, circumspect respect for the power that behavioral heuris-
tics and cognitive biases exert over naked rationality. The French version
of the fable by Jean de La Fontaine, La Cigale et la Fourmi,147 is cel-
ebrated for its moral ambiguity and its veiled critique of the fabulists own
financial imprudence.148 La Fontaine cast the fables insects as parties to a
failed lending transaction. Said the hungry cicada (never a grasshopper in
la version franaise): On insects honor, Ill repay you/Well before fall.
With interest, too!149 Mais non: Our antno willing lender she! Least
of her faults!150 When the ant at last tells her starving neighbor to dance
through the winter, it is far from clear which insect has the moral upper
footor the stronger grip on the human readerships collective thorax.
And morality is perhaps the most deeply emotional, least mechanistically
rational projection of the human mind at work.151
Among the branches of economics, environmental economics may har-
bor the richest trove of departures from strict rationality. The valuation of
environmental benefits, from individual specimens to entire populations
and ecosystem services, is beset by disagreements over methodology and
validity. Environmental decision-making rarely offers the numerical clarity
of financial problems. For instance, Mark Sagoff defends legal intervention
to save endangered species on strictly [m]oral, aesthetic, and spiritual
grounds, finding an instrumental or economic rationale to lie beyond
reach.152 But it demands almost deliberate disregard of economic real-
ity to insist on valuing polar bears, for instance, strictly on the basis of
their value for sport hunting and subsistence. Treating Ursa maritimus
as so much Arctic bushmeat sets the value of Canadas bear population at
$600,000, far below the estimated $6 billion in indirect and passive uses,
including bequest and existence value.153 In the United States, the polar
42 J.M. CHEN
bears greatest value may lie in its role in providing a basis for applying the
Endangered Species Act to problems of climate change that Congress has
persistently ignored.154
The incorporation of behavioral psychology into environmental eco-
nomics is, if nothing else, the story of an intellectual discipline that has
come to embrace the richness of analytical tools transcending austerely
beautiful but excessively rigid mathematical models. The long reaches of
the peaks of song, whether delivered by La Fontaines cicada or a human
master of music as mathematics made flesh, rebuff the formic formalist
of the French fable.155 If indeed the ant symbolizes the entomological
equivalent of poet Edwin Markhams [s]lave of the wheel of labor, a
nuanced approach to environmental as well as financial economics may
rightfully ask, what to [her]/Are Plato and the swing of Pleiades?156
Understanding the impact of investor behavior on the performance of an
individual portfolio or perhaps even the financial marketplace as a whole
indeed begins with the rift of dawn, the reddening of the rose.157
NOTES
1. See, e.g., Alenco Communications, Inc. v. FCC, 201 F.3d 608, 615 (5th
Cir. 2000) (recognizing that historical investments are sunk costs
and have no relevance to current costs or business decisions based on
them); Armen A.Alchian & William R.Allen, Exchange and Production
222 (3d ed. 1983) (observing that the cost of an item, onceacquired,
is irrelevant to any future decision); James D. Gwartney & Richard
L.Stroup, Economics 417419 (4th ed. 1982) (If they are to minimize
costs, business decision-makers must recognize the irrelevance of sunk
costs.); N.Gregory Mankiw, Principles of Economics 291 (1997) (The
irrelevance of sunk costs explains how real businesses make decisions.);
Robert S.Pindyck & Daniel L.Rubenfeld, Microeconomics 7.1, at 199
(2d ed. 1992) (arguing that sunk costs, though usually visible, should
always be ignored when making economic decisions); Richard A.Posner,
Economic Analysis of Law 1.1, at 7 (3d ed. 1986) ([C]ost to an econ-
omist is a forward-looking concept; costs already incurred do not
affect decisions on price and quantity.).
2. See Rudolf Conradi & August Conradi, Die Familie Padde 26 (1858)
(available at https://books.google.com/books?id=WNGVZMZethcC):
Whats past is past; we will speak no more of it.
3. Paul A.Samuelson & William D.Nordhaus, Economics 227 (19th ed.
2010).
4. See generally, e.g., Ziv Carmon & Dan Ariely, Focusing on the Forgone:
How Value Can Appear So Different to Buyers and Sellers, 27J.Consumer
Research 360370 (2000); Herbert J.Hovencamp, Legal Policy and the
Endowment Effect, 20J.Leg. Stud. 225247 (1991); Daniel Kahneman,
Jack L.Knetsch & Richard H.Thaler, Experimental Tests of the Endowment
Effect and the Coase Theorem, 98J.Pol. Econ. 13251348 (1990); Jack
L. Knetsch, The Endowment Effect and Evidence of Nonreversible
Indifference Curves, 79 Am. Econ. Rev. 12771284 (1989); Nathan
Novemsky & Daniel Kahneman, The Boundaries of Loss Aversion,
42J.Marketing Research 119128 (2005).
5. See Andrew Healy & Neil Malhotra, Myopic Voters and Natural Disaster
Policy, 103 Am. Pol. Sci. Rev. 387406, 396 (2009) (estimating the
total benefit of a dollar of preparedness spending as all future reduc-
tions in damage, while discounting those benefitsfor the fact that
resources invested today in other ways could have yielded their own
return and that preparedness investments will depreciate); cf. M.Ishaq
Nadiri & Ingmar Prucha, Estimation of Depreciation Rate of Physical and
R&D Capital in the U.S.Total Manufacturing Sector, 34 Econ. Inquiry
4356 (1996) (estimating that physical capital in American manufactur-
ing depreciates 5.9% per year). Combining Nadiri and Pruchas 5.9%
depreciation rate with their own estimate of a 4% annual interest rate,
Healy and Malhotra estimate the [net present value] of $1 of disaster
preparedness to be about $15.
6. See Ben Depoorter, Horizontal Political Externalities: The Supply and
Demand of Disaster Management, 56 Duke L.J. 101125, 103 (2006);
Howard Kunreuther, Mitigating Disaster Losses Through Insurance,
12J.Risk & Uncertainty 171187, 177 (1996).
7. See Charles Cohen & Eric Werker, The Political Economy of Natural
Disasters, 52J.Conflict Resolution 795819 (2008).
8. Jean-Pierre Benot & Juan Dubra, On the Problem of Prevention, 54 Intl
Econ. Rev. 787805, 787 (2013).
9. Richard A. Epstein, Catastrophic Responses to Catastrophic Risks,
12 J. Risk & Uncertainty 287308 (1996). See generally James Ming
Chen, Correlation, Coverage, and Catastrophe: The Contours of Financial
Preparedness for Disaster, 26 Fordham Envtl. L. Rev. 5694, 7985
(2014).
10. New State Ice Co. v. Liebmann, 285 U.S. 262, 311 (1932) (Brandeis, J.,
dissenting).
11. See generally Nicholas Barberis & Ming Huang, Mental Accounting, Loss
Aversion, and Individual Stock Returns, 56J.Fin. 12471292 (2001).
12. See Eric J. Allen, Patricia M. Dechow, Devin G. Pope & George Wu,
Reference-Dependent Preferences: Evidence from Marathon Runners
2329, 26 (July 18, 2014) (figure 9(b)) (available at http://faculty.chi-
cagobooth.edu/devin.pope/research/pdf/Website_Marathons.pdf).
44 J.M. CHEN
A comparison of the normalized pace of marathon runners final 2.195

kilometers relative to their pace through the first 40 kilometers as a func-
tion of their finishing times shows striking spikes in speed at 30-minute
increments between two hours, 40 minutes and five hours, 20 minutes.
This is powerful evidence of these runners desire to attain a round-num-
ber benchmark such as three, four, or five hours, or even half-hour inter-
vals in between. On the behavioral psychology of sports, see generally
L.John Wertheim & Sam Sommers, This Is Your Brain on Sports: The
Science of Underdogs, the Value of Rivalries, and What We Can Learn
from the T-Shirt Cannon (2016).
13. See Amos Tversky & Daniel Kahneman, Judgment Under Uncertainty:
Heuristics and Biases, 185 Science 11241131 (1974).
14. This use of unicorn appears to have originated in Aileen Lee, Welcome to
the Unicorn Club: Learning from Billion-Dollar Startups, TechCrunch
(Nov. 2, 2013) (available at http://techcrunch.com/2013/11/02/
welcome-to-the-unicorn-club) (describing U.S.-based technology com-
panies started since January 2003 andvalued at $1 billion by private or
public markets). See generally, e.g., Robert P.Bartlett III, A Founders
Guide to Unicorn Creation: How Liquidation Preferences in M&A
Transactions Affect Start-Up Valuation, in Research Handbook on
Mergers and Acquisitions (Steven Davidoff Solomon & Claire Hill eds.,
forthcoming 2016) (available at http://ssrn.com/abstract=2664236);
Keith C.Brown & Kenneth W.Wiles, In Search of Unicorns: Private IPOs
and the Changing Markets for Private Equity Investors and Corporate
Control, 27:3J.Applied Corp. Fin. 3448 (Summer 2015).
15. See Ralph Hertwig, Greg Barror, Elke U.Weber & Ido Erev, Decisions
from Experience and the Effect of Rare Events on Risky Choice, 15 Psych.
Sci. 534539 (2004); Richard H.Thaler & Eric J.Johnson, Gambling
with the House Money and Trying to Break Even: The Effects of Prior
Outcomes on Risky Choice, 36 Mgmt. Sci. 643660 (1990); cf. Kevin
Keasey & Philip Moon, Gambling with the House Money in Capital
Expenditure Decisions, 50 Econ. Letters 105110 (1996) (concluding
that prior gains prompt risk-seeking, but finding no evidence that prior
losses shift behavior toward risk aversion).
16. See, e.g., James J.Choi, David Laibson, Brigitte C.Madrian & Andrew
Metrick, Reinforcement Learning and Savings Behavior, 64J.Fin. 2515
2534 (2009); Ulrike Malmendier & Stefan Nagel, Depression Babies: Do
Macroeconomic Experiences Affect Risk Taking?, 126 Q.J.Econ. 373416
(2011); Michal Ann Strahilevitz, Terrance Odean & Brad M. Barber,
Once Burned, Twice Shy: How Nave Learning, Counterfactuals, and
Regret Affect the Repurchase of Stocks Previously Sold, 48 J. Marketing
Research 102120 (2011); cf. Farah Said, Uzma Afzal & Ginger Turner,
Risk Taking and Risk Learning After a Rare Event: Evidence from a Field
Experiment in Pakistan, 118J.Econ. Behav. & Org. 167183 (2015)
(documenting levels of risk aversion as a function of the number of floods
and the severity of losses experienced by rural Punjabis).
17. See generally Gina Nicolosi, Liang Peng & Ning Zhu, Do Individual
Investors Learn from Their Trading Experience?, 12J.Fin. Markets 317
336 (2009).
18. See Markus Glaser & Martin Weber, Why Inexperienced Investors Do Not
Learn: They Do Not Know Their Past Portfolio Performance, 4 Fin.
Research Letters 203216 (2007); cf. Amit Seru, Tyler Shumway &
Noah Stoffman, Learning by Trading, 23 Rev. Fin. Stud. 705739
(2010) (cautioning that a significant number of traders stop trading upon
discovering their lack of aptitude and that studies purporting to find posi-
tive learning effects without accounting for this attrition reflects survivor-
ship bias).
19. See generally, e.g., Richard H.Thaler, Mental Accounting and Consumer
Choice, 3 Marketing Sci. 199214 (1985); Richard H.Thaler, Toward a
Positive Theory of Consumer Choice, 1 J. Econ. Behav. & Org. 3960
(1980).
20. Shlomo Benartzi & Richard H. Thaler, Myopic Loss Aversion and the
Equity Premium Puzzle, 110 Q.J.Econ. 7392, 74 (1995).
21. See Amos Tversky & Daniel Kahneman, Rational Choice and the Framing
of Decisions, 59 J. Bus. S251-S278 (1986); Amos Tversky & Daniel
Kahneman, The Framing of Decisions and the Psychology of Choice, 211
Science 453481 (1981).
22. See Amos Tversky, Elimination by Aspects: A Theory of Choice, 79 Psych.
Rev. 281299 (1972).
23. Statman, Fisher & Anginer, supra note 122 (Chapter 1), at 21. See gener-
ally Baba Shiv & Alexander Fedorikhin, Heart and Mind in Conflict: The
Interplay of Affect and Cognition in Consumer Decision Making,
26J.Consumer Research 278292 (1999).
24. See Haim Levy & Amnon Rapoport, Experimental Tests of the Separation
Theorem and the Capital Asset Pricing Model, 7 Am. Econ. Rev. 500518
(1988); Hersh Shefrin & Meir Statman, Behavioral Portfolio Theory,
35J.Fin. & Quant. Analysis 127151, 142 (2000).
25. See Philippe Jorion, Mean/Variance Analysis of Currency Overlays, 50:3
Fin. Analysts J. 4856 (May/June 1994).
26. See id. at 49, 52.
27. See Victor DeMiguel, Lorenzo Garlappi & Raman Uppal, Optimal Versus
Nave Diversification: How Inefficient Is the 1/N Portfolio Design?, 22
Rev. Fin. Stud. 19151953 (2009). See generally Shlomo Benartzi &
46 J.M. CHEN
Richard H. Thaler, Nave Diversification Strategies in Defined

Contribution Plans, 91 Am. Econ. Rev. 7998 (2001).
28. See Samuel M.Hartzmark, The Worst, the Best, Ignoring All the Rest: The
Rank Effect and Trading Behavior, 28 Rev. Fin. Stud. 10241059
(2015).
29. See, e.g., Cass R. Sunstein & Richard Zeckhauser, Overreaction to
Fearsome Risks, 48 Envtl. & Resource Econ. 435449 (2011).
30. See generally Sitabhra Sinha, Arnab Chatterjee, Anirban Chakraborti &
Bikas K.Chakrabarti, Econophysics: An Introduction (2011). The inter-
action between finance and physics has come a long way since Louis Jean-
Baptiste Bacheliers pathbreaking books, such as Thorie de la Spculation
(1900), Calcul des Probabilits (1912), and Le Jeu, la Chance, et le
Hasard (1914). See James Owen Weatherall, The Physics of Wall Street:
A Brief History of Predicting the Unpredictable 1011 (2013) (report-
ing that Bacheliers thesis at La Sorbonne was poorly received because he
was trying to apply mathematics to a field with which mathematicians of
his time were unfamiliar).
31. See Shefrin & Statman, Behavioral Capital Asset Pricing Theory, supra
note 79 (Chapter 1), at 323.
32. Cf. supra 1.5, text accompanying notes 7984 (Chapter 1) (distin-
guishing between conventionally rational and behaviorally sensitive
accounts of finance).
33. See Phillipe J.S.De Brouwer, Behavioural Finance and Decision Making
in Financial Markets, in Principles of Modelling, Forecasting, and
Decision-Making 2444 (Wladislaw Milo & Piotr Wdowinski eds.,
2006); Phillipe J.S. De Brouwer, Maslowian Portfolio Theory: An
Alternative Formulation of the Behavioral Portfolio Theory, 9 J. Asset
Mgmt. 359365 (2009). De Brouwer has consolidated his views on the
subject in Philippe J.S. De Brouwer, Maslowian Portfolio Theory: A
Coherent Approach to Strategic Asset allocation (2012). Subsequent
citations to De Brouwer, Maslowian Portfolio Theory will refer to the
2009 article in the Journal of Asset Management.
34. See generally Abraham H.Maslow, A Theory of Human Motivation, 50
Psych. Rev. 370396 (1943).
35. See id. at 372376 (physiological needs, based on homeostasis and
appetites); id. at 376380 (safety).
36. See id. at 380381 (love); id. at 381382 (esteem).
37. De Brouwer, Maslowian Portfolio Theory, supra note 33, at 360.
38. Id. at 361; cf. Walther, supra note 144 (Chapter 1) (distinguishing
between utility from wealth and utility from emotional reactions to the
resolution of uncertainty over wealth).
39. See Charles M.Harvey, Prescriptive Models of Psychological Effects on Risk
Attitudes, 19 Annals Oper. Research 141170 (1989).
40. See infra 5 1., text accompanying notes 1418 (Chapter 5).
41. See De Brouwer, Maslowian Portfolio Theory, supra note 33, at 361.
42. Sanjiv Das, Harry Markowitz, Jonathan Scheid & Meir Statman, Portfolio
Optimization with Mental Accounts, 45J.Fin. & Quant. Analysis 311
334, 313 (2010).
43. Id. at 312. See generally Wilbur G.Lewellen, Ronald C.Lease & Gary
G. Schlarbaum, Portfolio Design and Portfolio Performance: The
Individual Investor, 32J.Econ. & Bus. 185197 (1980).
44. Das, Markowitz, Scheid & Statman, supra note 42, at 312.
45. De Brouwer, Maslowian Portfolio Theory, supra note 33, at 361; see also
Das, Markowitz, Scheid & Statman, supra note 42, at 313.
47. Philippe J.S.De Brouwer, Target-Oriented Investment Advice, 13J.Asset
Mgmt. 102114, 104 (2012).
48. Id. at 105.
49. Id. at 104.
51. Id.
52. Id. at 313.
53. Id. at 315; see also id. at 318.
54. Peter Brooks, Greg B. Davies & Robert E.D. Smith, A Behavioral
Perspective on Goal-Based Investing, Invs. & Wealth Monitor, Nov./Dec.
2015, at 1618 & 37, 16.
55. See Lola L.Lopes & Gregg C.Oden, The Role of Aspiration Level in Risky
Choice: A Comparison of Cumulative Prospect Theory and SP/A Theory,
43J.Math. Psych. 286313, 307 (1999).
56. See Das, Markowitz, Scheid & Statman, supra note 42, at 315 n.4; cf.
Robert B. Barsky, F. Thomas Juster, Miles S. Kimball & Matthew
D. Shapiro, Preference Parameters and Behavioral Heterogeneity: An
Experimental Approach in the Health and Retirement Study, 112
Q.J. Econ. 537579, 545 (1997) (defining risk tolerance in terms of
choices over risky behaviorsthe decisions to smoke and drink, to buy
insurance, to immigrate, to be self-employed, and to hold stock); Miles
S. Kimball, Claudia R. Sahm & Matthew D. Shapiro, Imputing Risk
Tolerance from Survey Responses, 103 J. Am. Stat. Assn 10281038,
1029 (2008) (observing that the subjective wording of measures for
risk tolerance may generate uninterpretable variation).
57. De Brouwer, Target-Oriented Investment Advice, supra note 47, at 112.
58. Id. at 103.
59. Douglas T.Kenrick, Vladas Griskevicius & Mark Schaller, Renovating the
Pyramid of Needs: Contemporary Extensions Built upon Ancient
Foundations, 5 Persp. on Psych. Sci. 292314, 292 (2010).
48 J.M. CHEN
60. See id. at 293.

61. Id. at 294.
62. A.H. Maslow, Motivation and Personality 21 (2d ed. 1970); accord
Kenrick, Griskevicius & Schaller, supra note 59, at 294.
63. Kenrick, Griskevicius & Schaller, supra note 59, at 296.
64. Id.
65. Id. See generally Martie G. Haselton & Daniel Nettle, The Paranoid
Optimist: An Integrative Evolutionary Model of Cognitive Bias, 10
Personality & Soc. Psych. Rev. 4766 (2006); Randolph M. Nesse,
Natural Selection and the Regulation of Defenses: A Signal Detection
Analysis of the Smoke Detector Principle, 26 Evol. & Human Behav.
88105 (2005); Paul Rozin & Edward B. Royzman, Negativity Bias,
Negativity Dominance, and Contagion, 5 Personality & Soc. Psych. Rev.
296320 (2001).
66. Emma Roller, After Paris, Who Passes the Commander-in-Chief Test?,
N.Y. Times, Nov. 17, 2015 (available at http://nyti.ms/1YeKhtZ)
(attributing this sort of situational risk aversion to Abraham Maslows
hierarchy of needs).
67. See Cameron Anderson & Adam D. Galinsky, Power, Optimism, and
Risk-Taking, 36 Eur. J.Soc. Psych. 511536 (2006).
68. See Jon K. Maner, Matthew T. Gaillliot, David A. Butz & B. Michelle
Peruche, Power, Risk, and the Status Quo: Does Power Promote Riskier or
More Conservative Decision Making?, 33 Personality & Soc. Psych. Bull.
451462 (2007).
69. See Eva Jonas, Daniel Sullivan & Jeff Greenberg, Generosity, Greed,
Norms, and DeathDifferential Effects of Mortality Salience on
Charitable Behavior, 35J.Econ. Psych. 4757 (2013).
70. Maslow, Theory of Human Motivation, supra note 34, at 382; accord
Kenrick, Griskevicius & Schaller, supra note 59, at 297.
71. Cf. Frank Luntz, Words That Work: Its Not What You Say, Its What
People Hear 119 (2007) (Why did the U.S.Army jettison be all that
you can be, surely one of the most widely known taglines in the world,
for the rather odd and uninspiring An army of one?).
73. Id. at 298; see also Robert Kurzban & C.Athena Aktipis, Modularity and
the Social Mind: Are Psychologists Too Selfish?, 11 Personality & Soc. Psych.
Rev. 131149 (2007).
75. See Abraham H.Maslow, Critique of Self-Actualization Theory, in Future
Visions: The Unpublished Papers of Abraham Maslow 2632 (Edward
Hoffman ed., 1996); Abraham H. Maslow, The Farther Reaches of
Human Nature, 1J.Transpersonal Psych. 19 (1969).
76. See Maberly & Pierce, supra note 37 (Chapter 1), at 30.
77. These titles come from Harvard Business Review alone, including the
journals related press: Teresa Amabile & Steve Kramer, What Makes
Work Worth Doing?, Harv. Bus. Rev. Online (Aug. 31, 2012) (available at
https://hbr.org/2012/08/what-makes-work-worth-doing); Teresa
Amabile & Steven J.Kramer, Inner Work Life: Understanding the Subtext
of Business Performance, Harv. Bus. Rev. (May 2007) (available at
https://hbr.org/2007/05/inner-work-life-understanding-the-subtext-
of-business-performance); Adam Grant, How Customers Can Rally Your
Troops, Harv. Bus. Rev. (June 2011) (available at https://hbr.
org/2011/06/how-customers-can-rally-your-troops). See generally
Teresa M.Amabile & Steven J.Kramer, The Progress Principle: Using
Small Wins to Ignite Joy, Engagement, and Creativity at Work (2011).
78. See Amy Wrzesniewski & Jane E. Dutton, Crafting a Job: Revisioning
Employees as Active Crafters of Their Work, 26 Acad. Mgmt. Rev. 179
201 (2001).
79. The seminal source arguing no to this proposition is Milton Friedman,
Capitalism and Freedom: Fortieth Anniversary Edition 119136 (2002;
1st ed. 1962).
80. See generally, e.g., C.B. Bhattacharya, Sankar San & Daniel Korschun,
Leveraging Corporate Social Responsibility: The Stakeholder Route to
Business and Social Value (2011).
81. Alexander Kempf & Peer Osthoff, The Effect of Socially Responsible
Investing on Portfolio Performance, 13 Eur. Fin. Mgmt. 908922 (2007);
see also Bryan W. Hustead & Jos de Jesus Salazar, Taking Friedman
Seriously: Maximizing Profits and Social Performance, 43J.Mgmt. Stud.
7591 (2006).
82. See Rob Bauer, Kees Koedijk & Rogr Otten, International Evidence on
Ethical Mutual Fund Performance and Investment Style, 29J.Banking &
Fin. 17511767 (2006).
83. See Luc Reeneborg, Jenke Ter Horst & Chendi Zhang, The Price of Ethics
and Stakeholder Governance: The Performance of Socially Responsible
Mutual Funds, 14 J. Corp. Fin. 302322 (2008). On the four-factor
FamaFrenchCarhart model, see infra 12.2.; Chen, Postmodern
Portfolio Theory, supra note 1 (Chapter 1), 3.3.
84. See Raphie Hayat & Roman Kraeussl, Risk and Return Characteristics of
Islamic Equity Funds, 12 Emerging Mkts. Rev. 189203 (2011) (finding
that Islamic equity funds underperform both Islamic and conventional
equity benchmarks); Saida Daly, Sonia Ghorbel-Zouari & Mohamed
Frikha, Islamic Financial Stability During the US Sub-Prime Crisis: Using
from Data Panel, 1J.Behav. Econ., Fin., Entrepreneurship, Accounting
& Transp. 112 (2013) (finding that adherence to Islamic banking prin-
50 J.M. CHEN
ciples had a mixed impact on bank stability, relative to conventional bank-

ing rules).
85. See Anastasia ORourke, The Message and Methods of Ethical Investing,
11J.Cleaner Production 683693 (2003).
86. See, e.g., Homayoon Shalchian, Kais Bourlah & Bouchra MZali, A
Multi-Dimensional Analysis of Corporate Social Responsibility: Different
Signals in Different Industries, 4 J. Fin. Risk Mgmt. 92109 (2015)
(finding that investors are sensitive to activities affecting the environ-
ment among mining and manufacturing companies and to employee
welfare among service companies).
87. See John Francis T.Diaz, Return and Volatility Performance Comparison
of Ethical and Non-Ethical Publicly-Listed Financial Service Companies,
13 thique et conomique 1224, 1820 (2016).
88. See id. at 21.
89. See Indrani De & Michelle Clayman, The Benefits of Socially Responsible
Investing: An Active Managers Perspective (July 9, 2014) (forthcoming
J.Investing) (available at http://ssrn.com/abstract=2464204).
90. See id.
91. See Chen, Postmodern Portfolio Theory, supra note 1 (Chapter 1), 2.6.,
at 1213.
92. Kathleen M.Eisenhardt & Claudia Bird Schoonhoven, Organizational
Growth: Linking Founding Team Strategy, Environment, and Growth
Among U.S.Semiconductor Ventures, 19781988, 35 Admin. Sci. Q. 504
529 (1990); accord Michael Beer, High Commitment High Performance:
How to Build a Resilient Organization for Sustained Advantage 101
(2009).
93. See Amy L. Kristof-Brown, Ryan D. Zimmerman & Erin C. Johnson,
Consequences of Individuals Fit at Work: A Meta-Analysis of Person-Job,
Person-Organization, Person-Group, and Person-Supervisor Fit, 58
Personnel Psych. 281342 (2005).
94. See generally Daniel H.Pink, Drive: The Surprising Truth About What
Motivates Us (2010).
95. Philip Bromiley, Kent D. Miller & Devaki Rau, Risk in Strategic
Management Research, in The Blackwell Handbook of Strategic
Management 259298, 259 (Michael A.Hitt, R.Wdward Freeman &
Jeffrey S.Harrison eds., 2006).
96. Cf. Shawn M.Riley, Market Valuation of Firm Investments in Training
and Human Capital Management (2011) (Ph.D. dissertation, University
of Illinois at Urbana-Champaign) (available at http://hdl.handle.
net/2142/29535) (conducting an event study of the economic impact
of information on firms investments in human capital and on firms man-
agerial policies).
97. See Theresa Welbourne & Alice Andrews, Predicting Performance of

Initial Public Offering Firms: Should HRM Be in the Equation?, 39 Acad.
Mgmt. J. 891919, 910911 (1996) (reporting five-year survival rates of
companies after their IPOs, based on differences in human resource man-
agement and employee reward systems).
98. Jeffrey Pfeffer, The Human Equation: Building Profits by Putting People
First 36 (1998).
99. Id. at 37.
100. Id. at 36.
101. Barry Schwartz, Rethinking Work, N.Y.Times, Aug. 28, 2015 (available
at http://nyti.ms/1KegWwu); see also id. (When employees have work
that they want to do, they are happier. And when they are happier, their
work is better, as is the companys bottom line.).
102. See generally, e.g., Gary S. Becker, Human Capital: A Theoretical and
Empirical Analysis, with Special Reference to Education (3d ed. 1993).
103. See generally, e.g., Albert O. Hirschman, Exit, Voice, and Loyalty:
Responses to Decline in Firms, Organizations, and States (1970); William
T.Turnley & Daniel C.Feldman, The Impact of Psychological Contract
Violation on Exit, Voice, Loyalty, and Neglect, 52 Human Relations 895
922 (1999); Michael J.Withey & William H.Cooper, Predicting Exit,
Voice, Loyalty, and Neglect, 34 Admin. Sci. Q. 521539 (1989).
104. For a vivid illustration of the impact of different accounting rules, see
United States v. Hill, 506 U.S. 546, 550551 (1993), which distin-
guished between cost depletion and percentage depletion methods
of accounting for wealth extracted in the form of minerals or other wast-
ing assets under I.R.C. 611, 613; Treas. Reg. 1.611-1 to 1.613-1.
Compare Paragon Jewel Coal Co. v. Commissioner, 380 U.S 624, 631
(1965) (contemplating a reasonable allowance for depletion of miner-
als exhausted in production) with Hill, 506 U.S at 551 (observing that
percentage depletion generously allows the taxpayer extracting miner-
alsto deduct a specified percentage of his gross income, even when his
prior depletion deductions have exceeded his investment).
105. See, e.g., Manfred Max-Neef, Economic Growth and Quality of Life: A
Threshold Hypothesis, 15 Ecol. Econ. 115118 (1995).
106. See, e.g., Robert Costanza etal., The Value of the Worlds Ecosystem Services
and Natural Capital, 387 Nature 253260 (1997); Karl-Gran Mler,
Sara Aniyar & sa Jansson, Accounting for Ecosystem Services as a Way to
Understand the Requirements for Sustainable Development, 105 PNAS
95019506 (2008); Sheng Zhao, Huasheng Hong & Luoping Zhang,
Linking the Concept of Ecological Footprint and Valuation of Ecosystem
Services: A Case Study of Economic Growth and Natural Carrying
Capacity, 15 Intl J.Sustainable Dev. & World Ecol. 448456 (2008).
52 J.M. CHEN
107. See, e.g., Ida Kubiszewski, Robert Costanza, Carol Franco, Philip Lawn,
John Talberth, Tim Jackson & Camille Aylmer, Beyond GDP: Measuring
and Achieving Global Genuine Progress, 93 Ecol. Econ. 5768 (2013).
108. See, e.g., Philip A.Lawn, A Theoretical Foundation to Support the Index of
Sustainable Economic Welfare (ISEW), Genuine Progress Indicator (GPI),
and Other Related Indexes, 44 Ecol. Econ. 105118 (2003).
109. See, e.g., Farhad Noorbaksh, A Modified Human Development Index, 26
World Dev. 517528 (1998); Ambuj D.Sagar & Adil Najam, The Human
Development Index: A Critical Review, 25 Ecol. Econ. 249264 (1998).
110. See Stefan Preisner, Gross National Happiness: Bhutans Vision of
Development and Its Challenges, in Indigeneity and Universality in Social
Science: A South Asian Response 212232 (Partha Nath Mukherji &
Chandan Sengupta eds., 2004). See generally, e.g., Philip Layard,
Happiness: Lessons from a New Science (2005); Economics and
Happiness: Framing the Analysis (Luigino Bruni & Pier Luigi Porta eds.,
2006).
111. Edward O.Wilson, Consilience: The Unity of Knowledge 237 (1998).
With respect to the notion that humans feel an innate emotional connec-
tion to other species, see generally Edward O. Wilson, Biophilia: The
Human Bond with Other Species (1984); Ursula Goodenough, The
Sacred Depths of Nature (1998).
112. See, e.g., Jia Wei Zhang, Ryan T.Howell & Ravi Iyer, Engagement with
Natural Beauty Moderates the Positive Relation Between Connectedness
with Nature and Psychological Well-Being, 38 J. Envtl. Psych. 5563
(2014).
113. Robert B.Zajonc, Feeling and Thinking: Preferences Need No Inferences,
35 Am. Psychologist 151175, 154 (1980); accord Statman, Fisher &
Anginer, supra note 122 (Chapter 1), at 20.
114. See Robert B.Zajonc, Mere Exposure: A Gateway to the Subliminal, 10
Current Directions in Psych. Sci. 224228 (2001).
115. See generally, e.g., Melissa L. Finucane, Ali Alhakami, Paul Slovic &
Stephen M. Johnson, The Affect Heuristic in Judgments of Risks and
Benefits, 13J.Behav. Decision Making 117 (2000); Paul Slovic, Ellen
Peters, Melissa L.Finucane & Donald G.MacGregor, Affect, Risk, and
Decision Making, 24 Health Psych. S35S40 (2005); Meir Statman,
Kenneth L.Fisher & Deniz Anginer, Affect in a Behavioral Asset-Pricing
Model, 64:2 Fin. Analysts J. 2029 (March/April 2008); R.B. Zajonc,
Feeling and Thinking: Preferences Need No Inferences, 35 Am. Psychologist
151175 (1980). Paul Slovic, Melissa Finucane, Ellen Peters & Donald
G. Macgregor, The Affect Heuristic, in Heuristics and Biases: The
Psychology of Intuitive Judgment 397420 (Thomas Gilovich, Dale
Griffin & Daniel Kahneman eds., 2002).
116. Paul Slovic, Melissa Finucane, Ellen Peters & Donald G.Macgregor, The
Affect Heuristic, in Heuristics and Biases: The Psychology of Intuitive
Judgment 397420, 397 (Thomas Gilovich, Dale Griffin & Daniel
Kahneman eds., 2002).
117. See, e.g., Baba Shiv & Alexander Fedorikhin, Heart and Mind in Conflict:
The Interplay of Affect and Cognition in Consumer Decision Making,
26J.Consumer Research 278292 (1999).
118. See, e.g., Graldine David, Kim Oosterlinck & Ariane Szafarz, Art Market
Inefficiency, 121 Econ. Letters 2325 (2013); Pter Erdos & Mihly
Ormos, Random Walk Theory and the Weak-Form Efficiency of the US Art
Auction Prices, 34J.Banking & Fin. 10621076 (2010).
119. See, e.g., Elroy Dimson & Christophe Spaenjers, Ex Post: The Investment
Performance of Collectible Stamps, 110J.Fin. Econ. 443458 (2011).
120. See, e.g., Luc Renneboog & Christophe Spaenjers, Buying Beauty: On
Prices and Returns in the Art Market, 59 Mgmt. Sci. 3653 (2013).
121. See Edward L. Glaeser, A Nation of Gamblers: Real Estate Speculation
and American History (Feb. 2013) (NBER Working Paper No. 18825)
(available at http://www.nber.org/papers/w18825).
122. See Orley Ashenfelter, How Auctions Work for Wine and Art, 3J.Econ.
Perspectives 2326 (1989).
123. See Malcolm Baker & Jeffrey Wurgler, Investor Sentiment and the Cross-
Section of Stock Returns, 61J.Fin. 16451680 (2006); Malcolm Baker &
Jeffrey Wurgler, Investor Sentiment in the Stock Market, 21 J. Econ.
Perspectives 129151 (2007); Gregory W. Brown & Michael T. Cliff,
Investor Sentiment and Asset Valuation, 78J.Bus. 405440 (2005).
124. Statman, Fisher & Anginer, supra note 113, at 20.
125. Id.
126. See, e.g., U.N. Conference on Environment and Development, Rio de
Janeiro, June 314, 1992, Rio Declaration on Environment and
Development, U.N.Doc. A/CONF. 151/26 (vol I), annex 1, principle
15 (Aug. 12, 1992) (Where there are threats of serious or irreversible
damage, lack of full scientific certainty shall not be used as a reason for
postponing cost-effective measures to prevent environmental degrada-
tion.); John Applegate, The Precautionary Preference: An American
Perspective on the Precautionary Principle, 6 Hum. & Ecol. Risk Assessment
413 (2000). For efforts to reconcile the precautionary principle with cost-
benefit analysis, see Daniel H.Cole, Reconciling Cost-Benefit Analysis with
the Precautionary Principle (March 5, 2012) (available online at https://
www.law.upenn.edu/blogs/regblog/2012/03/reconciling-cost-
benefit- analysis-with-the-precautionar y-principle.html ); Douglas
A. Kysar, It Might Have Been: Risk, Precaution and Opportunity Costs,
22J.Land Use & Envtl. L. 157 (2006).
54 J.M. CHEN
127. Compare Jrgen Haug, Thorsten Hens & Peter Woehrmann, Risk
Aversion in the Large and in the Small, 118 Econ. Letters 310313
(2013) with Miles S.Kimball, Precautionary Saving in the Small and in
the Large, 58 Econometrica 5373 (1990).
128. Lisa Grow Sun, Smart Growth in Dumb Places: Sustainability, Disaster,
and the Future of the American City, 2011 BYU L. Rev. 21572201,
21922193; see also id. at 2193 (Individuals who live in cities vulnerable
to natural disasters may adopt the attitude that every place is risky in
some way andview that vulnerability as just one of the many risks of
modern life.); cf. Justin Pidot, Deconstructing Disaster, 2013 BYU
L.Rev. 213257, 213 (observing that the public may fall into a danger-
ous state of complacency about environmental risks after a long period
of calm, as thoughnatural hazards no longer exist).
129. See generally Tristam McPherson, Moorean Arguments and Moral
Revisionism, 3:2J.Ethics & Soc. Phil. 124, 20 (June 2009); Stephen
P.Stitch & Richard E.Nisbett, Justification and the Psychology of Human
Reasoning, 47 Phil. Sci. 188202, 192193, 196197 (1988).
130. Pidot, supra note 128, at 138.
131. Lisa Grow Sun & Brigham Daniels, Mirrored Externalities, 90 Notre
Dame L. Rev. 135185, 161162 (2014); cf. Michele Landis Dauber,
The Sympathetic State: Disaster Relief and the Origins of the American
Welfare State 6 (2013) (tracing disaster policy and the modern welfare
state in general to the Depression-era imperative to respond to blame-
less suffering).
132. See generally, e.g., Hein Fennema & Peter Wakker, Original and
Cumulative Prospect Theory: A Discussion of Empirical Differences,
10J.Behav. Decision Making 5364 (1997).
133. See Christine Jolls, Cass R.Sunstein & Richard H.Thaler, A Behavioral
Approach to Law and Economics, 50 Stan. L. Rev. 14711550, 1477
1478 (1998).
134. See generally Markus K. Brunnermeier, Herding and Informational
Cascades, in Asset Pricing Under Asymmetrical Information: Bubbles,
Crashes, Technical Analysis, and Herding 147164 (2001); Timur Kuran
& Cass R.Sunstein, Availability Cascades and Risk Regulation, 51 Stan.
L.Rev. 683768 (1999).
135. See Lisa Grow Sun., Disaster Mythology and Availability Cascades, 23
Duke Envtl. L. & Poly F. 7392, 7781 (2012); Lisa Grow Sun.,
Disaster Mythology and the Law, 96 Cornell L.Rev. 11311207, 1150
1152 (2011).
136. See, e.g., Jan. Bonhoeffer & Ulrich Heininger, Adverse Events Following
Immunization: Perception and Evidence, 20 Current Opin. In Infectious
Diseases 237246 (2007); Mark B. Pepys, Science and Serendipity, 7
Clin. Med. 562578 (2007).
137. See, e.g., Aaron M. McCright & Riley E. Dunlap, Challenging Global
Warming as a Social Problem: An Analysis of the Conservative Movements
Counter-Claims, 47 Soc. Probs. 499522 (2000).
138. See generally Scott Plous, The Psychology of Judgment and Decision
Making 233 (1993); Maria Lewicka, Confirmation Bias: Cognitive Error
or Adaptive Strategy of Action Control?, in Personal Control in Action:
Cognitive and Motivational Mechanisms 233255 (Mirosaw Kofta,
Gifford Weary & Grzegorz Sedek eds., 1998); Raymond S.Nickerson,
Confirmation Bias: A Ubiquitous Phenomenon in Many Guises, 2 Rev.
Gen. Psych. 175220 (1998).
139. See Dan M.Kahan etal., The Polarizing Impact of Science Literacy and
Numeracy on Perceived Climate Change Risks, 2 Nature Climate Change
732735 (2012).
140. See generally Mary Douglas & Aaron B.Wildavsky, Risk and Culture: An
Essay on the Selection of Technical and Environmental Dangers (1982)
(propounding a cultural theory of risk); Paul Slovic, The Perception of
Risk (200) (propounding a psychometric paradigm for risk manage-
ment through public policy).
141. See Lennart Sjberg, World Views, Political Attitudes, and Risk Perception,
9 Risk: Health, Safety & Envt 137152 (1998) (arguing that cultural
cognition theory accounts for only a portion of the diversity in attitudes
toward risk).
142. See Robert R.M.Verchick, Culture, Cognition, and Climate, 2016 U.Ill.
L.Rev. (forthcoming); cf. Lisa Grow Sun, Disaggregating Disasters, 60
UCLA L.Rev. 884948, 887 (2013) (criticizing the framing of natural
and technological disasters within the narrative of war and national secu-
rity, as though those disasters involved an anthropomorphic enemy to
be demonized and defeated).
143. New State Ice Co. v. Liebmann, 285 U.S. 262, 311 (1932) (Brandeis, J.,
dissenting).
144. See, e.g., Aesops Fables 6566 (Laura Gibbs trans., 2003).
145. Proverbs 6:68 (Revised Standard Version); see also Proverbs 30:2425
(Four things on earth are small, but they are exceedingly wise: the ants
are a people not strong, yet they provide their food in the summer).
146. Jim Chen, Webs of Life: Biodiversity Conservation as a Species of Information
Policy, 89 Iowa L. Rev. 495608, 603 (2004) (quoting David Takacs,
The Idea of Biodiversity: Philosophies of Paradise 255 (1996)).
147. The original French version of La Cigale et la Fourmi, alongside a ser-
viceable if pedantically literal English translation, appears in Jean de La
Fontaine, Selected Fables/Fables Choisies: A Dual Language Book 23
(Stanley Appelbaum trans., 1997).
56 J.M. CHEN
148. See Andrew Calder, The Fables of La Fontaine: Wisdom Brought Down
to Earth 1824 (2001).
149. The Complete Fables of Jean de La Fontaine 5 (Norman R. Shapiro
trans., 2007).
150. Id. In the original French, these are the pivotal lines: Je vous paierai, lui
dit-elle,/Avant laot, foi danimal,/Intrt et principal./La fourmi
nest pas prteuse:/Cest l son moindre dfaut. La Fontaine, supra
note 147, at 2. In Appelbaums translation, Ill pay you/before har-
vest time, on my word as an animal/both interest and principal./The
ant wasnt the lending kind;/if she had any fault, it wasnt that one. Id.
at 3.
151. See, e.g., Jesse Graham, Brian A. Nosek, Jonathan Haidt, Ravi Iyer,
Spassena Koleva & Peter H. Ditto, Mapping the Moral Domain,
101J.Personality & Soc. Psych. 366385 (2011); Spassena P.Koleva,
Jesse Graham, Ravi Iyer, Peter H.Ditto & Jonathan Haidt, Tracing the
Threads: How Five Moral Concerns (Especially Purity) Help Explain
Culture War Attitudes, 46J.Research in Personality 184194 (2012).
152. Mark Sagoff, Muddle or Muddle Through? Takings Jurisprudence Meets
the Endangered Species Act, 38 Wm. & Mary L. Rev. 825993, 844
(1997); accord Chen, supra note 146, at 602608.
153. See coRessources Consultants, Evidence of the Socio-Economic Importance
of Polar Bears for Canada (2011) (available at http://www.registrelep.
gc.ca/document/default_e.cfm?documentID=2307); Leslie Richardson
& John Loomis, Total Economic Valuation of Endangered Species: A
Summary and Comparison of the United States and the Rest of the World
Estimates, in Conserving and Valuing Ecosystem Services and Biodiversity
2546 (K.N.Ninan ed., 2009). See generally John B.Loomis & Douglas
S.White, Economic Benefits of Rare and Endangered Species: A Summary
and Meta-Analysis, 18 Ecol. Econ. 197206 (1996).
154. See generally James Ming Chen, : Protecting Biodiversity
Against the Effects of Climate Change Through the Endangered Species
Act, 47 Wash. U.J.L. & Poly 1127 (2015).
155. See Edwin Markham, The Man with a Hoe, in The Man with a Hoe and
Other Poems 1518 (Doubleday & McClure 1899).
156. Id. at 16. Jean de La Fontaine referred to both insects in his fable, con-
sistent with the rules of the French language, by the feminine gender.
157. Id.
CHAPTER 3
Higher-Moment Capital Asset Pricing

andIts Behavioral Implications
3.1 The Conventional Capital Asset Pricing

Model
Mental accounting, as depicted in the preceding chapter, organizes poten-
tially bewildering financial information in ways that address humans
physiological, social, and emotional needs. These processes operate at the
level of the self and at the level of civilized society addressing risk man-
agement and resource allocation questions of global proportions. Having
introduced these simplifying frames of thought in purely qualitative terms,
this book will now present some of the quantitative tools that undergird
not only mathematical finance in the traditional sense, but also behavioral
finance as an extension beyond strictly rational considerations of risk and
return. After describing the capital asset pricing model (CAPM) in its con-
ventional form, this chapter will present a four-moment CAPM as a Taylor
series expansion of mean returns. Treating returns as a function of their
mean, variance, skewness, and kurtosis enables us to ascribe behavioral
significance to the odd and even moments of the distribution of returns.1
The conventional CAPM remains the dominant paradigm in financial
risk management2at least among practitioners, if not among scholars.3
Once upon a time, long ago, the hegemony of the CAPM could be
attributed mostly to its apparent ease of applicability and, to a lesser extent,
its empirical justifications.4 The latter excuse, at least, has withered away.
Despite evidence that beta, the CAPMs primary measure of risk, is not

DOI10.1007/978-3-319-32711-2_3
58 J.M. CHEN
positively related to returns on stock,5 much of contemporary financial

theory still hinges on the CAPM.With resilience worthy of cartoon char-
acters like Wile E.Coyote who have the ability to come back to original
shape after being blown to pieces, the CAPM persists because (a) the
empirical support for other asset-pricing models is no better, (b) the the-
ory behind the CAPM has an intuitive appeal that other models lack, and
(c) the economic importance of the empirical evidence against the CAPM
is ambiguous.6
It takes a better theory to kill an existing theory, and the financial
profession has yet to see [a] better theory.7 Even Eugene Fama, argu-
ably the economist who has done the most work to undermine the CAPM
and its theoretical underpinnings, has conceded that market profession-
als (and academics) still think about risk in terms of market .8 Legal
decision-makers are even more committed to the conventional CAPM in
their approaches to financial risk and asset pricing.9 In professional prac-
tice, if not in academic theory, the CAPM is alive and well.10
The CAPM expresses return on an asset as a function of risk, which
in turn can be expressed as volatility, beta, or some other measure drawn
from the second moment of the distribution of financial returns. The inde-
pendent development, particularly by William Sharpe and John Lintner,
of general models represent[ing] equivalent approaches to the problem
of capital asset pricing under uncertainty gave rise to what we recognize
today as the CAPM.11
The CAPM quantifies the premium demanded by the market for shoul-
dering that assets volatility over a benchmark represented by the return
on a risk-free investment:12
Ra = R f + a ( Rm R f )

where Ra, Rm, and Rf, respectively, represent returns on the asset, on the
broader market of comparable investments, and on a risk-free investment,
and where a represents the individual assets beta vis--vis the broader
market portfolio.13 This formula takes the form of a linear equation where
the return on an asset (Ra) is expressed as a function of the premium over
a risk-free baseline (RmRf).14
Modest algebraic rearrangement of the CAPM yields the following
relationship:
HIGHER-MOMENT CAPITAL ASSET PRICING... 59
Ra R f
Rm R f =
a

The left side of the foregoing equation represents the risk premium
demanded for the entire asset class represented by a particular segment
of the market.15 The risk premium is the difference between returns on
a specific investment or class of investments and some sort of risk-free
benchmark.16 This premium dictates a firms cost of capital. Capital asset
pricing, in its original incarnation, offered a solution to the problem of
determining the price that investors would demand for bearing risk in
excess of a risk-free alternative.17
Another common application of the CAPM compares an index of equi-
ties designed to track the Standard & Poors 500 against the putatively
risk-free baseline of short-term Treasury bills.18 This market-wide risk pre-
mium is equivalent to the risk-adjusted premium expressed on the right
side of the equationnamely, the risk premium for the asset vis--vis a
risk-free investment, divided by the individual assets beta.19
This ratio between risk-adjusted return and volatility bears closer exam-
ination. Recall that the foregoing equation is merely an algebraically refor-
mulated version of the basic CAPM:
Ra = R f + a ( Rm R f )

The ratio of (1) the premium over a baseline return to (2) the volatility
associated with that asset or portfolio demonstrates how market returns
are adjusted for a measure of risk such as volatility or beta:
Ra Rb
Treynor ratio =
a

Algebraic manipulation, in one direction or another, connects the

Treynor ratio of reward to volatility with the more general CAPM.Indeed,
the Treynor ratio restates the CAPM.Although mathematical congruence
undermines the contribution of the Treynor ratio to empirical tests of the
CAPM,20 this definitional unity does make the Treynor ratio a convenient
tool for evaluating and understanding the broader model of asset pricing.
60 J.M. CHEN
The Treynor ratio demonstrates that the general risk premium of a class
of investments is equivalent to the premium for a specific investment over
risk-free return, discounted by the volatility of returns on that specific
asset relative to returns on the benchmark class as a whole. In other words,
we can extrapolate the Treynor ratio from the CAPM, and the CAPM
from the Treynor ratio.
The Treynor ratio measures reward as return on an asset, Ra, above
some benchmark return, Rb, relative to the volatility of that assets return
as expressed by its beta, a.21 The benchmark return, Rb, often is, but need
not be, equivalent to the risk-free baseline, Rf. The Treynor ratio closely
resembles a generalized version of the Sharpe ratio of reward to variability
(as measured by the standard deviation of portfolio returns):22
Ra Rb
Generalized Sharpe ratio =

Both the Sharpe and Treynor ratios evaluate portfolio returns or port-
folio manager performance according to the relationship between returns
and some proxy for risk.23
3.2 Higher-Moment CAPM

asaTaylor Series Expansion
Mathematical analysis provides a way to harmonize the conventional

CAPM with behavioral economics. Overtly, behavioral interpretations of
moment-based theories of finance associate different moments with differ-
ent emotions.24 All fields of applied behavioral science, including finance,
require models whose primary or even exclusive purpose is to describe
economic behavior as undertaken by actual humans, as opposed to hypo-
thetical economic reason dictated by quantitative logic.
We can reduce the conceptual distance between conventional capital
asset pricing and behavioral economics by adopting a four-moment vari-
ant of the CAPM.If mean-variance optimization arises from the idea that
the investor prefers higher expected returns and lower risk, then higher-
moment CAPM presumes, ceteris paribus, investors prefer a high prob-
ability of an extreme event in the positive direction over a high probability
of an extreme event in the negative direction.25 On that assumption,
conventional mean-variance optimization represents merely a special case
of a more comprehensive model that accounts for asymmetry in returns

(as expressed by skewness) and the relative probability of extreme events
in the distribution of returns (as expressed by kurtosis).26 The goal, there-
fore, is to devise a theoretically coherent account of investor preferences
with respect to at least the first four momentsmean, variance, skewness,
and kurtosis.
One mathematically cogent way of dealing with higher moments in
the asset allocation is the use of the Taylor series expansion to derive an
approximation of the expected utility function.27 A higher-order Taylor
series expansion can simplify[] the asset allocation task28 and inform
optimal portfolio selection in the presence of higher-order moments
and parameter uncertainty.29 Specifically, to measur[e] the effect of
higher moments on asset allocation, we can approximate the expected
utility by a Taylor series expansion around the expected wealth.30
The Taylor series expansion for a function, f(x), that is infinitely differ-
entiable at value a takes the form of a power series:31
f (a) f ( a ) f ( a )
f ( x) = f (a) + ( x a) + ( x a) ( x a) +
2 3
+
1! 2! 3!
Exploiting multiple mathematical identitiesnamely, that (xa)0 and 0!

both equal 1, and the zeroth-order derivative of f(x) is f(x) itselfenables
us to express the Taylor series expansion in a more compact form:

f(
n)
(a)
f ( x) = ( x a)
n
n=0 n!
Accordingly, if an investors utility function is expressed in terms of

the wealth distribution, so that:
U ( w ) = U ( w ) f ( w ) dw

where f(w) is the probability distribution function of end-of-period

wealth, contingent upon the multivariate distribution of returns and
the weights of the portfolios components, then the infinite-order Taylor
series expansion of the utility function is:
62 J.M. CHEN
U(
n)
(w)(w w)
n

U (w) =
n =0 n!
where w = w = 1 + denotes the expected end-of-period wealth,

designates the vector of expected returns, and designates the vector of
portfolio weights.32
The application of this model to a simple set of financial returns is even
more straightforward. Let R and r denote simple (arithmetic) and loga-
rithmic (continuously compounded) returns respectively.33 By definition,
rt.=ln(1+Rt),34 where rt. designates the continuously compounded return
or log return of an asset and is defined to be the natural logarithm of
[the assets] gross return (1 + Rt).35 Computating log rather than arith-
metic returns allows us to express continuously compounded multiperiod
return [as] simply the sum of continuously compounded single- period
returns.36
Indeed, it is important for mathematical and not just esthetic reasons
to express returns in logarithmic form and calculate geometric rather than
arithmetic average returns. When returns are serially correlated, then
the arithmetic average can lead to misleading estimates .37 Arithmetic
average return exceeds its geometric counterpart; if the returns are log-
normally distributed, the difference between the two is one-half the vari-
ance of the returns.38 The computation of log returns not only overcomes
the difficulty of manipulating geometric averages, but also allows the
expression of continuously compounded multiperiod return [as] simply
the sum of continuously compounded single-period returns:39
rt ( k ) = ln 1 + Rt ( k ) = ln (1 + Rt ) (1 + Rt 1 ) (1 + Rt k +1 )
rt ( k ) = ln (1 + Rt ) + ln (1 + Rt 1 ) + + ln (1 + Rt k +1 )
rt ( k ) = rt + rt 1 + + rt k +1

Even more importantly, this expression of log returns lends itself to a

very simple and elegant Taylor series expansion. For the logarithmic func-
tion ln(1+x), the Taylor series expansion takes this form:40

xn
ln (1 + x ) = ( 1)
n +1
n =1 n
Generalizing the Taylor series expansion to account for ln(1 + x) at

x= yields:
x (x ) (x ) (x )
2 3 4
ln (1 + ) + + o ( x )
5
+
1 + 2 (1 + ) 3 (1 + ) 4 (1 + )
2 3 4

where o[(x)5] represents remaining terms of the fifth order and above.41
Inasmuch as time series also rely on logarithmic returns, a more elabo-
rate third-order (or higher) Taylor series expansion may be derived by
relat[ing] the discount factor to the marginal rate of substitution between
periods t and t + 1in a two-period economy.42
Somewhat optimistically, Javier Estrada leaps directly from this model
to a Taylor series expansion that consists exclusively of alternative central
statistical moments. He is partially correctand almost entirely correct if
we modify the definition of skewness and kurtosis. If we let and [rep-
resent] the mean and variance of R, then the conventional CAPM takes
the form of a Taylor series expansion of expected returns:43
x 1 2
r = ln (1 + R ) ln (1 + ) + + R2 ln (1 + R )
1 + 2 (1 + )2

where Rn(x) designates the remaining terms of the Taylor series expan-
sion beyond order n (which in this example is 2).44 Estrada omits the sec-
x
ond term of the series, , and takes no explicit account of the Taylor
1+
remainder.
More explicitly, Estrada proposes to interpret the remainder term as
direct implementations of skewness and kurtosis:45
1 2 1 Skew 1 Kurt
r = ln (1 + R ) ln (1 + ) +
2 (1 + )2 3 (1 + )3 4 (1 + )4

This Taylor series expansion directly provides [a]n approximate expected

utility [functions] based on mean, variance, skewness, and kurtosis for
an investor who displays a logarithmic utility function:46
64 J.M. CHEN
1 2 1 Skew 1 Kurt
AEU Kurt ln (1 + ) +
2 (1 + ) 3 (1 + ) 4 (1 + )4
2 3
Estradas interpretation of this Taylor series expansion is correct, but only

if one defines skewness and kurtosis, respectively, as the third and fourth
central moments rather than the third and fourth standardized moments.
Potential confusion arises from other sources adoption of the definition of
3
x 3
skewness as the third standardized moment, 1 = = 3 .47 Note

further that standard treatments of kurtosis subtract 3 from the fourth
standardized moment in order to express excess kurtosis by reference to the

kurtosis of a normal Gaussian distribution: 2 = 2 3 = 44 3.

In the interest of precision, we should restate Estradas interpretation
of the Taylor series expansion of log returns:
ln (1 + x ) at ( x = )
x (x ) (x ) (x )
2 3 4
ln (1 + ) + + o ( x )
5
+
1 + 2 (1 + ) 3 (1 + )3 4 (1 + )4
2
x 2 3 4
ln (1 + ) + + + o ( 5 )
1 + 2 (1 + )2 3 (1 + )3 4 (1 + )4

Therefore, this Taylor series expansion does correspond to definitions for

skewness and kurtosis,48 but only if we define skewness and kurtosis
as central higher moments in place of these terms traditional statistical
definitions as standardised central moments.49
If we insist, very modest rearrangement allows us to restate the Taylor
series expansion in terms of more traditional interpretations of skewness
and kurtosisnamely, 1, 2, and 2:
ln (1 + x ) at ( x = )
x 2 1 3 2 4
ln (1 + ) + + + o ( 5 )
1 + 2 (1 + ) 3 (1 + ) 4 (1 + )4
2 3
x 2 1 3 ( 2 + 3) 4
ln (1 + ) + + + o ( 5 )
1 + 2 (1 + )2 3 (1 + )3 4 (1 + )4

The advantage of defining skewness and kurtosis in this Taylor series

expansion as central moments, however, is the clarity with which the
x 2 3 4
expression, ln (1 + ) + + + o ( 5 ) ,
1 + 2 (1 + )2 3 (1 + )3 4 (1 + )4
demonstrates that the investors expected utility depends on all central
moments of the distribution of end-of-period wealth.50
Of even greater significance is what the Taylor series expansion implic-
itly says about the impact of skewness and kurtosis on investor welfare.
Under rather modest assumptionspositive marginal utility, decreasing
risk aversion at all wealth levels, and a strict consistency in the inves-
tors attitude toward a given statistical moment without regard to her or
his wealththe following inequalities hold:51
U(
n)
( w ) > 0 w, if n is odd and

U ( n)
( w ) < 0 w, if n is even.
Specifically, investors have positive preference for positive skewness

and negative preference for negative skewness.52 The Taylor series
expansion of any utility function shows that if the third derivative of the
utility function is positive, there is a preference for skewness.53 And the
preferences most commonly employed in economics and finance reveal
a positive third derivative.54 Decreasing absolute risk aversionwhich
in turn is a special case of hyperbolic absolute risk aversion55likewise
implies that investors like skewness.56 Decreasing absolute risk aver-
sion implies, as is intuitively appropriate, that individuals become more
willing to accept[] risky situations as they acquire greater wealth.57
Stated more plainly, richer investors are more willing to take risk in their
investment portfolios.58 The presence of a positive third derivative in the
most commonly employed models therefore predicts that investors are
more willing to indulge their taste for positively skewed outcomes as their
wealth grows.59
Combining what we instinctively know about variancethat inves-
tors dislike it, at least on the downsidewith this preference for pos-
itive skewness enables us to generalize to the next moment, kurtosis:
Consistent risk aversion, strict consistency of moment preference, and
positive preference for positive skewness imply negative preference for the
fourth statistical moment (kurtosis).60 Or in even simpler terms: inves-
tors like mean return and positive skewness and dislike variance and kur-
tosis.61 The odd moments, mean and skewness, advance returns, while
66 J.M. CHEN
the even moments produce a drag on expected compound return.62

The alternating treatment of odd- and even-numbered mathematical
moments represents a logical extension of an essential non-linear fea-
ture of observed investor behavior already capture[d] by the single-
sided treatment of semivariance: [M]ost investors perceive infrequent
large losses or shortfalls [to be] far more risky than more frequent smaller
losses or shortfalls.63
The adage that investors generally prefer high values for odd moments
and low ones for even moments reaches its greatest clarity in the extremes
of the distribution.64 Odd moments can be seen as a way to decrease
extreme values on the side of losses and increase them on the gains.65
Over the long haul, positive skewness indicates the presence of outsized
gains; it suggests the tantalizing possibility that certain holdings in the
portfolio will offer disproportionately large payouts, as though they were
winning lottery tickets.66 By contrast, even moments measure dispersion,
and therefore volatility, something undesirable that increases the uncer-
tainty of returns.67
Combining this insight with the basic, general definition of the Taylor
series expansion as a function of differentials, factorials, and polynomials,
f (a) f ( a ) f ( a )
f ( x) = f (a) + ( x a) + ( x a) + ( x a ) + , enables
2 3
1! 2! 3!
an even simpler approximation of expected returns or investor utility as a
preference function:68
1 (2) 1 3 1 4
U (w) U (w) + U ( w ) 2 + U ( ) ( w ) s3 + U ( ) ( w ) k 4
2! 3! 4!
where s and k designate skewness and kurtosis, again defined as cen-

tral moments. Consistent with our understanding of investor responses
to odd and even moments, the actual expected utility approximated by
this function depends positively on expected return and skewness and
negatively on variance and kurtosis.69 In this view of the approximate
preference function, the work of sorting expected return and skewness
from variance and kurtosis is performed by the sign of the odd- and even-
numbered derivatives of the utility function.
3.3 A Bridge Between Econometric and

Behavioral Views ofLow Volatility
The foregoing presentation of a four-moment CAPM will prove useful
at several junctures in this book. The Taylor series expansion constitutes
an important mathematical step in the formulation of the intertempo-
ral CAPM70 and of various accounts of risk aversion.71 In addition, the
fundamental insight of higher-moment capital asset pricingthat odd-
numbered statistical moments such as mean and skewness convey positive
investor utility, while even-numbered moments such as variance and kur-
tosis express competing forms of disutilityinforms the leading accounts
of risk-averse and risk-seeking behavior by investors: prospect theory72 and
SP/A theory (which in turn is a critical precursor of behavioral portfolio
theory).73
Before elaborating those facets of behavioral finance, however, I will
address a more immediate threat to the theoretical relationship between
risk and return. Contrary to the fundamental supposition that high returns
reward high risk, scholars of mathematical finance and of strategic man-
agement, almost entirely independent of each other, have discovered that
lower-risk companies actually deliver higher returns. This low-volatility
anomaly poses a serious intellectual challenge to the edifice of contempo-
rary financial thought.
Chapter 7 of Postmodern Portfolio Theory examines the low-volatility
anomaly in depth, albeit from the perspective of quantifying the relative
contributions of volatility and correlation, on either the upside or the
downside of expected returns, to departures from the conventional rela-
tionship between risk and return.74 The next chapter of this book will look
anew at this anomaly as a behavioral as well as an econometric enigma.
Notes
1. The companion volume to this book, Chen, Postmodern Portfolio Theory,
supra note 1 (Chapter 1), addresses these topics in Chapter 2, at 525
(conventional CAPM) and Chapters 1011, at 189213 (higher-moment
capital asset pricing).
2. See, e.g., Fischer Black, Capital Market Equilibrium with Restricted
Borrowing, 45J.Bus. 444455 (1972); Fischer Black, Michael C.Jensen &
Myron S.Scholes, The Capital Asset Pricing Model: Some Empirical Tests, in
Studies in the Theory of Capital Markets 79121 (Michael C.Jensen ed.,
68 J.M. CHEN
1972); John Lintner, Security Prices, Risk and Maximal Gains from
Diversification, 20J.Fin. 587615 (1965); John Lintner, The Valuation of
Risk Assets and the Selection of Risky Investments in Stock Portfolios and
Capital Budgets, 73 Rev. Econ. & Stats. 1337 (1965); Jan. Mossin,
Equilibrium in a Capital Asset Market, 34 Econometrica 768783 (1966);
William F. Sharpe, Capital Asset Prices: A Theory of Market Equilibrium
Under Conditions of Risk, 19 J. Fin. 425442 (1964); Jack L. Treynor,
Toward a Theory of Market Value of Risky Assets, in Asset Pricing and
Portfolio Performance: Models, Strategy and Performance Metrics 1522
(Robert A. Korajczyk ed., 1999); Jack L. Treynor & Fischer Black,
Corporate Investment Decisions, in Modern Developments in Financial
Management 310327 (Stewart C.Myers ed., 1976). See generally Bernell
K.Stone, Risk, Return, and Equilibrium: A General Single-Period Theory
of Asset Selection and Capital Market Equilibrium (1970); Eugene F.Fama
& Kenneth R. French, The Capital Asset Pricing Model: Theory and
Evidence, 18:3J.Econ. Persp. 2546 (Summer 2004).
3. See Levy, CAPM in the 21st Century, supra note 41 (Chapter 1), at 45
(describing the CAPM and its constituent concepts as probably the most
widely employed measures used by academic researchers and even more
intensively used by investment firms and practitioners).
4. Philip H.Dybvig & Jonathan E.Ingersoll, Jr., Mean-Variance Theory in
Complete Markets, 55J.Bus. 233251, 233 (1982).
5. See Eugene F.Fama & Kenneth R.French, The Cross-Section of Expected
Stock Returns, 47J.Fin. 427465 (1992); see also Marc R.Reinganum, A
New Empirical Perspective on the CAPM, 16 Fin. & Quant. Analysis 439
462 (1981).
6. Ravi Jagannathan & Zhenyu Wang, The Conditional CAPM and the Cross-
Section of Expected Returns, 51 J. Fin. 353, 4 (1996) (footnote
omitted).
7. Tim Koller, Marc Goedhart & David Wessels, Valuation: Measuring and
Managing the Value of Companies 261 (5th ed. 2010).
8. Eugene F. Fama, Efficient Capital Markets: II, 46 J. Fin. 15751617,
1593 (1991); accord Glenn N.Pettengill, Sridhar Sundaram & Ike Mathur,
The Conditional Relation Between Beta and Returns, 30J.Fin. & Quant.
Analysis 101116, 102 (1995).
9. See, e.g., In re American Classic Voyages Co., 367 B.R. 500, 513 n. 19
(D.Del. Bankr. 2007) (While there are other models to determine equity,
CAPM is probably the most widely used. [quoting Peter V.Pantaleo &
Barry W.Ridings, Reorganization Value, 51 Bus. Law. 419442, 433 n.52
(1996)]); cf. AEP Texas North Co. v. Surface Transp. Bd., 609 F.3d 432,
443 (D.C.Cir. 2010) (observing that courts do not sit aspanel[s] of
statisticians, but aspanels of generalist judges).
10. See Haim Levy, The CAPM Is Alive and Well: A Review and Synthesis, 16
Eur. Fin. Mgmt. 4371 (2009); cf. Levy, CAPM in the 21st Century, supra
note 41 (Chapter 1), at 22 (describing the CAPM and related models of
mean-variance optimization as still alive and kicking).
11. Eugene Fama, Risk, Return, and Equilibrium: Some Clarifying Comments,
23J.Fin. 2940, 40 (1968).
12. See Robert A. Korajczyk, Introduction, in Asset Pricing and Portfolio
Performance, supra note 2, at viii, xv.
13. See id.
14. See id.
15. See id.
16. See William F.Sharpe, Capital Asset Prices: A Theory of Market Equilibrium
Under Conditions of Risk, 19J.Fin. 425442, 426427 (1964).
17. See Franco Modigliani & Merton Miller, The Cost of Capital, Corporate
Finance, and the Theory of Investment, 48 Am. Econ. Rev. 261297 (1958).
18. See 1 Handbook of Quantitative Finance and Risk Management 1.5, at
333, 1012 (Cheng-Few Lee, Alice C.Lee & John Lee. eds., 2010).
19. See Korajczyk, supra note 12, at xv.
20. See Richard Roll, A Critique of the Asset Pricing Theorys Tests, 4 J. Fin.
Econ. 129176, 136 (1977).
21. See Treynor, supra note 2, at 1617.
22. See William F. Sharpe, Mutual Fund Performance, 39 J. Bus. 119138
(1966); William F. Sharpe, Adjusting for Risk in Portfolio Performance
Measurement. 1:2J.Portfolio Mgmt. 2934 (Winter 1975).
23. See J.D.Jobson & Bob M.Korkie, Performance Hypothesis Testing with the
Sharpe and Treynor Measures. 36J.Fin. 888908 (1981); Sharpe, Mutual
Fund Performance, supra note 22, at 121122.
24. See, e.g., Ole Hagen, Separation of Cardinal Utility and Specific Utility of
Risk in Theory of Choices Under Uncertainty, 3 Saertrykk av Statskonomisk
Tidsskrift 81107, 9299 (1969).
25. Campbell R.Harvey, John C.Liechty, Merrill W.Liechty & Peter Mller,
Portfolio Selection with Higher Moments, 10 Quant. Fin. 469485, 471
(2010); Gustavo M. de Athayde & Renato G. Flres, Jr., Finding a
Maximum Skewness PortfolioA General Solution to Three- Moments
Portfolio Choice, 28J.Econ. Dyn. & Control 13351352, 1342 (2004).
26. See Eric Jondeau & Michael Rockinger, Optimal Portfolio Allocation Under
Higher Moments, 12 Eur. Fin. Mgmt. 2955, 33 (2006).
27. Id. at 30.
28. Harvey, Liechty, Liechty & Mller, supra note 25, at 470.
29. Id. at 469.
30. Jondeau & Rockinger, supra note 26, at 33.
31. See George B.Arfken & Hans J.Weber, Taylors Expansion, in Mathematical
Methods for Physicists 5.6, at 303313 (3d ed. 1985); https://en.wikipedia.
70 J.M. CHEN
org/wiki/Taylor_series. Unless otherwise noted, background information on

the mathematics of the Taylor series expansion comes from these sources. The
special case of a Taylor series where a=0 is often designated a Maclaurin series.
33. Javier Estrada, Mean-Semivariance Behaviour: An Alternative Behavioural
Model, 3J.Emerging Mkt. Fin. 231248, 241 (2004).
34. See id.
35. Campbell, Lo & MacKinlay, supra note 5 (Chapter 1), at 11.
36. Id.
37. Rajnish Mehra & Edward C. Prescott, The Equity Premium Puzzle in
Retrospect, in 1 Handbook of the Economics of Finance, supra note 40
(Chapter 1), at 888936, 889.
38. Id. at 888.
40. See https://en.wikipedia.org/wiki/Taylor_series
41. See http://www.wolframalpha.com/input/?i=taylor+series+for+ln%281%
2Bx%29+at+x%3Dy
42. Campbell R. Harvey & Akhtar Siddique, Conditional Skewness in Asset
Pricing Tests, 55J.Fin. 12631295, 1269 (2000).
43. Estrada, An Alternative Behavioural Model, supra note 33, at 241.
44. See https://en.wikipedia.org/wiki/Taylor%27s_theorem
46. Id. at 246.
47. See, e.g., Leslie A.Balzer, Investment Risk: A Unified Approach to Upside
and Downside Returns, in Managing Downside Risk in Financial Markets:
Theory, Practice and Implementation 103155, 121 (Frank A.Sortino &
Stephen E.Satchell eds., 2001).
48. Cf. Jondeau & Rockinger, supra note 26, at 34 (adopting a functionally
equivalent definition of the Taylor series expansion of expected returns).
49. Id. at 34 n.5.
51. Id. at 34.
52. Robert C. Scott & Philip A. Howath, On the Direction of Perference for
Moments of Higher Order Than the Variance, 35 J. Fin. 915919, 917
(1980); see also Markus K. Brunnermeier, Christian Gollier & Jonathan
A. Parker, Optimal Beliefs, Asset Prices, and the Preference for Skewed
Returns, 97 Am. Econ. Rev. 159165 (2007).
53. Levy, CAPM in the 21st Century, supra note 3, at 61 n.4.
54. Id.
55. See generally infra 6.4, at 115117.
56. Levy, CAPM in the 21st Century, supra note 3, at 70.
57. Gary G.Venter, Utility with Decreasing Risk Aversion, 70 Casualty Actuarial
Socy Forum 144155, 147 (1983); see also Leonard R. Freifelder, A
Decision Theoretic Approach to Insurance Ratemaking 35 (1976).
58. See generally Yora Kroll, Moshe Leshno, Haim Levy & Yishay Spector,
Increasing Risk, Decreasing Absolute Risk Aversion and Diversification,
24J.Math. Econ. 537556 (1995); Liqun Liu & Jack Meyer, Decreasing
Absolute Risk Aversion, Prudence, and Increased Downside Risk Aversion,
3J.Risk & Uncertainty 243260 (2012).
59. Levy, CAPM in the 21st Century, supra note 3, at 70. Skewness prefer-
ences and positive third derivatives are related to third degree Stochastic
Dominance. Id. at 61 n.4. See generally G.A. Whitmore, Third-Degree
Stochastic Dominance, 60 Am. Econ. Rev. 457459 (1970).
60. Scott & Howath, supra note 52, at 91718.
62. Id.
63. Balzer, supra note 47, at 130.
64. Athayde & Flres, supra note 25, at 1336.
65. Id.
66. See Turan G.Bali, Nusret Cakici & Robert Whitelaw, Maxing Out: Stocks
as Lotteries and the Cross-Section of Expected Returns, 99J.Fin. Econ. 427
446 (2011).
67. Athayde & Flres, supra note 25, at 1336.
69. Id.
70. See infra 5.1 at 97.
71. See infra 6.1, at 112113; 6.3, at 114115.
72. See infra Chapters 8 and 9.
73. See infra Chapter 10.
74. See generally Chen, Postmodern Portfolio Theory, supra note 1 (Chapter
1), 7.17.10, at 107151.
CHAPTER 4
Tracking theLow-Volatility Anomaly

AcrossBehavioral Space
4.1 The Low-Volatility Anomaly

andBowmans Paradox
In his popular guide to asset allocation, neurologist-turned-financial-

analyst William J. Bernstein offers a bit of jarring advice to investors:
Good companies are usually bad stocks; bad companies are usually good
stocks.1 Bernsteins practical prescription stems from an academic insight:
Growth opportunities are usually the source of high betas.2 In principle,
these high betas should impart higher risk and higher returns to growth
stocks: [B]ecause growth options tend to be most valuable in good
times and have implicit leverage, which tends to increase beta, they con-
tain a great deal of systematic risk.3 Nevertheless, even though growth
options hinge upon future economic conditions and must be riskier than
assets in place, the historical pattern cuts in the opposite direction:
[G]rowth stocks earn lower average returns than value stocks.4 From
these observations flows Bernsteins advice to the individual investor:
Favor a value approach in your stock and mutual fund choices.5
The best indicator of a stocks characterization as a value or growth
stock is the ratio of its price to its book (P/B) value.6 The inverse of the
so-called P/B ratio, or the book-to-market equity ratio, captures the
relative corporate distress factor and is thus a risk variable that needs to
be compensated.7 To the extent that the willingness to commit capital to

DOI10.1007/978-3-319-32711-2_4
74 J.M. CHEN
undervalued, distressed firms commands a risk premium, then the pres-

ence of trouble within a corporation may explain the superior performance
of low-beta portfolios.8
Such an explanation would address one of the most spectacular fail-
ures of traditional financial models. The tendency of value stocks [to]
earn higher expected returns than growth stocks poses as a troublesome
anomaly for rational expectations.9 The entire point of an efficient capital
market is to reward the assumption of risk with returns. A contempora-
neous relationship between factor loadings and risk premia is the founda-
tion of a cross-sectional risk-return relationship, and has been exploited
from the earliest tests of the CAPM.10 Ceteris paribus, risky assets have
high returns. Safe assets dont.
If only actual markets behaved so simply. This simple empirical propo-
sitionthat returns follow risk in a straightforward mannerhas been
hard to support on the basis of the history of U.S. stock returns.11 Indeed,
the most widely used measures of risk point rather strongly in the wrong
direction.12 For instance, in a survey of stock returns from 1926 to 1971,
Robert Haugen and James Heins concluded: [O]ver the long run, stock
portfolios with lesser variance in monthly returns have experienced greater
average returns than their riskier counterparts.13
More recent studies confirm the presence of a low-volatility anomaly.
Stocks exhibiting the highest levels of volatility have abysmally low aver-
age returns.14 The inversion of returns on low- versus high-volatility
stocks has been detected across numerous historical periods and in mar-
kets around the world.15 That low beta is high alpha is a robust historical
pattern.16 The presence of returns as low as negative 0.02% per month
in the quintile of stocks exhibiting the highest levels of volatility is not
merely a puzzle.17 Abysmal returns in the most volatile quantile are the
natural and predictable consequence of a statistically significant negative
price of risk of approximately 1% per annum charged against inno-
vations in aggregate volatility.18 Because it so strikingly challenges the
basic notion of a risk-return tradeoff, the long-term outperformance of
low-risk portfolios is quite possibly the greatest anomaly in finance.19
Writing from the perspective of accounting rather than finance, the stra-
tegic management literature has identified the same phenomenon. In his
epochal 1980 article, A Risk/Return Paradox for Strategic Management,
Edward H. Bowman observed that the riskiest firms provided the low-
est returns.20 This work, buttressed by two further articles confirming the
unexpected relationship between high risk and low return,21 uncovered a
TRACKING THELOW-VOLATILITY ANOMALY ACROSS BEHAVIORAL SPACE 75
negative slope between risk and return, albeit through accounting data
rather than the cross section of financial returns.22 Bowmans findings pre-
sented a paradox because they implie[d] that managers accept higher risk
at the same time that they expect lower returns.23 What is now known as
Bowmans paradox joins the low-volatility anomaly in contradicting the
established financial economics tenet that higher risk must be compensated
by higher return in order to motivate agents to undertake risky ventures.24
For the first 12 years after the 1980 publication of Bowmans original
article documenting low returns among risky firms, academic literature on
mathematical finance made no citation of Bowmans paradox.25 To this
day, finance and strategic management, as academic disciplines, continue
to maintain different perspectives on risk.26 In finance, market forces
determine the relationship between risk and return.27 By contrast, strate-
gic management assumes that risk management strategies are proprietary,
inaccessible, and illiquid and that any one firms strategic opportunity or
innovation may have little value to other firms.28
The strategic management literature evaluating Bowmans paradox in
its native environment and on its own terms has generated considerable
controversy while reaching no consensus.29 It is nevertheless evident
that the strategic management literature is striving to explain the same
phenomenon that the financial literature calls the low-volatility anomaly.30
At least one source tests Bowmans paradox differently in bull markets
and in bear markets,31 in harmony with the bifurcation of beta depicted
in Chapters 4 through 7 of Postmodern Portfolio Theory.32 At a more
abstractly theoretical level, strong correlation between empirical measures
of market risk (the subject of financial economics) and accounting risk
(the subject of strategic management) permit comparisons between the
low-volatility anomaly and Bowmans paradox, and between the bodies of
knowledge brought to bear on both phenomena.33
Three streams of literature seek to explain Bowmans paradox. One
treats the paradox as the product of misspecification.34 Another ascribes
its origins to theories of organizational strategy and behavior.35 Section
2.3s discussion of abnormal returns from firms with high corporate and
social responsibility ratings fits within the tradition emphasizing strategy
and behavior. The third stream, consisting of works aligning Bowmans
paradox with prospect theory, comes closest to harmonizing this bedrock
of strategic management theory with mathematical finance through the
behavioral sciences.36 Section 9.3 of this book will take a closer look at this
line of thinking.
76 J.M. CHEN
4.2 Beta asaComposite Measure ofVolatility

andCorrelation
So far, we have spoken of risk as if it subsisted entirely in the second

moment of the distribution of returns. Implicitly, we have assumed that
volatility, or the standard deviation of returns, serves as a good proxy for
risk. These assumptions comport with the simplest interpretation of the
four-moment CAPM specified in 3.2 which suggests that even-numbered
moments embody fear, while odd-numbered moments represent hope or
even greed.
A behaviorally sophisticated assessment of risk measures derived from
the second moment of the distribution of returns, however, warrants
a closer look at beta. This basic measure of systematic risk commands
extended attention in Postmodern Portfolio Theory. Without rebuilding the
entire mathematical apparatus outlined in Chapters 3 through 6 of that
book, I will now redirect that analysis toward the specific goal of address-
ing the low-volatility anomaly. Behavioral finance provides multiple, non-
mutually exclusive solutions to this puzzle. What these solutions have in
common is the bifurcation of beta along some sort of spatial or temporal
dimension. The balance of this chapter will evaluate beta according to
two spatial dichotomies. Chapter 5 will then examine beta and the low-
volatility anomaly in light of the intertemporal CAPM.
As the basic measure of systematic risk in the conventional CAPM,
beta remains the most explanatory element of the risk premium in most
asset pricing models.37 Beta retains a place in nearly all financial mod-
els despite the considerable evidence adduced against standard deviation
and its variations as measures of risk in non-normal distributions, if
only because relatively little effort has been made to devise a better risk
measure.38 Even models that purportedly deprecate beta as an insignifi-
cant financial factor continue to treat beta as an important explanatory
variable, despite denying beta its traditional status as the main explana-
tory variable.39
Formally, beta is the product of (1) the ratio of asset-specific volatility
to market-wide volatility and (2) the correlation between returns on that
asset and market-wide returns is the beta of that asset:
sa s cov ( a,m ) cov ( a,m )

ba = r ( a,m ) = a =
sm sm s as m s m2

When beta is broken down into these components, it is readily understood

as correlated relative volatility.40 First, relative volatility is the ratio of asset-
specific volatility to market-wide volatility: s a s m . Second, beta reflects the
correlation between an asset or asset class and the market as a whole: (a, m).
As between these components of beta, volatility commands the lions
share of popular and academic attention. It is volatility rather than cor-
relation that typically (or stereotypically) strikes fear in investors hearts.
Consider, for example, the Chicago Board Options Exchange (CBOE)s
celebrated and vaunted Volatility Index, or VIX.41 The VIX forecasts the
expected movement in the Standard & Poors (S&P) 500 Index over the
next 30-day period according to the square root of the par variance swap
rate.42 Although it is often characterized as the investor fear index,43
VIX is nothing more mysterious than a measure of implied volatility
derived from the prices on the CBOEs volatility options that traders
are willing to pay, given their forward-looking expectations of volatility.
Misunderstandings of VIX and other volatility measurements support the
sentiment that we dont quite know what were talking about when we
talk about volatility.44
One thing we do know is that implied volatility, as derived from option
prices, generally exhibit[] much more pronounced asymmetry than
other measures of volatility.45 Appropriately enough, it is exaggerated fear
of volatility, especially on the downside of expected returns, that looms
largest in nearly all behavioral accounts of finance. Downside volatility
threatens to inflict the very sort of loss that humans fear most: losing
ground relative to a fixed reference point.46 Humans subjectively measure
their welfare in relative rather than absolute terms.47
Volatility per se poses a direct threat to financial expectations. The
nearly instinctive reaction of individual and even institutional investors
to the slightest perception of news48 breathes life into John Maynard
Keyness observation that financial transactions invariably contain an ele-
ment of caprice.49 Critical decisions are made on impulse rather than
calculation; investors often set aside their own elaborate calculations of
probabilities and proceed instead on gut feeling.50
By contrast, correlation tightening undermines diversification strategies
adopted for the specific purpose of taming volatility. Unlike the instinc-
tive and immediate fear of loss associated with volatility, comprehending
the surprising or even paradoxical impact of correlation tightening
on portfolio management requires further reflection.51 In contrast
with downside volatility, tightening correlation during market downturns
creates a distinct sort of peril, one as insidious as it is subtle. Correlations
78 J.M. CHEN
across asset classes, relatively stable under ordinary market conditions,

tighten under stress.52 The very portfolios that are diversified to with-
stand volatility might suffer greater damage as correlations rise during
period[s] of severe market turmoil.53 Perhaps the most familiar illustra-
tion arises from the demand for safe haven assets such as gold and US
Treasury bonds during market downturns.54 Paradoxically, the adoption
of negative interest rates by central banks as a tool for making macro-
economic policy may increase safe haven demand for sovereign debt by
signaling extremely low economic expectationsand thereby drive rates
even deeper into negative territory.55
Correlation risk almost certainly includes the probability that liquid-
ity will evaporate when everyone heads for the exit at once.56 In extreme
circumstances, when financial shocks cross national borders, global con-
tagion poses a special threat to portfolio diversification strategies that
aim to minimize exposure to any single market.57 The financial literature
distinguishes between two types of contagion; identifying the right type
of contagion guides the proper policy response. Shift contagion occurs
where the interdependencies between pairs of markets increase during a
crisis.58 Normal interdependencies from pre-existing market linkages,
such as goods trade, financial flows, or exposure to common shocks can
become[] unstable during an episode of high volatility.59 By contrast,
pure contagion reflects excess contagion suffered during a crisis that is
not explained by market fundamentals or common shocks.60 Pure conta-
gion arises from idiosyncratic shocks being transmitted to other countries
through channels that could not have been identified before the event.61
When pure contagion causes multiple markets to decline simultaneously,
policies such as capital controls aimed at breaking market linkages are
unlikely to succeed.62 Taking aim at risks specific to each country is like-
lier to stop pure contagion.63
4.3 Downside Volatility andCorrelation

Tightening inEmerging Markets
Global investing offers a vivid illustration of the difference between volatil-
ity and correlation as indicators of investment risk. This difference becomes
evident in a simple comparison between developed and emerging markets.
Emerging markets have higher average returns and volatility than devel-
oped markets.64 Conventional mean-variance optimization compares
returns from emerging markets with their volatility to assess the value
of the trade-off of higher volatility for higher return. At sufficiently high
levels of volatility, an investor may not find it worthwhile to venture the

risk from exposure to emerging markets. But low correlations between
emerging markets and developed markets point with little ambiguity in
the direction of portfolio investment opportunities.65 Consequently, the
inclusion of emerging market assets in a mean-variance efficient port-
folio will significantly reduce portfolio volatility and increase expected
returns.66
Moreover, emerging markets are typically treated as the canonical
example of markets with negatively skewed returns.67 Variation among
emerging markets arises from their failure to be fully integrated with global
finance, due to market liquidity, political risk, and other [f]actors such
as taxes.68 Risk variables as diverse as size, value, momentum, and single-
sided and conventional measures of volatility and beta have an impact on
returns which varies from country to country.69
Evidence from emerging markets confirms the value of evaluating beta
differently, according to whether returns fall on the positive or negative
side relative to investor expectations. For instance, one study has found
that the conventional CAPM generated a positive and significant relation-
ship between portfolio betas and returns in only one market, Mexico.70
By contrast, Argentina, Brazil, and Chile react[ed] more to downs than
ups markets [sic], thus confirming a non-symmetrical conditional rela-
tionship between portfolio beta and returns.71 A contemporaneous study
of international stock markets throughout the 1990s similarly found a
positive relationship between beta and realized excess returns when
those returns exceed the risk-free rate of interest, and a correspondingly
negative relationship with beta when realized market excess returns
fall below the risk-free rate.72 This bifurcated relationship among interna-
tional stock markets also appears on a monthly basis. Specifically, beta is
positively related to returns during months when returns are positive, and
negatively related in months when returns are negative.73 Such demonstra-
tions that beta is significantly related to realized returns in both up and
down marketsalbeit in different directionsshow that beta is still a
good measure of risk and can still inform investors in making optimal
investment decisions.74
Even more importantly, emerging market evidence also suggests that
it is correlation tightening that drives nearly all of the difference between
downside beta and its conventional counterpart. In an illustrative 2002
study, Javier Estrada found that average downside beta in emerging mar-
kets was 50% larger than average beta.75 In Estradas survey of emerging
markets from 1988 through 2001, emerging markets exhibit[ed] more
80 J.M. CHEN
downside volatility than relative volatility.76 In 2007, Estrada published a

parallel survey of developed markets over the same 19882001 period.77
Notably, Estrada calculated standardized skewness for all markets in his
surveys. Contrary to the usual characterization of emerging markets, most
of the emerging markets in Estradas 2002 and 2007 studies were posi-
tively skewed.78
The strength of downside betas explanatory power prompted
Estrada to conclude that downside beta should replace beta as the single
explanatory variable of the cross section of stock returns.79 Through an
admittedly cursory meta-analysis that I conducted in Postmodern Portfolio
Theory,80 I have shown that the explanatory power of downside beta
resides predominantly in the ability of this statistic to report increases in
correlation among markets as they decline.
Put another way, beta combines the volatility of an asset relative to
a benchmark with the correlation between the asset and its benchmark.
Closer examination of beta in declining markets (especially in emerging
economies) shows that much of the increase in systematic risk when mar-
kets come under pressure arises from the correlation component. The
real basis for the price premium commanded by emerging market equities
therefore arises not from their higher absolute levels of volatility (contrary
to the implication of the casual description of VIX as a fear index), but
rather from the heightened vulnerability of this asset class to correlation
tightening in times of crisis.
4.4 Pricing andPredicting Correlation Risk

By no means is correlation tightening a unique property of emerging
markets. Quite the contrary. Because the Normal distribution con-
sistently underestimates the probability of (positive or negative) large
returns, extreme events, in both directions, are much more likely to
occur than a Normal distribution would predict.81 Correlation in crashes,
however, differs from correlation in booms. In their canonical study of
correlation under extreme conditions in France, Germany, the UK, and
the USA, Franois Longin and Bruno Solnik discovered that conditional
correlation among these developed markets strongly increases, but
only in bear markets.82 By contrast, conditional correlation does not
seem to increase in bull markets.83
Longin and Solniks empirical distinction between bear and bull markets
has potential implications for asset allocation and portfolio construction.84
One such implication affects the impact of asymmetric correlation on
extremely negative returns in otherwise weakly related markets. Where

the correlation structure of large returns is asymmetric, such that
[c]orrelation tends to decrease with the absolute size of the threshold for
positive returns but tends to increase for negative returns, it necessar-
ily follows that the probability of having large losses simultaneously on
two markets is much larger than would be suggested under the assump-
tion of multivariate normality.85
Actual evidence of differences in correlation on either side of mean
returns reinforces the supposition that [c]orrelation asymmetries are far
greater for extreme downward moves.86 Andrew Ang and Joseph Chen
have found a swing exceeding 3 percentage points (from 8.48 to 11.61%)
between (1) observed correlations between narrower, asset classbased
portfolios and the broader market, relative to correlations implied by a
normal distribution of returns, and (2) the average difference between
the same correlations [c]onditional on just downside moves.87
One study has suggested that correlation risk is directly priced in equity
options, both on individual equities and on indexes (such as VIX) that
track systematic, stochastic correlation risk across the market as a whole.88
A study of Chinese equity markets has reached similar conclusions.89
Moreover, stock market returns follow average correlation rather than
average variance or volatility.90 The fact that both individual options and
index options reflect correlation risk suggests a diminished role for either
component of volatilitythe systematic volatility of the overall market
or idiosyncratic volatility unique to individual securitiesin affecting the
cross section of returns.
Effective diversification requires low or even negative correlation
between assets. Idiosyncratic risk, the phenomenon that propels both
raw volatility and volatility relative to a broader benchmark, is the very
reason for diversification.91 Inasmuch as changes in market volatility
represent[] a deterioration in investment opportunities, investors as
[r]isk-averse agents [will] demand stocks that hedge against that risk.92
[A]n asset [that] tends to move downward in a declining m arket more
than it moves upward in a rising market is an unattractive asset to hold,
because it tends to have very low payoffs precisely when the wealth of
investors is so low.93 Consequently, it is not return or volatility alone, but
also correlation, that informs portfolio construction and asset allocation.
If [c]orrelations conditional on downside movements exceed cor-
relations implied by a normal distribution, or even correlations under
other market conditions, such that all stocks tend to fall together as
the market falls, the value of diversification may be overstated to the
82 J.M. CHEN
extent of the failure to tak[e] the increase in downside correlations into

account.94 Because they undermine diversification, portfolio theorys
standard response to idiosyncratic risk, changes in correlation under
stressed market conditions inject a distinct and dangerous sort of risk in its
own right, wholly apart from volatility. Correlation tightening can disrupt
managerial strategies that assume lower levels of correlation among asset
classes, or even negative correlation so that certain holdings can hedge
against declines in others.
In principle, this interest in diversification may, on its own, justify
investments in high-volatility assets whose returns lag behind those of the
broader universe of tradable instruments. A high-volatility component of
a broader portfolio may lower risk by provid[ing] insurance against bad
events, especially by delivering returns during sharp downturns.95 Stocks
with high upside potential relative to downside risk tend[] to pay off
more when an investors wealth is already high.96 Such stocks are not
as desirable as stocks that pay off when the market decreases.97 If assets
with high sensitivities to market volatility risk do indeed provide hedges
against market downside risk, then higher demand for assets with high
systematic volatility loadings should increase[] their price and lower[]
their average return.98
Actual evidence, however, sometimes points in the opposite direction.
Once again, we confront the confounding effects of the low-volatility
anomaly. High-volatility portfolios perform at their worst in precisely
those periods when an insurance payment would have been most wel-
come, such as the downturns of 197374 and 20002002, the crash of
1987, and the financial crisis that began in the fall of 2008.99 Rising
correlation in stressed markets poses an even greater threat to portfolios
consciously designed to weather ordinary volatility.100
Poor performance by high-beta portfolios has been confirmed in stud-
ies of the ten months with the sharpest downward market movements in
the six decades after 1932101 and of all periods of market distress, defined
as declines in the S&Ps 500 Index exceeding 10%, in the quarter cen-
tury preceding Fama and Frenchs attack on beta.102 In 2011, Malcolm
Baker, Brendan Bradley, and Jeffrey Wurgler not only found the low-beta
anomaly across diverse market conditions, but also detected a meaning-
ful difference between up and down markets.103 Although the low-beta
anomaly persists in all market environments on a capital asset pricing
model market-adjusted basis, it differs on either side of mean returns:
[H]igh-beta stocks earned higher (lower) total returns than did low-beta
stocks in up (down) markets.104
As between the discrete components of beta, it is correlation tightening

rather than relative volatility that drives the low-volatility anomaly. A 2006
study by Andrew Ang, Joseph Chen, and Yuhang Xing, controlling for
downside correlation, showed that stocks with high [downside] volatility
tend to have low returns, which is exactly opposite to the high aver-
age return effect predicted by standard portfolio theory.105 These, after
all, are the stocks with the highest levels of volatility and correspondingly
abysmal[] returns.106
By contrast, holding downside volatility constant so that correlation
accounts for changes in beta as returns fall below their expected levels
creates an average 5% annualized difference in returns on a month-after-
month basis between the tenth and first decile portfolios, sorted on past
downside correlation.107 If upside gains fail to offset outsized losses during
downturns, as one might expect in negatively skewed markets, these results
may explain at least some of the mechanics of the low-volatility anomaly.
4.5 Evidence Against aCorrelation Risk Premium

As between the two components of beta, correlation risk appears to
outweigh relative volatility. This conclusion arises from Javier Estradas
emerging market data (reviewed in 4.3) and from a broader examina-
tion of financial theory and evidence ( 4.4). Correlation risk does appear
to stand on its own and to carry a price premium wholly apart from the
premium associated with volatility. Estradas emerging market data leave
little room for attributing returns from those markets to any notion of
volatility, idiosyncratic or systematic, rather than correlation tightening in
declining markets.
One source, however, suggests otherwise. Zhanhui Chen and Ralitsa
Petkova have argued that the low-volatility anomaly can be explained by dif-
ferences in idiosyncratic volatility, wholly apart from the average correlation
component of aggregate market variance.108 After evaluating portfolios
sorted by size and idiosyncratic volatility [f]or the period from July 1996
to December 2009, Chen and Petkova concluded that only exposure to
average variance (and not correlation) is priced in stock market returns.109
Chen and Petkova emphasize that the price of average variance tracks the
size and value factors and, critically, is negative.110 Finding that average
variance predicts lower future market returns and higher future market vari-
ance, their study concluded that [e]xposure to average correlation is not
an important determinant of average returns.111 Whereas high average
variance unequivocally worsens the investors risk-return trade-off and
84 J.M. CHEN
commands a risk premium, high correlation more ambiguously predicts

[both] higher future market returns and higher future market variance.112
Part of the problem may stem from Chen and Petkovas interpreta-
tion of contemporaneous research by Joost Driessen, Pascal Maenhout,
and Grigory Vilkov into the pricing of correlation risk within options on
equity indexes and individual equities.113 Chen and Petkova read Driessen,
Maenhout, and Vilkov as concluding that individual options are not sig-
nificantly exposed to correlation risk.114 But Driessen, Maenhout, and
Vilkov found evidence of priced correlation based on prices of index and
individual variance risk.115 They likewise concluded that [c]orrelation
risk exposure explains the cross-section of index and individual option
returns well.116 It is therefore problematic for Chen and Petkova to align
themselves with Driessen, Maenhout, and Vilkov by proclaiming that
average correlation is not priced in the cross section of assets sorted by
idiosyncratic risk.117
Despite these differences, I draw two lessons from Chen and Petkovas
findings. First, Chen and Petkova apply no single-sided measure of risk;
they adhere to average variance and average correlation. Portfolios orga-
nized according to traits defined by average returns give potentially
misleading guidance on asset prices, if only because broad sampling
understates dispersion in the summary statistics.118 Some consideration of
differences in idiosyncratic volatility, to say nothing of systematic volatility
or of correlation, on either side of mean returns may clarify the meaning
of their results. Volatility and correlation differ in up and down markets.
Absent some consideration of those differences, inferences drawn about
the impact of individual measures of risk on returns and prices are likely to
be premature. At the very least, the analysis is incomplete.
Second, regardless of the proper interpretation of their results, Chen
and Petkova have confirmed the basic methodology outlined in Chapters 3
through 6 of Postmodern Portfolio Theory, that of decompos[ing] aggre-
gate market variance as the product of average stock variance [times]
average stock correlation.119 Whatever the true significance of their
results may be, Chen and Petkovas effort to disentangle[] the here-
tofore combined effects of average variance and average correlation on
stock returns much more clearly highlights the role of average vari-
ance and correlation in explaining the low-volatility anomaly.120 To the
best of [their] knowledge or mine, these effects have not been docu-
mented before.121
Notes
1. William J.Bernstein, The Intelligent Asset Allocator: How to Build Your
Portfolio to Maximize Returns and Minimize Risk 174 (2000).
2. Mark Grinblatt & Sheridan Titman, Financial Markets and Corporate
Strategy 392 (2d ed. 2001).
3. Id.
4. Lu Zhang, The Value Premium, 60J.Fin. 67103, 67 (2005).
5. Bernstein, supra note 1, at 174.
6. Id. See generally Benjamin Graham & David L. Dodd, Security Analysis
548558 (6th ed. 2008).
7. David Morelli, Beta, Size, Book-to-Market Equity and Returns: A Study
Based on UK Data, 17J.Multinatl Fin. Mgmt. 257272, 267 (2007). See
generally K.C.Chan & Nai-Fu Chen, Structural and Return Characteristics
of Small and Large Firms, 46J.Fin. 14671484 (1991).
8. See Malcolm Baker, Brendan Bradley & Jeffrey Wurgler, Benchmarks as
Limits to Arbitrage: Understanding the Low-Volatility Anomaly, 67:1 Fin.
Analysts J. 4054, 46 (Jan./Feb. 2011).
9. Zhang, The Value Premium, supra note 4, at 67.
10. Andrew Ang, Joseph Chen & Yuhang Xing, Downside Risk, 19 Rev. Fin.
Stud. 11911239, 1193 (2006).
11. Baker, Bradley & Wurgler, supra note 8, at 40.
12. Id. (emphasis in original).
13. Robert A. Haugen & A. James Heins, Risk and the Rate of Return on
Financial Assets: Some Old Wine in New Bottles, 10J.Fin. & Quant. Analysis
775784, 782 (1975) (emphasis added).
14. Andrew Ang, Robert J.Hodrick, Yuhang Xing & Xiaoyan Zhang, The Cross-
Section of Volatility and Expected Returns, 61J.Fin. 259299, 296 (2006);
accord Baker, Bradley & Wurgler, supra note 8, at 43; see also Andrew Ang,
Robert J. Hodrick, Yuhang Xing & Xiaoyan Zhang, High Idiosyncratic
Volatility and Low Returns: International and Further U.S.Evidence, 91J.Fin.
Econ. 123 (2009).
15. See, e.g., David C.Blitz & Pim van Vliet, The Volatility Effect: Lower Risk
Without Lower Return, 34:1J.Portfolio Mgmt. 102113 (Fall 2007); Roger
Clarke, Harindra de Silva & Steven Thorley, Minimum-Variance Portfolios in
the U.S.Equity Market, 33:1J.Portfolio Mgmt. 1024 (Fall 2006); Andrea
Frazzini & Lasse Heje Pedersen, Betting Against Beta, 111 J. Fin. Econ.
125 (2014); Robert A.Haugen & Nardin L.Baker, The Efficient Market
Inefficiency of Capitalization-Weighted Stock Portfolios, 17:3 J. Portfolio
Mgmt. 3540 (Spring 1991); cf. Javier Estrada & Ana Paula Serra, Risk and
Return in Emerging Markets: Family Matters, 15J.Multinatl Fin. Mgmt.
257272, 267 (2004) (finding, counterintuitively, that low risk portfo-
lios in emerging markets outperformhigh-risk portfolios over 20years,
at least when portfolios are rebalanced every 10years).
86 J.M. CHEN

17. Ang, Hodrick, Xing & Zhang, The Cross-Section of Volatility and Expected
Returns, supra note 14, at 297.
18. Id. at 260 (emphasis added); see also id. (If the price of aggregate volatility
is negative, stocks with large, positive sensitivity should have low average
returns.).
19. Baker, Bradley & Wurgler, supra note 8, at 43; see also Jonathan Fletcher,
On the Conditional Relationship Between Beta and Return in International
Stock Returns, 9 Intl Rev. Fin. Analysis 235245, 240 (2000).
20. Edward H.Bowman, A Risk/Return Paradox for Strategic Management,
21 Sloan Mgmt Rev. 1733 (1980).
21. See Edward H.Bowman, Risk Seeking by Troubled Firms, 23 Sloan Mgmt.
Rev. 3342 (1982); Edward H. Bowman, Content Analysis of Annual
Reports for Corporate Strategy and Risk, 14 Interfaces 6171 (1984).
22. Manuel Nez Nickel & Manuel Cano Rodriguez, A Review of Research on
the Negative Accounting Relationship Between Risk and Return: Bowmans
Paradox, 30 Omega 118, 1 (2002). Omega describes itself as The
International Journal of Management Science.
23. Id. at 2.
24. Id.
25. Id.
26. Bromiley, Miller & Rau, supra note 95 (Chapter 2), at 259.
27. Id.
28. Id.
29. Nickel & Rodriguez, supra note 22, at 2; see also Gerry McNamara & Philip
Bromiley, Risk and Return in Organizational Decision Making, 42 Acad.
Mgmt. J. 330339, 330 (1999).
30. See, e.g., Sayan Chatterjee, Michael H. Lubatkin & William S. Schulze,
Toward a Strategic Theory of Risk Premium: Moving Beyond CAPM, 24
Acad. Mgmt. Rev. 556567 (1999); Avi Fiegenbaum & Howard Thomas,
Dynamic and Risk Management Perspectives on Bowmans Risk-Return
Paradox for Strategic Management: An Empirical Study, 7 Strat. Mgmt. J.
394407 (1986); Rajaram Veliyath & Stephen P.Ferris, Agency Influences
on Risk Reduction and Operating Performance: An Empirical Investigation
Among Strategic Groups, 39J.Bus. Research 219230 (1997).
31. See Moon K. Kim & Badr E. Ismail, An Accounting Analysis of the Risk-
Return Relationship in Bull and Bear Markets, 7 Rev. Fin. Econ. 173182
(1998).
32. See generally Chen, Postmodern Portfolio Theory, supra note 1 (Chapter 1),
4.17.10, at 40151.
33. See William Beaver, Paul Kettler & Myron Scholes, The Association Between
Market Determined and Accounting Determined Risk Measures, 45
Accounting Rev. 654682 (1970); William Beaver & James Manegold, The
Association Between Market-Determined and Accounting-Determined
Measures of Systematic Risk: Some Further Evidence, 10 J. Fin. & Quant.
Analysis 231284 (1975); Robert G.Bowman, The Theoretical Relationship
Between Systematic Risk and Financial (Accounting) Variables, 34 J. Fin.
617630 (1979).
34. See, e.g., Joachim Henkel, The Risk-Return Paradox for Strategic
Management: Disentangling True and Spurious Effects, 30 Strat. Mgmt. J.
287303 (2009); Benjamin M. Oviatt & Alan D. Bauerschmidt, Business
Risk and Return: A Test of Simultaneous Relationships, 37 Mgmt. Sci.
14051423 (1991); Robert M.Wiseman & Philip Bromiley, Risk-Return
Associations: Paradox or Artifact? An Empirically Tested Explanation, 12
Strat. Mgmt. J. 231241 (1991).
35. See, e.g., Torben J.Andersen, Jerker Denrell & Richard A.Bettis, Strategic
Responsiveness and Bowmans Risk-Return Paradox, 28 Strat. Mgmt. J.
407429 (2007); David B. Jemison, Risk and the Relationship Among
Strategy, Organizational Processes, and Performance, 33 Mgmt. Sci. 1087
1101 (1987).
36. See generally Nickel & Rodriguez, supra note 22, at 45.
37. M.A. Bellelah, M.O. Bellelah, H. Ben Ameur & R. Ben Hafsia, Does the
Equity Premium Puzzle Persist During Financial Crisis? The Case of the
French Equity Market, Research in Intl Bus. & Fin. (2015) (preprint at 14)
(available at http://dx.doi.org/10.1016/j.ribaf.2015.02.018).
38. Guy Kaplanski, Traditional Beta, Downside Risk Beta and Market Risk
Premiums, 44 Q.Rev. Econ. & Fin. 636653, 637 (2004).
39. Levy, CAPM in the 21st Century, supra note 41 (Chapter 1), at 4 (emphasis
in original).
40. See Martin L. Leibowitz, Anthony Bova & P. Brett Hammond, The
Endowment Model of Investing: Return, Risk, and Diversification 14 (2010)
(defining beta as the correlation between the asset (or portfolio) return and
the market return, multiplied by the ratio of their volatilities); Michael
B.Miller, Mathematics and Statistics for Financial Risk Management 198,
213, 292 (2d ed. 2014) (defining beta as the product of correlation between
the returns on two assets and the ratio of their volatilities); Shannon P.Pratt
& Roger J.Grabowski, Cost of Capital: Applications and Examples 305306
(4th ed. 2010).
41. See generally Chicago Board Options Exchange, The CBOE Volatility
IndexVIX (2009) (available online at http://www.cboe.com/micro/
vix/vixwhite.pdf).
42. See Menachem Brenner & Dan Galai, New Financial Instruments for
Hedging Changes in Volatility, 45:4 Fin. Analysts J. 6165 (July/Aug.
88 J.M. CHEN
1989); Menachem Brenner & Dan Galai, Hedging Volatility in Foreign

Currencies, 1:1 J. Derivatives 5358 (Fall 1993); Matthew T. Moran,
Review of the VIX Index and VIX Futures, 7:5J.Indexes 1619 (Oct./Nov.
2004); Robert E.Whaley, Trading Volatility: At What Cost?, 40:1J.Portfolio
Mgmt. 95108 (Fall 2013); Robert E. Whaley, Understanding the VIX,
35:3J.Portfolio Mgmt. 98105 (Spring 2009).
43. See Robert E. Whaley, The Investor Fear Gauge, 26:3 J. Portfolio Mgmt.
1217 (Spring 2000).
44. Daniel G.Goldstein & Nicholas Nassim Taleb, We Dont Quite Know What
We are Talking About When We Talk About Volatility, 33:4 J. Portfolio
Mgmt. 8486 (Summer 2007).
45. Yueh-Neng Lin & Ken Hung, Is Volatility Priced?, 9 Annals Econ. & Fin.
3975, 41 (2008). See generally David S.Bates, Post-87 Crash Fears in the
S&P Futures Options Market, 94J.Econometrics 191238 (2000); Bjrn
Eraker, Do Stock Prices and Volatilities Jump? Reconciling Evidence from Spot
and Option Prices, 59J.Fin. 13671404 (2004); Guojun Wu & Zhijie Xiao,
A Generalized Partially Linear Model of Asymmetrical Volatility,
9J.Empirical Fin. 287319 (2002).
46. See Kahneman, Thinking, Fast and Slow, supra note 11 (Chapter 1), at 281.
47. See, e.g., Gary Charness, David Masclet & Marie Claire Villeval, The Dark
Side of Competition for Status, 60 Mgmt. Sci. 3855 (2014); Simon Dato &
Petra Nieken, Gender Differences in Competition and Sabotage, 100J.Econ.
Behav. & Org. 6480 (2014); Thomas Dohmen, Armin Falk, Klaus
Flessbach, Uwe Sunde & Bernd Weber, Relative Versus Absolute Income, Joy
of Winning, and Gender: Brain Imaging Evidence, 95J.Pub. Econ. 279
285 (2011); Camellia M. Kuhnen & Agnieszka Tymula, Feedback, Self-
Esteem, and Performance in Organizations, 58 Mgmt. Sci. 94113 (2012);
Mark Sheskin, Paul Bloom & Karen Wynn, Anti-Equality: Social Comparison
in Young Children, 130 Cognition 152156 (2014).
48. See Brad M. Barber & Terrance Odean, All That Glitters: The Effect of
Attention and News on the Buying Behavior of Individual and Institutional
Investors, 21 Rev. Fin. Stud. 785818 (2008).
49. John Maynard Keynes, A Treatise on Probability 23 (1921); accord Robert
J.Shiller, Irrational Exuberance 253 (3d ed. 2015).
50. Shiller, supra note 49, at 253.
51. Leibowitz, Bova & Hammond, supra note 40, at 265.
52. See Malcolm P. Baker & Jeffrey Wurgler, Comovement and Predictable
Relations Between Bonds and the Cross-Section of Stocks, 2 Rev. Asset Pricing
Stud. 5787 (2012).
53. Leibowitz, Bova & Hammond, supra note 40, at 265.
54. See, e.g., Thomas J. Flavin, Clara E. Morley & Ekaterini Panopoulou,
Identifying Safe Haven Assets for Equity Investors Through an Analysis of the
Stability of Shock Transmission, 33J.Intl Fin. Mkts., Insts. & Money 137
154 (2014); cf. John Francis T.Diaz, Do Scarce Precious Metals Equate to
Safe Harbor Investments? The Case of Platinum and Palladium, 2016 Econ.
Research Intl 2361954 (arguing that platinum but not palladium provides
a financial safe haven because only platinum has a symmetric volatility
response to shocksbecause negative and positive shocks have equal effects
onreturns and volatilities).
55. See generally, e.g., Mark Burgin & Gunter Meissner, Negative Probabilities
in Financial Modeling, 58 Wilmott 6065 (2012); Carina Moselund Jensen
& Morten Spange, Interest Rate Pass-Through and the Demand for Cash at
Negative Interest Rates, Danmarks Nationalbank Monet. Rev., 2d quarter
2015, at 5566.
56. See Lasse Heje Pedersen, When Everyone Runs for the Exit, 5 Intl J.Cent.
Banking 177179 (2009).
57. See generally Geert Bekaert, Campbell R. Harvey & Angela Ng, Market
Integration and Contagion, 78J.Bus. 3969 (2005).
58. Thomas J. Flavin & Ekaterini Panopoulou, Detecting Shift and Pure
Contagion in East Asian Equity Markets: A Unified Approach, 15 Pac. Econ.
Rev. 401421, 401 (2010).
59. Id. See generally Toni Gravelle, Maral Kichian & James Morley, Detecting
Shift-Contagion in Currency and Bond Markets, 68J.Intl Econ. 409423
(2006).
60. Flavin & Panopoulou, supra note 58, at 401402.
61. Id. at 402. See generally Marcello Pericoli & Massimo Sbracia, A Primer in
Financial Contagion, 17J.Econ. Surveys 571608 (2003).
62. Flavin & Panopoulou, supra note 58, at 402.
63. See id.
64. Campbell R.Harvey, Predictable Risk and Returns in Emerging Markets, 8
Rev. Fin. Stud. 773816, 780 (1995).
65. Id. at 781; see also Vihang R.Erruzna, Emerging Markets: New Opportunity
for Improving Global Portfolio Performance, 39:5 Fin. Analysts J. 5158
(Sept./Oct. 1983).
66. Harvey, Predictable Risk and Returns in Emerging Markets, supra note 64,
at 811.
67. See, e.g., Christopher J.Adcock & Karl Shutes, An Analysis of Skewness and
Skewness Persistence in Three Emerging Markets, 6 Emerging Mkts. Rev.
396418 (2005) (reporting significant skewness in daily returns on stocks
in Kenya, Poland, and the Czech Republic); Soonsung Hwang & Christian
S.Pedersen, Asymmetrical Risk Measures When Modelling Emerging Markets
Equities: Evidence for Regional and Timing Effects, 5 Emerging Mkts. Rev.
109128 (2004).
90 J.M. CHEN
68. Harvey, Predictable Risk and Returns in Emerging Markets, supra note 64, at
787; see also id. at 801 (concluding that local information materially affects
returns in emerging markets, while most of the variation in developed markets
is driven by global information variables rather than local information);
Geert Bekaert & Campbell R. Harvey, Emerging Equity Market Volatility,
43J.Fin. Econ. 2977 (1997); Geert Bekaert & Campbell R.Harvey, Time-
Varying World-Market Integration, 50J.Fin. 403444 (1995). On the impact
of foreign exchange markets on stock prices, especially in emerging markets,
see Geert Bekaert & Robert J.Hodrick, Characterizing Predictable Components
in Excess Returns on Equity and Foreign Exchange Markets, 47J.Fin. 467509
(1992); Richard Roll, Industrial Structure and the Comparative Behavior of
International Stock Market Indexes, 47J.Fin. 341 (1992). On the impact of
foreign speculators, see Geert Bekaert & Campbell R. Harvey, Foreign
Speculators and Emerging Equity Markets, 55J.Fin. 565613 (2000).
69. Estrada & Serra, supra note 15, at 268.
70. Eduardo A. Sandoval & Rodrigo N. Saens, The Conditional Relationship
Between Portfolio Beta and Return: Evidence from Latin America, 41
Cuadernos de Economa 6589, 75 (2004).
71. Id. at 82.
72. Gordon Y.N.Tang & Wai C.Shum, The Conditional Relationship Between
Beta and Returns: Recent Evidence from International Stock Markets, 12
Intl Bus. Rev. 109126, 110 (2003).
73. See Fletcher, On the Conditional Relationship Between Beta and Return in
International Stock Markets, supra note 19.
74. Tang & Shum, The Conditional Relationship Between Beta and Returns,
supra note 72, at 110.
75. Javier Estrada, Systematic Risk in Emerging Markets: The D-CAPM, 3
Emerging Mkts. Rev. 365377, 374 (2002).
76. Id.; see also id. at 378 (table A1) (tabulating summary statistics).
77. See Javier Estrada, Mean-Semivariance Behavior: Downside Risk and Capital
Asset Pricing, 16 Intl Rev. Econ. & Fin. 169185, 175176 (2007) (exhibit 1).
78. See Estrada, Systematic Risk in Emerging Markets, supra note 75, at 378;
Estrada, Downside Risk and Capital Asset Pricing, supra note 77, at 175176.
79. Estrada, Downside Risk and Capital Asset Pricing, supra note 77, at 177.
80. See Chen, Postmodern Portfolio Theory, supra note 1 (Chapter 1), 7.6, at
122126.
81. Felipe M. Aparicio & Javier Estrada, Empirical Distributions of Stock
Returns: European Securities Markets, 199095, 7 Eur. J.Fin. 121, 1516
(2001). See generally, e.g., Joseph Chen, Harrison Hong & Jeremy C.Stein,
Forecasting Crashes: Trading Volume, Past Returns, and Conditional
Skewness in Stock Prices, 61J.Fin. Econ. 345391 (2001); Eugene F.Fama,
The Behavior of Stock Market Prices, 38 J. Bus. 34105 (1965); Amado
Peir, Skewness in Financial Returns, 23J.Banking & Fin. 847862 (1999);

Michael A. Simkowitz & William L. Beedles, Diversification in a Three-
Moment World, 13 J. Fin. & Quant. Analysis 927941 (1978); J. Clay
Singleton & John Wingender, Skewness Persistence in Common Stock Returns,
21J.Fin. & Quant. Analysis 335341 (1986).
82. Franois Longin & Bruno Solnik, Extreme Correlation of International
Equity Markets, 56J.Fin. 649676, 650 (2001).
83. Id. at 651.
84. Id.
85. Id. at 669670.
86. Andrew Ang & Joseph Chen, Asymmetric Correlations of Equity Portfolios,
63J.Fin. Econ. 443494, 469 (2002).
87. Id.
88. See Joost Driessen, Pascal J. Maenhout & Grigory Vilkov, The Price of
Correlation Risk: Evidence of Equity Options, 64J.Fin. 13771496 (2009).
89. See Yiwen Deng, Chen Liu & Zhenlong Zheng, The Price of Correlation
Risk: Evidence from Chinese Stock Market, 4 China Fin. Rev. Intl 343359
(2014).
90. See Joshua M. Pollet & Mungo Wilson, Average Corelation and Stock
Market Returns, 96J.Fin. Econ. 364380 (2010).
91. See generally John Y.Campbell, Martin Lettau, Burton G.Malkiel & Yexiao
Xu, Have Individual Stocks Become More Volatile? An Empirical Exploration
of Idiosyncratic Risk, 56J.Fin. 143 (2001).
Returns, supra note 14, at 260. See generally John Y.Campbell, Intertemporal
Asset Pricing Without Consumption Data, 83 Am. Econ. Rev. 487512
(1993); Joseph Y.Campbell, Understanding Risk and Return, 104J.Pol.
Econ. 298345 (1996).
93. Ang, Chen & Xing, supra note 10, at 1191.
94. Ang & Chen, supra note 86, at 444 (reporting an 11.6 percent increase in
downside correlation); see also id. at 450 (showing graphically the economic
cost of ignoring or miscalculating downside correlation).
95. Baker, Bradley & Wurgler supra note 8, at 43.
96. See Ang, Chen & Xing, supra note 10, at 1199.
97. Id.
Returns, supra note 14, at 260. See generally Gurdip Bakshi & Nikunj
Kapadia, Delta-Hedged Gains and the Negative Market Volatility Risk
Premium, 16 Rev. Fin. Stud. 527566 (2003).
92 J.M. CHEN
100. See Leibowitz, Bova & Hammond, supra note 40, at 235, 265.
101. See Louis K.C. Chan & Josef Lakonishok, Are Reports of Betas Death
Premature?, 19:4J.Portfolio Mgmt. 5162 (Summer 1993).
102. See Kevin Grundy & Burton G. Malkiel, Reports of Betas Death Have
Been Greatly Exaggerated, 22:3J.Portfolio Mgmt. 3644 (Spring 1996);
cf. Fama & French, The Cross-Section of Expected Stock Returns, supra
note 5 (Chapter. 3).
104. Id.; cf. R. Burr Porter, Semivariance and Stochastic Dominance: A
Comparison, 64 Am. Econ. Rev. 200204 (1974) (finding that portfolios
with below-target semivariance showed statistic dominance over their
benchmark, but that porfolios with below-mean semivariance did not).
105. Ang, Chen & Xing, supra note 10, at 1228.
106. See sources cited supra note 14.
107. Ang, Chen & Xing, supra note 10, at 1228 n.15.
108. See Zhanhui Chen & Ralitsa Petkova, Does Idiosyncratic Volatility Proxy
for Risk Exposure, 25 Rev. Fin. Stud. 27452787, 2746 (2012).
110. Id.
111. Id. at 2747.
112. Id.
113. See Driessen, Maenhout & Vilkov, supra note 88.
114. Chen & Petkova, supra note 108, at 2747 (emphasis added); see also id.
at 2750 (observing that Driessen, Maenhout, and Vilkov, supra note 88,
show that individual options are not exposed to correlation risk, while
index options are.).
115. Driessen, Maenhout & Vilkov, supra note 88, at 1377 (abstract) (empha-
sis added).
116. Id. (abstract) (emphasis added).
117. Chen & Petkova, supra note 108, at 2747.
118. See Kent Daniel & Sheridan Titman, Evidence on the Characteristics of
Cross Sectional Variation in Stock Returns, 52J.Fin. 133 (1997).
119. Chen & Petkova, supra note 108, at 2746. See generally Chen, Postmodern
Portfolio Theory, supra note 1 (Chapter. 1), 4.17.10, at 40151.
120. Chen & Petkova, supra note 108, at 2747.
121. Id.
CHAPTER 5
The Intertemporal Capital Asset Pricing

Model: Hedging Investment Risk
Across Time
5.1 The Intertemporal Capital Asset Pricing

Model
The previous chapter traced the nuances of the low-volatility anomaly
across behavioral space. Specifically, it explored whether examining beta
on either side of mean returns or separately evaluating its relative volatility
and correlation components might offer insight into why low-volatility
stocks offer higher returns. An even fuller explanation of the mechanics
of the low-volatility anomaly lies in the work of John Campbell. That
explanation, in turn, traces its origins to Robert Mertons intertemporal
CAPM.1
Intertemporal CAPM differs from its conventional counterpart in
explicitly acknowledging the possibility that investors must account for
consumption decisions not merely in the present, but across an indefinite
time horizon based on the relationship between current period returns
and returns that will be available in the future.2 Intertemporal asset pric-
ing theory shares a key trait with time series analysis:3 Both methods seek
to counteract the unrealistic assumption that investors live for only one
period and, as such, design strictly static portfolios.4 These techniques
reject one of the suppositions on which the conventional CAPM rests:
Agents preferences depend only on the mean and variance of consump-
tion of a single good at a single date.5

DOI10.1007/978-3-319-32711-2_5
94 J.M. CHEN
Indeed, it is quite proper to question whether time series estimates of rel-

ative risk aversion and other intertemporal economic analyses are directly
comparable to cross-section estimates.6 Studies of that sort compare the
decisions of low and high wealth individuals at a point in time in order
to infer their degree of risk aversion.7 They therefore assume that the
degree of risk aversion in any one person stays constant over time.8 By
contrast, intertemporal CAPM, while drawing no firm conclusions about
differences in risk aversion among individuals on the basis of differences
in wealth, consciously assumes that any one person over time may have
different aversions and affinities toward risk.
At a minimum, intertemporal CAPM extends the reach of portfolio
theorys asset allocation and pricing solutions to future market conditions
not adequately anticipated by current models of risk.9 Mertons approach
to intertemporal asset pricing also informs a more general and dynamic
approach to prices and interest rates across the entire economy.10 General
equilibrium models bridge the intellectual progression from the Arrow
Debreu model of complete markets11 to the concept of risk neutrality
animating the BlackScholes option pricing model.12 Relaxing the assump-
tions in continuous-time general equilibrium modelsso as to accom-
modate variations such as the presence of multiple but finite risky assets,
the possibility of recessionary and expansionary economic states, and
differences in the magnitude of shocks to the economyfacilitates even
broader and more realistic application of intertemporal asset pricing to
broader economic questions.13
Because it explicitly distinguishes between current and future portfo-
lios, intertemporal CAPM represents a conceptual departure from two
of the foundational ideas of modern portfolio theory, James Tobins
notion of liquidity preference and the closely related mutual fund separa-
tion theorem. In concert, these expressions of modern portfolio theory
suggest that a single, mean-variance optimized portfolio along the effi-
cient frontier suffices to hold the entirety of any investors wealth and
to address all of her or his financial needs. Just as investors can use cash
(or leverage) to balance risk against liquidity needs,14 any portfolio can
be decomposed into separate, individually efficient mutual funds or sub-
portfolios.15 In principle, the mean and variance of any efficient portfolio
can be duplicated through the combination of two other funds or portfo-
lios lying along the efficient frontier. An investor may achieve perfect effi-
ciency at any point of the frontier by investing in a combination of these
two and only these two funds. Lying at the heart of [modern portfolio]
THE INTERTEMPORAL CAPITAL ASSET PRICING MODEL 95
theory, the separation theorem holds that all investorsirrespective of

preferencewill choose portfolios that are two simplicial combinations
of two given funds or portfolios.16 The separation theorem implies that
the set of efficient portfolios is contained in a line and is spanned by any
two efficient portfolios.17 This assumption is quite fragile; any departure
from the presumed convexity of efficient portfolios defeats the separation
theorem.18
Taking account of the possibility that investors might have distinct pref-
erences for risk at different times, even life stages, intertemporal CAPM
explicitly contemplates two distinct components in an investors demand
for a financial asset. The impact of current investment choices on future
lifestyles forms the basis of the life cycle model of consumption, invest-
ment, and saving,19 which will figure prominently in this books discus-
sion of behaviorally sensitive solutions to the equity premium puzzle.20
Part of the investors demand reflects immediate, single-period demand
for a risky asset as a function of the investors tolerance for risk.21 The
other component reflects the investors demand for the asset as a vehicle
to hedge against unfavorable shifts in the investors opportunity set.22
Since the intertemporal CAPM adopts the simplifying assumption that
investors derive all their income from capital gains sources,23 an unfavor-
able shift in future investment opportunity translates into a loss of future
consumption.
The smoothing of consumption across time that this model implies
is not the traditional type of maintenance of a constant level of consump-
tion, but rather an attempt to minimize the (unanticipated) variability
in consumption over time.24 The financial manifestation of the general
economic equivalence of future investment and future consumption takes
its most obvious form in the standard practice of enabling investors to
choose between reinvesting dividends and taking them as cash.25 Future
volatility impairs future investment or consumption, and risk-averse inves-
tors are willing to pay some price to hedge against risks spanning a time
horizon beyond the immediate future.26
The simplest hedge against the risk of a deteriorating investment
opportunity set consists of hold[ing] stocks that have high returns
when the market volatility is higher than expected.27 The intertemporal
CAPM predicts that [h]igh demand for these stocks, whose returns
are highly correlated with innovations in market volatility, will lower
[their] required returns.]28 For some other investors, protection against
that decline takes the form of portfolio insurance, which ensures that
96 J.M. CHEN
the value of [an] insured portfolio, at some specified date, will not fall
below some specified level, even though it is well understood that under
almost all circumstances, a simple portfolio insurance strategy is inconsis-
tent with expected-utility maximization.29
Sequence of returns risk, a particularly pernicious threat to retirement
security,30 especially when retirement coincides with a particularly sharp
market downturn,31 has inspired the creative use of put options and other
hedges to reduce vulnerability to poor returns at the beginning of an
investors retirement.32 Structured financial products perform this func-
tion, at a price, for other investors.33 Still other investors maintain distinct
subportfolios that accept reduced returns in exchange for a hedge against
future declines in investment or consumption opportunities. The inter-
temporal CAPM can be readily adapted to address this option.
Under the intertemporal CAPM, the return on an investors overall
portfolio return must reflect the systematic risk of the overall market as
well as the systematic risk of the hedged portfolio:34
a p = rf + b m ( a m - rf ) + b h ( a h - rf )

where a designates a vector of expected returns, rf indicates the risk-

free rate, and subscripts p, m, and h identify, respectively, either the vector
of expected returns or the value for beta associated with the investors
overall portfolio, the broader market in its current state, and the investors
future-oriented, hedged portfolio. The expected return at equilibrium in
this model compensates investors for bearing market (systematic) risk
and for bearing the risk of unfavorable shifts in the [future] investment
opportunity set.35
Without loss of generality, we may recalibrate the intertemporal
CAPMs measure of future consumption to reflect aggregate real con-
sumption as computed [by] an instantaneously additive price index,36
which would provide a more accurate gauge of future consumption
than standard market portfolio measures, which typically fail to cover
human capital, real estate, and consumer durables.37 This is yet another
variation on the theme of Richard Rolls second critique, which assailed
the conventional CAPM for its failure to reflect the full portfolio of eco-
nomic resourcesfinancial or otherwiseavailable to investors.38
Other potential modifications abound. Conceptually, nothing prevents
the application of intertemporal CAPM to multifactor models such as
Eugene Fama and Kenneth Frenchs three-factor model,39 the four-factor

extension incorporating Mark Carharts momentum principle, or the four-
moment CAPM.Chapter 3 of Postmodern Portfolio Theory discusses Fama
and Frenchs three-factor model as the foundational theory of contempo-
rary mathematical finance.40 Momentum figures prominently in 12.2
and 12.3 of this book, which treats the FamaFrench three-factor model
as a special case of the FamaFrenchCarhart four-factor model. Section
3.2 specified four-moment CAPM and explored its behavioral significance.
At the risk of being tedious, I shall restate the four-moment CAPM as
a Taylor series expansion of r=ln(1+x)at(x=):41
x-m s2 g 1s 3 b 2s 4
r = ln (1 + m ) + - + - + o ( m5 )
1 + m 2 (1 + m )2 3 (1 + m )3 4 (1 + m )4

If an investors overall portfolio consists of a market-oriented core

designed to maximize current return and a hedged portfolio providing
protection against shortfalls in future investment or consumption oppor-
tunities, such that the expected return from the overall portfolio and its
two components are designated rp, rm, and rh, intertemporal asset
pricing theory would define rp as the weighted sum of the Taylor series
expansions of rm and rh:
rp = wm rm + wh rh

where coefficients wm and wh, respectively, designate the weights of the

market-oriented core and the hedged portfolio.
Since Mertons formulation of the basic model in 1973, the intertem-
poral CAPM has figured prominently in studies of asset allocation and
asset pricing.42 Retirement investing represents perhaps the simplest and
(at least for individual investors) most practically important application
of intertemporal CAPM. Saving for retirement is essentially a problem
of portfolio design and asset allocation over the time horizon of a career.
Projection bias, however, may lead a worker to miscalculate her or his
future consumption preferences.43 As a result, that worker may decide
over time to consumer more and save less than she or he might have origi-
nally planned.44
98 J.M. CHEN
Poor investment outcomes traceable to a workers youth, whether

attributable to a failure to save, imprudent asset allocation, or just bad
luck, can in principle be offset by future labor. The option of working
more each day or deeper into years otherwise reserved for retirement may
be viewed as a component of a lifelong portfolio. But with old age comes
less health, less vigor, less employer demand, and (brutally enough) fewer
years of life. In the long run, it bears repeating, we are all dead.45 An
older worker simply does not have the option to work more, acquire skills,
or switch professions. Evaluating these realities through as a problem of
intertemporal portfolio choice confirms the standard investment advice,
epitomized by the target-date retirement fund, that individual investors
should scale back on risk as they age.46
From the 1990s onward, John Campbell has formulated far more
sophisticated implementations of Mertons original model of intertempo-
ral asset pricing.47 One such effort divides beta in a uniquely and simul-
taneously spatial and temporal way. In so doing, Campbells extension
of the intertemporal CAPM offers a behaviorally sensitive way of under-
standing the low-volatility anomaly.
5.2 Bad Beta, Good Beta

In their 2004 article, Bad Beta, Good Beta, John Campbell and Tuomo
Vuolteenaho applied intertemporal asset pricing theory to address the
challenge that Eugene Fama and Kenneth French had posed to the
conventional CAPMnamely, that small stocks and value stocks have
delivered higher average than their betas can justify.48 For the intertem-
poral CAPM to outperform the traditional static CAPM in explaining
why value stocks outperform growth stocks, particularly among smaller
stocks, abnormally low returns on small growth stocks must be off-
set by intertemporal hedging value in protecting future investment
opportunities.49
We can describe this state of affairs in terms better suited to the per-
spective of a risk-averse investor. The fraction of the equity market for
which Fama and Frenchs three-factor model predicts lower returns rela-
tive to small-cap and value stocks may be considered a component of an
altogether distinct subportfolio. The primary purpose of that subportfolio
is not maximum return for a given level of volatility (as classical methods
of mean-variance optimization under modern portfolio theory would dic-
tate). Rather, lower-return securities provide a hedge against unanticipated
erosion in the ability of the market portfolio to satisfy future investment
or consumption. If large-cap and growth stocks offer risk reduction on

an intertemporal basis relative to the higher returns (and correspondingly
higher risk) of small-cap and value stocks, then a risk-averse investor may
build a hedged portfolio from large-cap and growth stocks whose primary
purpose is to minimize future shortfalls. Return as such is a secondary
concern. Earning a bit while hedging risk, as they say in Louisiana, is for
lagniappe.50
According to Bad Beta, Good Beta, the difference between large-
cap and small-cap stocks arises from the decomposition of returns into
changes in cash flow versus changes in discount rates:
An increase in expected future cash flows is associated with a capital gain

today, while an increase in discount rates is associated with a capital loss
today. [W]ith a given dividend stream, higher future returns can be gen-
erated only by future price appreciation from a lower current price. These
return components can also be interpreted approximately as permanent and
transitory shocks to wealth. Returns generated by cash-flow news are never
reversed subsequently, whereas returns generated by discount-rate news are
offset by lower returns in the future.51
The complete separation of news affecting returns into cash-flow news

and discount-rate news implies that total market beta consists of the
sum of the cash-flow beta and the discount-rate beta:52
b a , m = b a , CF + b a , DR

where the subscripts cf and dr, respectively, indicate cash-flow and

discount-rate beta. This decomposition of beta into separate, quantifiable
cash-flow and discount-rate components is consistent with the defini-
tion of asset prices as the dynamic present value of future dividends53 and
with the definition of expected returns as a dual function of growth and
discounting.54
Although the complete separation of unexpected stock returns into
one category associated with changes in expectations of future cash flows
and another category associated with changes in expectations of future
discount rates arises from an accounting identity rather than a behav-
ioral model, unmistakably behavioral consequences flow from this distinc-
tion.55 [C]onservative long-term investors are more averse to cash-flow
risk than to discount rate risk.56 This aversion represents an extension of
the long-standing recognition that market prices reflect macroeconomic
100 J.M. CHEN
variables such as the spread between long-term and short-term interest

rates, the spread between high- and low-grade bonds, and expected versus
unexpected inflation.57
To extend (and enliven) the analogy, beta based on these distinct
sources of risk takes on the traits of the good and bad variants of
cholesterol.58 [B]ad cash-flow beta inflicts unequivocal, irreversible loss,
while good discount-rate beta, though not good in absolute terms,
is better in relation to the other type of beta in that discount-rate beta
gives some basis for hope in improved future returns.59 Intertemporal
asset pricing theory suggests that the bad cash-flow beta should have a
higher price of risk than the good discount-rate beta.60
It turns out that there is a striking difference in the beta composition of
value and growth stocks: The market betas of growth stocks are dispro-
portionately composed of discount-rate betas rather than cash-flow betas.
The opposite is true for value stocks.61 As a result, growth stocks, whose
betas are primarily of the good variety and accordingly carr[y] a low
premium, have lower returns, on average, than value stocks.62 Indeed, in
the post-1963 markets where Campbell and Vuolteenaho demonstrated
the connection between growth stocks and good beta, growth stocks
exhibited negative CAPM alphas.63 Conversely, value stocks and small
stocksthe highest-returning equity classes in Fama and Frenchs three-
factor modelhave considerably higher cash-flow betas than growth stocks
and large stocks, which in turn explain[s] their higher average returns.64
Bad beta and good beta also correlate with macroeconomic data.
The size, value, and momentum factors of the FamaFrenchCarhart four-
factor model appear to project future changes in gross domestic product.65
In particular, returns on value stocks are correlated to shocks to forecasts
in GDP growth.66 The cash-flow fundamentals at the microeconomic level
of differential returns on a subset of equities reflect cash-flow fundamen-
tals in the economy at large.67
These findings lend themselves to straightforward behavioral interpre-
tations and equally unambiguous advice for different classes of investors.
[S]ufficiently risk-averse long-term investors should view the high
average returns on value stocks and small stocks as appropriate compen-
sation for risk rather than a justification for systematic tilts toward these
types of stocks.68 By contrast, long-term institutional investors [such] as
pension funds might take a different view in assessing the risks of port-
folio tilts toward value and small stocks.69
In a 2015 update of Bad Beta, Good Beta, John Campbell and three
coauthors found that value stocks and small-cap stocks continue to have
higher cash-flow betas and higher discount-rate betas relative to growth

stocks and large-cap stocks, respectively.70 Their results were extremely
similar to those of the original 2004 version of Bad Beta, Good Beta,
despite the use of a richer, heteroskedastic vector autoregression time
series model.71 One new finding, however, is that value stocks, while
continu[ing] to have much lower volatility betas, now exhibit a spread
in volatility beta[] that is even greater than before.72 As a result, growth
stocks not only hedge news about future real stock returns, which was
Campbells primary conclusion in 2004, but also hedge news about the
variance of the [future] market return.73
Campbells new findings not only make economic sense; they also
have the potential to explain the low-volatility anomaly, specifically the
puzzling finding that high idiosyncratic-volatility stocks have lower aver-
age returns than low idiosyncratic-volatility stocks:74
High idiosyncratic volatility increases the value of growth options, which is

an important effect for growing firms with flexible real investment oppor-
tunities, but much less so for stable, mature firms. Valuable growth options
in turn imply high betas with aggregate volatility shocks. Hence high idio-
syncratic volatility naturally raises the volatility beta for growth stocks more
than for value stocks.75
As growing firms with flexible opportunities [become] more prevalent

in modern markets, the effect of high idiosyncratic volatility has become
stronger in modern sample[s] of the market.76
5.3 Addressing theLow-Volatility Anomaly

Through Spatial andTemporal Bifurcations ofBeta
John Campbells intertemporal interpretation of bad and good beta
revives our discussion of the long-term outperformance of low-risk
portfolios, which arguably remains the greatest anomaly in finance.77
We have now identified three distinct explanations for the low-volatility
anomaly. In 4.3 of this book and 7.6 of Postmodern Portfolio Theory,78
I argued that a closer examination of Javier Estradas emerging market
data suggested that correlation tightening in declining market conditions
accounted for nearly all of the changes in downside beta in those markets.
This observation gives rise to the inference that those equities returns
reflect the heightened risk of downside correlation between emerging
markets and the rest of the universe of tradable securities.
102 J.M. CHEN
By contrast, as I described in 4.5, Zhanhui Chen and Ralitsa Petkova

have denied any role to correlation. Instead, they assert that only expo-
sure to average variance is priced in stock market returns.79 In their
2015 reexamination of bad and good betas, John Campbell and his
coauthors agreed, in principle, that the idiosyncratic volatility effect can
be explained by aggregate volatility risk.80 Campbell and his coauthors
observed, however, that Chen and Petkova do not use a theoretically
motivated volatility risk factor.81 Earlier work by Petkova with Lu Zhang
has likewise examined the premium paid on value stocks relative to growth
stocks, albeit without reference to Campbells distinction between cash-
flow innovations and discount-rate innovations.82 For its part, Campbells
work, in 2004 and in 2015, has taken no explicit position on the contribu-
tion (if any) of correlation risk to the cross-section of stock market returns.
It is not certain that there is any tension between Javier Estradas emerg-
ing market data and John Campbells application of intertemporal CAPM to
value stocks, small-cap stocks, and the low-volatility anomaly. Emerging mar-
ket stocks embody many of the risk factors that account for higher returns
in small-cap stocks.83 The strong propensity of emerging market stocks to
move in unison with downward trends in developed market returns reduces
or eliminates their value as hedges against downside loss. On the other hand,
emerging market stocks carry immense amounts of foreign exchange risk
that is largely absent from developed equity markets.84 Foreign exchange
innovations would appear closer in spirit, if not entirely identical to, news
on discount rates. Unlike disappointments in cash flow, which irrevocably
reduce the value of a portfolio, bad foreign exchange news can be overcome
by the time a risk-averse investor must liquidate an emerging market position
in order to satisfy future investment or consumption demands.
Notwithstanding the differences in these narratives, the intertemporal
CAPM is supple enough to accommodate any of these explanations of
the low-volatility anomaly. In every setting, the pivotal factor is the pres-
ervation of an investors interest in future consumption or investment.
Consistent with Postmodern Portfolio Theorys characterization of volatility
and correlation on the downside as behaviorally distinct responses to finan-
cial sinking, fast and slow, idiosyncratic volatility and correlation tight-
ening pose distinct threats to future cash flows.85 In Campbells account,
a lower dose of bad cash-flow beta, at least relative to good discount-
rate beta, warrants lower returns for growth stocks, large stocks, and any
other asset that reduces the risk to future returns. Correlation tightening
expresses a more subtle concern that certain assets will contribute little
to diversification strategies if their correlation with the overall portfolio
increases more dramatically in declining markets. But as Campbells 2015
update to Bad Beta, Good Beta acknowledged, even the more intuitive
impact of idiosyncratic volatility on future returns requires consideration
of more subtle influences on the variance of those future returns.
These three approaches, typified by the work of John Campbell,
Zhanhui Chen, and Ralitsa Petkova, and Javier Estrada, ultimately bifur-
cate beta in three radically different ways. Bad Beta, Good Beta and
Campbells 2015 update evaluate the difference between cash-flow and
discount-rate beta. The application of intertemporal asset pricing theory
by Campbell and his coauthors demonstrates that the distinction between
cash-flow and discount-rate beta actually follows a temporal boundary
separating tomorrow from today. Chen and Petkova, on the one hand,
and Estrada, on the other, split beta along two different spatial bound-
aries. Betas relative volatility and correlation components are not only
distinct from one another and from the composite, unconditional form of
beta used in the conventional CAPM.Instead, the relative volatility and
correlation components of beta assume different shapes on the upside and
downside of mean returns.
It should come as no surprise, then, that such divergent approaches
to beta should yield such diverse answers to the low-volatility anomaly
(Table 5.1). In light of the foregoing, perhaps it is enough to recognize,
as Campbell has, that volatility betas vary with multiple stock characteris-
tics, and that techniques that take this into account may be more effective
in explaining the low-volatility anomaly.86 Fully reconciling the bifurcation
of beta across time, as between volatility and correlation, and on either
side of mean returns would bring this financial puzzle into a more refined,
multidimensional focus.
Table 5.1 Three ways to bifurcate beta, with three divergent explanations for the
low-volatility anomaly
Bifurcation of beta Source(s)
Upside beta versus downside beta Javier Estrada, Mean-Semivariance Behavior:

Downside Risk and Capital Asset Pricing (2007)
Average volatility versus average Zhanhui Chen & Ralitsa Petkova, Does
correlation ( versus ) Idiosyncratic Volatility Proxy for Risk Exposure?
(2012)
Discount rate versus cash flow John Campbell & Tuomo Vuolteenaho, Bad
(implicitly current versus future Beta, Good Beta (2004); Campbell etal., An
investment or consumption under Intertemporal CAPM with Stochastic Volatility
intertemporal CAPM) (2015)
104 J.M. CHEN
Notes
1. See Robert C.Merton, An Intertemporal Capital Asset Pricing Model, 41
Econometrica 867887 (1973). Notably, Zhanhui Chen and Ralitsa
Petkova describe their evaluation of idiosyncratic volatility, systematic vola-
tility, and correlation as being motivated by the intertemporal
CAPM.Chen & Petkova, supra note 108 (Chapter 4), at 2746.
2. See Merton, Intertemporal CAPM, supra note 1, at 870.
8.18.5, at 155172.
4. Ravi Jagannathan & Zhenyu Wang, The Conditional CAPM and the Cross-
Section of Expected Returns, 51J.Fin. 353, 4 (1996).
5. Philip H. Dybvig, Distributional Analysis of Portfolio Choice, 61 J. Bus.
369394, 369 (1988).
6. George G.Szpiro, Risk Aversion: An Alternative Approach, 68 Rev. Econ.
& Stat. 156159, 156 n.1 (1986).
7. Id.
8. Id.
9. See, e.g., Robert C.Merton, Optimum Consumption and Portfolio Rules in
a Continuous-Time Model, 3 J. Econ. Theory 373413 (1971); Jan.
Mossin, Optimal Multiperiod Portfolio Policies, 41J.Bus. 215229 (1968).
10. See, e.g., John C.Cox, Jonathan E.Ingersoll, Jr. & Stephen A.Ross, An
Intertemporal General Equilibrium Model of Asset Prices, 53 Econometrica
363384 (1985); Cox, Ingersoll & Ross, A Theory of the Term Structure of
Interest Rates, supra note 81 (Chapter 1); Robert E.Lucas, Jr., Asset Prices
in an Exchange Economy, 46 Econometrica 14291445 (1978).
11. See Kenneth J.Arrow & Grard Debreu, Existence of an Equilibrium for a
Competitive Economy, 22 Econometrica 265290 (1954) (describing
markets whose prices allow aggregate supplies of every commodity in the
economy to satisfy aggregate demand); Linel W. McKenzie, On the
Existence of General Equilibrium for a Competitive Economy, 27
Econometrica 5471 (1959). See generally Kartik B.Athreya, The Modern
Macroeconomic Approach and the Arrow-Debreu-McKenzie Model, in Big
Ideas in Macroeconomics: A Nontechnical View 1146 (2013); John
Geanakoplos, Arrow-Debreu Model of General Equilibrium, 1 The New
Palgrave: A Dictionary of Economics 116124 (Peter Newman, John
Eatwell & Murray Milgate eds., 1987).
12. See Fischer Black & Myron S.Scholes, The Pricing of Options and Corporate
Liabilities, 81J.Pol. Econ. 637654 (1973).
13. Robert J.Elliot, Hong Miao & Jin Yu, General Equilibrium Asset Pricing
Under Regime Switching, 2 Communications on Stochastic Analysis 445
458, 445446 (2008); accord, e.g., Abraham Lioui & Patrice Poncet,
General Equilibrium Pricing of Nonredundant Forward Contracts,
23J.Futures Mkts. 817840 (2003); Julien Hugonnier, Erwan Morellec
& Suresh Sundaresan, Irreversible Investment in General Equilibrium
(June 2005); Jae W.Sim, Uncertainty, Irreversible Investment and General
Equilibrium (Aug. 2007) (available at http://web.stanford.edu/group/
SITE/archive/SITE_2007/segment_8/sim_UIRRGE.pdf).
14. See James Tobin, Liquidity Preference as Behavior Towards Risk, 67 Rev.
Econ. Stud. 6586 (1958). See generally John Hicks, Liquidity, 72 Econ.
J. 787802 (1962).
15. See David Cass & Joseph E.Stiglitz, The Structure of Investor Preferences
and Asset Returns, and Separability in Portfolio Allocation, 2 J. Econ.
Theory 122160 (1970); Robert C.Merton, An Analytic Derivation of
the Efficient Portfolio Frontier, 7 J. Fin. & Quant. Analysis 18511872
(1972); Stephen A.Ross, Mutual Fund Separation and Financial Theory
The Separating Distributions, 17J.Econ. Theory 254286 (1978).
16. Philip H. Dybvig & Stephen A. Ross, Portfolio Efficient Sets, 50
Econometrica 15261546, 1526 (1982). The term simplicial refers to a
simplex, which in algebraic topology is the generalization of a tetrahedral
region of space to n dimensions. See generally James R.Munkres, Elements
of Algebraic Topology 1.11.2, at 214 (1993) (Simplices and
Simplicial Complexes and Simplicial Maps); F. Buekenhoutt &
M.Parker, The Number of Nets of the Regular Convex Polytopes in Dimension
4, 186 Discrete Math. 6994 (1998) (describing a simplex as a
hypertetrahedron).
17. Dybvig & Ross, supra note 16, at 1526.
18. See id.
19. See, e.g., Martin Browning & Thomas F.Crossley, The Life-Cycle Model of
Consumption and Saving, 15 J. Econ. Persp. 322 (2001); cf. Hersh
M.Shefrin & Richard H.Thaler, The Behavioral Life-Cycle Hypothesis, 26
Econ. Inq. 609643 (1988).
20. See infra 7.5, at 148150; cf. infra 9.49.5, at 230234 (presenting
proposed solutions to the equity premium puzzle from the point of view of
prospect theory).
21. Merton, Intertemporal CAPM, supra note 1, at 876.
22. Id.
23. Id. at 875.
24. Id. at 877.
25. See generally John Y. Campbell & Robert J. Shiller, The Dividend Price
Ratio and Expectations of Future Dividends and Discount Factors, 1 Rev.
Fin. Stud. 195228 (1988); John Y.Campbell & Robert J.Shiller, Stock
Prices, Earnings, and Expected Dividends, 43J.Fin. 66176 (1988).
106 J.M. CHEN
26. See Ang, Hodrick, Xing & Zhang, The Cross-Section of Volatility and
Expected Returns, supra note 14 (Chapter 4), at 260; George Chacko &
Luis M.Viceira, Dynamic Consumption and Portfolio Choice with Stochastic
Volatility in Incomplete Markets, 18 Rev. Fin. Stud. 13691402, 1370
(2005).
27. Bo Young Chang, Peter Christoffersen & Kris Jacobs, Market Skewness
Risk and the Cross Section of Stock Returns, 107J.Fin. Econ. 4668, 64.
28. Id.
29. Michael J. Brennan & Eduardo S. Schwartz, Time-Invariant Portfolio
Insurance Strategies, 43J.Fin. 283299, 283 (1988). See generally Michael
J. Brennan & Ray Solanki, Optimal Portfolio Insurance, 16 J. Fin. &
Quant. Analysis 279300 (1981); Michael J. Brennan & Eduardo
S. Schwartz, Portfolio Insurance and Financial Market Equilibrium,
62J.Bus. 455472 (1989); Simon Benninga & Marshall Blume, On the
Optimality of Portfolio Insurance, 40J.Fin. 13411352 (1985).
30. See generally, e.g., Anup K. Basu, Brett Doran & Michael E. Drew,
Sequencing Risk: The Worst Returns in Their Worst Order, 4 JASSA: Finasia
J.Applied Fin. 713 (2013); Larry R.Frank, John B.Mitchell & David
M. Blanchett, Probability-of-Failure-Based Decisions Rules to Manage
Sequence Risk in Retirement, 24:11J.Fin. Planning 4480 (Nov. 2011);
Matthew B.Kenigsberg, Prasenjit Dey Mazumdar & Steven Feinschreiber,
Return Sequence and Volatility: Their Impact on Sustainable Withdrawal
Rates, 2:2J.Retirement 8198 (Fall 2014).
31. See generally, e.g., Robert Argento, Victoria L.Bryant & John Sabelhaus,
Early Withdrawals from Retirement Accounts During the Great Recession,
33 Contemp. Econ. Poly 116 (2015); Robert L.Clark & John Sabelhaus,
How Will the Stock Market Crash Affect the Choice of Pension Plans?, 62:3
Natl Tax J. 120 (Sept. 2009); Christopher R.Tamborini, Patrick Purcell
& Howard M. Iams, The Relationship Between Job Characteristics and
Retirement Savings in Defined Contribution Plans During the 20072009
Recession, 136 Monthly Labor Rev. 316 (May 2013); cf. Gordon B.Pye,
The Effect of Emergencies on Retirement Savings and Withdrawals,
23:11J.Fin. Planning 5762 (Nov. 2010).
32. See W.V.Harlow & Keith C.Brown, Improving the Outlook for a Successful
Retirement: A Case for Using Downside Hedging, 3:3J.Retirement 3550
(Winter 2016).
33. Under former Rule 434 of its regulations implementing the Securities Act
of 1933, the Securities and Exchange Commission defined structured secu-
rities as securities whose cash flow characteristics depend upon one or
more indices or that have embedded forwards or options or securities where
an investors investment return and the issuers payment obligations are
contingent on, or highly sensitive to, changes in the value of underlying
assets, indices, interest rates or cash flows. Prospectus Delivery Requirements

in Firm Commitment Underwritten Offerings of Securities for Cash, 17
C.F.R. 230.434 (repealed); accord United States Securities and Exchange
Commission, Staff Summary Report on Issues Identified in Examinations of
Certain Structured Securities Products Sold to Retail Investors 3 n.3 (July 27,
2011) (available at https://www.sec.gov/news/studies/2011/ssp-study.
pdf). See generally Mehraj Mattoo, Structured Derivatives: A Handbook of
Structuring, Pricing and Investor Applications (1996). For a sense of the
diversity and complexity of the issues surrounding structured securities, see
Wolfgang Breuer & Achim Perst, Retail Banking and Behavioral Financial
Engineering: The Case of Structured Products, 31J.Banking & Fin. 827
844 (2007); Pavel A.Stoimenov & Sascha Wilkens, Are Structured Products
Fairly Priced? An Analysis of the German Market for Equity-Linked
Instruments, 29J.Banking & Fin. 29712993 (2005).
34. Cf. Merton, Intertemporal CAPM, supra note 1, at 882 (equation 34).
35. Id.
36. Douglas T.Breeden, An Intertemporal Asset Pricing Model with Stochastic
Consumption and Investment Opportunities, 7J.Fin. Econ. 265296, 267
(1979).
37. Id. at 292.
38. See Roll, supra note 20 (Chapter 3), at 155. See generally Chen, Postmodern
Portfolio Theory, supra note 1 (Chapter 1), 3.3, at 2931.
39. See Fama & French, The Cross-Section of Expected Stock Returns, supra note 5
(Chapter 3).
4.14.4, at 4158.
41. See supra 3.2, at 6066.
42. See, e.g., Eugene F.Fama, Multifactor Portfolio Efficiency and Multifactor
Asset Pricing, 31J.Fin. & Quant. Analysis 441465 (1996).
43. See George Loewenstein, Ted ODonoghue & Matthew Rabin, Projection
Bias in Predicting Future Utility, 118 Q.J. Econ. 12091248, 1230
(2003).
44. See id.
45. John Maynard Keynes, A Tract on Monetary Reform 80 (1924).
46. See Ronald J.Balvers & Douglas W.Mitchell, Autocorrelated Returns and
Optimal Intertemporal Portfolio Choice, 43 Mgmt. Sci. 15371551 (1997);
Ronald J. Balvers & Douglas W. Mitchell, Efficient Gradualism in
Intertemporal Portfolios, 24J.Econ. Dynamics & Control 2138 (2000);
see also sources cited supra notes 3031 (describing sequence-of-returns
risk).
47. See John Y. Campbell, Intertemporal Asset Pricing Without Consumption
Data, 83 Am. Econ. Rev. 487512 (1993).
108 J.M. CHEN
48. John Y.Campbell & Tuomo Vuolteenaho, Bad Beta, Good Beta, 94 Am.
Econ. Rev. 12491275, 1249 (2004).
49. Id. at 12511252.
50. Mark Twain, Life on the Mississippi, in Mississippi Writings 217616,
489490 (Library of America ed. 1982) (1st ed. 1883) (We picked up
one excellent worda word worth travelling to New Orleans to get; a
nice, limber, expressive, handy wordlagniappe. They pronounce it
lanny-yapIt is the equivalent of the thirteenth roll in a bakers dozen.
It is something thrown in, gratis, for good measure.); accord http://
etymonline.com/index.php?term=lagniappe; see also Life on the
Mississippi, supra, at 71 (Chapter 2) (the English were trading beads and
blankets to [native Americans] for a consideration, and throwing in civili-
zation and whiskey, for lagniappe).
51. Campbell & Vuolteenaho, supra note 48, at 1252.
52. Id. at 1257.
53. See Ravi Bansal & Amir Yaron, Risks for the Long Run: A Potential
Resolution of Asset Pricing Puzzles, 59J.Fin. 14811509 (2004).
54. See John Cochrane, Discount Rates, 66J.Fin. 10471108 (2011).
56. Id. at 12521253.
57. See generally Nai-Fu Chen, Richard Roll & Stephen A. Ross, Economic
Forces and the Stock Market, 59J.Bus. 383403 (1986).
58. See Campbell & Vuolteenaho, supra note 48, at 1250.
59. Id.
60. Id. at 1271; cf. Robert J. Shiller, Do Stock Prices Move Too Much to Be
Justified by Subsequent Changes in Dividends?, 71 Am. Econ. Rev. 421436
(1981).
62. Id. at 1271.
63. Id.
64. Id.
65. See Jimmy Liew & Maria Vassalou, Can Book-to-Market, Size and
Momentum Be Risk Factors That Predict Economic Growth?, 57 J. Fin.
Econ. 221245 (2000).
66. See Maria Vassalou, News Related to Future GDP Growth as a Risk Factor
in Equity Returns, 68J.Fin. Econ. 4773 (2003); cf. Randolph B.Cohen,
Christopher Polk & Tuomo Vuolteenaho, The Value Spread, 58 J. Fin.
609641 (2003).
68. Id. at 1272.
69. Id.
70. See John Y.Campbell, Stefano Giglio, Christopher Polk & Robert Turley,
An Intertemporal CAPM with Stochastic Volatility (June 2015) (available
at http://scholar.harvard.edu/campbell/files/cgpt_volatilityrisk
_20150619.pdf).
71. Id. at 29. On the vector-autoregressive time series methodology, see gen-
erally John Y. Campbell & Robert J. Shiller, Cointegration and Tests of
Present Value Models, 95J.Pol. Econ. 10621088 (1987).
72. Campbell, Giglio, Polk & Turley, supra note 70, at 30.
73. Id. (emphases added).
74. Id. (citing Ang, Hodrick, Xing & Zhang, The Cross-Section of Volatility and
Expected Returns, supra note 14 (Chapter 4)).
76. Id.
77. Baker, Bradley & Wurgler, supra note 8 (Chapter 4), at 43.
78. See Chen, Postmodern Portfolio Theory, supra note 1 (Chapter 1), 7.6, at
122126.
79. Chen & Petkova, supra note 108 (Chapter 4), at 2746.
80. Campbell, Giglio, Polk & Turley, supra note 70, at 32 n.17.
81. Id.
82. See Ralitsa Petkova & Lu Zhang, Is Value Riskier Than Growth?, 78J.Fin.
Econ. 187202 (2005).
83. Harvey, Predictable Risk and Returns in Emerging Markets, supra note 64
(Chapter 4), at 787 (attributing variation among emerging markets to
market liquidity, political risk, and other [f]actors such as taxes, which
in the aggregate keep emerging markets from becoming fully integrated
with the global financial system); see also id. at 801 (concluding that local
information materially affects emerging markets, whereas developed mar-
kets tend to be driven by global information variables rather than local
information); cf. Bekaert & Harvey, Foreign Speculators and Emerging
Equity Markets, supra note 68 (Chapter 4).
84. See Geert Bekaert & Robert J. Hodrick, Characterizing Predictable
Components in Excess Returns on Equity and Foreign Exchange Markets,
47 J. Fin. 467509 (1992); Richard Roll, Industrial Structure and the
Comparative Behavior of International Stock Market Indexes, 47 J. Fin.
341 (1992); cf. Bekaert & Harvey, Emerging Equity Market Volatility,
supra note 68 (Chapter 4); Bekaert & Harvey, Time-Varying World-Market
Integration, supra note 68 (Chapter 4).
85. See Chen, Postmodern Portfolio Theory, supra note 1 (Chapter 1),
7.17.10, at 107151.
CHAPTER 6
Risk Aversion
6.1 The ArrowPratt Measures ofRisk Aversion;

theCoefficient ofAbsolute Risk Aversion
If only because protective instincts are pervasive among humans and

prized in many social settings,1 behavioral finance demands a cred-
ible account of risk aversion.2 That account begins with the decline of
expected utility theory.3 Behavioral economics arose as a response to the
limitations of conventional game theory and expected utility theory.4
Behavioral economics adds a host of considerations that elude these
conventional models of utility and risk.5 Because conventional defini-
tions of risk aversion hold the key to solving behavioral challenges such
as the equity risk premium and the equity premium puzzle,6 I will now
propound some of the foundations of expected utility theory. I start by
presenting the absolute and relative versions of the ArrowPratt mea-
sures of risk aversion, named for Kenneth Arrow7 and John Pratt.8 These
measures are also known as the coefficients of absolute and relative
risk aversion.9
Arrows specification of the coefficient of absolute risk aversion begins
with this question: How much compensation would a risk-averse inves-
tor demand in exchange for accepting an actuarially fair gamble, condi-
tional upon her or his existing wealth?10 Let w represent existing wealth.
Furthermore, let z represent a gamble with these payouts:

DOI10.1007/978-3-319-32711-2_6
112 J.M. CHEN
+z with probability p
z with probability 1p
If U(w) represents our utility function, then we can describe risk aver-
sion in mathematical terms by solving this equation:

U ( w ) = U w + z

We start by expanding the right side to account for the conditional

value of z:
U ( w ) = p U ( w + z ) + (1 p ) U ( w z )

The key to solving this equation lies in performing the Taylor series
expansion of the terms U(wz) on the right side. Recall from 3.2 the
formula for a Taylor series expansion for a function, f(x), that is infinitely
differentiable at value a:11
f (a) f ( a ) f ( a )
f ( x) f (a) + ( x a) + ( x a) ( x a)
2 3
+ +
1! 2! 3!
Applying this definition to our risk-aversion equation yields the follow-

ing expansion:
U(w) U ( w )
U ( w ) = p U ( w ) +
1!
(w + z w) +
2!
( w + z w ) + o z3
2
( )

U(w) U ( w )
+ (1 p ) U ( w ) +
1!
(w z w) +
2!
( w z w ) + o z3
2
( )

Simplifying somewhat:
U ( w )
U ( w ) = p U ( w ) + U ( w ) z +
2
z2 + o z3 ( )

U ( w )
+ (1 p ) U ( w ) U ( w ) z +
2
z2 + o z3 ( )

RISK AVERSION 113
It should already be apparent that the odd-numbered terms of the

Taylor series expansions of both terms on the right side of the equation
will cancel out. Truncating the Taylor series expansions after their second-
order terms and reorganizing provides a reasonably tractable approxima-
tion of the utility function:
1
U ( w ) p U ( w ) + U ( w ) z + U ( w ) z 2
2
1
+ (1 p ) U ( w ) U ( w ) z + U ( w ) z 2
2
In turn, multiplying out and adding all terms yields a final approxima-
tion of the utility function:
1
U ( w ) U ( w ) + ( 2 p 1) U ( w ) z + U ( w ) z 2
2
All that remains in this exercise is to solve for p in terms of the first and
second derivatives of the utility function:
1
2 p U ( w ) z U ( w ) z U ( w ) z 2
2
1 1 U ( )
w
p z
2 4 U(w)

This approximation of p lends itself to a very intuitive interpretation.

The default value for p is . The only meaningful variable affecting the
U ( w )
value of p is . If we define A(w), our coefficient of absolute risk
U(w) U ( w )
aversion, as the negative of this ratio, such that A ( w ) = , then p
U(w)
may be expressed entirely in terms of the constant ; the positive value of
the bet, z; and the coefficient of absolute risk-aversion payoff, A(w):
1 1
p + z A( x)
2 4
114 J.M. CHEN
6.2 The Coefficient ofRelative Risk Aversion

Repeating the foregoing exercise by defining the gamble z as a propor-
tion of initial wealth w enables us to state p in terms of a coefficient of
relative
risk aversion. Specifically, let us define the possible outcomes for
z in the following terms:
+zx with probability p

zx with probability 1p
Proceeding from this definition, we can approximate p in terms of the

constant , the positive payoff z, and R(w), the coefficient of relative risk
aversion:12
1 1
p + z R (w)
2 4
U ( w )
R ( w ) = w
U(w)

6.3 Pratts Risk-Averse Insurance Premium

John Pratts closely related definition of risk-aversion
approaches the
problem as one of insurance. Let us define gamble z as a fair bet. It yields
+z with probability , and the opposite outcome z with equal probability
. The question now is that of the insurance premium that a risk-averse
agent would pay to avoid z , despite the zero expected value of that bet.
In formal terms:13
U w + z ( w, z ) = U ( w + z ) = U ( w )

Pratts solution to this problem lay in approaching ( w, z ) on the

assumption of normally distributed z . Specifically: z = 0 and z2 > 0 .
The Taylor series expansion of two equivalent expressions of the relevant
utility functions as z2 0 reveals:14
RISK AVERSION 115
U ( w ) = U ( w ) U ( w ) + O 2 ( )
1
U ( w + z ) = U ( w ) + z2 U ( w ) + o ( )
2
z
2
where, consistent with Taylor series notation, O() means terms of order
at most and o() means terms of smaller order than.15
These equations open a path to approximating the risk-averse insurance
premium in terms of the derivatives of the utility function and the vari-
ance of the unwanted gamble:
1
U ( w ) . U ( w ) = U ( w ) + z2 . U ( w )
2
1
. U ( w ) = z . U ( w )
2
2
1 U ( w )
= z2 .
2 U ( w )

Since we have already defined the coefficient of absolute risk aversion as

the negative of ratio of the second derivative of the utility function to its
U ( w )
first derivative, A ( w ) = , we can likewise restate the risk-averse
U(w)
insurance premium in terms of A(w):16
1
= z2 A ( w )
2
6.4 Hyperbolic Absolute Risk Aversion

U ( w )
Defining the coefficient of absolute risk aversion as A ( w ) = facili-
U(w)
tates the development of an entire family of measures of risk aversion,
known as hyperbolic absolute risk aversion (HARA) or linear risk toler-
ance.17 HARA arises from the description of A(w), heretofore defined as
116 J.M. CHEN
a differential equation, as providing a measure of absolute risk aversion

that is positive and hyperbolic in consumption:18
U ( w ) 1
A(w) = = >0
U(w) w b
+
1 a

subject to these restrictions:
aw
1; a > 0; + b > 0; b = 1 if =
1
To avoid the unwieldiness of the foregoing formula, we may wish to

define a measure of risk tolerance, T(w), as the reciprocal of A(w):
1 U(w) w b
T (w) = = = +
A(w) U ( w ) 1 a

Whence arises the name linear risk tolerance, since risk tolerance T(w) is
linear in form and risk aversion A(w) is hyperbolic.19
Solving A(w) or T(w) as differential equations defines the correspond-
ing utility function, U(w), as a member of the HARA class of utility func-
tions if and only if U(w) has the form:20

1 aw
U (w) = + b
1

A wide range of utility functionsenabling risk aversion to be expressed

as increasing, decreasing, or constant, in absolute or in relative termscan
be defined as special cases of HARA (Table 6.1).21
Decreasing absolute risk aversion, or A(w) < 0, occurs only if
< < 1.22 This intuitively plausible condition, which suggests that
investors are more willing to take on risk as their wealth increases, plays
a pivotal role in establishing skewness preference as a function of the
positive third derivative within the four-moment CAPM drawn from the
Taylor series expansion of the distribution of returns.23
RISK AVERSION 117
Table 6.1 Special cases of hyperbolic absolute risk aversion
Condition(s) Resulting utility function
=1 Linear (risk-neutral)
=2 Quadratic
Exponential: U(w)=eaw
<1; a=1 Power: U(w)=w/
<1; a=1; b=0 Isoelastic, with constant relative risk
aversion
0; a=1 Logarithmic
0; a=1; b=0 Constant relative risk aversion:
U(w)=ln w
The final line of Table 6.1 also bears further explanation. For condi-
tions 0, a=1, and b=0, the resulting utility function can be defined
only as a limit:
w 1
U ( w ) = lim

0
The application of LHpitals rule24 makes it clear that the utility func-
tion for constant relative risk aversion (CRRA) takes logarithmic form:25
f ( x) f ( x)
lim = lim
x c g ( x) x c g ( x )

w 1 w ln w
U ( w ) = lim = lim = ln w

0 0 1
Or simply: U(w)=ln w.
The scale-invariant nature of CRRA makes it the preference func-
tion of choice in much of the literature on growth and in Real Business
Cycle theory.26 For purposes of solving the equity premium puzzle that
will be the subject of Chapter 7, the scale-invariant property of CRRA
enables us to compare investors within the aggregate market without
needing to make complex and highly contestable judgments on the impact
of individual wealth on risk aversion.27 A logarithmic utility function guar-
antees optimization through maximizing the geometric mean of returns.28
118 J.M. CHEN
Permitting the assumption that investors have heterogeneous expecta-

tions, however, means that different investors face different subject
[mean-variance] efficient set[s] and, hence, hold different portfolios, in
violation of modern portfolio theorys two-fund separation theorem.29
6.5 A Comparison withScale-Invariant Models

ofFinancial Returns
The intellectual allure of CRRA is evocative of the burst of academic

interest in scale-invariant stable Paretian distributions during the 1960s
and 1970s.30 A Lvy stable distribution can have skewness and fat tails
and obeys scaling properties.31 In 1963, both Benoit Mandelbrot32 and
Eugene Fama33 hypothesized that stock prices follow a stable distribution.
Fama later asserted that [t]he presence, in general of leptokurtosis in
the empirical distributions seems indisputable.34 Mandelbrot speculated
that financial markets have infinite variance, perhaps even infinite expected
value.35
Stable distributions have fallen out of favor in contemporary finance
because they make financial modeling so difficult.36 For any value of
shape parameter <2, the stable distribution has infinite variance.37 The
infinite variance of most stable distributions not only violates empirical
observations and logic that dictates finite variance, but also significantly
complicates the task of risk estimation and limits the practical applica-
tion of the stable distribution.38 The stable distribution defeats the use of
the sample standard deviation as a meaningful measure of risk, since
standard deviation, as the sample size increases, does not converge to
any given value.39
In practice, financial forecasting is more tractable than Mandelbrot and
Fama suggested. [S]ample estimates of variance and higher moments
converge rather than increase as sample size increases.40 Moreover, evi-
dence of abnormality is much weaker for long-horizon returns than for
short-horizon returns.41 Empirical evidence marshaled in the early 1970s
did show fatter tails than a normal distribution would have permitted, but also
finite varianceand therefore finite standard deviation as a well behaved
function of scalein contradiction to the stable Paretian distribution.42
As a matter of theory, it has always been possible to reject the sugges-
tion that a finding of returns are not normally distributed implies that
it has infinite variance.43 The refusal to to accept [those] underlying
assumptions led to the contrary assumption that all distributions have
finite means and variances.44 Financial analysis today prefers distributions
RISK AVERSION 119
with finite second moments of returns, and often finite higher moments
as well.45 Reliance on stable distributions arose from a previous gen-
erations self-imposed methodological constraints, which dictated an
all-or-nothing choice between normally distributed or infinitely variable
returns.46 That brittle statistical assumption is a matter of choice, not a
reflection of immutable truth.
6.6 Risk Aversion, Risk Tolerance, andTheir

Relationship totheSharpe andKappa Ratios
The ArrowPratt methodology yields measures of absolute risk aversion
and risk tolerance that are stated as differential equations:47
U ( w )
A(w) =
U(w)

1 U(w)
T (w) = =
A(w) U ( w )

These measures are related to traditional measures of portfolio risk in

an intriguing way. In the conventional CAPM, the Sharpe ratio is the slope
of the capital allocation line:48
E ( Rp ) R f
E ( Rc ) = R f + c
p

rp rf
Sharpe ratio = =
p

The ArrowPratt measures suggest that risk aversion may be defined as

the marginal return that an investor demands in exchange for accepting
each additional unit of risk. In that event, differentiation of the Sharpe
ratio provides a plausible definition of risk aversion:
d
As ( w ) =
d
120 J.M. CHEN
The risk-free rate or the target return, being defined as a constant, disappears
upon differentiation. This definition of risk aversion is indistinguishable from
the first derivative of the ratio of financial returns signal-to-noise ratio, which
is the ratio of mean to standard deviation.
The ArrowPratt measures of risk aversion bear a close relationship to
a mathematically related family of single-sided risk measures examined in
6.8 through 6.10 of Postmodern Portfolio Theory.49 Unlike the Sharpe
ratio, which measures only the sign and magnitude of the average risk
premium relative to the risk incurred in achieving it, these single-sided
risk measures focus[] [instead] on the likelihood of not meeting some
target return.50 These measures share a primary goal of anticipating the
reaction of investors to losses and a secondary goal of retaining sensitivity
to prospective gains. In other words, the postmodern quest in portfolio
theory is heightened vigilance against downside risk, without complete
sacrifice of upside potential.
William Shadwick and Con Keating have developed what they call
the omega measure, consisting of the ratio of the first-order upper and
lower partial moments of the distribution of returns, each about the target
return, :51
1 F ( x ) dx
( ) =

F ( x ) dx

where is the target return and F(x) is the cumulative distribution func-
tion of returns about that target. Or, in terms of partial and complete
moments of the probability density function:52
( ) f ( x ) dx
( ) =

( ) f ( x ) dx

In an even more general formulation of single-sided risk measures, Paul

Kaplan and James Knowles have restated the omega ratio in terms of lower
partial moments:53
RISK AVERSION 121

+ ( ) f ( x ) dx
( ) =

=
+1
( ) f ( x ) dx ( ) f ( x ) dx

This formulation connects omega to a more general family of downside

risk-adjusted performance measures called kappa:

n ( ) =

n
( ) f ( x ) dx

By this definition, the omega ratio is 1+1.54 A third measure, designated

by its creators as the Sharpe-omega ratio,55 turns out to be, straightfor-
wardly, 1.56
A variant of the kappa ratio allows us to extending this analogy to risk
measures based on higher moments. The simplest definition of nth-order
kappa is the ratio of expected return in excess of a target (the same numer-
ator as the Sharpe ratio) to the normalized nth lower partial moment of
the distribution of returns:57

n =
n
n

It is very easy to define a variant of kappa as the ratio of the mean less
target return to any normalized complete moment of order n:

n =
n
n

In that event, differentiating both the numerator and the denominator

of modified kappa generates an analogous measure of risk aversion cor-
responding to the measure based on the first derivative of the Sharpe
ratio:
122 J.M. CHEN
d d
A ( w ) = =
d n
n n d n

The corresponding measures of risk tolerance within this framework are

either the first derivative of the coefficient of variation (in the case of the
Sharpe measure) or the corresponding version of that derivative, adjusted
according to the order of the modified kappa ratio:
d
Ts ( w ) =
d
n d n
T ( w ) =
d
6.7 The Allais Paradox

The ArrowPratt model is by no means the last word on risk aversion.
Although the preceding exercise demonstrates how portfolio theory
and its extensions can be harmonized with expected utility theory and
its measures of risk aversion, a sense of discomfort persists. Expected
utility theorys account of risk aversion does not explain the full range
of human responses to risk.58 Although humans do tend on balance to
be risk averse, risk-seeking behavior does exist. Any credible account of
behavioral economics must therefore explain both risk aversion and risk-
seeking. The balance of this chapter will explore the Allais paradox and
the St. Petersburg paradox, two puzzles that figured prominently in the
decline of expected utility theory and the willingness of economists and
psychologists, whether working together or independently, to find alter-
native accounts of human decision-making under risk and uncertainty.59
The difference between risk aversion and risk-seeking is readily and
intuitively illustrated in graphic terms (Figs. 6.1 and 6.2). Risk aversion is
portrayed by a concave curve where the value of the utility function, u(x),
is less than x for any value of x within the relevant range. Risk-seeking,
by contrast, is depicted by a convex curve where the value of u(x) actually
exceeds x. The risk premium is the difference between u(x) as the expected
value of a prospect and x as its certainty equivalent. Risk-averse individuals
RISK AVERSION 123
Fig. 6.1 A utility curve for a risk-averse individual
have a negative risk premium. By contrast, for risk-seeking individuals, the

risk premium is positive.
Early experimental studies of risk aversion reported that preferences are
not uniform, let alone uniformly concave, throughout the entire domain of
losses and gains.60 Economists continue to disagree whether absolute risk
aversion increases or decreases with wealth.61 Despite the allure of models
presuming decreasing absolute risk aversion, hypotheses positing either
increasing or constant absolute risk aversion continue to hold sway.62
The most visible failures of expected utility theory lie in its extrapola-
tion of extreme risk aversion from modest gambles to large gambles. The
original FriedmanSavage utility function, which pioneered the notion
that an individuals utility function might vary according to her level of
124 J.M. CHEN
Fig. 6.2 A utility curve for a risk-seeking individual
wealth, suggested that the same person might be risk averse when poor
(as indicated by the purchase of insurance) and risk seeking upon gaining
more wealth (as suggested by a willingness to gamble).63 Harry Markowitz
protested that the FriedmanSavage utility function, taken to its logical
conclusion, implied that poor people would never buy lottery tickets and
middle-income people would never buy insurance against small losses.64
Empirical evidence against both of these propositions, especially regard-
ing lottery expenditures by the poor, is overwhelming.65 To move past
both the FriedmanSavage utility function and Markowitzs critique of
it, we need a credible account of an everyday paradox: putatively rational
individuals routinely play the lottery and buy insurance, with no sense of
contradiction.66
A subtler version of this puzzle involves the admittedly bourgeois
dilemma in which an affluent albeit not wealthy family holds securities in a
taxable brokerage account (including bonds paying lower rates of interest
than the rate charged by the mortgage lender), but would not think of
liquidating its portfolio in order to prepay its home loan. This issue arises
with some regularity in popular writing on personal finance.67 Prepaying
the mortgage guarantees some return in the form of a more quickly retired
RISK AVERSION 125
home loan. On the other hand, investing in the stock market offers returns
that often but not invariably exceed those on the family home.
Any practical resolution of this problem is confounded by the sheer
size and illiquidity of the typical households financial stake in residential
real estate. Owner-occupied housing, especially in the form of a single-
family home, represents a significant portion of many households overall
financial portfolio.68 Consumption demand for housing imposes a port-
folio constraint on the typical households optimal holdings of financial
assets, since the households demand for real estate is overdetermined
in the sense that the optimal level of housing for consumption purposes
may differ from the optimal share of housing in a households portfolio.69
Especially among younger and poorer households, house price risk crowds
out stock market investments.70 Relative to other households, households
with mortgage debt are 9.8 % less likely to own stocks and 37.3 % less
likely to own bonds.71 The effect on bonds is especially strong since debt
repayment and bonds serve as substitute assets72 and since repayment
of outstanding debt almost always yields a higher rate of return than
short-term Treasury bills or long-term government bonds.73
Maurice Allais identified a specific instance in which seemingly plau-
sible choices by the same rational person were logically impossible under
expected utility theory.74 Imagine that our hypothetically rational person
were given a choice of either of these payouts:
A A 100%chance of receiving $1 million

B A 1%chance of receiving nothing, plus
an 89%chance of receiving $1 million, plus
a 10%chance of receiving $5 million
Most people would choose A over B. In terms of expected utility theory,

that choice implies:
U (1) > 0.01U ( 0 ) + 0.89U (1) + 0.1U ( 5 )

0.11U (1) > 0.01U ( 0 ) + 0.1U ( 5 )

Now consider presenting a choice between these payouts to the same

rational person:
126 J.M. CHEN
C An 89%chance of receiving nothing

An 11%chance of receiving $1 million
D A 90%chance of receiving nothing
A 10%chance of receiving $5 million
Picking D, as Allais expected most people would choose, implies:
0.89U ( 0 ) + 0.11U (1) < 0.90U ( 0 ) + 0.1U ( 5 )

0.11U (1) < 0.01U ( 0 ) + 0.1U ( 5 )

But choosing both A and D poses a serious problem for expected

utility theory, because those choices dictate opposite outcomes for the
inequality, 0.11 U(1) 0.01 U(0) + 0.1 U(5). This violation of the
independence axiom, a foundational premise of the FriedmanSavage
utility function,75 is known as the Allais paradox.76 Allaiss identification
of a certainty effect in risky choice would eventually lead to the con-
cept of decision weights, a pivotal step in the evolution of behavioral
economics.77 [I]nvestors do not employ objective probabilities but
rather subjective values called decision weights.78
6.8 The St. Petersburg Paradox

Matthew Rabin has generalized many of the instances in which expected
utility theory generates an implausible or paradoxical result. Such failures
often take this form: If an expected-utility maximizer always turns down
modest-stakes gamble X, she will always turn down large-stakes gamble
Y.79 Expected utility theory dictates that a person who turns down
gambles where she loses $100 or gains $110, each with 50 % probabil-
ity would necessarily reject 5050 bets of losing $20,000 or gaining
any sum.80 These are implausible degrees of risk aversion, sufficient in
their departure from reality to undermine an expected-utility framework
where turning down a modest-stakes gamble means that the marginal
utility of money must diminish very quickly for small changes in wealth.81
One vivid demonstration of expected utility theorys implausibility showed
how a person who is indifferent as between receiving a certain $7 and a
$0/$21 gamble with even odds for either outcome would prefer $7 over
any gamble of any size where the probability of winning is less than 40%.82
RISK AVERSION 127
Most spectacularly, perhaps, expected utility theory struggles to explain

the St. Petersburg paradox, a problem attributed to the eighteenth-
century mathematician, Daniel Bernoulli.83 The paradox consists of a
game of chance involving a single player and the tossing of a fair coin. The
pot starts at $2 and doubles every time a head appears. But the first time
a tail appears, the game ends, and the player collects the pot. The players
winnings equal $2 if a tail first appears on the first toss, $4 on the second,
$8 on the third, and so forth. In other words, the player wins $2k for k
tosses until a tail appears; k Z, k1. The player cannot lose and wins
a minimum of $2. What is a fair price to pay for admission to this game?
In fact, as long as the casino offering this game has unlimited resources
and will keep playing as long as the coin shows heads, the actuarial value
of the game is infinite:
1 1 1
E= 2 + 4 + 8 +
2 4 8
E = 1 + 1 + 1 +
E=

Given infinite expected value, a player should pay any price to play this
game. The paradox arises from the fact that most humans would refuse to
pay even a modest amount to play.84
Technically, the St. Petersburg paradox may be characterized as a reverse
martingale:85 The gambler surrenders her or his entire stake up front as
the price of entry into the game, in exchange for a chance at a stream of
wealth whose expected value is infinite.86 The instinctive aversion to this
game arguably reflects an intuitive grasp of gamblers ruin, the certainty
with which a bettor with finite wealth playing against a foe with infinite
wealth will eventually lose everything, even with fair odds.87 But the value
of the game remains infinite, and human players routinely refuse to play
it. The St. Petersburg paradox therefore presents a particularly extreme
illustration of the failure of conventional risk-aversion models expected
outcome criterion to account for risk.88
Paul Samuelson pinpoints the technical flaw that prevents many utility
functions from being able to provide a plausible explanation for the St.
Petersburg paradox. To provide consistent rankings of risky prospects, a
utility function must be bounded from above and from below.89 In addi-
tion, in their original article outlining prospect theory, Daniel Kahneman
and Amos Tversky credited Harry Markowitz with a more sophisticated
128 J.M. CHEN
understanding of risk than traditional models of risk aversion permitted:

Markowitz was the first one to propose that utility be defined on gains
and losses rather than final asset positions . Markowitz also noted the
presence of risk seeking in preferences among positive as well as negative
prospects, and he proposed a utility function with has convex and concave
regions in both the positive and the negative domains.90
Utility functions following upper and lower bounds and reflecting both
concavity and convexity are absent in the HARA family.91 In particular,
the widely used HARA specifications of constant absolute and constant
relative risk aversion fail to satisfy these criteria. Of course, satisfying
Samuelsons and Markowitzs criteria comes at a modest price of its own:
Those who are tempted to insist upon both an upper and lower bound for
utility must sacrifice the property of utility concavity and general risk aver-
sion everywhere. They gain the property that, for any defined probability
distribution of outcomes, a transitive ordering is well defined.92
There are many ways to put an upper and lower bound on utility and to
relax the assumption that preferences are invariably concave. Many cumu-
lative distribution functions exhibit well-behaved properties that make
them good candidates for modeling utility. They are positive, have posi-
tive first derivativesthe probability densitiesand if they are unimodal,
have negative second derivatives beyond the mode.93 As we will see in
8.4 and 8.5, the value function in prospect theory, one of the most impor-
tant expressions of behavioral economics, may be rendered in precisely this
sort of sigmoid form.
Notes
1. Cf. Michael Lewis, The Blind Side 167 (Michael Lewis afterword, 2009)
(1st ed. 2006) (describing future professional football player Michael
Ohers superlative score for protective instincts on an aptitude test).
2. See generally, e.g., Levy, CAPM in the 21st Century, supra note 41 (Chapter
1), at 2362.
3. See generally Paul J.H. Schoemaker, The Expected Utility Model: Its
Variants, Purpose, Evidence and Limitations, 20 J. Econ. Lit. 529563
(1982).
4. Compare John Neumann & Oskar Morgenstern, Theory of Games and
Economic Behavior (1953); and William J. Baumol, The Neumann-
Morgenstern Utility IndexAn Ordinalist View, 59 J. Pol. Econ. 6166
(1951) with, e.g., Peter C. Fishburn, Nonlinear Preference and Utility
RISK AVERSION 129
Theory (1988); Colin F. Camerer, An Experimental Test of Several

Generalized Utility Theories, 2J.Risk & Uncertainty 61104 (1989); and
Mark J. Machina, Choices Under Uncertainty: Problems Solved and
Unsolved, 1J.Econ. Persp. 121154 (1987).
5. See, e.g., Paul J.H.Schoemaker, Determinants of Risk-Taking: Behavioral
and Economic Views, 6J.Risk & Uncertainty 4973, 53 (1993).
6. See infra 7.17.9, at 137182.
7. Kenneth J.Arrow, The Theory of Risk Aversion, in Aspects of the Theory of
Risk Bearing: Yrj Jahnsson Lectures 2844 (1965), reprinted in Kenneth
J. Arrow, The Theory of Risk Aversion, in Essays in the Theory of Risk-
Bearing 90109 (1971).
8. John W.Pratt, Risk Aversion in the Small and in the Large, 32 Econometrica
122136 (1964).
9. Risk aversion and risk measures may be rendered dynamically. See generally
Emanuela Rosazza Gianin, Risk Measures Via g-Expectations, 39 Ins.:
Math. Econ. 1934 (2006).
10. See generally Arrow, The Theory of Risk Aversion, supra note 7.
11. See https://en.wikipedia.org/wiki/Taylor_series
12. See, e.g., Carl P.Simon & Lawrence E.Blume, Mathematics for Economists
363 (1994); Mossin, supra note 9 (Chapter 5), at 215.
13. See Pratt, supra note 8, at 124.
14. See id. at 125.
15. Id. at 125 n.3
16. See id. at 125.
17. See Jonathan E.Ingersoll, Jr., Theory of Financial Decision Making 3940
(1987).
18. Merton, Optimum Consumption, supra note 9 (Chapter 5), at 389.
19. See Ingersoll, supra note 17, at 39.
20. See Merton, Optimum Consumption, supra note 9 (Chapter 5), at 389.
21. See Ingersoll, supra note 17, at 3940; Merton, Optimum Consumption,
supra note 9 (Chapter 5), at 389 & n.19. For illustrative applications of
specialized risk aversion models, see Atanu Saha, Flexible Utility Form, 75
Am. J.Agric. Econ. 905913 (1993); Atanu Saha, C.Richard Shumway &
Hovav Talpaz, Joint Estimation of Risk Preference Structure and Technology
Using Expo-Power Utility, 76 Am. J. Agric. Econ. 173184 (1994);
Danyang Xie, Power Risk Aversion Utility Functions, 1 Annals Econ. & Fin.
265282 (2000).
22. See Merton, Optimum Consumption, supra note 9 (Chapter 5), at 389
n.19.
23. See generally supra 3.2, at 6066.
24. See A.E.Taylor, LHospitals Rule, 59:1 Am. Math. Monthly 2024 (Jan.
1952).
25. See Ingersoll, supra note 17, at 40.
130 J.M. CHEN
26. Rajnish Mehra, The Equity Premium: Why Is It a Puzzle?, 59:1 Fin. Analysts
J. 5469, 57 (Jan./Feb. 2003); Mehra & Prescott, The Equity Premium in
Retrospect, supra note 37 (Chapter 3), at 901. On real business cycle theory,
see Thomas F.Cooley, Frontiers of Business Cycle Research (1995); David
Romer, Real-Business-Cycle Theory, in Advanced Macroeconomics 189237
(4th ed. 2011); Robert J. Hodrick & Edward C. Prescott, Postwar
U.S. Business Cycles: An Empirical Investigation, 29 J. Money, Credit &
Banking 116 (1997); Finn E. Kydland & Edward C. Prescott, Time to
Build and Aggregate Fluctuations, 50 Econometrica 13451370 (1982);
George W.Stadler, Real Business Cycles, 32J.Econ. Lit. 17501783 (1994);
Lawrence H.Summers, Some Skeptical Observations on Real Business Cycle
Theory, 10:4 Fed. Reserve Bank Minneapolis Q.Rev. 2327 (Fall 1986).
27. For different approaches to capital asset pricing given the presence of het-
erogeneous investor beliefs, see Anat R. Admati, A Noisy Rational
Expectation for Multi-Asset Securities Markets, 53 Econometrica 629658
(1985); Suleyman Basak, Asset Pricing with Heterogeneous Beliefs,
29 J. Banking & Fin. 28492881 (2005); Peter DeMarzo & Costis
Skiadas, Aggregation, Determinacy, and Informational Efficiency for a
Class of Economics with Asymmetric Information, 80J.Econ. Theory 123
152 (1998); Jerome Detemple & Shashidhar Murthy, Intertemporal Asset
Pricing with Heterogeneous Beliefs, 62J.Econ. Theory 294320 (1994);
Haim Levy, Moshe Levy & Golan Benita, Capital Asset Pricing with
Heterogeneous Beliefs, 79 J. Bus. 13171353 (2006); John Lintner, The
Aggregation of Investors Diverse Judgments and Preferences in Purely
Competitive Security Markets, 4J.Fin. & Quant. Analysis 347400 (1969);
Joseph T.Williams, Capital Asset Prices with Heterogeneous Beliefs, 5J.Fin.
Econ. 219239 (1977).
28. See Henry Allen Latan, Criteria for Choice Among Risky Ventures,
67J.Pol. Econ. 144155 (1959); Harry M.Markowitz, Investment for the
Long Run: New Evidence for an Old Rule, 31J.Fin. 12731286 (1976).
29. Levy, CAPM in the 21st Century, supra note 41 (Chapter 1), at 172; see
sources cited supra notes 1518 (Chapter 5) (describing the separation
theorem).
30. See generally Paul Lvy, Calcul des Probabilits (1925); Benoit Mandelbrot,
The Pareto-Lvy Law and the Distribution of Income, 1 Intl Econ. Rev.
79106 (1960); Benoit Mandelbrot, Stable Paretian Random Functions and
the Multiplicative Variation of Income, 29 Econometrica 517543 (1961);
Paul A.Samuelson, Efficient Portfolio Selection for Pareto-Lvy Investments,
2J.Fin. & Quant. Analysis 107122 (1967); Chen, Postmodern Portfolio
Theory, supra note 1 (Chapter 1), 14.2, at 262263.
31. James A.Xiong & Thomas M.Idzorek, The Impact of Skewness and Fat
Tails on the Asset Allocation Decision, 67:2 Fin. Analysts J. 2335, 24
(March/April 2011).
RISK AVERSION 131
32. See Benoit B. Mandelbrot, The Variance of Certain Speculative Prices,

36J.Bus. 394419 (1963).
33. See Eugene F. Fama, Mandelbrot and the Stable Paretian Hypothesis,
36J.Bus. 420429 (1963).
34. Eugene F.Fama, The Behavior of Stock-Market Prices, 38J.Bus. 34105,
42 (1965).
35. Benoit B.Mandelbrot, The Fractal Geometry of Nature 337338 (1983).
36. Campbell, Lo & MacKinlay, supra note 5 (Chapter 1), at 18. But cf.
S. Ortobelli L & S.T. Rachev, Safety-First Analysis and Stable Paretian
Approach to Portfolio Choice Theory, 34 Math. & Computer Modeling
10371072 (2001) (using the stable Paretian distribution to model heavy-
tailed return distributions and comparing that model to a safety-first
approach). On Roys safety-first principle, see generally infra 10.4.
37. See https://en.wikipedia.org/wiki/Stable_distribution
38. Xiong & Idzorek, supra note 31, at 24.
39. Aparicio & Estrada, supra note 81 (Chapter 4), at 2 & n.5.
41. Id.
42. R.R.Officer, The Distribution of Stock Returns, 67J.Am. Stat. Assn 807
812 (1972).
43. M.S.Feldstein, Mean-Variance Analysis in the Theory of Liquidity Preference
and Portfolio Selection, 36 Rev. Econ. Stud. 512, 5 n.2 (1969).
44. Id.
46. Harry M. Markowitz, Portfolio Selection: Efficient Diversification of
Investments, at ix (2d ed. 1991) (1st ed. 1959); accord Juan Salazar
& Annick Lambert, Fama and MacBeth Revisited: A Critique, 1 Aestimatio
4871, 53 n.13 (2010).
47. See supra 6.4, at 115117.
48. See Andr F.Perold, The Capital Asset Pricing Model, 18J.Econ. Persp. 3,
1012 (2004); cf. Korajczyk, supra note 12 (Chapter 3), at xv (defining beta
as the slope of the capital allocation line and the risk-free return (Rf) as a
constant that defines the functions y-intercept). See generally Chen,
Postmodern Portfolio Theory, supra note 1 (Chapter 1), 2.5, 2.8, at 913,
1517.
1), 6.86.10, at 9499.
50. Paul D.Kaplan & James A.Knowles, Kappa: A Generalized Downside Risk-
Adjusted Performance Measure 2 (2004) (available at http://corporate.morn-
ingstar.com/NO/documents/MethodologyDocuments/ResearchPapers/
KappaADownsideRisk_AdjustedPerformanceMeasure_PK.pdf).
132 J.M. CHEN
51. See William F.Shadwick & Con Keating, A Universal Performance Measure,
6:3J.Performance Measurement 5984 (Spring 2002).
52. See Kaplan & Knowles, supra note 50, at 15.
53. See id.
54. See id. at 3.
55. See Hossein Kazemi, Thomas Schneeweis & Raj Gupta, Omega as a
Performance Measure (June 15, 2003) (available at http://faculty.fuqua.
duke.edu/~charvey/Teaching/BA453_2006/Schneeweis_Omega_as_a.
pdf) (developing a closely related measure called Sharpe-Omega).
56. See Kaplan & Knowles, supra note 50, at 3 n.1
57. See id. at 15.
58. See, e.g., Stephen A.Ross, Some Stronger Measures of Risk Aversion in the
Small and in the Large with Applications, 49 Econometrica 621639
(1981); cf. Denis Conniffe, The Flexible Three Parameter Utility Function,
8 Annals Econ. & Fin. 5763, 58 (2007) (proposing a flexible three
parameterutility function capable of encompass[ing] other systems of
utility functions including the hyperbolic absolute risk aversion (HARA)
family and incorporat[ing] properties, such as subsistence and satura-
tion, that elude other utility functions).
59. See generally Lola L.Lopes, Psychology and Economics: Perspectives on Risk,
Cooperation, and the Marketplace, 45 Ann. Rev. Psych. 197227, 199203
(1994).
60. See, e.g., Robert Libby & Peter C. Fishburn, Behavioral Models of Risk
Taking in Business, 15 J. Accounting Research 272292 (1977); Ralph
O.Swalm, Utility TheoryInsights into Risk Taking, 44 Harv. Bus. Rev.
123136 (1966).
61. Compare, e.g., Pratt, supra note 8, at 122 (characterizing utility functions
with decreasing absolute risk aversion as logical candidates for describ-
ing actual human behavior) with Joseph G. Eisenhauer, Risk Aversion,
Wealth, and the DARA Hypothesis: A New Test, 3 Intl Advances Econ.
Research 4653 (1997) (presenting evidence that absolute risk aversion
affirmatively increases with wealth).
62. Charles Wolf & Larry Pohlman, The Recovery of Risk Preferences from
Actual Choices, 51 Econometrica 843850, 847 (1983); accord Saha,
63. See Milton Friedman & L.J.Savage, Utility Analysis of Choices Involving
Risk, 56J.Pol. Econ. 279304 (1948).
64. Harry Markowitz, The Utility of Wealth, 60J.Pol. Econ. 151158 (1952);
accord Shefrin & Statman, Behavioral Portfolio Theory, supra note 24
(Chapter 2), at 128.
65. See generally infra 9.1, at 217218.
66. See Elie Applebaum & Eliakim Katz, Market Constraints as a Rationale for
the Friedman-Savage Utility Function, 89J.Pol. Econ. 819825 (1981);
RISK AVERSION 133
Philip R.P.Coelho & James E.McClure, Social Context and the Utility of
Wealth: Addressing the Markowitz Challenge, 37J.Econ. Behav. & Org.
305314 (1998); Nils H.Hakansson, Friedman-Savage Utility Functions
Consistent with Risk Aversion, 84 Q.J.Econ. 472487 (1970); cf. Roger
Hartley & Lisa Farrell, Can Expected Utility Theory Explain Gambling, 92
Am. Econ. Rev. 613624 (2002). For legal scholarship exploring the fun-
damental economic challenge that lotteries pose to assumptions such as
wealth maximization and risk aversion, see Lloyd R.Cohen, The Lure of the
Lottery, 36 Wake Forest L.Rev. 705745 (2001); Edward J.McCaffery,
Why People Play Lotteries and Why It Matters, 1994 Wis. L.Rev. 71122.
67. See Matthew Amster-Burton, Paying Down Your Mortgage vs. Investing More
(Nov. 22, 2011) (available at http://blog.mint.com/investing/paying-
down-your-mortgage-vs-investing-more-112011); Michael Kitces, Why Is It
Risky to Buy Stocks on Margin But Prudent to Buy Them On Mortgage?
(Oct. 24, 2011) (available at https://www.kitces.com/blog/Why-Is-It-
Risky-To-Buy-Stocks-On-Margin-B ut-Prudent-To-Buy-Them-On-
Mortgage); Michael Kitces, Why Keeping a Mortgage and a Portfolio May Not
Be Worth the Risk (May 2, 2012) (available at https://www.kitces.com/
blog/why-keeping-a-mortgage-and-a-portfolio-may-not-be-worth-
the-risk).
68. See William Nelson Goetzmann, The Single Family Home in the Investment
Portfolio, 6 J. Real Estate Fin. & Econ. 201222 (1993); cf. John
Y.Campbell, Household Finance, 61J.Fin. 15531604, 15641565 (2006)
(describing the share of asset classesranging from safe assets to vehicles,
real estate, private business, and public equityamong households in the
USA by percentile in the distribution of total assets); Alessandro Bucciol &
Raffaele Mianiaci, Household Portfolio Risk, 19 Rev. Fin. 739783 (2015)
(updating Campbells survey of the distribution of asset classes in household
portfolios according to their percentile rank within the overall population);
Luigi Guiso & Paolo Sodini, Household Finance: An Emerging Field, 2
Handbook of the Economics of Finance 13971532, 14061417 (George
M.Constantinides, Milton Harris & Ren M.Stulz eds., 2013) (same).
69. Marjorie Flavin & Takashi Yamashita, Owner-Occupied Housing and the
Composition of the Household Portfolio, 92 Am. Econ. Rev. 345362, 345
(2002). See generally Roger G.Ibbotson & Laurence B.Siegel, Real Estate
Returns: A Comparison with Other Investments, 12 Real Estate Econ. 219
242 (1989).
70. See Joo F.Cocco, Portfolio Choice in the Presence of Housing, 19 Rev. Fin.
Stud. 535567 (2005).
71. See Thomas A. Becker & Reza Shabani, Outstanding Debt and the
Household Portfolio, 23 Rev. Fin. Stud. 29002934, 2918, 2920 (2010).
72. Id. at 2908; see also Zvi Bodie, Robert C.Merton & William F.Samuelson,
Labor Supply Flexibility and Portfolio Choice in a Life-Cycle Model,
134 J.M. CHEN
16J.Econ. Dynamics & Control 427449 (1992) (reasoning that the lack
of correlation between income from labor and equity returns leads to low
levels of bond ownership); cf. Stephen J.Davis, Felix Kubler & Paul Willen,
Borrowing Costs and the Demand for Equity over the Life Cycle, 88 Rev.
Econ. & Stat. 348362, 356357 (2006) (observing that bonds and
income from labor are substitutes in the sense that neither is correlated
with investment income); Rui Yao & Harold H. Zhang, Optimal
Consumption and Portfolio Choice with Risky Housing and Borrowing
Constraints, 18 Rev. Fin. Stud. 197239, 212 (2005) (same).
73. Becker & Shabani, supra note 71, at 2931.
74. See M.Allais, Le Comportement de lHomme Rationnel Devant le Risque:
Critique des Postulats et Axiomes de lcole Amricaine, 21 Econometrica
503546, 525528 (1953). In accord with Levy, CAPM in the 21st
Century, supra note 41 (Chapter 1), at 3940, I have converted Allaiss
choice of currency from the (old) French franc to the modern US dollar
and reduced the sums in his paradox by two orders of decimal magnitude.
Inasmuch as the French franc traded at levels between 119.1 and 350 and
eventually 493.7 francs to the dollar between the institution of the Bretton
Woods foreign exchange system and the 1960 introduction of the new
franc, this 100:1 reduction is not entirely out of line. It also appears that
the old franc was worth roughly two-hundredths of a euro in 2007 terms.
See https://en.wikipedia.org/wiki/French_franc
75. See, e.g., Daniel Ellsberg, Risk, Ambiguity, and the Savage Axioms, 75
Q.J.Econ. 643669 (1961); Howard Raiffa, Risk, Ambiguity, and the Savage
Axioms: Comment, 75 Q.J. Econ. 690694 (1961); Paul Slovic & Amos
Tversky, Who Accepts Savages Axioms?, 19 Behav. Sci. 368374 (1974).
76. See, e.g., Expected Utility Hypotheses and the Allais Paradox: Contemporary
Discussions of Decisions Under Uncertainty with Allais Rejoinder
(Maurice Allais & Ole Hagen eds., 2013); David M.Kreps, Notes on the
Theory of Choice 192 (1988); Maurice Allais, An Outline of My Main
Contributions to Economic Science, 87 Am. Econ. Rev. 312 (1997); Adam
Oliver, A Quantitative and Qualitative Test of the Allais Paradox Using
Health Outcomes, 24J.Econ. Psych. 3548 (2003).
77. Levy, CAPM in the 21st Century, supra note 41 (Chapter 1), at 306.
78. Id.
79. Matthew Rabin, Risk Aversion and Expected-Utility Theory: A Calibration
Theorem, 68 Econometrica 12811292, 1282 (2000).
81. Id.; accord Olof Johansson-Stenman, Risk Aversion and Expected Utility of
Consumption over Time, 68 Games & Econ. Behav. 208219 (2010); see
also Mark J. Machina, Expected Utility Theory Without the Independence
Axiom, 50 Econometrica 277323 (1982).
RISK AVERSION 135
82. See Bengt Hansson, Risk Aversion as a Problem of Conjoint Measurement,

in Decision, Probability, and Utility: Selected Readings 136158 (Peter
Grdenfors & Nils-Eric Sahlin eds., 1988).
83. See Daniel Bernoulli, Exposition on a New Theory on the Measurement of
Risk, 22 Econometrica 2236 (1954) (1st ed. 1738) (Louise Sommer
trans.). Bernoulli learned about this problem from his cousin, Nicolas, in
1713. See Correspondence of Nicolas Bernoulli Concerning the St. Petersburg
Game (Jan. 1, 2013) (Richard J. Pulskamp trans.) (available at http://
cerebro.xu.edu/math/Sources/NBernoulli/correspondence_petersburg_
game.pdf). I have adapted the description of the paradox from https://
en.wikipedia.org/wiki/St._Petersburg_paradox. Notably, Bernoullis rec-
ommended strategy of maximizing the geometric mean of expected out-
comes is mathematically equivalent to the Kelly criterion for optimizing
the size of a series of bets. See sources cited supra note 118 (Chapter 1).
84. See, e.g., Ole Peters, The Time Resolution of the St Petersburg Paradox, 369
Phil. Trans. Royal Socy 49134931 (2011); Marc Oliver Rieger & Mei
Wang, Cumulative Prospect Theory and the St. Petersburg Paradox, 28
Econ. Theory 665679 (2006).
85. See generally David Williams, Probability with Martingales 93105 (1991).
86. See Michael Mitzenmacher & Eli Upfal, Probability and Computing:
Randomized Algorithms and Probabilistic Analysis 298 (2005).
87. See generally J.W.F.Edwards, Pascals Problem: The Gamblers Ruin, 51
Revue Internationale de Statistique 7379 (1983).
89. See Paul A.Samuelson, St. Petersburg Paradoxes: Defanged, Dissected, and
Historically Described, 15J.Econ. Lit. 2455, 35 (1977).
90. Daniel Kahneman & Amos Tversky, Prospect Theory: An Analysis of Decision
Under Risk, 47 Econometrica 263291, 276 (1979); accord Levy, CAPM
in the 21st Century, supra note 41 (Chapter 1), at 301.
91. See supra 6.4, at 115117.
92. Samuelson, St. Petersburg, supra note 89, at 36.
93. Denis Conniffe, The Generalised Extreme Value Distribution as Utility
Function, 38 Econ. & Soc. Rev. 275288, 275 (2007); see also Marco
LiCalzi & Annamaria Sorato, The Pearson System of Utility Functions, 172
Eur. J.Oper. Research 560573 (2006).
CHAPTER 7
The Equity Risk Premium andtheEquity

Premium Puzzle
7.1 The Equity Risk Premium

All of finance rests on the proposition that investors dislike risk and demand
higher returns as compensation for bearing risk. In behavioral terms, the
equity risk premium may be regarded as the additional rate of return that
risk-averse investors, as a class, demand in exchange for the burden of
bearing volatility and the attendant risk of downside loss. Although one
study has concluded that the replacement of standard deviation in the
conventional CAPM by a downside risk measure would advise investors
to lower the stock allocations within their portfolios,1 another study sug-
gests that investors reliance on fixed-income positions vastly exceeds the
allocation that any strictly rational, utilitarian evaluation of risk in equity
investing would ever counsel.2 Given the presence of a sizeable equity
premium, why indeed should a substantial fraction of investable wealth
[be] invested in fixed income instruments?3
But the extent of this premium comes as an econometric shock and
presents a persistent, still unresolved puzzle. In their original 1985 for-
mulation of the equity premium puzzle, Rajnish Mehra and Edward
Prescott found that the average real annual yield on equity in hypo-
thetically competitive pure exchange economies should be a maximum

DOI10.1007/978-3-319-32711-2_7
138 J.M. CHEN
of four-tenths of a percent higher than that on short-term debt, a sur-

rogate for the risk-free rate.4 This result stood in sharp contrast to the
six percent premium observed in Mehra and Prescotts actual survey of
American economic history.5
In revisiting the problem nearly two decades later, Mehra and Prescott
observed: The historical U.S. equity premium (the return earned by a
risky security in excess of that earned by a relatively risk free U.S.T-bill)
is an order of magnitude greater than can be rationalized in the context
of the standard neoclassical paradigm of financial economics.6 The equity
premium puzzle cannot be dismissed lightly, since much of our economic
intuition is based on the very class of models that fall so dramatically when
confronted with financial data.7 The puzzles persistence underscores
the failure of paradigms central to financial and economic modeling
to capture the characteristic that appears to make stocks comparatively
so risky.8 Because the equity premium is so practically important, and
because the equity premium puzzle continues to defy theoretical under-
standing,9 this book will contextualize its consideration of risk aversion by
taking a detailed look at the equity risk premium.
There is some dispute over the magnitude of the equity risk premium,
but not enough to nullify its existence or to diminish its practical and the-
oretical significance. One survey has estimated that American stocks from
1802 through 2012 delivered an average annual inflation-adjusted return
of 6.6%, as compared with 3.6% on bonds over the same period.10 This
evaluation is not out of line with other professional and academic esti-
mates. A 2003 study has estimated that American equities have exceeded
the return on Treasury bonds by a geometrically compounded average of
4% per year.11
In the 2014 iteration of their periodic evaluation of the equity risk pre-
mium, John Graham and Campbell Harvey have forecast a total return on
stocks of 6.43%, good for an equity risk premium of 3.73% (which in turn
represents to a level only slightly higher than the long-term average, after a
crisis-induced spike peaking in February 2009).12 Globally, the equity risk
premium hovers around 33% on a geometric mean basis, or approxi-
mately 45% on an arithmetic basis.13 In fairness, a different survey of
stock market growth around the world from 1926 to 1996 found a global
median of 0.8% per year, in stark contrast with an annual real growth rate
of 4.3% for the USA.14
THE EQUITY RISK PREMIUM ANDTHEEQUITY PREMIUM PUZZLE 139
The sheer length of these studies ameliorates concerns that the equity
risk premium is subject to shifts in probability distribution aptly described
as structural breaks, or systematic disruptions in the relationship
between risk and return, including the magnitude of the equity risk pre-
mium.15 The presence of structural breaks in the risk-return relationship
is arguably the financial manifestation of punctuated equilibria in evolu-
tionary biology, the idea that speciation, far from being gradual, uniform,
and constant, occurs in rare and geologically rapid spurts.16 Other social
sciences17 (including computational linguistics and the anthropological
analysis of mythology)18 have consciously incorporated the idea of punc-
tuated equilibria. It remains an open question whether the importation
of this insight from evolutionary biology will significantly enhance finan-
cial understanding of structural breaks in the equity premium, speculative
bubbles, and other similar phenomena.
In fairness, advocates of fixed-income investing have marshaled some
evidence contradicting the magnitude and durability of an equity risk pre-
mium. Robert Arnott argues that historic expectations of 8% real returns
on stock and a 5% risk premium over bonds are not supported by evidence
of long-run dividend growth and inflation expectations and concluding
instead that the long-term, forward-looking risk premium may be near
zero or even negative.19 In response to the provocative question Bonds:
why bother? Arnott argues that bonds have beaten stocks for the past
40 yearsor at least have delivered real returns in excess of those on
stocks during certain 10-, 20-, and 40-year periods.20
Despite this contrary line of argument, Rajnish Mehras 2003 survey
of the equity risk premium around the world retains its power to surprise
(Table 7.1).21 As Mehra took pains to point out, these five markets (USA,
Table 7.1 The equity risk premium in major markets during part or all of the
twentieth century
Country Period Mean real return (in percentage points) Equity risk premium
(in percentage points)
Market index Relatively riskless
security
USA 18892000 7.9 1.0 6.9

UK 19471999 5.7 1.1 4.6
Japan 19701999 4.7 1.4 3.3
Germany 19781997 9.8 3.2 6.6
France 19731998 9.0 2.7 6.3
140 J.M. CHEN
UK, Japan, Germany, France) comprised more than 85% of the worlds
capitalized equity value.22 The equity risk premium prevails around the
world. In absolute terms, it is substantial.
Or stated somewhat differently: The terminal (real) value, as of 2000, of
$1 invested in stocks in 1926 would be $266.47, while the same $1 invested
in Treasury bills that year would be worth $1.71 in real terms by 2000,
assuming reinvestment of all dividends and interest and zero taxes.23 Those
numbers imply an average equity risk premium of 7.1% from 1976 to 2000
(7.8% annual real return on stocks, versus 0.7% on T-bills).24 Comparable
back-of-the-envelope calculations report ratios of 66:1 for all-stock ver-
sus all-bond portfolios from 1925 to 199725 and up to 7:1 for retirement
accounts built over 40-year time frames between 1876 and 1900.26
7.2 A Cautious Stroll off Wall Street

Although the equity premium puzzle is a quantitative puzzle,27 the
comparably persistent failure of investors to exploit the equity premium
also poses a serious behavioral puzzle and a corresponding challenge for
public policy. Without suggesting that questions of behavioral finance
should be (much less could be) segregated from the rest of mathematical
finance, I merely wish to acknowledge the qualitative difference between
the theoretical value of properly specified econometric models and the
practical value of discovering the motivation of individual investors. And
one great practical puzzle in mature economies such as the USA, given the
size and durability of the equity risk premium, is why any investor opts out
of stock ownership.
In 1995, a decade after Mehra and Prescott challenged fellow econo-
mists to solve the equity premium puzzle, surveys of data presented in
the Panel Study of Income Dynamics pegged the proportion of stock-
holders among American households, including indirect holdings
through pension funds and IRAs, at no higher than 36.8%.28 After
another two decades of tumultuous change in the financing of retirement
in the USA, during which privately funded pensions shifted away from
the defined benefit model of a bygone era,29 not even higher participa-
tion in defined contribution plans has made stockholders of a majority
of Americans. The paradoxical combination of an equity premium (and a
considerable one at that) with widespread failure to exploit it gives rise to
a stock market participation puzzle: [E]ven though the stock market
has a high mean return and a low correlation with other household risks,
many households have historically been reluctant to allocate any money to

it.30 Even in affluent households, rates of stock market nonparticipation
are far from trivial. [A]t the 80th percentile of the wealth distribution in
the USA, the fraction of households that hold neither private business
assets nor public equity is just under 10%.31
It has long been understood, contrary to the separation theorem and
other precepts of modern portfolio theory,32 that most investors do not
hold the market-wide portfolio.33 The failure to hold well-diversified port-
folios knows no national boundaries; it also affects international investors.34
The Federal Reserves 2014 survey of consumer finances revealed that
the total participation in the stock market among American households,
including both direct ownership of stock and indirect ownership through
retirement plans and the like, fell from a peak of 53.2% in 2007 to 48.8%
in 2013.35 The participation rate fell below one-half even though nearly
four-fifths of upper-middle-class households (defined as the 50th through
89.9th percentile) with adults aged 35 through 64 (regarded as prime-
age families whose heads, in general, have finished their education and
have established careers, but are too young to retire fully) participated in
a retirement plan of some sort.36 A mere 13.8% of American households
hold stocks directly.37 And even among those households that do hold
securities, either directly or indirectly, for retirement purposes or in more
liquid, taxable accounts, financial advisors38 and mutual fund managers39
counsel conservative investors to hold more bonds than stocks, in square
contradiction not only of the equity risk premium, but also of Tobins
two-fund separation theorem,40 which prescribes maintaining a constant
ratio between stocks and bonds and varying the level of the risk-free asset
as the tool for managing risk aversion.
The American nightmare of societal underinvestment in the stock mar-
ket arises in part from the American dream of mass home ownership. The
US policy of using tax deductions to favor highly leveraged purchases of
owner-occupied housing41 suppresses societal rates of stock market par-
ticipation.42 One survey shows that households owing mortgage debt are
10% less likely to own stocks.43 Debt service is sufficiently high that 25.8%
of such households arguably should forgo equity market participation, in
the sense that repaying their outstanding debt obligations [offers] an
after-tax interest rate higher than the average after-tax return to stock
ownership.44 Homeownership distorts household finance in other odd
ways, not least of which is the failure of homeowners to adopt rational
approaches toward mortgage refinancing.45
142 J.M. CHEN
All of this has occurred against the backdrop of dire predictions of lack
of preparedness for retirement.46 Among the ways in which the federal
government hopes to exploit behavioral science insightsdefined as
research findings from behavioral economics and psychology about
how people make decisionsa desire to help Americans accumulate
additional savings ranks very high.47 Boston Colleges Center for
Retirement Research estimates that financial retirement insecurity plagues
more than half of the American population;48 other sources sound an
even louder alarm.49 The Congressional Research Service estimates that
the average household in which the head was aged 6267in 2010 had
$341,417 in total retirement assets (defined as individual retirement
accounts and other defined contribution vehicles).50 The median house-
hold in this category had $150,000in total retirement assets.51 Even at
an arguably unsustainable withdrawal rate of 46%, the higher $341,417
figure generates between $13,657 and $20,485in annual gross income.
Other sources of retirement income come at a steep price. Early claiming
of Social Security benefits locks in a lower lifetime annuity; it increased
during the recession of 20072009.52 Reverse mortgages can help retirees
unlock their home equity, but these expensive, highly complex products
are viewed as a retirement planning tool of last resort.53
It may be mere coincidence that the 2013 rate of stock market partici-
pation, direct and indirect, reported by the Federal Reserve (48.8 %) is
almost exactly the mirror image of that years at-risk percentage reported
by Boston Colleges National Retirement Risk Index (52 %). Although
other factorsranging from outright poverty to a behavioral tendency
to consign retirement savings behind other financial priorities54plainly
affect preparedness for retirement,55 aversion to stock market participation
contributes to American social insecurity. Even accounting for inertia,56
the impact of differences in financial literacy on stock market participa-
tion,57 and the possibility of simple misunderstanding that leaves many
people confused about the relative safety of different investments over
long horizons,58 we are left to explain the presence of a nontrivial num-
ber of individuals [who] are highly averse to investing more than a small
percentage of their financial assets in stocks.59 The question is no lon-
ger, strictly speaking, why many wealthy households hold no stock at
all,60 but rather why so few hold stocksperiod.61
Social insecurity in retirement is hardly the only economic conse-

quence flowing from the equity risk premium and the failure of many
investors to exploit it. For every microeconomic phenomenon, there is
a macroeconomic consequence, and vice versa. Economic models that
generate[] values of the equity premium consistent with observed asset
prices suggest that the welfare costs of systematic risk should be com-
parable in magnitude to the welfare benefits of economic growth.62 It
therefore follows that the most direct and obvious implication of a large
risk premium for equity is that the economic cost of the systematic risk in
returns to equity is also large.63
We can extend this simple logic with a simple form of the conventional
CAPM.If the return on any corporation is given by the basic CAPM for-
mula, ra=rf + (rmrf), the equity premium may be expressed as rmrf.
A larger value for rmrf implies a larger return on firm assets, since rais-
ing rmrf, ceteris paribus, raises ra. Assuming constant mean cash flows,
the payback period for recouping investment in a capital asset is the
reciprocal of ra.64 The return-driven compression of the payback period
heightens the risk of short-termism in corporate decision-making, par-
ticularly if investors misperceive the riskiness of long-term investments
(say those with returns more than five years in the future) and, as a
consequence, demand a return that is disproportionately large for firms
undertaking such investments.65 At an extreme, especially if this logic is
coupled with the rather uncontestable observation that adverse selection
precludes insurance against systematic risk in labor income,66 the presence
of an equity premium could be wielded as a justification for public owner-
ship of equity, with the goal of eliminating the premium and optimizing
social welfare.67
One need not countenance, let alone endorse, nationalization to rec-
ognize the practical and intellectual significance of the problem. Resolving
the equity premium puzzle would not only address a long-standing chal-
lenge to the theoretical cogency of mathematical finance, but also con-
tribute to the formulation of better policies for individual savers, financial
advisors, pension plan managers, and the Social Security administration.
Two things persist. One is the quantitative puzzle. The other is a more
practical, even demographic, form of persistence. A significant fraction of
the population continues to resist participation in the equity market. The
balance of this chapter will address both forms of persistence.
144 J.M. CHEN
7.3 The Equity Premium Puzzle

Armed with 6.4s presentation of a utility function modeling CRRA,
we are now prepared to tackle the equity premium puzzle. The following
presentation is a highly compressed version of descriptions of the puzzle
by Rajnish Mehra and Edward Prescott.68
Consider a simple economy represented by a typical household that
orders its preferences according to the following function:

E0 b tU ( ct ) , 0 < b < 1
t =0
where:
E0() = expectation operator, conditional on information available at t = 0
= a subjective time discount factor describing the households
impatience as its unwillingness to defer consumption
U() = a concave, monotonically increasing, continuously differentia-
ble utility function
ct = per capital consumption
We further restrict U() to the class of CRRA utility functions:
c1-a
U ( c, a ) = , 0 <a <
1-a
Consistent with the description of CRRA utility functions in 6.4,69 U(c, )

is defined as a logarithmic function according to its limit as approaches
1. Substituting U(c, ) into our original preference function yields:
c1-a
E0 b t , 0 < a < , 0 < b < 1
t =0 1 - a
Mehra and Prescott assume that this stylized economy has a single
equity share, whose share price is pt and which represents a claim on pro-
ductivity from the stochastic process {yt}. The application of a fundamental
pricing relationship70 allows the expected gross return on equity to be
expressed in the familiar form of the conventional CAPM:
-U ( ct +1 ) , Re, t +1
Et ( Re, t +1 ) = R f , t +1 + cov t
Et U ( ct +1 )
This relationship allows the equity premium to be easily computed as

Et(Re, t+1)Rf, t+1. In the words of Mehra and Prescott:
Expected asset returns equal the risk-free rate plus a premium for bearing
risk, which depends on the covariance of the asset returns with the marginal
utility of consumption. Assets that covary positively with consumption
that is, assets that pay off in states when consumption is high and marginal
utility is lowcommand a high premium because these assets destabilize
consumption.71
The question is whether covariance between the assets and the marginal
utility of consumption is large enough to justify the observed equity
premium.72
By adopting a model proposed by Andrew Abel,73 Mehra and Prescotts
2003 restatement of the equity premium puzzle subjected their original
1985 presentation of the puzzle74 to certain simplifying assumptions:
The growth rate of consumption (xt) and the growth rate of divi-
dends (zt) are both identically and independently distributed (i.i.d.).
Those growth rates (consumption, dividends) are jointly lognor-
mally distributed.
As a consequence of those assumptions, gross return on equity is
also i.i.d.
As a further consequence, gross return on equity (Re, t) and the growth
rate of consumption are also jointly lognormally distributed.75
These assumptions enable the expected return on equity to be written

in terms of the growth rates of dividends and of consumption:
E t ( z t +1 )
Et ( Re, t +1 ) =
b Et ( zt +1 xt-+a1 )

There is an an alogous expression of the risk-free return:

146 J.M. CHEN
1
Re,t +1 =
b Et ( xt-+a1 )

Applying the properties of the lognormal distribution and imposing

the model equilibrium condition that x = znamely, that the return
on equity be perfectly correlated with the growth rate of consumption
yields the following expression of the equity premium as a function of the
risk-free rate, our original coefficient of relative risk aversion (), and the
variance of the growth rate of consumption:76
ln E ( Re ) - ln R f = as x2

Since variance in the growth rate of consumption is determined empiri-

cally from historic econometric data, all that remains is the calculation of
, the coefficient of relative risk aversion. The trouble is that the variance
of the growth rate is far too small (approximately 0.00125, or 1/800) to
account for an observed equity risk premium in the neighborhood of six
percentage points, absent outlandish, implausible values for .77 In con-
crete terms, historically observed equity premiums imply that the typical
American investor would forgo a gamble that would pay either $50,000
or $100,000 with equal probability, in favor of a guaranteed payout of
$51,858.78 In other words, accepting this seemingly airtight account of
the equity risk premium leads to the unbelievable conclusion that a typical
American investor, having received $50,000, would prefer to enhance that
payout by another $1,858, guaranteed, rather than gambling the marginal
$1,858 on a coin flip worth $50,000 if it is correctly called.
7.4 Another Puzzle, andaChallenge

Locked within the equity premium puzzle is another enigma. Mehra and
Prescotts original formulation of the equity premium puzzle assumed
that an individual who was averse to variation in levels of consumption
at a particular point of time would necessarily be averse to consump-
tion variation over [all] time.79 In response, Larry Epstein and Stanley
Zin defined a more flexible general expected utility function that
defines independent parameters for the coefficient of risk aversion and the
e lasticity of intertemporal substitution.80 Epstein and Zins utility function

can be stated as follows:

{ ( )
U ct , Et (U t +1 ) = (1 - b ) ct1-a /q + b Et U t1+-1a }
1/q q / (1-a )
where ct is consumption at time t, > 0 is the coefficient of relative

risk aversion, and 0 < < 1 is the impatience parameter.81 The variable
representing elasticity of intertemporal substitution, > 0, does not
1 - a 82
directly appear. Instead, contributes to the definition of q .
1 - 1 /y
By defining independently, Epstein and Zins general expected utility
function allows the elasticity of intertemporal substitution to differ from
the reciprocal of risk aversion.83 (That quantity, it is worth recalling from
6.4, represents the coefficient of risk tolerance under hyperbolic absolute
risk aversion models.)84 This flexibility enables a model of consumption
utility to postulate high risk aversion [and] resolve the equity premium
puzzle without driving the elasticity of intertemporal substitution to an
unreasonably low value.85
Although Mehra and Prescott have acknowledged that Epstein and
Zins general expected utility approach would resolve the risk-free rate
component of the equity premium puzzle,86 or at least ameliorate it,87
they argue that the separate parameterization of risk aversion and the
elasticity of intertemporal substitution relies too heavily on unobservable
variables to provide a complete solution.88 Mehra and Prescott have also
argued that evidence reporting low elasticity of intertemporal substitution
undermines Epstein and Zins claim to generality.89
Whether the elasticity of intertemporal substitution in fact fairly small
is subject to some dispute.90 What is certain is that given extremely small
values for the elasticity of intertemporal substitution, investors have an
overpowering preference for a flat consumption path.91 In light of his-
torical upward drift in productivity and, consequently, in consumption as
well, an overwhelming preference for a flat consumption path implies an
extremely strong desire to borrow from the future.92 The only remedy for
that problem is a low or even negative rate of time preference,93 some-
thing unlikely to be supported by a body of economic literature where the
only real debate pits exponential against hyperbolic discounting.94 We are
148 J.M. CHEN
left to reconcile theories of asset pricing with a counterfactually high real

interest rate.95 Indeed, Mehra and Prescotts application of their own
restatement of the original equity premium puzzle under the lognormal
distribution resulted in a barely plausible risk-free rate of 12.7 percent!96
The prevalence of much lower risk-free interest rates in real life gives rise
to a corollary of the equity premium puzzle: the risk-free rate puzzle.97
The equity premium puzzle may be justifiably regarded as a singularly
stiff behavioral challenge to the theoretical integrity of portfolio allocation
and asset pricing theories. For reasons akin to those provided in 4.1 for
describing the low-volatility anomaly as possibly the greatest anomaly in
finance,98 we have a strong case for treating the equity premium puzzle
as the foundational problem of behavioral finance. The persistence of the
equity premium shares this much with the long-term outperformance
of low-risk portfolios: both phenomena fundamentally challenge[] the
basic notion of a risk-return tradeoff.99
Narayana Kocherlakota goes even further in characterizing the scien-
tific significance of the problem. The failure to solve the risk-free rate
puzzle indicates that we do not know why people save even when returns
are low, suggesting that models of aggregate savings behavior are omit-
ting some crucial element.100 For its part, [t]he equity premium puzzle
demonstrates that we do not know why individuals are so averse to the
highly procyclical risk associated with stock returns.101 Together, the two
puzzles signal large gaps in our understanding of the macroeconomy.102
As noted in 1.2, scientific revolutions invariably begin with puzzles
defying the logic of existing normal science. As the low-volatility anom-
aly prompted efforts to describe the dynamics of abnormal markets, the
puzzles of the risk-free rate and the equity risk premium may inspire more
credible accounts of risk aversion and investor behavior.
7.5 Proposed Solutions totheEquity Premium

Puzzle; Habit Formation andtheLife Cycle
Hypothesis
Although there are very ambitious efforts to solve the equity premium
puzzle by reconciling the intertemporal CAPM with models of general
equilibrium,103 I will present proposed answers that focus on the intertem-
poral CAPM.Mehra and Prescotts definition of the equity risk premium
makes it clear that almost any solution to the puzzle involves the fear of
losing future consumption.104 Even without consideration of general equi-

librium models, habit formation explanations of the equity premium
puzzle offer a rich, behaviorally nuanced account of the risk-averse view
of equity investing.
Building upon the observation that consumers cultivate tastes gradually
and, in turn, that past economic experiences can influence future consump-
tion,105 George Constantinides has suggested that rational management
of consumption opportunities over time can explain the equity premium
puzzle.106 Put more simply, once consumers acquire certain tastes and
habits, it is hard to change courseeven if income or wealth suddenly
fallsand, in a strictly rational sense, dictate abandonment of expensive
preferences.107 The underlying psychology resembles the endowment
effect, which describes how people demand much more to surrender an
object than they would be willing to pay to acquire it.108
Constantinides model can be summarized through a simple modifica-
tion of the utility function used to describe the household preference-
ordering function in Mehra and Prescotts stylized economy. Habit
appears in a model that assumes that utility is affected not only by current
consumption but also by past consumption.109 To capture[] the notion
that utility is a decreasing function of past consumption and marginal util-
ity is an increasing function of past consumption,110 we can incorporate
habit into the model as preferences with a one period lag:111
( ct + s - l ct + s -1 )
1 -a

U ( c ) = Et b s
,l >0
s =0 1-a
where is a parameter that captures the effect of past consumption.112

Setting = 1 fixes the subsistence level and transforms the formula for
constant relative risk aversion into a special case of this habit function:113
(c - x )
1-a
u (c) =
1-a
where x defines the fixed subsistence level.

The effect of defining preference orderings according to past consump-
tion makes the agent extremely averse to consumption risk even when
the risk aversion is small.114 The connection between habit formation
150 J.M. CHEN
and intertemporal asset pricing is transparent: Investors who (correctly)

anticipate that economic shocks will not likely prompt them to surrender
expensive tastes and habits will structure their portfolios accordingly.115 If
anything, habit formation leads people to exaggerate and overestimate the
extent to which their future tastes will resemble their current tastes. This
projection bias leads people to excess consumption in earlier stages of
life and to systematic shifts from savings to consumption relative to their
original plans.116
The life cycle theory of consumption and saving117 maintains that the
level of savings depends on the age of consumers, and hence on the demo-
graphic structure of society rather than on the level of family income.118
In a life cycle economy, precautionary saving is a function of the vari-
ability of income rather than its absolute level: [A]s we move up the
income distribution, a higher and higher fraction of people are there on a
temporary basis, with high transitory income, and thus a temporarily high
saving ratio.119 The fleeting nature of high income explains why savings
rates rise more rapidly with income among farmers or small-business
proprietors, whose income tends to be relatively volatile.120
At the macroeconomic level, the saving ratio of an economy as a
whole should be constant over the long run (provided that the rate of
growth does not change), but will vary pro-cyclically over the business
cycle.121 One observation of the life cycle hypothesis proves to be of cen-
tral importance to the habit formation explanation of the equity premium
puzzle: Over the business cycle, as over the life-cycle, consumption is
smoother than income.122
Nevertheless, habit formation may fall slightly short of fully solving
the equity premium puzzle. The fact that the vast majority of Americans
hold no stocks outside their pension wealth or tax-sheltered retirement
accounts renders quite tenuous the link that Constantinides draws
between stock returns and consumption.123 Confirming the suspi-
cion that a pure habit formation solution to the equity premium puzzle
may not quite capture the right behavioral intuitions is the compara-
tive strength of the housing-related wealth effect over its stock-related
counterpart:124 Increases in housing wealth promote more consumption
(perhaps 0.4% more consumption for every 10% increase in housing
wealth), while similar increases in financial wealth have much lower or
negligible effects.125 Indeed, the condition of the housing market in the
USA drives a host of economic decisions, including homeownership,

the share of the overall household portfolio allocated to the home, and
the use of leverage in home purchases.126 Expected increases in home
prices, the propensity to trade up to a more expensive house, and the
reduction of other risky assets in the household portfolio are all cor-
related with each other.127 The presence of different wealth effects for
different sources of wealth testifies, once again, to the behavioral power
of mental accounting.128
7.6 Catching UpwiththeJoneses

A simple but powerful extension of Constantinides habit formation model
consists of defining the utility of consumption not in absolute terms, but
relative to average per capital consumption:129
(c )
1-a
/ Ctg-1
U ( c ) = Et b
t
s
,a > 0
s =0 1-a
where Ct1 is the lagged aggregate consumption.130 Consumer prefer-

ences are now defined as the ratio of current, individual consumption
to lagged, aggregate consumption. For this reason, this model is called
catching up with the Joneses, not merely keeping up with them.131
Mehra and Prescott credit John Campbell132 with devising a model of try-
ing to catch up with lagged, aggregate consumption that is equivalent
to Abels original catching up with the Joneses model, except that log-
normal distribution of the growth rate of consumption does not ensure
lognormal distribution of returns on assets.133
Because humans measure their happiness in relative rather than abso-
lute terms,134 and because conspicuous consumption signals social sta-
tus,135 sometimes so compellingly that public displays of discretionary
income spur others to be unhappy with their own standards of living136
and thereby inflict macroeconomic harm in the form of distorted sav-
ings rates,137 the intuition behind catching up with the Joneses carries
extraordinary power:
152 J.M. CHEN
[O]nes utility depends not on the absolute level of consumption but on

how one is doing relative to others. The effect is that, once again, an indi-
vidual can become extremely sensitive and averse to consumption variation.
Equity may have a negative rate of return, which can result in personal con-
sumption falling relative to others consumption. Equity thus becomes an
undesirable asset relative to bonds. Since average per capita consumption is
rising over time, the induced demand for bonds helps mitigate the risk-
free puzzle.138
The further insight that investors care not only about overall volatil-
ity, but also about the temporal distribution of that volatility139 directly
connects the habit formation explanation of the equity risk premium to
John Campbells observation that growth stocks provide a hedge not
only against declines in future real stock returns, but also against the
variance of [those] return[s].140 Combining habit formation with inter-
temporal models of consumption and risk aversion thus eases some of
the tension in the equity premium puzzle.141 Defining the market port-
folio according to societal or peer group per capita consumption har-
monizes the catching-up-with-the-Joneses models with conventional
mean-variance optimization and the CAPM.142 Even more simply, habit
formation provides a straightforward reason for a phenomenon that more
elaborate theories of risk aversion struggle to explain: Why do rich people
save as much as they do?143 Because its good to be rich, and even better
to stay that way.144
7.7 Macroeconomic Disaster andPersonal Peril

Catching up with the nice, fashionable things that the Joneses bought
last seasonremember, Andrew Abels model measures lagged consump-
tionmay be a luxury if the economy has tanked and taken your job
down with it. This is the instinct underlying prominent modifications of
Constantinides basic habit formation model and Abels catching up with
the Joneses model. Correlation, as it turns out, affects more than trad-
able capital assets. Job loss and other individually scaled economic disas-
ters are likelier to happen during recession.145
Habit formation models of consumption can account for economic
downturns by [i]ncorporat[ing] the possibility of recession as a state
variable:146
Incorporat[ing] the possibility of recession as a state variable. In this

model, the risk aversion of investors rises dramatically when the chances of a
recession increases . Since risk aversion increases precisely when consump-
tion is low, it generates a precautionary demand for bonds that helps lower
the risk-free rate. This model is consistent with both consumption and asset
market data.147
Under plausible assumptions about risk aversion, income that a potentially

unemployed investor might use to sustain consumption during recession
may be worth roughly twice as much as the same amount of income at the
margin under normal economic conditions.148 More succinctly: The larger
the risk premium, the larger the marginal value of recession state income.149
An even more extreme version of this narrative projects truly night-
marish conditions, under which war, revolution, or some other calam-
ity utterly devalues equity investments.150 This may prove too much.
Although the outbreak of global warfare will undoubtedly wreak havoc
on equity investmentsthe opening month of World War II (September
1939), for instance, could be classified as a 9 event151events that cata-
clysmic are likely to devastate holdings of sovereign debt. This is especially
true if a bondholder has cast her or his lot, for example, with Czarist
Russia in 1918 or Nationalist China in 1949.152
Greater plausibility lies in scaling the catastrophic scenario downward,
toward smaller but likelier events. Indeed, given the innate cognitive ten-
dency to overestimate low-probability events and underestimate high-
probability events,153 we should affirmatively ponder likelier events that
escape notice precisely because they are too commonplace. Investors and
ordinary humans should heed that sound diagnostic aphorism regarding
horses, zebras, and unicorns: When you hear hoofbeats, think of horses,
not zebras.154 Though zebras lurk in exotic corners of the planet, you are
far likelier to encounter horses. And unicorns, sad as it may be to admit,
are purely fanciful fabrications.155 At least in mature economies and stable
polities, personal unemployment is far likelier than confiscation of prop-
erty or repudiation of sovereign debt by a rogue government.
The greatest risk that holding stocks poses to individual investors may
arise from the correlation of their portfolios with labor-related sources of
income. In principle, as Robert Mertons original formulation of the inter-
temporal CAPM recognized, we can always redefine wealth as capitalized
future wage income.156 There are few if any practical ways for investors
[to] issue shares against future income, however, and as a result, any
154 J.M. CHEN
consideration of wage income will cause systematic effects on portfo-

lio and consumption decisions.157 Consequently, intertemporal and life
cycle models of investment and consumption typically assume that future
earnings cannot be capitalized.158 (In fact, it is possible to securitize future
royalty income, but most people simply arent David Bowie.)159
It is one thing to treat labor income, conceptually, as a substitute for
bond income as a source of diversification for an equity portfolio160
until an economic crisis strikes and tightens the correlation between
labor income and portfolio returns.161 Loss of income from individual
labor, whether through job loss or a slowdown affecting an investors
own business, is highly correlated with an economic downturn.162 [T]he
returns to human capital and physical capital are very highly correlated in
developed economies, even though the growth rates of labor and capital
income are not highly correlated.163 As if the failure of investors to look
past their own countries stock markets were not debilitating enough,164
the strong correlation between returns on human capital and on domestic
stocks requires a substantial short position in domestic marketable assets
to hedge human capital risk.165
Worse still, other tools for smoothing consumption or replacing income
are unavailable. Even if [d]iscussion of how investment returns correlate
with incomes or with home prices were not foreign to most peoples
thinking, there are few tools for offsetting the risks that the market
poses to labor income and home equity, which account for the great
bulk of most peoples wealth.166 Whatever the future holds for con-
tinuous workout mortgages, a home lending instrument with a built-in,
ongoing mechanism for reducing the principal owed if home values fall,167
there are few if any workable solutions for hedging against the loss of
labor income.168 Adverse selection and moral hazard defeat private insur-
ance against job loss.169 Younger workers, whose earnings lie mostly in the
future, are subject to a borrowing constraint that undermines otherwise
optimal investments in equity.170 This borrowing constraint exacerbates
risk aversion among consumers who already confront a labor market with
incomplete consumption insurance.171
This dynamic drives the life cycle economic model. Broad economic
cycles arise from the interaction of retirement saving by the young and the
shedding of assets by older individuals who have shifted to pure consump-
tion after leaving the work force: The total wealth in the economy depends
on the length of retirement, and in simple cases, the ratio of a countrys
wealth is a half of the average length of retirement . More generally,
the ratio of wealth to income is lower the faster the rate of growth of the
economy, and is at its largest when the rate of growth is zero.172 And even
individuals who have succeeded in saving for retirement remain vulnerable
to bad timing in the guise of sequence-of-returns risk.173
The answer to the equity premium puzzle thus lies in idiosyncratic
income shocks from job loss or other major personal disasters that
consumers cannot hedge[] against or insure away:174
Equities and related pro-cyclical investments exhibit the undesirable feature

that they drop in value when the probability of job loss increases, as, for
instance, in recessions. In economic downturns, consumers thus need an
extra incentive to hood equities ; the equity premium can be rationalized
as [that] added inducement.175
This model unites risk aversion with finance, labor economics, and
macroeconomics.176 More broadly, the life cycle hypothesis has informed
the grandest issues in economics, such as retirement policy, the impact
of stock market returns on saving and spending, and thrift [as] the well-
spring of growth or simply its consequence.177 This branch of economics
explains why the equity risk premium behaves in counter-cyclical fash-
ion: the risk premium is highest in a recession since equities are a poor
hedge against the potential loss of employment.178
Moreover, the countercyclical nature of the equity risk premium is con-
sistent with one account of the value premium within three- and four-factor
models of stock market returns.179 In bad times, value firms are burdened
with more unproductive capital, finding it more difficult to reduce their
capital stocks than growth firms do.180 Worse, these previously incurred
commitments of capital are harder to reverse, since the general economic
principle of costly reversibility implies that firms face higher costs in cut-
ting than in expanding capital.181 Coupled with growth firms relative
ease in expanding capital investments during good times,182 value firms
disproportionately greater challenge of reallocating unproductive capital
during recessions accounts for value firms greater cyclicality riskand the
higher returns commanded by equity stakes in these firms.183
Habit formation, the life cycle hypothesis, and kindred branches of eco-
nomics thus provide the most persuasive answer available to the equity pre-
mium puzzle. In line with Rolls second critique of the CAPM, this literature
shows that a tradable investment portfolio represents only part of larger
portfolio of resources, including labor as a noncapitalizable, uninsurable
156 J.M. CHEN
variable annuity.184 The correlation between returns on the investment

portfolio and returns on personal labor becomes the driving factor.
7.8 Familiarity Breeds Irrationality

The progression from habit formation to catching up with the Joneses
and ultimately to the recession-and-unemployment account of the equity
risk premium culminates in an impeccable nugget of advice on personal
finance: Never bet your retirement on a single asset.185 Given how often
individual investors uphold this advice only in the breach, one of the
most treacherous foes of diversification (and of personal finance and
retirement security more generally) is the employee stock ownership
plan.186 Familiarity breeds irrationality: In a highly localized variant of the
home bias that keeps investors in domestic stock,187 employees invest in
company stock because they know many more stories about their own
companies than the rest of the investing universe.188 The urge to catch
up with the Joneses exaggerates the already irresistible behavioral pull
of the home bias: Investing close to home provides greater assurance of
calibrating future consumption according to that of the peer group most
familiar to the investor.189
Although home bias assuredly arises from investor competence, confi-
dence, and comfort,190 it also serves as a variation on the more common
theme of failing to diversify. Under that assumption, home bias may arise
from unfounded optimism that a concentrated portfolio of stocks reflect-
ing the investors personal knowledge promises true upside potential.191
Among individual investors, undiversified portfolios tend to be more
positively skewed (and therefore promise lower returns) than diversified
portfolios.192 This preference for upside potential sheds light on prospect
theory and the deeper phenomenology of behavioral economics.193
Whether home bias actually gives investors a chance to exploit geo-
graphic and/or professional familiarity or is yet another source of
performance-destroying cognitive bias is a subject of intense debate.194 Not
even Scandinavian studies can reach agreement on this issue: A Swedish
study has credited investors with profits from individualized geographic or
professional knowledge,195 while a Norwegian study has found that inves-
tors who buy stock on firms within their own industries realize negative
returns.196 It is almost certain, however, that many investors fail to take
advantage of the full benefits of diversification, choosing to overinvest
in company stock, local stocks, familiar stocks, and domestic companies,
in a bid to feel safe, but at a cost of increased volatility in their invest-

ment returns.197 In an era when rising correlation across global markets
has swamped any savings from the reduction in search costs for foreign
investments, the home bias may be here to stay.198
All of which, confusingly enough, suggests that we may have come as
close to a solution to the equity premium puzzle as finance permits, up
to the inefficient frontier of investor bias and behavior: [E]ven though
per capita consumption growth is poorly correlated with stock returns,
investors require a hefty premium to hold stocks over short-term bonds
because stocks perform poorly in recessions, when an investor is more
likely to be laid off.199
7.9 Gaudeamus Igitur:The familiar but curious

economics of university endowments
I hasten to note a critical departure from this narrative. Certain institu-

tional investors, at least in theory, do not face the biological and emotional
constraints on human behavior. It is arguably hard to see why the habit-
formation model should apply to pension funds [and] endowments.200
But universities, among other institutions, do depend on endowment
payouts as part of their income stream.201 The financial crisis of 200809
was the worst since endowments became an important part of U.S. uni-
versity funding, and it caught the institutions and investors by surprise.202
Guided by the Endowment Model of portfolio management that David
Swensen perfected at Yale,203 many universities had sought refuge in illiq-
uid alternatives and, for many years before the crisis, achieved enormous
double-digit annual returns and very rapid[] endowment growth.204
The party ended when those universities endowments exotic invest-
ments, especially commodities such as oil, agriculture or lumber, crashed
inlock-step with global stock markets.205
The resulting decline in university endowments created an opportu-
nity to examine how these not-for-profit institutions treated endowment
income under financial distress.206 To be sure, individuals, unlike universities
and other institutions, are Mortal Men [and Women] doomed to die.207
In the words of Gaudeamus Igitur, a traditional academic anthem, Post
molestam senectutem / Nos habebit humus.208 David Swensen admitted as
much in steering individual investors away from the Endowment Model
he devised for Yale and toward a very traditional approach emphasizing
158 J.M. CHEN
diversification, portfolio rebalancing, and vigilance against excessive fees.209

By contrast, the essentially infinite time horizon enjoyed by universi-
ties, charitable foundations, pension funds, and other institutional inves-
tors gives rise to a modest version of the equity premium puzzle: Given
the historical equity premium, why would any institutional investor adopt
a conservative asset allocation (such as the conventional 60% to 40% split
between stocks and bonds) rather than invest[ing] a higher proportion
in stocks?210
Despite their immunity from biological mortality, institutions may man-
age their investments in a way comparable to the way in which households
and individual human beings balance income from labor and from finan-
cial investments. Institutions do have memory and social sensibility, after
all; institutional performance is sensitive to aspiration levels influenced
by an institutions own historical expectations and by societal expecta-
tions.211 Moreover, there is one respect in which universities and operat-
ing foundations operate at a relative disadvantage vis--vis their mortal
counterparts.212 Whereas an individual saving for retirement arguably
should care solely about terminal wealth (roughly defined as the size of
the annuity that can be purchased at retirement),213 [t]ransitory fluctua-
tions in portfolio value impose upon universities not only the psychic
cost [of] seeing the value of the endowment fall, but also the very real
cost of cutting back programs if there is a cash flow reduction for a period
of years.214
The rough equivalence between individual portfolios and university
endowments, at any rate, was once conventional wisdom at the highest
levels of the economic academy. Well before either the bursting of the
technology bubble in 2000, let alone the financial crisis of 200809, lead-
ing economists had devised theories of university endowments remark-
ably similar to household approaches to financial wealth. Fischer Black,
Robert Merton, and James Tobin all viewed endowment payouts as part
of universities overall income stream. They differed only subtly in treating
those payouts as simply part of consumption income (Black),215 as a way
of smoothing cash flow (Tobin),216 or as a hedge against disturbances to
other sources of revenue (Merton).217
Although Black and Tobin wrote well before Mehra and Prescott pre-
sented the equity premium puzzle, their theoretical accounts, as well as
Mertons, are consistent with the habit formation resolutions of that prob-
lem. Moreover, all three accounts may be harmonized with more general
theories of precautionary saving as a tool for smoothing consumption into
the financially uncertain future.218 More recent examinations of univer-

sity endowments focus on spending rules keyed to these endowments
value.219 Most universities (and many other charitable endowments)
observe spending limits defined as x% of an n-year moving average of the
value of the endowment, where n is typically five or less.220 Although
the purpose of such moving averages is to smooth out stock market
fluctuations, spending limitations triggered by a sudden drop or a long
bear market can have a pronounced effect on a university.221 The decline
in endowment value forces the university to choose between the com-
peting goals of maximizing the present value of spending over an infinite
horizon, and maintaining a steady operating budget.222
An even more recent and comprehensive study of endowment pay-
outs by universities has negated any suggestion that universities treat their
endowments as human households might.223 Relative to their published
payout policies, endowments hold payout amounts in place after posi-
tive shocks, but actively reduce their payouts after negative shocks.224 This
asymmetric pattern supports the endowment hoarding hypothesis, which
posits that university and endowment leadership care about endowment
size above and beyond the endowments contribution to university opera-
tions.225 Put crassly, size does matter. The typical university cares about
the sheer size of its endowments as a symbol of status and prestige,
wholly apart from its ability to fund operations.226 The maintenance of a
formidable university endowment is often viewed as an object in its own
right, rather than as simply a means to an end.227
What inspires such hoarding behavior by universities? Inasmuch as
endowment size does affect the salaries paid to university presidents,228
the value of the endowment at the beginning of a presidents tenure might
provide a benchmark. And consistent with the endowment hoard-
ing hypothesis, reduction in endowment payouts in response to negative
shocks is entirely driven by those universities whose current endowment
value is close (within 10%) to the value of the endowment when the presi-
dent started.229 This presidential penalty, so to speak, can be precisely
quantified: [F]or universities close to the current presidents benchmark,
a negative endowment shock equal to 10 percent of the universitys bud-
get leads to a 37 to 39 basis point reduction in the endowments pay-
out rate.230 Universities may be putatively not-for-profit organizations,
but their presidents are by no means immune to overconfidence, empire-
building instincts, and the other behavioral pitfalls that hound their cor-
porate equivalents, chief executive officers.231 One campus, one president,
160 J.M. CHEN
one master of the academic universe. The president does bestride the cam-
pus like a colossus,232 or even like Sauron, J.R.R.Tolkeins ominous Lord
of the Rings:
One Ring to bring them all and in the darkness bind them
In the Land of [Knowledge] where the Shadows lie.233
All of this would be cause for bemusement, but for the very real human
impact of reductions in university endowment payouts. A negative endow-
ment shock equivalent to a 10 % reduction in the universitys budget
translates to a reduction of 5% in tenure-system faculty and in support
employees such as secretaries.234 Tellingly, the original National Bureau of
Economic Research working paper revealing these academic foibles bore
the title, Why I Lost My Secretary: The Effect of Endowment Shocks
on University Operations.235 In contrast to large effects on other
employees, however, university administrators are largely unaffected
when the university endowment, like a blindsided quarterback on the var-
sity football team, is thrown for a loss.236 If anything, their ranks actually
increase by 1.4 percent following [a] negative shock.237
Habit formation, so it seems, does affect universities. Just not the same
way it affects real human beings. Vivat academia, indeed. Alma mater
floreat / Quae nos educavit.238
Notes
1. See Henk Grootveld & Winfried Hallerbach, Variance vs. Downside Risk:
Is There Really That Much Difference?, 114 Eur. J. Oper. Research
304319, 315 (1999).
2. See Guido Baltussen, Thierry Post & Pim Van Vliet, Downside Risk
Aversion, Fixed Income Exposure, and the Value Premium Puzzle,
36J.Banking & Fin. 33823398 (2012).
3. Id. at 3383.
4. Rajnish Mehra & Edward C.Prescott, The Equity Premium: A Puzzle,
15J.Monetary Econ. 145161, 146 (1985). On pure exchange econo-
mies, see generally Robert E. Lucas, Jr., Asset Prices in an Exchange
Economy, 46 Econometrica 14291445 (1978).
5. Mehra & Prescott, The Equity Premium, supra note 4, at 146.
6. Mehra, supra note 26 (Chapter 6), at 54; Mehra & Prescott, The Equity
Premium in Retrospect, supra note 37 (Chapter 3), at 889.
9. See Narayana R. Kocherlakota, The Equity Premium: Its Still a Puzzle,
34J.Econ. Lit. 4271 (1996).
10. See Jeremy J.Siegel, Stocks for the Long Run: The Definitive Guide to
Financial Market Returns and Long-Term Investment Strategies 6 (5th
ed. 2013).
11. See Roger G. Ibbotson & Peng Chen, Long-Run Stock Returns:
Participating in the Real Economy, 59:1 Fin. Analysts J. 8898 (Jan./
Feb. 2003).
12. See John R.Graham & Campbell R.Harvey, The Equity Risk Premium in
2014 (April 7, 2014) (available at http://ssrn.com/abstract=2422008).
13. Elroy Dimson, Paul Marsh & Mike Staunton, The Worldwide Equity
Premium: A Smaller Puzzle, in Handbook of the Equity Risk Premium
467529, 468 (Rajnish Mehra ed., 2008).
14. Philippe Jorion & William N. Goetzmann, Global Stock Markets in the
Twentieth Century, 54 J. Fin. 953980, 954955, 978 (1999); accord
Shiller, Irrational Exuberance, supra note 89 (Chapter 1), at 217.
15. See ubo Pstor & Robert F. Stambaugh, The Equity Premium and
Structural Breaks, 56J.Fin. 12071239 (2001).
16. See Niles Eldredge & Stephen Jay Gould, Punctuated Equilibria: An
Alternative to Phyletic Gradualism, in Models in Paleobiology 82115
(Thomas J.M. Schopf ed., 1972), reprinted in Niles Eldredge, Time
Frames 193223 (1985); Stephen Jay Gould & Niles Eldredge,
Punctuated Equilibria: The Tempo and Mode of Evolution Reconsidered, 3
Paleobiology 115151, 145 (1975). See generally Stephen Jay Gould,
The Structure of Evolutionary Theory (2002).
17. See, e.g., Claudio Cioffi-Revilla, The Political Uncertainty of Interstate
Rivalries: A Punctuated Equilibrium Model, in The Dynamics of
Enduring Rivalries 6497 (Paul Diehl ed., 1998); Daniel A.Levinthal,
The Slow Pace of Rapid Technological Change: Gradualism and
Punctuation in Technological Change, 7 Indus. & Corp. Change 217247
(1998); Andrs Tilcsik & Christopher Marquis, Punctuated Generosity:
How Mega-Events and Natural Disasters Affect Corporate Philanthropy in
U.S.Communities, 58 Admin. Sci. Q. 111148 (2013).
18. See Quentin D. Atkinson, Andrew Meade, Chris Venditti, Simon
J. Greenhill & Mark Pagel, Languages Evolve in Punctuational Bursts,
319 Science 588 (2008); Dan Dediu & Stephen C.Levinson, Abstract
Profiles of Structural Stability Point to Universal T
endencies, Family-
Specific Factors, and Ancient Connections Between Languages, 7 PLoS
162 J.M. CHEN
One E45198 (2012); Julien dHuy, Les Mythes volueraient par

Ponctuations. 252 Mythologie Franaise 812 (2013).
19. See Robert D. Arnott & Peter L. Bernstein, What Risk Premium Is
Normal?, 58:2 Fin. Analysts J. 6485 (March/April 2002).
20. See Robert Arnott, Bonds: Why Bother?, 12:3 J. Indexes 1017 (May/
June 2009) , at 10 (available at http://www.indexuniverse.com/publica-
tions/journalofindexes/joi-articles/5710-bonds-why-bother.html and
h t t p : / / w w w. e t f . c o m / d o c s / m a g a z i n e / 2 / 2 0 0 9 _
149.pdf); cf. Lubo Pstor & Robert F.Stambaugh, Are Stocks Really Less
Volatile in the Long Run?, 67J.Fin. 431 (2012) (observing that while
reversion to the mean strongly reduces long-horizon variance, this stabi-
lizing effect is neutralized and overcome by the investors uncertainties,
especially uncertainty about the expected return).
21. See Mehra, supra note 26 (Chapter 6), at 5455 (tables 1 and 2).
22. See id. at 54.
23. See id.
24. The general formula for the geometric average growth rate is
mg = t vt v0 -1, where t indicates the number of years and v indicates the
x
real value of the compounding stock or bond portfolio after x years. In
this example, t = 74, and v74 for the stock and bond portfolios in 2000,
respectively, is $266.47 and $1.71.
25. See Jeremy J.Siegel & Richard H.Thaler, The Equity Premium Puzzle,
11J.Econ. Persp. 191200, 19192 (1997).
26. See Benartzi & Thaler, Myopic Loss Aversion, supra note 20 (Chapter 2), at
76-77 (reporting unpublished work by Thomas MaCurdy & John
Shoven, Accumulating Pension Wealth with Stocks and Bonds (Stanford
Working Paper, Jan. 1992)).
27. Mehra, supra note 26 (Chapter 6), at 60 (emphasis in original); Mehra &
Prescott, The Equity Premium in Retrospect, supra note 37 (Chapter 3), at
909 (emphasis in original).
28. Michael Haliassos & Carol C.Bertaut, Why Do So Few Hold Stocks?, 105
Econ. J. 11101129, 1110 (1995)
29. Compare Dora L.Costa, The Evolution of Retirement, in The Evolution
of Retirement: An American Economic History, 1880-1990, at 631
(Dora L. Costa ed., 1998) with James Poterba, Joshua Rauh, Steven
Venti & David Wise, Defined Contribution Plans, Defined Benefit Plans,
and the Accumulation of Retirement Wealth, 91J.Pub. Econ. 20622086
(2007).
30. Nicholas Barberis, Ming Huang & Richard H. Thaler, Individual
Preferences, Monetary Gambles, and Stock Market Participation: A Case for
Narrow Framing, 96 Am. Econ. Rev. 10691090, 1070 (2006) (empha-
sis added).
31. Campbell, Household Finance, supra note 68 (Chapter 6), at 1564.

32. See sources cited supra notes 1518 (Chapter 5) (describing the separa-
tion theorem).
33. See Richard A. Cohn, Wilbur G. Lewellen & Gary G. Schlarbaum,
Individual Investor Risk Aversion and Investment Portfolio Composition,
30J.Fin. 605620, 614615 (1975).
34. See Gordon Y.N.Tang & Wai Cheong Shum, The Relationships Between
Unsystematic Risk, Skewness and Stock Returns During Up and Down
Markets, 62 Intl Bus. Rev. 523541 (2003).
35. See Changes in U.S.Family Finances from 2010 to 2013: Evidence from the
Survey of Consumer Finances, 100:4 Fed. Reserve Bull. 141, 18 (Sept.
2014) (available at http://www.federalreserve.gov/pubs/bulle-
tin/2014/pdf/scf14.pdf). See generally id. at 1621 (especially boxes 6
and 7).
36. See id. at 20 (box 7).
37. See id. at 16.
38. See Niko Canner, N.Gregory Mankiw & David Weil, An Asset Allocation
Puzzle, 87 Am. Econ. Rev. 181191 (1997); Oussama Chakroun,
Georges Dionne & Amlie Dugas-Sampara, Empirical Evaluation of the
Asset-Allocation Puzzle, 100 Econ. Letters 304307 (2008).
39. See Kenneth L.Fisher & Meir Statman, Investment Advice from Mutual
Fund Companies, 24:1J.Portfolio Mgmt. 925 (Fall 1997).
40. See Tobin, supra note 14 (Chapter 5).
41. See I.R.C. 163(h); William G. Gale, Jonathan Gruber & Seth
I. Stephens-Davidowitz, Encouraging Homeownership Through the Tax
Code, 115 Tax Notes 11711189, 1171 (2007) (observing that the
mortgage interest deduction does little if anything to encourage home
ownership, but does operate to raise the price of housing and land and
to encourage people who do buy homes to borrow more and to buy
larger homes than they otherwise would); Edward McCaffery, Cognitive
Theory and Tax, 41 UCLA L. Rev. 18611947, 19181920 (1994)
(describing the failure of taxpayers to understand that the financial ben-
efits of the mortgage interest deduction are capitalized into housing
prices); Lilian V. Faulhaber, The Hidden Limits of the Charitable
Deduction: An Introduction to Hypersalience, 92 B.U.L. Rev. 13071348,
13361339 (2012) (describing the salience of the mortgage interest
deduction as a contributor to the elevated and arguably excessive con-
sumption of residential housing by Americans).
42. See Cocco, supra note 70 (Chapter 6).
43. See Becker & Shabani, supra note 71 (Chapter 6), at 2918, 2920.
44. Id. at 2927.
45. See Campbell, Household Finance, supra note 70 (Chapter 6), at 157785.
164 J.M. CHEN
46. See generally James M. Poterba, Retirement Security in an Aging

Population, 104 Am. Econ. Rev. 130 (2014).
47. Executive Order No. 13,737, Using Behavioral Science Insights to Better
Serve the American People, 80 Fed. Reg. 56,36556,367, 56,365 (Sept.
18, 2015). See generally Social and Behavioral Sciences Team, Annual
Report (Sept. 2015) (Executive Office of the President, National Science
and Technology Council) (available at https://www.whitehouse.gov/
sites/default/files/microsites/ostp/sbst_2015_annual_report_
final_9_14_15.pdf).
48. See Alicia H.Munnell, Wenliang Hou & Anthony Webb, NRRI Update
Shows Half Still Falling Short (Center for Retirement Research Issues in
Brief, No. 1420, Dec. 2014) (available at http://crr.bc.edu/wp-con-
tent/uploads/2014/12/IB_14-20-508.pdf).
49. See, e.g., Nari Rhee, The Retirement Savings Crisis: Is It Worse Than We
Think? 1 (June 2013) (reporting a median retirement account balance of
$3,000 for all US households and $12,000 for households near retire-
ment, a collective savings gap ranging from $6.8 to $14 trillion, and
retirement risk for 6592 % of the American population) (available at
http://www.nirsonline.org/storage/nirs/documents/Retirement%20
Savings%20Crisis/retirementsavingscrisis_final.pdf). A 2015 update of
this source reports a $50,000 median retirement account balance for
households with retirement assets and a $2500 median for all households
with heads aged 2564. In the cohort where heads of household are aged
5564, median retirement account balances rise to $104,000 for house-
holds with retirement accounts and $14,500 for all households. The dis-
parity between account balances is more readily understood in light of the
estimate that 45% of US households do not have a retirement account. See
Nari Rhee & Ilana Boivie, The Continuing Retirement Savings Crisis 10
(March 2015) (available at http://laborcenter.berkeley.edu/pdf/2015/
RetirementSavingsCrisis.pdf).
50. See John J.Topoleski, U.S.Household Savings for Retirement in 2010, at
16 (July 23, 2013) (Congressional Research Serv. Report R43057)
(available at http://fas.org/sgp/crs/misc/R43057.pdf).
51. See id.
52. See generally Jason J.Fichtner, John W.R.Phillips & Barbara A.Smith,
Retirement Behavior and the Global Financial Crisis, in Reshaping
Retirement Security: Lessons from the Global Financial Crisis 81100
(Raimond Maurer, Olivia S.Mitchell & Mark Warshawsky eds., 2012).
53. See generally Roberta Leviton, Reverse Mortgage Decision-Making,
13 J. Aging & Soc. Poly 116 (2002); David W. Rasmussen, Isaac
F.Megbolugbe & Barbara A.Morgan, The Reverse Mortgage as an Asset
Management Tool, 8J.Aging & Soc. Poly 173194 (1997).
54. See Shefrin & Thaler, Behavioral Life-Cycle Hypothesis, supra note 19
(Chapter 5); Richard H.Thaler & Hersh M.Shefrin, An Economic Theory
of Self-Control, 89J.Pol. Econ. 392406 (1981).
55. See Richard H.Thaler & Shlomo Benartzi, Save More Tomorrow: Using
Behavioral Economics to Increase Employee Saving, 112 J. Pol. Econ.
S164S187 (2004); Steven A. Sass & Jorge D. Ramos-Mercado, Are
Americans of All Ages and Income Levels Shortsighted About Their
Finances? (Center for Retirement Research Issues in Brief, No. 159. May
2015) (available at http://crr.bc.edu/wp-content/uploads/2015/05/
IB_15-9.pdf).
56. See John Beshears, James J.Choi, David Laibson & Brigitte C.Madrian,
The Importance of Default Options for Retirement, in Lessons from
Pension Reform in the Americas 5987 (Stephen J.Kay & Tapen Sinha
eds., 2007) (finding that the default option dictates investment choices
for 80 % of participants in 401(k) plans); Olivia S. Mitchell, Gary
R. Mottola, Stephen P. Utkus & Takeshi Yamaguchi, The Inattentive
Participant: Portfolio Trading Behavior in 401(k) Plans (June 2006)
(University of Michigan Retirement Research Center, Working Paper
2006115) (available at http://www.mrrc.isr.umich.edu/publications/
papers/pdf/wp115.pdf) (finding that 80% of 401(k) participants con-
ducted no trades within two years and 11% conducted only one trade).
57. See, e.g., Douglas B.Bernheim, Patrick J.Bayer & John Karl Scholz, The
Effects of Financial Education in the Workplace: Evidence from a Survey of
Employers, 47 Econ. Inquiry 605624 (2009); Maarten van Rooij,
Annamaria Lusardi & Rob Alessie, Financial Literacy and Stock Market
Participation, 101 J. Fin. Econ. 449472 (2011); Maarten van Rooij,
Annamaria Lusardi & Rob Alessie, Financial Literacy, Retirement
Planning and Household Wealth, 122 Econ. J. 449478 (2012). See gen-
erally John Y. Campbell, Restoring Rational Choice: The Challenge of
Consumer Financial Regulation (Jan. 2016) (available at http://scholar.
harvard.edu/files/campbell/files/elylecturejan182016.pdf) (arguing
that the complexity of financial products compounds pervasive financial
ignorance and disables many households from effective management of
their financial affairs).
58. Benartzi & Thaler, Myopic Loss Aversion, supra note 20 (Chapter 2), at 77
(quoting Thomas MaCurdy and John Shoven).
59. Sundn & Surette, supra note 149 (Chapter 1), at 210.
60. N.Gregory Mankiw & Stephen P.Zeldes, The Consumption of Stockholders
and Nonstockholders, 29J.Fin. Econ. 97112, 110 (1991).
61. Haliassos & Bertaut, supra note 28, at 1110 (reporting that the propor-
tion of nonstockholding households in the USA is remarkably stable
through time).
166 J.M. CHEN
62. Simon Grant & John Quiggin, The Risk Premium for Equity: Implications
for Resource Allocation, Welfare and Policy, 45 Austral. Econ. Papers
253265, 259 (2006).
63. Id.
64. See id. at 262.
65. Id.; accord David Miles, Testing for Short Termism in the UK Stock Market,
103 Econ. J. 13791396 (1993).
66. Few if any insurance companies are willing to underwrite policies promis-
ing steady income, lest such policies attract shirkers and malingerers. See,
e.g., N.Gregory Mankiw, The Equity Premium and the Concentration of
Aggregate Shocks, 17J.Fin. Econ. 211219 (1986).
67. See Grant & Quiggin, supra note 62, at 26465; Simon Grant & John
Quiggin, Public Investment and the Risk Premium for Equity, 70
Economica 118 (2003).
68. See Mehra, supra note 26 (Chapter 6), at 5758 (equations (1) through
(16)); id. at 6768 (Appendix A); Mehra & Prescott, The Equity Premium
in Retrospect, supra note 37 (Chapter 3), at 900909.
69. See supra 6.4, at 117.
70. See John B.Donaldson & Rajnish Mehra, Comparative Dynamics of an
Equilibrium Intertemporal Asset Pricing Model, 51 Rev. Econ. Stud.
491508 (1984); Edward C. Prescott & Rajnish Mehra, Recursive
Competitive Equilibrium: The Case of Homogeneous Households, 48
Econometrica 13651379 (1980)
73. See Andrew B.Abel, Stock Prices Under Time Varying Dividend Risk: An
Exact Solution in an Infinite-Horizon General Equilibrium Model,
22J.Monetary Econ. 375394 (1988).
74. See Mehra & Prescott, The Equity Premium Puzzle in Retrospect, supra
note 37 (Chapter 3), at 928933; Mehra & Prescott, The Equity
Premium: A Puzzle, supra note 4, at 150156.
75. See Mehra, supra note 26 (Chapter 6), at 58; Mehra & Prescott, The
Equity Premium in Retrospect, supra note 37 (Chapter 3), at 903. For
more background on the lognormal distribution, see generally J.Aitchison
& J.A.C.Brown, The Lognormal Distribution, with Special Reference to
Its Use in Econometrics (1957); Hal Forsey, The Mathematicians View:
Modelling Uncertainty with the Three Parameter Lognormal, in Managing
Downside Risk in Financial Markets 5158 (Frank A.Sortino & Stephen
E.Satchell eds., 2001); K.Krishnamoorthy & Thomas Mathew, Inferences
on the Means of Lognormal Distributions Using Generalized p-Values and
Generalized Confidence Intervals, 115 J. Stat. Planning & Inference

103121 (2003); Eckhard Limpert, Werner A.Stahel & Markus Abbt,
Log-Normal Distributions Across the Sciences: Keys and Clues, 51
BioScience 341 (2001); Ulf Olsson, Confidence Intervals for the Mean of
a Log-Normal Distribution, 13:1J.Stat. Educ. (2005); Xiao-Hua Zhou,
Sujuan Gao & Siu L. Hui, Methods for Comparing the Means of Two
Independent Log-Normal Samples, 53 Biometrics 11291135 (1997).
Equity Premium in Retrospect, supra note 37 (Chapter 3), at 904.
Disputes over precise measures within Mehra and Prescotts analysis have
been trivial and unlikely to alter the equity premium puzzle. See, e.g.,
Kocherlakota, supra note 9, at 68 (setting the relatively riskless rate at 1%
instead of 0.8 %, and finding the variance of that rate to be 0.00308
instead of 0.0032).
78. See Mankiw & Zeldes, supra note 60, at 105 (describing the value of
$51,858 as implausible and concluding as a consequence that the
level of risk aversion necessary to generate the observed equity premium
is too large to be believable).
80. See Larry G.Epstein & Stanley E.Zin, Substitution, Risk Aversion, and
the Temporal Behavior of Consumption and Asset Returns: A Theoretical
Framework, 57 Econometrica 937-969 (1989); Larry G. Epstein &
Stanley E.Zin, Substitution, Risk Aversion, and the Temporal Behavior of
Consumption and Asset Returns: An Empirical Analysis, 99J.Pol. Econ.
263286, 265272 (1991).
81. See Campbell, Asset Pricing at the Millennium, supra note 7 (Chapter 1),
at 1538.
82. See id.
83. See id. at 1544.
84. See supra text accompanying note 19 (Chapter 6).
85. Campbell, Asset Pricing at the Millennium, supra note 7 (Chapter 1), at
1544.
86. Mehra & Prescott, The Equity Premium in Retrospect, supra note 37
87. Mehra, supra note 26 (Chapter 6), at 61.
168 J.M. CHEN
1544. Compare Robert E. Hall, Intertemporal Substitution in
Consumption, 96J.Pol. Econ. 221273 (1988) and John Y.Campbell &
N. Gregory Mankiw, Consumption, Income, and Interest Rates:
Reinterpreting the Time-Series Evidence, in National Bureau of Economic
Research Macroeconomics Annual 185215, 19899 (Olivier Jean
Blanchard & Stanley Fischer eds., 1989) (reporting low elasticity) with
Orazio P.Attansio & Guglielmo Weber, Consumption Growth, the Interest
Rate, and Aggregation, 60 Rev. Econ. Stud. 631649 (1993) and Paul
Beaudry & Eric van Wincoop, The Intertemporal Elasticity of Substitution:
An Exploration Using a U.S.Panel of State Data, 63 Economica 495512
(1996) (reporting greater elasticity).
1544.
92. Id.
93. Id.
94. See, e.g., Steffen Andersen, Glenn W.Harrison, Morten Lau & Elisabet
E. Rutstrm, Discounting Behavior: A Reconsideration, 71 Eur. Econ.
Rev. 1533 (2014); Shane Frederick, George Loewenstein & Ted
ODonoghue, Time Discounting and Time Preference: A Critical Review,
40 J. Econ. Lit. 351401 (2002); Ariel Rubinstein, Economics and
Psychology? The Case of Hyperbolic Discounting, 44 Intl Econ. Rev.
12071216 (2003); cf. Daniel Read, Is Time-Discounting Hyperbolic or
Subadditive?, 23J.Risk & Uncertainty 532 (2001).
1544.
97. See Philippe Weil, The Equity Premium Puzzle and the Risk-Free Rate
Puzzle, 24J.Monetary Econ. 401421 (1989).
98. See Baker, Bradley & Wurgler, supra note 8 (Chapter 4), at 43.
99. Id.
100. Kocherlakota, supra note 9, at 44.
101. Id.
102. Id.
103. See, e.g., David K. Backus & Allan W. Gregory, Theoretical Relations
Between Risk Premiums and Conditional Variances, 11J.Bus. & Econ.
Stat. 177185 (1993); John Y. Campbell, Intertemporal Asset Pricing
Without Consumption Data, 83 Am. Econ. Rev. 487512 (1993);
Lawrence R.Gloston, Ravi Jagannathan & David Runkle, On the Relation
Between the Expected Value and the Volatility of the Nominal Excess Return
on Stocks, 48J.Fin. 17791901 (1993); John T.Scruggs, Resolving the
Puzzling Intertemporal Relation Between the Market Risk Premium and

Conditional Market Variance: A Two-Factor Approach, 53 J. Fin.
575603 (1998); cf. Francisco Azeredo, The Equity Premium: A Deeper
Puzzle, 10 Annals Fin. 347373 (2014) (arguing that traditional mea-
sures of consumption before 1930 understate the extent of serial correla-
tion in the annual growth rate of consumption of services and nondurable
goods, implying that some general equilibrium models may generate
negative equity premiums for reasonable levels of risk aversion). With
respect to general equilibrium models, see sources cited supra note 10
(Chapter 5)
104. See Ang, Hodrick, Xing & Zhang, The Cross-Section of Volatility and
Expected Returns, supra note 14 (Chapter 4), at 262263 (observing that
future volatility impairs future investment returns and therefore future
consumption opportunities); Chacko & Viceira, supra note 26 (Chapter 5),
at 1370 (same); cf. Das, Markowitz, Scheid & Statman, supra note 42
(Chapter 2), at 311312 (noting that any mean-varianceefficient portfo-
lio enables the separation of production and consumption functions).
105. See, e.g., James S. Duesenberry, Income, Saving, and the Theory of
Consumer Behavior (1949); Gary S. Becker & Kevin M. Murphy, A
Theory of Rational Addiction, 96 J. Pol. Econ. 675700 (1988); cf.
Markowitz, The Utility of Wealth, supra note 64 (Chapter 6) (building an
inflection point into an otherwise concave utility function, at a point indi-
cating customary wealth, where an investor would seek to preserve
status quo wealth and consumption).
106. George M.Constantinides, Habit Formation: A Resolution of the Equity
Premium Puzzle, 98J.Pol. Econ. 519543, 521 (1990); see also Wayne
Ferson & George M.Constantinides, Habit Persistence and Durability in
Aggregate Consumption: Empirical Tests, 39 J. Fin. Econ. 199240
(1991).
107. See generally Philip H.Dybvig, Duesenberrys Ratcheting of Consumption:
Optimal Dynamic Consumption and Investment Policies Given Intolerance
for Any Decline in Standard of Living, 62 Rev. Econ. Stud. 287313
(1995); Roger Mason, The Social Significance of Consumption: James
Duesenberrys Contribution to Consumer Theory, 34 J. Econ. Issues
553572 (2000).
108. See Daniel Kahneman, Jack L. Knetsch & Richard H. Thaler, The
Endowment Effect, Loss Aversion, and Status Quo Bias, 5J.Econ. Persp.
193206, 193 (1991); Niels van de Ven, Marcel Zeelenberg & Eric van
Dijk, Buying and Selling Exchange Goods: Outcome Information, Curiosity
and the Endowment Effect, 26J.Econ. Psych. 459468 (2005); sources
cited supra note 4 (Chapter 2).
170 J.M. CHEN
110. Id.
111. Mehra & Prescott, The Equity Premium in Retrospect, supra note ,37
(Chapter 3), at 913 (equation 30); see also Mehra, supra note 26 (Chapter
6), at 61 (equation 21).
(Chapter 3), at 913 (equation 31).
115. See Suresh M.Sundaresan, Intertemporally Dependent Preferences and the
Volatility of Consumption and Wealth, 2 Rev. Fin. Stud. 7389 (1989).
116. See George Loewenstein, Ted ODonoghue & Matthew Rabin, Projection
Bias in Predicting Future Utility, 118 Q.J.Econ. 12091248 (2003).
117. See Albert Ando & Franco Modigliani, The Life Cycle Hypothesis of
Saving: Aggregate Implications and Tests, 53 Am. Econ. Rev. 5584
(1963); Franco Modigliani & Richard H.Brumberg, Utility Analysis and
Aggregate Consumption Functions: An Attempt at Integration, in 2 The
Collected Papers of Franco Modigliani: The Life Cycle Hypothesis of
Saving 128197 (Andrew B. Abel ed., 1980); Franco Modigliani &
Richard H. Brumberg, Utility Analysis and the Consumption Function:
An Interpretation of Cross-Sectional Data, in Post Keynesian Economics
388-436 (Kenneth K. Kurihara ed., 1954). See generally Browning &
Crossley, supra note 19 (Chapter 5).
118. Mauro Baranzini, Modiglianis Life-Cycle Theory of Savings Fifty Years
Later, 58 Banco Nazionale del Lavoro Q.Rev. 109172, 109 (2005).
119. Angus Deaton, Francisco Modigliani and the Life-Cycle Theory of
Consumption, 58 Banco Nazionale del Lavoro Q. Rev. 91107, 95
(2005).
120. Id. at 9596.
121. Id. at 96.
122. Id.
123. Benartzi & Thaler, Myopic Loss Aversion, supra note 20 (Chapter 2), at
78; see supra 7.2, at 140143.
78.
125. See Karl E.Case, John M.Quigley & Robert J.Shiller, Comparing Wealth
Effects: The Stock Market Versus the Housing Market, 5:1 Advances in
Macroecon., art. 1, at 26 (2005) (available at http://www.econ.yale.
edu/~shiller/pubs/p1181.pdf). See generally, e.g., Kul B. Bhatia, Real
Estate Assets and Consumer Spending, 102 Q.J.Econ. 437443 (1987);
Richard K. Green, Stock Prices and House Prices in California: New
Evidence of a Wealth Effect?, 32 Regional Sci. & Urban Econ. 775783

(2002).
126. See Marcel Fischer & Michael Z. Stamos, Optimal Life Cycle Portfolio
Choice with Housing Market Cycles, 26 Rev. Fin. Stud. 23112352
(2013).
127. See Stefano Corradin, Jos L. Fillat & Carles Vergara-Alert, Optimal
Portfolio Choice with Predictability in Housing Prices and Transaction
Costs, 27 Rev. Fin. Stud. 823880 (2014).
128. See Laurence Levin, Are Assets Fungible? Testing the Behavior of Life-Cycle
Savings, 36J.Econ. Org. & Behav. 5983 (1998); Richard H.Thaler,
Saving, Fungibility and Mental Accounts, 4 J. Econ. Persp. 193206
(1990).
(Chapter 3), at 914 (equation 32).
131. Andrew B.Abel, Asset Prices Under Habit Formation and Catching Up
with the Joneses, 80 Am. Econ. Rev. 3842, 38 n.1 (1990); accord Mehra
& Prescott, The Equity Premium in Retrospect, supra note 37 (Chapter
3), at 914 n.36.
15451548.
133. The Equity Premium in Retrospect, supra note 26 (Chapter 6), at
915916. See generally sources cited supra note 75 (describing the prop-
erties of the lognormal distribution).
134. See, e.g., Gary Charness, David Masclet & Marie Claire Villeval, The Dark
Side of Competition for Status, 60 Mgmt. Sci. 3855 (2014); Simon Dato
& Petra Nieken, Gender Differences in Competition and Sabotage,
100 J. Econ. Behav. & Org. 6480 (2014); Thomas Dohmen, Armin
Falk, Klaus Flessbach, Uwe Sunde & Bernd Weber, Relative Versus
Absolute Income, Joy of Winning, and Gender: Brain Imaging Evidence,
95 J. Pub. Econ. 279285 (2011); Camellia M. Kuhnen & Agnieszka
Tymula, Feedback, Self-Esteem, and Performance in Organizations, 58
Mgmt. Sci. 94113 (2012); Mark Sheskin, Paul Bloom & Karen Wynn,
Anti-Equality: Social Comparison in Young Children, 130 Cognition
152-156 (2014); Lior Jacob Strahilevitz, Absolute Preferences and
Relative Preferences in Property Law, 160 U. Pa. L. Rev. 21572187
(2012).
135. See, e.g., Thorstein Veblen, The Theory of the Leisure Class: An Economic
Study of Institutions 68101 (1902) (formulating the notion of con-
spicuous consumption); Ori Heffetz, A Test of Conspicuous Consumption:
Visibility and Income Elasticities, 93 Rev. Econ. & Stat. 11011117
172 J.M. CHEN
(2011); Erzo Luttmer, Neighbors as Negatives: Relative Earnings and

Well-Being, 120 Q.J.Econ. 9631002 (2005); Roger Mason, Conspicuous
Consumption: A Literature Review, 18 Eur. J.Marketing 2639 (1984);
Andrew B. Trigg, Veblen, Bourdieu, and Conspicuous Consumption,
35J.Econ. Issues 99115 (2001); Paurav Shukla, Status Consumption
in Cross-National Context: Socio-Psychological, Brand and Situational
Antecedents, 27 Intl Marketing Rev. 108129 (2010).
136. See, e.g., Mark Grinblatt, Matti Keloharju & Seppo Ikaheimo, Social
Influence and Consumption: Evidence from the Automobile Purchases of
Neighbors, 90 Rev. Econ. & Stat. 735753 (2008); Dirk Krueger &
Fabrizio Perri, Does Income Inequality Lead to Consumption Inequality?
Evidence and Theory, 73 Rev. Econ. Stud. 163193 (2006).
137. See Duesenberry, supra note 105, at 27 (calling this phenomenon the
demonstration effect); cf. Michael Bordo & Christopher Meissner,
Does Inequality Lead to a Financial Crisis?, 31 J. Intl Money & Fin.
21472161 (2012).
138. Mehra, supra note 26 (Chapter 6), at 61; accord Mehra & Prescott, The
139. Constantinides, supra note 106, at 521.
140. Campbell, Giglio, Polk & Turley, supra note 70 (Chapter 5), at 30
(emphases added).
141. See Donald J. Meyer & Jack Meyer, Risk Preferences in Multi-Period
Consumption Models, the Equity Premium Puzzle, and Habit Formation
Utility, 8J.Monetary Econ. 14971515 (2005).
142. See Moshe Levy & Haim Levy, Keeping Up with the Joneses and Optimal
Diversification, 58J.Banking & Fin. 2938 (2015).
143. See Kimball, Precautionary Saving, supra note 127 (Chapter 2); cf. Louis
Eeckhoudt & Harris Schlesinger, On the Utility Premium of Friedman
and Savage, 105 Econ. Letters 4648 (2009) (suggesting that a plausibly
risk-averse reaction to increased wealth is to engage in precautionary
saving).
144. See Christopher D.Carroll, Portfolios of the Rich, in Household Portfolios
389430 (Luigi Guiso, Michael Haliassos & Tullio Jappelli eds., 2002);
Christopher D.Carroll, Why Do the Rich Save So Much?, in Does Atlas
Shrug? The Economic Consequences of Taxing the Rich 465-484 (Joel
B.Slemrod ed., 2000).
145. See John Y.Campbell & John Cochrane, By Force of Habit: A Consumption
Based Explanation of Aggregate Stock Market Behavior, 107J.Pol. Econ.
205251 (1999).
147. Id.
148. See Grant & Quiggin, supra note 62, at 260.
149. Id. at 262.

150. See Thomas A.Reitz, The Equity Risk Premium: A Solution, 22J.Monetary
Econ. 117131 (1988); Robert J.Barro, Rare Disasters and Asset Markets
in the Twentieth Century, 121 Q.J.Econ. 823-866 (2006).
151. See Juan Salazar & Annick Lambert, Fama and MacBeth Revisited: A
Critique, 1 Aestimatio 4871, 64 (2010).
152. See Rajnish Mehra & Edward C.Prescott, The Equity Rate Premium: A
Solution?, 22J.Monetary Econ. 133136 (1988).
153. See, e.g., Colin Farrell Camerer & Teck H.Ho, Violations of the Betweenness
Axiom and Nonlinearity in Probability, 8J.Risk & Uncertainty 167196
(1994).
154. John G. Sotos, Zebra Cards: An Aid to Obscure Diagnoses 1 (2d ed.
2006) (crediting this aphorism to Dr. Theodore Woodward, who so
instructed medical interns at the University of Maryland School of
Medicine during the 1940s); see also Pascal James Imperato, Medical
Detective 13, 18 (1979).
155. See Gil Grantmore, The Phages of American Law, 36 U.C.Davis L.Rev.
455504, 494495 (2003). The purely mythological unicorn should not
be confused with billion-dollar unicorns, which are venture-funded firms
that have reached the $1 billion threshold in private or public valuation.
See sources cited supra note 14 (Chapter 2). Those unicorns have mytho-
logical properties, to be sure, but some do warrant a ten-digit valuation.
156. Merton, Intertemporal CAPM, supra note 1 (Chapter 5), at 875 n.23.
157. Id.
158. See, e.g., John Heaton & Deborah J. Lucas, Evaluating the Effects of
Incomplete Markets on Risk Sharing and Asset Pricing, 104J.Pol. Econ.
443487 (1996); Luis M. Viceira, Optimal Portfolio Choice for Long-
Horizon Investors with Nontradable Labor Income, 56 J. Fin. 433470
(2001).
159. See Jennifer Burke Sylva, Bowie Bonds Sold for Far More Than a Song:
The Securitization of Intellectual Property as a Super-Charged Vehicle
for High Technology Financing, 15 Santa Clara High Tech. L.J.
195230 (1999); Nicole Chu, Note, Bowie Bonds: A Key to Unlocking
the Wealth of Intellectual Property, 21 Hastings Communications &
Entertainment L.J. 469499 (199899); Dan McCrum, A Short
History of the Bowie Bond, Fin. Times Alphaville, Jan. 11, 2016 (avail-
able at http://ftalphaville.ft.com/2016/01/11/2149761/a-short
-history-of-the-bowie-bond).
160. See sources cited supra note 72 (Chapter 6).
161. See generally supra 4.3, at 7880 (discussing correlation tightening).
162. See generally Kjetil Storesletten, Chris I.Telmer & Amir Yaron, Cyclical
Dynamics in Idiosyncratic Labor Market Risk, 112J.Pol. Econ. 695717
(2004).
174 J.M. CHEN
163. Marianne Baxter & Urban J.Jermann, The International Diversification

Puzzle Is Worse Than You Think, 87 Am. Econ. Rev. 170180, 170
(1997) (emphasis in original).
164. See Kenneth R.French & James M.Poterba, International Diversification
and International Equity Markets, 81 Am. Econ. Rev. 222226, 222223
(1991) (reporting that American investors hold approximately 94% of
their financial assets in American securities, even though American securi-
ties comprise only 48% of investable securities worldwide, and that the
domestic share of portfolios in Germany, Japan, and the UK each exceeds
85%); Anders Karlsson & Lars L.Norden, Home Sweet Home: Home Bias
and International Diversification Among Individual Investors,
31J.Banking & Fin. 317333 (2007); Norman Strong & Xinzhong Xu,
Understanding the Equity Home Bias: Evidence from Survey Data. 85
Rev. Econ. & Stat. 307312 (2003); Linda L.Tesar & Ingrid Werner,
Home Bias and High Turnover. 14 J. Intl Money & Fin. 467492
(1995); Bruno Solnik & Luo Zuo, A Global Equilibrium Asset Pricing
Model with Home Preference, 58 Mgmt. Sci. 273292 (2011).
165. Baxter & Jermann, supra note 163, at 170.
166. Shiller, Irrational Exuberance, supra note 89 (Chapter 1), at 234.
167. See Robert J.Shiller, The Subprime Solution: How Todays Global Crisis
Happened and What to Do About It 157170 (2008); Robert J.Shiller,
Fafal M. Wojakowski, M. Shahid Ebrahim, Mark B. Shackleton,
Mitigating Financial Fragility with Continuous Workout Mortgages,
85J.Econ. Behav. & Org. 269285 (2013).
168. See John Heaton & Deborah Lucas, Evaluating the Effects of Incomplete
Markets on Risk Sharing and Asset Pricing, 104J.Pol. Econ. 668712
(1996); Luis M. Viceira, Optimal Portfolio Choice for Long-Horizon
Investors with Nontradable Labor Income, 56J.Fin. 433470 (2001).
169. See George M. Constantinides & Darrell Duffie, Asset Pricing with
Heterogeneous Consumers, 104J.Pol. Econ. 219240 (1996); Mankiw,
supra note 66.
170. See George M. Constantinides, John B. Donaldson & Rajnish Mehra,
Junior Cant Borrow: A New Perspective on the Equity Premium Puzzle,
107 Q.J.Econ. 269296 (2002).
171. See Alon Brav, George M.Constantinides & Christopher C.Geczy, Asset
Pricing with Heterogeneous Consumers and Limited Participation:
Empirical Evidence, 110J.Pol. Econ. 793824 (2002).
172. Deaton, supra note 119, at 92.
173. See sources cited supra notes 3032 (Chapter 5).
176. Cf. Valery Polkovnichenko, Life-Cycle Portfolio Choice with Additive Habit
Formation Preferences and Uninsurable Labor Income Risk, 20 Rev. Fin.
Stud. 83124 (2007) (predicting that households with low-to-moderate
wealth may increase the share of stocks in their financial holdings as wealth
increases and suggesting that younger people may consequently have
more conservative portfolios than conventional retirement planning
would otherwise predict or counsel).
177. Deaton, supra note 119, at 93.
Premium in Retrospect, supra note 37 (Chapter 3), at 917; cf. Bucciol &
Miniaci, supra note 68 (Chapter 6) (finding that ex ante risk in household
portfolios rises and falls with the business cycle).
179. See infra 12.2, at 305307.
180. Lu Zhang, The Value Premium, 60J.Fin. 67103, 68 (2005).
181. Id. On costly reversibility, see generally Andrew B. Abel & Janice
C. Eberly, A Unified Model of Investment Under Uncertainty, 84 Am.
Econ. Rev. 13691384 (1994); Andrew B. Abel & Janice C. Eberly,
Optimal Investment with Costly Reversibility, 63 Rev. Econ. Stud. 581-
593 (1996); Valerie A. Ramsey & Matthew D. Shapiro, Displaced
Capital: A Study of Aerospace Plant Closings, 109J.Pol. Econ. 958992
(2001).
182. See Zhang, supra note 180, at 68.
183. See id. at 9092.
184. See Roll, supra note 20 (Chapter 3), at 155; Chen, Postmodern Portfolio
Theory, supra note 1 (Chapter 1), 3.3, at 2931.
185. On the behavioral economics of retirement savings, see generally Shlomo
Benartzi & Richard H. Thaler, Heuristics and Biases in Retirement
Savings Behavior, 21J.Econ. Persp. 81104 (2007); Shlomo Benartzi &
Richard H. Thaler, Behavioral Economics and the Retirement Savings
Crisis, 339 Science 11521153 (2013).
186. See, e.g., Shlomo Benartzi, Richard H.Thaler, Stephen P.Utkus & Cass
R.Sunstein, The Law and Economics Company Stock in 401(k) Plans, 50
J.L. & Econ. 4579 (2007); James M. Poterba, Employer Stock and
401(k) Plans, 93 Am. Econ. Rev. 398404 (2003); Andrew Stumpff &
Norman Stein, Repeal Tax Incentives for ESOPs, 125 Tax Notes 337340,
339340 (2009). For summaries of studies showing that employee stock
ownership plans do little in themselves to improve productivity, see Brett
McDonnell, ESOPs Failures: Fiduciary Duties When Managers of
Employee-Owned Companies Vote to Entrench Themselves, 2000 Colum.
176 J.M. CHEN
Bus. L.Rev. 199260, 235; Ezra S.Field, Note, Money for Nothing and
Leverage for Free, 97 Colum. Bus. L.Rev. 740785, 752 n.82 (1997).
187. See sources cited supra note 164; cf. Joshua D. Coval & Tobias
J.Moskowitz, Home Bias at Home: Local Equity Preference in Domestic
Portfolios, 54J.Fin. 20452073 (1999) (documenting a strong prefer-
ence for locally headquartered firms, particularly small firms with high
levels of leverage that produce nontraded goods). See generally Gur
Huberman, Familiarity Breeds Investment, 14 Rev. Fin. Stud. 659680
(2001) (documenting investors preference for geographically proximate
stocks within their domestic portfolios).
189. See Beni Lauterbach & Haim Reisman, Keeping Up with the Joneses and
the Home Bias, 10 Eur. Fin. Mgmt. 225234 (2004).
190. See John R. Graham, Campbell R. Harvey & Hai Huang, Investor
Competence, Trading Frequency, and Home Bias, 55 Mgmt. Sci.
10941106 (2009).
191. See Meir Statman, The Diversification Puzzle, 60:4 Fin. Analysts J. 4453
(July/Aug. 2004).
192. See Marie-Hlne Broihanne, Maxime Merli & Patrick Roger,
Diversification, Gambling and Market Forces, Rev. Quant. Fin. &
Accounting (Jan. 29, 2015) (available at http://link.springer.com/arti-
cle/10.1007/s11156-015-0497-1); Goetzmann & Kumar, supra note
130 (Chapter 1); Todd Mitton & Keith Vorkink, Equilibrium
Undiversification and the Preference for Skewness, 20 Rev. Fin. Stud.
1255-1288 (2007); Valery Polkovnichenko, Household Portfolio
Diversification: A Case for Rank-Dependent Preferences, 18 Rev. Fin.
Stud. 14671502 (2005).
193. See generally infra Chaps. 8 and 9.
194. Compare Zoran Ivkovi & Scott Weisbenner, Local Does as Local Is:
Information Content of the Geography of Individual Investors Common
Stock, 60J.Fin. 267306 (2005) (asserting that home bias does translate
into positive returns) with Mark S. Seasholes & Ning Zhu, Individual
Investors and Local Bias, 65J.Fin. 19872010 (2010) (concluding that
individual investors do not realize excess returns on local stocks).
195. See Massimo Massa & Andrei Simonov, Hedging, Familiarity, and
Portfolio Choice, 19 Rev. Fin. Stud. 633685 (2006).
196. See Trond M.Dskeland & Hans K.Hvide, Do Individual Investors Have
Asymmetric Information Based on Work Experience?, 66J.Fin. 10111041
(2011).
197. Brad M.Barber & Terrance Odean, The Behavior of Individual Investors,
in 2 Handbook of the Economics of Finance, supra note 68 (Chapter 6),
at 15331570, 1563; see also id. at 1563 n.7 (observing that any return
advantage from investing inlocal stockswould need to be sufficiently

large to cover transaction costs and the under-diversification that would
result from a locally concentrated portfolio).
198. See Haim Levy & Moshe Levy, The Home Bias Is Here to Stay, 47J.Banking
& Fin. 2940 (2014).
78.
201. See generally Stephen G. Dimmock, Background Risk and University
Endowment Funds. 94 Rev. Econ. & Stat. 789799 (2012).
202. Howard Wolinsky, The Crash Reaches the University: The Global Financial
Crisis Threatens Private and Public University Funding in the USA and
Europe, 10 EMBO Reports 209211, 209 (2009) (European Molecular
Biology Organization).
203. See generally David F.Swensen, Pioneering Portfolio Management: An
Unconventional Approach to Institutional Investment (2000).
204. Wolinsky, supra note 202, at 209; see Josh Lerner, Antoinette Schoar &
Jialan Wang, Secrets of the Academy: The Drivers of University Endowment
Success, 22J.Econ. Persp. 207222 (2008); Keith C.Brown, Lorenzo
Garlappi & Cristian Tiu, Asset Allocation and Portfolio Performance:
Evidence from University Endowment Funds, 13J.Fin. Markets 268-294
(2010); cf. Charles T.Clotfelter, The Familiar but Curious Economics of
Higher Education: Introduction to a Symposium, 13J.Econ. Persp. 312
(1999).
205. Wolinsky, supra note 202, at 209.
206. See generally Peter Conti-Brown, Scarcity Amidst Wealth: The Law,
Finance, and Culture of Elite University Endowments in Financial Crisis,
63 Stan. L.Rev. 699749 (2011).
207. J.R.R.Tolkein, The Lord of the Rings: 50th Anniversary, One-Volume
Edition 50 (2005).
208. https://en.wikipedia.org/wiki/Gaudeamus_igitur. Literally, these lines
mean: After a troubling old age / the earth will have us. Id.
209. See generally David F.Swensen, Unconventional Success: A Fundamental
Approach to Personal Investment (2005).
87.
211. See Henrich R.Greve, Performance, Aspirations, and Risky Organizational
Change, 43 Admin. Sci. Q. 5886 (1998).
89.
178 J.M. CHEN
213. Id. Individuals seeking to leave a legacy, presumably to their own off-
spring, might focus on the amount of wealth they can transmit to suc-
ceeding generations. There is at least indirect evidence that married
women (presumably with children or with an expectation of having chil-
dren) are more sensitive to this consideration than any other demo-
graphic group among individual investors. See Sundn & Surette, supra
note 149 (Chapter 1), at 209; supra 1.6.
89.
215. See Fischer Black, The Investment Policy Spectrum: Individuals,
Endowment Funds, and Pension Funds, 32:1 Fin. Analysts J. 2331 (Jan./
Feb. 1976).
216. See James Tobin, What Is Permanent Endowment Income?, 64 Am. Econ.
Rev. 427432 (1974).
217. See Robert C. Merton, Optimal Investment Strategies for University
Endowment Funds, in Continuous Time Finance 649674 (rev. ed.
1992); Robert C.Merton, Optimal Investment Strategies for University
Endowment Funds, in Studies of Supply and Demand in Higher Education
211242 (Charles T.Clotfelter & Michael Rothschild eds., 1993).
218. See Miles Kimball & Philippe Weil, Precautionary Saving and Consumption
Smoothing Across Time and Possibilities, 41J.Money, Credit & Banking
245284 (2009); cf. Kimball, supra note 143.
219. See, e.g., Gregory P. Ho, Haim A. Mozes & Pavel Greenfield, The
Sustainability of Endowment Spending Levels: A Wake-Up Call for
University Endowments, 37:1J.Portfolio Mgmt. 133146 (Fall 2010).
89.
221. Id.
222. Id.; see also id. (The fact that stocks have outperformed bonds over every
twenty-year period in history is cold comfort after a decade of zero nomi-
nal returns.).
223. Jeffrey R. Brown, Stephen G. Dimmock, Jun-Koo Kang & Scott
J.Weisbenner, How University Endowments Respond to Financial Market
Shocks: Evidence and Implications, 104 Am. Econ. Rev. 931962 (2014).
224. See id. at 932933.
225. Id. at 954; cf. Raymond Fishman & R.Glenn Hubbard, Precautionary
Savings and the Governance of Nonprofit Organizations, 89J.Pub. Econ.
22312243 (2005) (exploring the incentives motivating nonprofit orga-
nizations to hoard assets).
226. Brown, Dimmock, Kang & Weisbenner, supra note 223, at 933 (quoting
Conti-Brown, supra note 206, at 704).
227. Henry Hansmann, Why Do Universities Have Endowments?, 19J.Legal

Stud. 342, 38 (1990); accord Brown, Dimmock, Kang & Weisbenner,
228. See Ronald G.Ehrenberg, John J.Cheslock & Julia Epifantseva, Paying
Our Presidents: What Do Trustees Value, 25 Rev. Higher Educ. 1537
(2001).
229. Brown, Dimmock, Kang & Weisbenner, supra note 223, at 955; see also
id. at 933.
230. Id. at 955.
231. See Ulrike Malmendier & Geoffrey Tate, CEO Overconfidence and
Corporate Investment, 60J.Fin. 26612700 (2005); Ulrike Malmendier
& Geoffrey Tate, Who Makes Acquisitions? CEO Overconfidence and the
Markets Reaction, 89J.Fin. Econ. 2043 (2008); Ulrike Malmendier,
Geoffrey Tate & Jon Yan, Overconfidence and Early-Life Experiences: The
Effect of Managerial Traits on Corporate Financial Policies, 66 J. Fin.
16871733 (2011); cf. Ulrike Malmendier & Stefan Nagel, Depression
Babies: Do Macroeconomic Experiences Affect Risk-Taking? 126 Q.J.Econ.
373416 (2012). See generally Itzhak Ben-David, John R. Graham &
Campbell R. Harvey, Managerial Miscalibration, 128 Q.J. Econ.
15471584 (2013).
232. Cf. William Shakespeare, Julius Caesar, act I, sc. 2, ll. 136139, in The
Oxford Shakespeare: The Complete Works 599626, 603 (Stanley Wells
& Gary Taylor eds., 1986) (Why, man, he doth bestride the narrow
world / Like a Colossus, and we petty men/Walk under his huge legs,
and peep about/To find ourselves dishonourable graves.).
233. Tolkein, supra note 207, at 50.
234. See Brown, Dimmock, Kang & Weisbenner, supra note 223, at 958.
235. NBER Working Paper 15,861 (April 2010) (available at http://www.
nber.org/papers/0015861).
236. Brown, Dimmock, Kang & Weisbenner, supra note 223, at 958.
237. Id.
238. Once again, I quote the traditional academic anthem, Gaudeamus Igitur.
In Mark Sugars translation of the Latin lyrics, Long live our academy!
May our alma mater thrive, a font of education. https://en.wikipedia.
org/wiki/Gaudeamus_igitur
CHAPTER 8
Prospect Theory
8.1 Comprehensive Accounts ofBehavioral

Finance
This chapter explores prospect theorys depiction of the impact of fear and
greed on financial markets. Prospect theory posits that the human evalua-
tion of uncertain gains and losses departs from a purely rational account of
expected utility in three crucial ways. First, humans pay heed to a reference
point, whether it is the price paid for a security or a target rate of return.
Second, humans hate losing more than they like winning. Third, humans
grow less sensitive to the magnitude of changes in welfare as gains or losses
increase. Critically, diminished sensitivity applies to gains as well as losses.
The resulting fourfold pattern of human responses to uncertainty pro-
vides a far more complete and persuasive account of risk-averse as well as
risk-seeking behavior.
Prospect theory opens the door to more comprehensive accounts of
behavioral finance. The balance of this book will engage those accounts.
Chapter 9 applies prospect theory to a wide range of practical problems
in finance, from a systematic preference for skewness and lottery-like pay-
outs to behaviorally sophisticated explanations for Bowmans paradox and
for the equity premium puzzle. Chapter 10 presents an even more com-
prehensive account of decision-making under uncertainty. SP/A theory
transforms the dynamics of hope and fear into a fully behavioral theory of
portfolio design.

DOI10.1007/978-3-319-32711-2_8
182 J.M. CHEN
Over time, behavioral frames of mind and modes of investing exact

their cost. Many of the abnormalities in financial markets arise from the
interaction of rational, informed traders with behaviorally motivated noise
traders in heterogeneous markets. Cycles of fear and greed systematically
erode investor returns. As Chapter 11 demonstrates, gaps in investor per-
formance can be measured through a new statistic, . Investor behavior
not only propels market momentum; it eventually triggers momentum
crashes. Chapter 12 offers insight into the dynamic interplay of abnor-
mal markets and emotionally driven, putatively irrational investors. This
interaction not only spurs the rise and eventual bursting of speculative
bubbles, but also yields a richer, fuller account of financial decision-mak-
ing under uncertainty.
8.2 Responding toAnomalies

inExpected Utility Theory
[P]rospect theory is widely viewed as the best available description

of how people evaluate risk in experimental settings.1 Prospect theory
provided the principal basis for Daniel Kahnemans 2002 Nobel Prize in
Economics2 and has inspired a deeper body of work in behavioral eco-
nomics.3 As is true of any revolutionary scientific development, prospect
theory began as a response to the shortcomings of its intellectual prede-
cessor, expected value theory.4 The ArrowPratt measures of risk aversion
prescribe utility functions that are concave everywhere.5 Quadratic util-
ity,6 a special case of hyperbolic absolute risk aversion,7 once held sway
throughout much of the economic literature,8 even though it implies
in contradiction to observed behavior that a wealthier investor is less
likely to choose a risky asset over a riskless asset.9 It is not enough that qua-
dratic utility fails to satisfy the primary requirement [that] a reasonable
utility function [be] monotonic[]; satiation in wealth along with the
implausible property of increasing aversion should be enough to under-
mine reliance on quadratic utility.10
In their 1979 article that laid the initial foundations for prospect the-
ory, Daniel Kahneman and Aaron Tversky identified several phenomena
which violate[d] tenets of expected utility theory.11 Their examina-
tion of the relative attractiveness of insurance undermined expected
utility theorys supposition that the utility function for money is con-
cave everywhere.12 Contrary to expected utility theorys assumption of
PROSPECT THEORY 183
constant risk aversion, people often prefer insurance programs that offer
limited coverage with low or zero deductibles over comparable policies
that offer higher maximal coverage with higher deductibles.13 Formally
speaking, purchasers of insurance act as though reducing the probability
of a loss from p to p/2 is less valuable than reducing the probability of that
loss from p/2 to 0.14 This preference is particularly intriguing because
all insurance is, in a sense, probabilistic insofar as even the most avid
buyer of insurance remains vulnerable to many financial and other risks
which his policies do not cover.15
Put another way, there is no such product as comprehensive, all-risks,
zero-deductible insurance. Preferences within the realm of insurance reveal
an intuitive aversion to insurance policies whose coverage is probabilistic
rather than complete, and which therefore fail to entirely eliminate[] the
element of risk.16 It is quite evident that the assumed concavity of the
utility function fails to adequately capture[] this intuitive notion of
risk.17
Within the nomenclature of prospect theory, that intuitive notion of
risk is called the certainty effect.18 The certainty effect enables prospect
theory to account for the Allais paradox.19 In a situation where winning
is possible but not probable, most people choose the prospect that offers
the larger gain.20 The possibility of a certain outcome within a mixed
prospect renders certainty a special case of a broader isolation effect,
whereby the evaluation of distinct components of alternative choices can
produce inconsistent preferences.21 The obverse of the certainty effect
is ambiguity aversion, or discomfort arising from the lack of knowledge of
the probabilities underlying a gamble.22
Moreover, upon examining not only positive prospects, or pros-
pects that involve no losses, but also the impact of reversing the signs
of the outcomes so that gains are replaced by losses, Kahneman and
Tversky identified a reflection effect.23 The significance of the reflection
effect lies in its implication that risk aversion in the positive domain is
accompanied by risk seeking in the negative domain.24 Combining the
certainty effect with the reflection effect dramatically changes the nature
of risk aversion. In the positive domain, the certainty effect contributes
to a risk averse preference for a sure gain over a larger gain that is merely
probable.25
By contrast, [i]n the negative domain, the certainty effect leads to
a risk-seeking preference for a loss that is merely probable over a smaller
184 J.M. CHEN
loss that is certain.26 Because the reflection effect eliminates aversion for
uncertainty or variability for losses, the same psychological principlethe
overweighting of certaintyfavors risk aversion in the domain of gains
and risk seeking in the domain of losses.27 In other words, certainty
increases the aversiveness of losses as well as the desirability of gains.28
The resulting division between risk aversion in the domain of gains and
risk seeking in the domain of losses has been observed, of all places, in the
behavior of pigeons.29
8.3 The Value Function

Prospect theorys value function implements many of the observations
first made by Harry Markowitz in his 1952 article, The Utility of Wealth.30
Critically, both Markowitz and prospect theory calculate utility according
to gains and losses rather than final asset positions.31 Noting the
presence of risk seeking in preferences among positive as well as among
negative prospects, Markowitz had proposed a utility function which
has convex and concave regions in both the positive and the negative
domains.32 Prospect theory accordingly adopts a value function that
is concave for gains and convex for losses.33 Moreover, prospect theory
treats changes of wealth, rather than final asset positions that include cur-
rent wealth, as the true carriers of value or utility.34 This conclusion,
a repudiation of expected utility theory, represents the cornerstone of
prospect theory as an alternative theory of risky choice.35
Treating changes in wealth or welfare, rather than final states, as
carriers of value harmonizes prospect theory with basic principles of
perception and judgment.36 Human responses to attributes such as
brightness, loudness, or temperature account for the past and present
context of experience.37 As with physical stimuli, so with non-sensory
attributes such as health, prestige, and wealth.38 Physical and fiscal attri-
butes alike are evaluated in relative rather than absolute terms, according
to an adaptation level, or reference point.39 In economic terms, the
same level of wealth may imply abject poverty for one person and
great riches for anotherdepending on their current assets.40
The magnitude of change, whether physical or economic, also matters.
Many sensory and perceptual dimensions share the property that the
psychological response is a concave function of the magnitude of physi-
cal change.41 Again, by analogy to physical perception: [I]t is easier to
PROSPECT THEORY 185
discriminate between a change of 3 and a change of 6 in room tem-

perature, than it is to discriminate between a change of 13 and a change
of 16.42 In principle, the same change in sensitivity applies to the
evaluation of monetary changes.43
Kahneman and Tverskys own summary of prospect theorys value
function thus identifies three distinctive properties. The value function is
(i) defined on deviations from the reference point;

(ii) generally concave for gains and commonly convex for losses; and
(iii) steeper for losses than for gains.44
Stripped of its nuances, prospect theory predicts that people generally

make risk-averse decisions when choosing between options that appear to
be gains and risk-seeking decisions when choosing between options that
appear to be losses.45 These predictions arise from three basic features of
human beings core cognitive system.46
First, all decision-making takes place relative to a neutral reference
point or adaptation level.47 Outcomes exceeding this reference point are
gains. Outcomes below the reference point are losses.48 As one business
manager succinctly summarized this instinct: Risk is the prospect of not
meeting the target rate of return. If you are one hundred percent sure
of making the target return, then it is a zero risk proposition.49
Notably, the reference point carries financial implications at a general
level and at a specific level. At a high level of abstraction, the fact that deci-
sions pivot on a reference point implies the presence of financial inertia,
or a strong bias in favor of the status quo.50 With respect to specific assets,
decision-making relative to the purchase price raises the possibility that
investors derive utility from realized gains and losses precisely at the
moment of sale, in a magnitude corresponding to the size of the gain
or loss realized.51 The former effect is called status quo bias; the latter is
called realization utility.
Second, loss aversion means that losses, when directly weighted or com-
pared against gains, loom larger.52 This is one respect in which prospect
theory preserves the premises of expected utility theory. Prospect theory
assumes that most individuals, as an expression of innate risk aversion, fear
potential losses more than they covet potential gains. We can state this
principle in increasingly simple ways. Within a formal economic frame-
work, loss aversion eliminates the neutrality of the dynamic aggregation
186 J.M. CHEN
rules people follow.53 The Supreme Court of the USA long ago rendered
this sentiment in more straightforward terms: Threat of loss, not hope of
gain, is the essence of economic coercion.54 Even more plainly, [l]osing
hurts worse than winning feels good.55
Third and finally, diminishing sensitivity applies to upward and down-
ward perceptions and to evaluation of changes of wealth. In concert, these
three principlesneutral reference point, loss aversion, diminishing sen-
sitivitycan be illustrated through a graph showing an asymmetrical sig-
moid curve whose inflection point occurs at the neutral adaptation level,
whose steeper slope below the adaptation level demonstrates loss aversion,
and whose declining rate of change in both directions reflects diminishing
sensitivity to gains and to losses.56
Accounting for all three of these properties generates a value function
as shown in Fig. 8.1.
Value
109
100
90
40 5060
$ 40 50 6 $
(Losses) (Gains)
200
222
234
Fig. 8.1 Prospect theorys value function

PROSPECT THEORY 187
8.4 Flagging Prospect Theory:

Log-Logistic Distribution
If prospect theory had a flag, this image would be drawn on it.57 In
their 1992 paper outlining their cumulative modification of prospect
theory,58 Amos Tversky and Daniel Kahneman defined the value function
v(x) as a two-part power function of the form:59
x if
x0
(x) ifx<0
Tversky and Kahneman estimated these parameters as , 0.88;
2.25.60 Tversky and Kahneman reported that the median exponent
of the value function was 0.88 for both gains and losses, which indicated
diminishing sensitivity to extreme outcomes in both domains.61 A sub-
sequent survey of studies estimating the value functions exponent observ-
ing that [d]espite [the] wide range of estimates for and , empirical
applications of cumulative prospect theory have generated a very inter-
esting result: These two parameters are either equal to each other or
very close to each other, such that .62
By contrast, by fixing the value of the coefficient at 2.25, Tversky and
Kahneman found pronounced loss aversion.63 That coefficient describes
the slope of the loss function relative to that of the gain function.64
The ratio of these slopes is a measure of loss aversion.65 Consistent with
the original estimate of 2.25, empirical measures of have hovered around
2, which mean[s] that the disutility of giving something up is [roughly]
twice as great as the utility of acquiring it.66 Not only does losing hurt
worse than winning feels good; we can say with some confidence that los-
ing hurts roughly twice as much. The 2.25 figure is consistent with other
measures of loss aversion estimated in very different contexts.67 These
contexts include the celebrated experiment that identified the endowment
effect, in which students given a Cornell University insignia mug valued
the mug at 2.5 times the level of students who were invited to choose
between the mug and a cash prize.68
Although Tversky and Kahneman specified prospect theorys value
function in conditional terms, it is possible to express the value function
through a single, unconditional equation. Putting the Kronecker delta
function to a rather unusual use69 enables us to draw cumulative prospect
theorys value function using Tversky and Kahnemans original parameters:
188 J.M. CHEN
v ( x ) -l sgn ( x ) x , a b
dx, - x a
Substituting the values 0.88, 2.25 into the foregoing formula gen-
erates the image in Fig. 8.2.70
On the other hand, if Tversky and Kahnemans specification of the
value function is merely intended as a general guide to depicting prospect
theorys interpretation of utility, risk aversion, and risk seeking, then there
are numerous other ways to represent the value function. If only for con-
venience, let us restate how prospect theorys asymmetrical utility function
illustrates three basic features of human beings core cognitive system:71
1.All decision-making takes place relative to a neutral reference point,

or adaptation level. Outcomes exceeding this reference point are
gains. Outcomes below the reference point are losses.
2.Loss aversion means that losses, when directly weighted or compared
against gains, loom larger.
3.Diminishing sensitivity applies to upward and downward percep-
tions and to evaluation of changes of wealth.
In concert, these three principlesneutral reference point, loss aver-

sion, diminishing sensitivitycan be illustrated through a graph s howing
Fig. 8.2 Prospect theorys value function according to Tversky and Kahnemans
original parameters
PROSPECT THEORY 189
an asymmetrical sigmoid curve whose inflection point occurs at the neutral

adaptation level, whose steeper slope below the adaptation level demon-
strates loss aversion, and whose declining rate of change in both direc-
tions reflects diminishing sensitivity to gains and to losses.72 These are the
properties of a well-behaved cumulative distribution function (CDF) of an
asymmetrical probability distribution.73
Like any other model of human behavior, prospect theory ultimately
dissolves into empirical, measurable questions of fear, greed, and subjec-
tive value.74 Like any other foundation for economic analysis, prospect
theory invites modeling, at least in the initial stages, on a parametric basis.
I will describe a surprisingly simple and supple method for parametrically
modeling prospect theory with closed-form expressions and elementary
functions. It accomplishes this task by transforming the CDF of the log-
logistic distribution.75 In economics, where the log-logistic distribution
provides a simple model for the distribution of wealth or income, this
distribution is often called the Fisk distribution.76
In plainer language, I will demonstrate how to draw the flag of pros-
pect theory as parsimoniously as possible. I will draw prospect theorys
flag with the simplest available mathematical functions and a minimal
amount of algebraic manipulation. The resulting formula can be expressed
with exactly two parameters. That formula can be readily modified to fit
empirical data garnered in support of nearly any hypothesis informed by
prospect theory.
A glance at the flag of prospect theory suggests that some sort of
sigmoid curve, bounded by upper and lower asymptotes, can serve as
the basis for a simple parametric model. Those are the signature prop-
erties of the CDF of a probability distribution. But not every CDF will
do. Prospect theory posits a steeper slope that reflects the natural human
aversion toward losses beyond the neutral reference point. To reflect that
instinct, a persuasive model of prospect theory should place its inflection
point at the origin. As losses deepen or gains accumulate, humans exhibit
diminishing sensitivity. That property comes quite naturally to any CDF,
which by definition can never be negative nor exceed 1.
The log-logistic distribution is the probability distribution of a nonneg-
ative random variable whose logarithm follows the logistic distribution.
The log-logistic distribution can be defined in terms of two parameters,
scale parameter and shape parameter . Both parameters must be positive:
190 J.M. CHEN
a > 0; b > 0
Scale parameter is also the median. When > 1, a log-logistic distri

bution is unimodal. Its dispersion decreases as increases.
The CDF of the log-logistic distribution can be rendered in closed
form:
1
F ( x; a , b ) = -b
x
1+
a
b
x
Multiplying both the numerator and the denominator by yields:
a
b
x
a
F ( x; a , b ) = b
x
1+
a
Multiplying both the numerator and the denominator again, this time by
, yields an even more elegant expression of the log-logistic distributions
CDF:
xb
F ( x; a , b ) =
a b + xb
This final formula allows us to parameterize prospect theorys subjec-

tive value function with ease. We can start by designating extremely simple
values for scale parameter and shape parameter . Let =1 and =5.
Substitution of those values for the two parameters generates the plot in
Fig. 8.3.77
Although the displacement of the curve into quadrant I (x>0, y>0)
reduces the visual impact of this curve, it does exhibit an inflection point
at x=1.0 and y=0.5. In other words, we have an appropriately shaped
curve over the range 0<x<2, but with a reference point inconveniently
displaced +1.0 units along the x-axis and +0.5 units along the y-axis.
PROSPECT THEORY 191
Fig. 8.3 Flagging prospect theory: the cumulative distribution function of the
log-logistic distribution
To recenter this graph at the origin, we need to perform three simple

mathematical tasks. First, recall that scale parameter is the median of the
distribution. We should therefore add to x in the expression x , which
appears in both the numerator and the denominator. This has the effect of
shifting the entire graph to the left by (which after all is the median of
this distribution), and align the reference point with x=0:
(x +a )
b
F ( x; a , b)=
a + (x +a )
b b
Substituting our stipulated values of 1 for scale parameter and 5 for

shape parameter generates the formulaic and graphic output depicted in
Fig. 8.4.78
Second, we should subtract 0.5 from the formula in its entirety so that
we shift the entire graph down by 0.5, which is the value of f(x; , ) at
x=0, no matter what the value of scale parameter and shape parameter.
(x +a )
b
1
F ( x; a , b)= -
a + (x +a )
b
b 2

192 J.M. CHEN
Fig. 8.4 Flagging prospect theory: shifting left
Fig. 8.5 Flagging prospect theory: shifting down
The resulting plot is depicted in Fig. 8.5.79

The foregoing graph aligns the inflection point with the origin and
exhibits steeper slope to the immediate left of the origin than to the imme-
diate right. The only problem, such as it is, is that the upper and lower
asymptotes, respectively, are at f(x; , )=0.5 and f(x; , )=0.5. Our
third and final transformation therefore consists of the simple expedient
of multiplying the entire formula by 2. The application of this stretching
coefficient has the effect of transforming the flag of prospect theory so
that its vertical scale matches its horizontal scale:
PROSPECT THEORY 193
2 (x +a )
b
F ( x; a , b ) = -1
a b + (x +a )
b

Figure 8.6 shows how our flag of prospect theory looks like after all three
transformations.80
We have therefore accomplished our goal of parameterizing a basic
model of prospect theorys utility function in closed form with elementary
functions and using only integers, at least in the first instance, to define
scale parameter and shape parameter . By placing both x and F(x; , )
within a range between 1 and 1, we have specified a model that can rep-
resent actual gains and losses along the x-axis and subjective value along
the y-axis.
8.5 Flagging Prospect Theory: Two-Parameter

Lognormal Distribution
A very modest modification allows us to depict the same flag of prospect
theorys value function using the CDF of the lognormal distribution.81 A
lognormal distribution characterizes the distribution of a random positive
variable x whose logarithm is normally distributed. Therefore, the lognor-
mal distribution is to the normal, Gaussian distribution as the log-logistic
distribution is to the logistic distribution. Unlike the log-logistic, the log-
Fig. 8.6 Flagging prospect theory: a parametric depiction of the value function
according to the log-logistic distribution
194 J.M. CHEN
normal distribution does not have a CDF that can be written in closed
form with elementary functions. That function, however, can be written
with erf, the error function, where erf(x) denotes:82
x
2
erf ( x ) =
2
e - t dt
p 0
The probability density function of the two-parameter lognormal dis-

tribution is as follows:
1 ( ln x - m )2
f ( ln x; m , s ) = exp - , x > 0
xs 2p 2s 2

The CDF of the two-parameter lognormal distribution takes the following

form:
1 1 ln x - m
F ( ln x; m , s ) = + erf

2 2 s 2
The relationship between the log-logistic distributions and

parameters and the lognormal distributions and parameters is quite
straightforward. A log-logistic distribution where =1 corresponds to a
lognormal distribution where =0. On those assumptions, it turns out
that =3/, and =3/.83 The lognormal distribution corresponding
to the log-logistic distribution whose parameters are =1 and =5 is
therefore:
3 5 25 ( ln x )2
f ln x; m = 0, s = = exp - , x > 0
5 x 6p 6

The log-logistic distribution corresponding to that lognormal distribu-

tion may be specified as:
PROSPECT THEORY 195
5x 4
f ( x; a = 1, b = 5 ) = , x>0
(1 + x )
2
5

Plotting the two probability density functions together in Fig. 8.7 shows
the closeness of the two distributions.84
All that remains is a straightforward application of the three operations
outlined in 8.4 to the CDF of the lognormal distribution: (1) Add 1 to
x so that the CDF is shifted left along the x-axis. (2) Subtract from the
value of the equation so that the CDF is shifted down along the y-axis. (3)
Multiply by 3 to rescale the flag so that it displays entirely within the
range of 1x1 and 1y1. The resulting function for the CDF
of a lognormal distribution with parameters =0 and =3/5 takes the
following form:
ln ( x + 1) - m
F ( ln x; m , s ) = erf
s 2
3 5 ln ( x + 1)
F ln x; m = 0, s = = erf
5 6

Fig. 8.7 A comparison of the probability density functions for the log-logistic
(red) and the two-parameter lognormal (blue) distributions
196 J.M. CHEN
Fig. 8.8 Flagging prospect theory: comparing the log-logistic and two-parameter
lognormal representations of prospect theorys value function
2 ( x + 1)
5
A plot of that function alongside F ( x; a = 1, b = 5 ) = -1 ,

1 + ( x + 1)
5
the prospect theory flag derived in 8.4 from the CDF of the log-
logistic distribution, is depicted in Fig. 8.8.85
Whether rendered for a log-logistic or a lognormal distribution, the
methodology outlined here satisfies all three conditions imposed by pros-
pect theory: (1) an inflection point at the reference point, recentered at
the origin, (2) plus a steeper slope for subjective value assigned to losses
vis--vis the comparatively modest slope for subjective value ascribed to
gains, and (3) diminishing sensitivity to losses and to gains. One can there-
fore illustrate prospect theory in vivid, graphic form, with an extremely
parsimonious mathematical apparatus.
8.6 Cumulative Prospect Theory

The conclusion to Kahneman and Tverskys 1979 article conceded that
the original formulation of prospect theory fell short of a fully adequate
account of decision-making processes under risk and uncertainty as
complex phenomena.86 Kahneman and Tversky did succeed in using
the simple overweighting of small probabilities and the S-shaped value
function to explain the typical conditions under which lottery tickets
and insurance policies are sold.87
PROSPECT THEORY 197
Kahneman and Tverskys original formulation of prospect theory,

however, could not fully account for instances in which the purchase of
insurance extends to the medium range of probabilities, and small
probabilities of disaster are sometimes entirely ignored.88 In addition to
pure attitudes toward uncertainty and money, a more comprehensive
theory of insurance behavior should consider such factors as the value
of security, social norms of prudence, [and] the aversiveness of a large
number of small payments spread over time.89
The shortcomings in prospect theorys original formulation lay in the
theorys approach to the weighting of probabilities through a monotonic
transformation of outcome probabilities.90 This approach did not invari-
ably satisfy stochastic dominance, an assumption many theorists wished
to retain.91 Nor was original prospect theory readily extended to pros-
pects with a large number of outcomes.92 Adopting advances achieved
in the literature of behavioral economics that responded to their origi-
nal formulation of prospect theory,93 Tversky and Kahneman embraced
the rank dependent or cumulative functional as their expression of the
weighting function.94 The practical effect of this change was dramatic:
Instead of transforming each probability separately, cumulative prospect
theory transforms the entire cumulative distribution function.95
Cumulative prospect theorys combination of this revised weighting
function with the original formulations value function, which would
remain concave for gains, convex for losses, and steeper for losses than
for gains, now offered a complete response to the failure of expected util-
ity theory.96 In addition to providing a more persuasive account of human
beings different evaluations of gain and losses, cumulative prospect
theory could now provide[] a unified treatment of both risk and uncer-
tainty.97 With appropriate humility, however, Tversky and Kahneman
have conceded that cumulative prospect theory, notwithstanding greater
generality, remains unlikely to be accurate in detail.98 The new formu-
lation leaves decision weights subject to influence by the formulation
of prospects, as well as [by] the number, the spacing and the level of
outcomes.99
Nevertheless, cumulative prospect theory does satisfy five major phe-
nomena of choice, which set a minimal challenge that must be met by
any adequate descriptive theory of choice.100
First, framing effects. A behavioral account of decision-making under
risk or uncertainty must explain how variations in the framing of options,
198 J.M. CHEN
especially in terms of gains or losses, can yield systematically different

preferences.101
Second, nonlinear preferences. A persuasive theory of behavioral eco-
nomics must accommodate the Allais paradox, which show[ed] that the
difference between probabilities of 0.99 and 1.00 has more impact on
preferences than the difference between 0.10 and 0.11.102
Third, source dependence. A persons willingness to bet on an uncer-
tain event depends not only on the degree of uncertainty but also on its
source.103 For example, consistent with home bias,104 people often pre-
fer a bet on a [vague] event in their area of competence over a [clear] bet
on an otherwise random event.105
Fourth, risk seeking. Despite the generally assumed prevalence of risk
aversion, risk-seeking choices are consistently observed in assessments
of human behavior.106 Two manifestations of risk seeking are especially
interesting. Within the domain of gains, people often prefer a small prob-
ability of winning a large prize over the expected value of that prospect.107
People likewise engage in risk seeking within the domain of losses when
they must choose between a sure loss and a substantial probability of a
larger loss.108 It is nearly self-evident that both behaviors constitute risk
seeking in the sense that they represent decisions to forgo the expected
value of a gain or a loss, in favor of an uncertain outcome that is worth less
according to a strictly probabilistic evaluation.
Fifth and finally, loss aversion. Risk aversion, though far from universal,
remains the default frame of human emotion. Losses continue to loom
larger than gains.109 The observed asymmetry between gains and losses
is far too extreme to be explained by income effects or by decreasing risk
aversion.110
8.7 The Weighting Function

The value of a prospect f in cumulative prospect theory is a function of
the value of each prospect, which in turn is a function of its nominal value,
adjusted by the subjective weight assigned to its probability. A grossly
oversimplified version of this function follows:111
n
V(f)= p i v ( xi )
i =- m
PROSPECT THEORY 199
where indexing variable i tracks prospects in ranked order from m

to n, v(xi) is the value function defined in 8.4 as a two-part power
function,112 and i represents the value of a prospect. i is defined as p i+
if i0 and p i- if i<0.113
Empirical studies have consistently suggested an inverse-S-shaped
weighting function, which is concave below and convex above some fixed
p*<0.40. Such an S-shaped function is consistent with risk seeking for
low probability gains and risk aversion for medium and high probability
gains.114 The pivotal step in defining the weighting function is defining
the weights to be assigned to probability p of all prospects i in the domain
of gains and in the domain of losses. Let w+(p) represent the weight in the
domain of gains and w(p) designate the weight in the domain of losses.
Tversky and Kahneman set out to define w(p) using only one param-
eter capable of reporting weighting functions with both concave and
convex regions.115
The following formulas, assigning different variables ( and ) for the
domain of gains and the domain of losses, achieve this goal:116
pg
w+ ( p ) = 1/ g
pg + (1 - p )g

d
p
w- ( p ) = 1/ d
pd + (1 - p )d

As shape parameter or approaches 0, the stretching of the weighting

function at either end of its range, 0 or 1, is greater. For =1 or =1, the
weighting function w(p)=p.117
In their 1992 article, Tversky and Kahneman estimated values for
these parameters: 0.61 and 0.69.118 The two weighting functions,
for gains in blue and for losses in red, appear in Fig. 8.9 alongside a
golden line representing perfectly risk-neutral weighting.119
The curvature of the weighting functions depicts the attitude of
most humans toward low, moderate, and high probabilities. [F]or
both positive and negative prospects, people overweight low probabili-
ties and underweight moderate and high probabilities.120 The modest
200 J.M. CHEN
Fig. 8.9 Prospect theorys weighting function according to Tversky and

Kahnemans original parameters
difference between the shape parameters for the domain of gains,

0.61, and the domain of losses, 0.69, would produce weight-
ing functions for gains and for losses [that] are quite close.121 But
because the shape parameter for gains calls for greater distortionthat
is, < the weighting function for gains is slightly more curved
than the weighting function for losses.122 Accordingly, risk aversion
for gains is more pronounced than risk seeking for losses, for moderate
and high probabilities.123
Empirical tests of prospect theory sometimes report lower levels of dis-
tortion in the weighting of probabilities. A 2009 study of stock option
data from the USA, for instance, generated values of =0.84 and =0.76,
yielding probability weighting functions for gains and losses generally
of the form suggested by cumulative prospect theory, but more lin-
ear than that estimated by Tversky and Kahnemans original laboratory
experiments (Fig. 8.10).124
Curiously, this field study of option prices reversed the usual relation-
ship between gains and losses. Tversky and Kahneman found that even
though the weighting functions for gains and for losses are quite close,
the former is slightly more curved than the latter (i.e., <).125 This
studys finding of a slightly higher value for the loss function reversed the
conventional supposition that risk aversion for gains is more pronounced
than risk seeking for losses.126
PROSPECT THEORY 201
Fig. 8.10 Replotting prospect theorys weighting function according to the

results of a 2009 field study
8.8 The Fourfold Pattern

By way of summary, we can now consolidate what prospect theory has
taught us about behavioral finance. The asymmetrical utility curve that
emerges from prospect theorys reevaluation of conventional accounts
of expected economic utility leads to some seeming contradictions.127 In
mixed gambles, for instance, where a decision-maker may realize either a
gain or a loss, loss aversion leads to extreme, even costly risk aversion. This
is perhaps the most familiar conclusion of prospect theory, the one most
readily summarized by the slogan, losing hurts worse than winning feels
good.128
But prospect theory predicts affirmatively risk-seeking behavior in other
circumstances. When a decision-maker is confronted with nothing but
bad choicesspecifically, those where a sure loss is compared to a
larger loss that is merely probablediminishing sensitivity to losses will
generate a greater willingness to absorb risk.129
Prospect theory therefore rests on two principal insights. First, humans
attach values to gains and losses rather than to wealth.130 Second, humans
making decisions assign weights to outcomes [that] are different from
probabilities.131 [A]ny good descriptive model of decision-making
under uncertainty should accommodate both gain-loss asymmetry and
nonlinearity in probability.132 The combination of these two heuristics
generates a distinctive pattern of preferences that Kahneman and Tversky
have called the fourfold pattern.133 Prospect theorys fourfold pattern
of risk attitudes [has] emerge[d] as a major empirical generalization about
202 J.M. CHEN
Table 8.1 Prospect theorys fourfold pattern of risk attitudes

The fourfold Gains Losses
pattern
High E.g., a 95% chance to win E.g., a 95% chance to lose $10,000
probability $10,000 leads to risk aversion leads to risk seeking (rogue trading
(certainty (annuities and sinecures) and other reckless gambles)
effect)
Low E.g., a 5% chance to win E.g., a 5% chance to lose $10,000
probability $10,000 leads to risk seeking leads to risk aversion (insurance)
(possibility (lotteries)
effect)
choices and risk.134 The fourfold pattern arises from the interaction of
prospect theorys value and weighting functions (Table 8.1). The shapes
of those functions imply risk-averse and risk-seeking preferences, respec-
tively, for gains and for losses of moderate or high probability.135 As long
as the outcomes are not extreme, the shape of the weighting functions
(one each for the domain of gains and the domain of losses) favors risk
seeking for small probabilities of gains and risk aversion for small probabili-
ties of loss.136
Let us examine more closely each of the four corners in prospect theo-
rys fourfold pattern. Three of these four behavioral possibilities had long
been understood before Kahneman and Tversky; prospect theory merely
provided the means by which to describe them formally.137 The cell at top
left describes how risk aversion leads people to lock in a sure gain below
the expected value of a gamble. Annuities work on this principle, as do
employment guarantees in unionized trades or on tenure-protected uni-
versity faculties.
The cell at lower right describes insurance: Individuals will pay much
more than the expected value of a loss to insure themselves against the
disturbing prospect of a catastrophic loss. On the flip side of that transac-
tion, insurance companies can pool risks assigned to them by risk-averse
policyholders and profit from the spread between expected losses and
premium payments. These risk-averse decisions reflect one of the core
instincts within prospect theory. Prospect theorys defensive account res-
onates with mental accounting, other behaviorally sensitive approaches
to portfolio design and asset pricing, and postmodern portfolio theorys
early emphasis on downside risk measures.138
PROSPECT THEORY 203
But there is also a risk-seeking side to this account of human behav-

ior. Lotteries routinely exploit the possibility effect. When the poten-
tial payout is enormous, ticket buyers become indifferent to their
miniscule chances of winning. This is the behavioral pattern reflected
by the lower left cell. The longing for lotteries, as a manifestation of
skewness preference, is consistent with the four-moment capital asset
pricing model, where the Taylor series expansion of expected returns
predicts negative investor reactions to even-numbered moments and
positive reactions to odd-numbered moments.139 The desire for invest-
ment opportunities with lottery-like payouts, ranging from actual lot-
teries to initial public offerings, warrants deeper examination in the
next chapter.140
What Kahneman and Tversky found truly surprising was the fourth
possibility, the one described in the risk-seeking cell at upper right of
Table8.1s summary of prospect theorys fourfold pattern. When humans
face the high probability of severe losses, they engage in affirmatively risk-
ier behavior. Prospect theory identifies two reasons for this sudden shift
in strategy.141 First, diminishing sensitivity means that humans react very
adversely to a sure loss: [T]he reaction to a loss of $900 is more than
90% as intense as the reaction to a loss of $1,000.142 The existence of a
loss, not its magnitude, is pivotal. Second and perhaps even more signifi-
cant, humans assign a much lower decision weight to an extreme loss than
its rationally expected value as calculated by the laws of probability. The
certainty effect, coupled with diminishing sensitivity, enhances the averse-
ness of a sure loss and reduces the averseness of the gamble.
This is the ugly corner of human decision-making where other-
wise responsible parties find themselves tempted to take risks that can
turn[] manageable failures into disasters.143 Rogue traders who
have amassed appalling losses let it all ride on a single reckless trade.
That gamble may destroy a systemically important financial institution.
A brief recitation of infamous rogue trading episodes suffices to illus-
trate the point. Bruno Iksil, better known as the London Whale,
inflicted a multibillion-dollar loss on J.P.Morgan Chase in 2012.144 Jon
Corzine destroyed MF Global in 2013 by mismanaging $1.5 billion.145
Nick Leeson destroyed Barings Bank in 1995 with a single series of
ill-fated trades.146 And literally, volumes have been written about Long-
Term Capital Management, whose multibillion-dollar failure in 1998
required the intervention of the Federal Reserve to stave off a global
financial crisis.147 Long-Term Capital Management, at least, illustrates
204 J.M. CHEN
a curious sort of negative skewness preference at odds with the more

common preference for investments offering positively skewed, lottery-
like returns.148 Contrary to those instances in which investors gam-
ble pennies to win dollars, stories such as that of Long-Term Capital
Management illustrate the opposite problem of gambl[ing] dollars to
win a succession of pennies.149
Looming defeat is the very corner of the fourfold pattern where
investment advisors, pension fund managers, and other practitioners of
behavioral finance should be most vigilant. Because defeat is so diffi-
cult to accept, chief executive officers and field marshals suffer from a
comparable inability to cut their losses and salvage what is left of their
companies and armies.150 As Kahneman and Tversky presciently recog-
nized in their original 1979 article, a person who has not made peace
with his losses is likely to accept gambles that would be unacceptable
under other circumstances.151
And once agents choose a losing course of action, prospect theory
predicts escalating commitment to that losing strategy instead of cautious
retreat.152 Treating the escalation of commitment as an artifact of innate
decision-making processesas opposed to self-justification, or an agents
psychological or sociological need to preserve the illusion of not hav-
ing erred153suggests that humans are likelier to double down on los-
ing strategies across a very wide range of circumstances.154 Whatever the
underlying mechanism, the dynamics of defeat leave little surprise that
truly large corporate failures result from downward spirals, the accumu-
lated progression of decisions whose eventual, final outcome is economic
death.155 As the costs and the likely futility of mitigating action increase,
humans find their own heuristics shoving their collective decision-mak-
ing processes further onto the frontier of desperation where risk-averse
acts such as hedging and insurance lose their appeal and yield ground
to active risk seeking. The rapid, automatic decision-making system that
has propelled humanity from Pleistocene competitiveness to Holocene
dominance and Anthropocene hegemony156 contains the seeds of its own
destruction.157
For their part, Amos Tversky and Daniel Kahneman have taken a more
sanguine view of human behavior. They have always posited prospect theory
as a descriptive, not a normative theory.158 In response to the convic-
tion that only rational behavior can survive in a competitive environment,
Tversky and Kahneman have observed that people can spend a lifetime
in a competitive environment without acquiring a general ability to avoid
PROSPECT THEORY 205
framing effects or to apply linear decision weights.159 Far from being cha-
otic and intractable, the human choices under risk and uncertainty are
orderly, even if they are not always rational in the traditional sense.160
Notes
1. Nicholas C. Barberis, Thirty Years of Prospect Theory in Economics: A
Review and Assessment, 27J.Econ. Persp. 173196, 173 (2013).
2. See Daniel Kahneman, Maps of Bounded Rationality: A Perspective on
Intuitive Judgment and Choice (Dec. 8, 2002) (available at http://www.
nobelprize.org/nobel_prizes/economic-sciences/laureates/2002/kahne-
mann-lecture.pdf), published sub nom. Daniel Kahneman, Maps of Bounded
Rationality: Psychology for Behavioral Economics, 93 Am. Econ. Rev. 1449
1475 (2003). Amos Tversky presumably would have shared Kahnemans
Nobel Prize, but he had flunked a core eligibility criterion by dying in 1996.
3. For Daniel Kahnemans own summary of prospect theory, see Kahneman,
Thinking, Fast and Slow, supra note 11 (Chapter 1), at 278288. See also,
e.g., Peter P.Wakker, Prospect Theory: For Risk and Ambiguity (2010).
5. See generally supra 6.16.3, at 111115.
6. See generally, e.g., Steven T.Buccola, Portfolio Selection Under Exponential
and Quadratic Utility, 7 W.J. Agric. Econ. 4251 (1982); Steven
D.Hanson & George W.Ladd, Robustness of the Mean-Variance Model
with Truncated Probability Distributions, 73 Am. J.Agric. Econ. 436445
(1991); Liping Liu, A New Foundation for the Mean-Variance Analysis,
158 Eur. J.Oper. Research 229242 (2004).
7. See generally supra. 6.4, at 115117.
9. Id. at 84.
10. Robert A. Collins & Edward E. Gbur, Quadratic Utility and Linear
Mean-Variance: A Pedagogic Note, 13 Rev. Agric. Econ. 289291, 289
(1991).
11. Kahneman & Tversky, Prospect Theory, supra note 90 (Chapter 6), at 264.
12. Id. at 269.
13. Id. (citing Victor R.Fuchs, From Bismark to Woodcock: The Irrational
Pursuit of National Health Insurance, 19 J.L. & Econ. 347359 (1976)).
15. Id.
16. Id.
17. Id.
18. See id. at 265267.
19. See id.
206 J.M. CHEN
20. Id. at 267.

21. Id. at 271.
22. See Daniel Ellsberg, Risk, Ambiguity, and the Savage Axioms, 75
Q.J.Econ. 643669, 650651 (1961); Craig R.Fox & Amos Tversky,
Ambiguity Aversion and Comparative Ignorance, 110 Q.J. Econ. 585
603 (1995); Chip Heath & Amos Tversky, Preference and Belief:
Ambiguity and Competence in Choice Under Uncertainty, 4 J. Risk &
Uncertainty 528 (1991).
24. Id. (citing C.Arthur Williams, Jr., Attitudes Toward Speculative Risks as
an Indicator of Attitudes Toward Pure Risk, 33J.Risk & Ins. 577586
(1966); Peter C. Fishburn & Gary A. Kochenberger, Two-Piece von
Neumann-Morgenstern Utility Functions, 10 Decision Scis. 503518
(1979)).
26. Id. at 268269.
27. Id. at 269.
28. Id.
29. See Shannon L. Hamm & Sara J. Shettleworth, Risk Aversion Among
Pigeons, 13 J. Experimental Psych: Animal Behav. Processes 376383
(1987).
30. See Harry Markowitz, The Utility of Wealth, 60J.Pol. Econ. 151158
(1952).
32. Id.
33. Id. at 276.
35. Id.
36. Id. at 275.
37. Id.
38. Id.
39. Id. (citing Henry Helson, Adaptation-Level Theory (1964)).
41. Id. at 276.
42. Id.
43. Id.
44. Id. at 277.
45. Chris Guthrie, Prospect Theory, Risk Preference, and the Law, 97 Nw.
U.L. Rev. 11151163, 1116 (2003).
46. See Kahneman, supra note 3 (Chapter 8), at 282.
47. See John W. Payne, Dan J. Laughhunn & Roy Crum, Translation of
Gambles and Aspiration Levels in Risky Choice Behavior, 26 Mgmt. Sci.
10391060 (1980); John W.Payne, Dan J.Laughhunn & Roy Crum,
Further Tests of Aspiration Level Effects in Risky Choice Behavior, 27
PROSPECT THEORY 207
Mgmt. Sci. 953958 (1987); Christopher P. Puto, The Framing of

Buying Decisions, 14J.Consumer Research 301315 (1987).
48. See Peter C. Fishburn, Mean-Risk Analysis with Risk Associated with
Below Target Returns, 67 Am. Econ. Rev. 116126 (1977).
49. James C.T. Mao, Survey of Capital Budgeting: Theory and Practice,
25 J. Fin. 349360, 353 (1970); see also Robert Libby & Peter
C. Fishburn, Behavioral Models of Risk Taking in Business Decisions: A
Survey and Evaluation, 15J.Accounting Research 272292 (1977).
50. See William Samuelson & Richard Zeckhauser, Status Quo Bias in Decision
Making, 1J.Risk & Uncertainty 759 (1988).
51. Nicholas Barberis & Wei Xiong, Realization Utility, 104J.Fin. Econ.
251271, 251 (2012); see also Jonathan E. Ingersoll Jr. & Lawrence
J.Jin, Realization Utility with Reference-Dependent Preferences, 26 Rev.
Fin. Stud. 723767 (2013).
52. See generally Daniel Kahneman, Jack L.Knetsch & Richard H.Thaler,
The Endowment Effect, Loss Aversion, and Status Quo Bias, 5 J. Econ.
Persp. 193206 (1991).
53. Benartzi & Thaler, Myopic Loss Aversion, supra note 20 (Chapter 2), at 74.
54. United States v. Butler, 297 U.S. 1, 82 (1936).
55. Lewis Grizzard, Gettin It On: A Down-Home Treasury 72 (1990); accord
Joe Garagiola, Its Anyones Ballgame 109 (1988). Lewis Grizzard (1946
94) was a humor columnist for the Atlanta Journal-Constitution.
56. See Kahneman & Tversky, Prospect Theory, supra note 90 (Chapter 6), at
282283.
57. Id. at 282.
58. Amos Tversky & Daniel Kahneman, Advances in Prospect Theory:
Cumulative Representation of Uncertainty, 5 J. Risk & Uncertainty
297323 (1992).
59. Id. at 309.
60. Id. at 311.
61. Id.
62. Levy, CAPM in the 21st Century, supra note 41 (Chapter 1) at 398.
63. Tversky & Kahneman, Advances in Prospect Theory, supra note 58
65. Id.
66. Id.; cf. id. at 83 (calculating a loss aversion factor of 2.77 in an evaluation
of prospect theorys contribution to solving the equity premium puzzle).
67. Id. at 79 n.5.
68. See Daniel Kahneman, Jack Knetsch & Richard H.Thaler, Experimental
Tests of the Endowment Effect and the Coase Theorem, 98 J. Pol. Econ.
13251348 (1990).
208 J.M. CHEN
69. See https://en.wikipedia.org/wiki/Kronecker_delta. The Kronecker

delta function reports the following values for arguments i and j:
ij=0 if ij
ij=1 if i=j
Though arguments i and j are usually restricted to positive integers,
Kronecker delta function, in principle, may accept any discrete rational
number for these arguments. See id.
70. See http://www.wolframalpha.com/input/?i=plot+2.25%5EKronecker
Delta%28x%2C-abs%28x%29%29*sign%28x%29*abs%28x%29%5E0.88+
for+-1%3Cx%3C1&f=1
71. See Kahneman, Thinking, Fast and Slow, supra note 11 (Chapter 1), at 282.
72. See Haim Levy & Zvi Wiener, Prospect Theory and Utility Theory:
Temporary and Permanent Attitudes About Risk, 68 J. Econ. & Bus.
123 (2002) (concluding that a positive risk premium reflecting decreas-
ing absolute risk aversion is consistent with prospect theorys sigmoid
utility function).
73. See Denis Conniffe, The Generalised Extreme Value Distribution as Utility
Function, 38 Econ. & Soc. Rev. 275288, 275 (2007); Marco LiCalzi &
Annamaria Sorato, The Pearson System of Utility Functions, 172 Eur.
J.Oper. Research 560573 (2006).
74. Cf. R.H. Coase, The Problem of Social Cost, 3 J.L. & Econ. 144, 43
(1960) (recognizing that all problems of welfare economics must ulti-
mately dissolve into a study of aesthetics and morals).
75. See generally http://en.wikipedia.org/wiki/Log-logistic_distribution
76. See generally Peter R.Fisk, The Graduation of Income Distributions, 29
Econometrica 171185 (1961); cf. Teun Kloek & Herman K. van Dyk,
Efficient Estimation of Income Distribution Parameters, 8J.Econometrics
6174 (1978).
77. I have derived all expressions and graphs from Wolfram Alpha, http://
www.wolframalpha.com. In my first example, I inserted this command
into Wolfram Alpha: plot x^5/(1+x^5), 0<x<2.
78. The operative Wolfram Alpha input is this command: plot (x + 1)^5/
(1+(x+1)^5), 1<x<1.
79. I entered the following command in Wolfram Alpha: plot (x + 1)^5/
(1+(x+1)^5)5, 1<x<1.
80. I entered the following command in Wolfram Alpha: plot 2(x+1)^5/
(1+(x+1)^5)1, 1<x<1.
81. This discussion of the properties of the lognormal distribution draws
upon background sources on the lognormal distribution. See generally
sources cited supra note 75 (Chapter 7).
82. See https://en.wikipedia.org/wiki/Error_function
83. See Arabin Kumar Dey & Debasis Kundu, Discriminating Between the
Log-Normal and Log-Logistic Distributions, 39 Communications in Stat.:
PROSPECT THEORY 209
Theory & Methods 280292 (2009) ( 3 and appendix); cf. Fahim

Ashkar & Franois Aucoin, Discriminating Between the Lognormal and
the Log-Logistic Distributions for Hydrological Frequency Analysis,
17J.Hydrological Engg 160167, 161 (2012) (observing that discrim-
ination between the lognormal and log-logistic distributions is equiva-
lent to discrimination between the [normal] and [logistic] distributions);
Elsyed A. Elsherpieny, Sahar A.N. Ibrahim & Noha U.M.M. Radwan,
Discriminating Between Weibull and Log-Logistic Distributions, 2 Intl
J.Innovative Research in Sci., Engg & Technol. 33583371 (2013).
84. I entered the following command in Wolfram Alpha: plot 5/
(x*sqrt(6*pi))*exp.(25*ln(x)^2/6) and 5*x^4/(1 + x^5)^2 for
0<x<2.
85. I entered the following command in Wolfram Alpha: plot erf(5*ln(x+1)/
sqrt(6)) and 2*(x+1)^5/(1+(x+1)^5)1 for 1<x<1.
87. Id.
88. Id.
89. Id.
90. Tversky & Kahneman, Advances in Prospect Theory, supra note 58, at
299.
91. Id.
92. Id.
93. See Arjan B. Berkelaar, Roy Kouwenberg & Thierry Post, Optimal
Portfolio Choice Under Loss Aversion, 86 Rev. Econ. & Stat. 973998
(2004); John Quiggin, A Theory of Anticipated Utility, 3J.Econ. Behav.
& Org. 323343 (1982); David Schmeidler, Subjective Probability and
Expected Utility Without Additivity, 57 Econometrica 571587 (1989);
J.A. Weymark, Generalized Gini Inequality Indices, 1 Math. Soc. Scis.
409430 (1981); Menahem E.Yaari, The Dual Theory of Choice Under
Risk, 55 Econometrica 95115 (1987).
299.
95. Id.
96. Id. at 298.
97. Id.
98. Id. at 317.
99. Id.
100. Id. at 298.
101. Id. (citing Tversky & Kahneman, Rational Choice and the Framing of
Decisions, supra note 21 (Chapter 2)).
102. Tversky & Kahneman, Advances in Prospect Theory, supra note 58, at 298.
103. Id.
210 J.M. CHEN
104. See supra 7.7, at 154; 7.8, at 156; source cited supra note 164 (Chapter 7).
105. Tversky & Kahneman, Advances in Prospect Theory, supra note 58, at 298
(citing Heath & Tversky, supra note 22).
298.
107. Id.
108. Id.
109. Id.
110. Id.
111. See id. at 300; Benartzi & Thaler, Myopic Loss Aversion, supra note 20
(Chapter 2), at 79.
309.
113. See id. at 300.
114. George Wu & Richard Gonzalez, Curvature of the Probability Weighting
Function, 42 Mgmt. Sci. 16761690, 1677 (1996).
310; accord Benartzi & Thaler, Myopic Loss Aversion, supra note 20
(Chapter 2), at 80 (acknowledging cumulative prospect theorys one-
parameter approximation of its weighting function); see also John
Quiggin, On the Optimal Design of Lotteries, 58 Economica 116 (1991).
116. See Tversky & Kahneman, Advances in Prospect Theory, supra note 58, at
309.
117. See Nicholas Barberis & Ming Huang, Stocks as Lotteries: The Implications
of Probability Weighting for Security Prices, 98 Am. Econ. Rev. 2066
2100, 2071 (2008) (recognizing that setting the shape parameter to
=1corresponds to no probability weighting at all).
118. See Tversky & Kahneman, Advances in Prospect Theory, supra note 58, at
312.
119. See http://www.wolframalpha.com/input/?i=plot+x%5E.61%2F%28x%
5E.61%2B%281-x%29%5E.61%29%5E%281%2F.61%29+and+x%5E.69%
2F%28x%5E.69%2B%281-x%29%5E.69%29%5E%281%2F.69%29+and+x
+for+x%3D0+to+1&f=1
312.
121. Id.
122. Id.
123. Id.
124. Gregory Gurevich, Doron Kliger & Ori Levy, Decision-Making Under
UncertaintyA Field Study of Cumulative Prospect Theory, 33J.Banking
& Fin. 12211229, 12241225 (2009); see also id. at 1226 (table 5)
(comparing field data from 30 stock samples with Tversky and
Kahnemans 1992 laboratory results). Figure 8.10 was generated at
PROSPECT THEORY 211
http://www.wolframalpha.com/input/?i=plot+x%5E.84%2F%28x%5E
.84%2B%281-x%29%5E.84%29%5E%281%2F.84%29+and+x%5E.76%2
F%28x%5E.76%2B%281-x%29%5E.76%29%5E%281%2F.76%29+and+x
+for+x%3D0+to+1
312.
126. Id.
127. See Kahneman, supra note 3, at 285.
128. See sources cited supra note 55.
130. Id. at 316317.
131. Id. at 317.
132. Wu & Gonzalez, supra note 114, at 1676.
133. Kahneman, supra note 3, at 317; accord Wu & Gonzalez, supra note
114, at 1676 (Together, these two regularities capture the fourfold pat-
tern of risk attitudes: risk aversion for most gains and low probability
losses, and risk seeking for most losses and low probability gains.); see
also Tversky & Kahneman, Advances in Prospect Theory, supra note 58, at
306 (The most distinctive implication of prospect theory is the fourfold
pattern of risk attitudes.).
307.
135. Id. at 306.
136. Id.
138. See Chen, Postmodern Portfolio Theory, supra note 1 (Chapter 1),
5.16.11, at 59105.
139. See supra 3.2, at 6066.
140. See infra 9.19.2, at 215224.
142. Id.
143. Id. at 319.
144. See Stephen Gandel, How JPMorgan Made Its Multi-Billion Dollar
Blunder, CNN Money (May 15, 2012) (available at http://finance.for-
tune.cnn.com/2012/05/15/jpmorgan-london-whale-blunder).
145. See Roger Parloff, How MF Globals Missing $1.5 Billion Was Lostand
Found, CNN Money (Nov. 15, 2013) (available at http://features.blogs.
fortune.cnn.com/2013/11/15/mf-global-joncorzine).
146. See The Economy: How Leeson Broke the Bank, BBC News Business (June
22, 1999) (available at http://news.bbc.co.uk/2/hi/business/375259.
stm).
147. See, e.g., Nicholas Dunbar, Inventing Money: The Story of Long-Term
Capital Management and the Legends Behind It (2000); Laurent
L. Jacque, Global Derivative Debacles: From Theory to Malpractice
212 J.M. CHEN
245273 (2010); Roger Lowenstein, When Genius Failed: The Rise and
Fall of Long-Term Capital Management (2000).
148. See infra 9.1, at 215222.
149. Nassim Nicholas Taleb, Bleed or Blowup? Why Do We Prefer Asymmetric
Payoffs?, 5J.Behav. Fin. 27, 2 (2004).
150. Kahneman, supra note 3, at 319.
152. See Glen Whyte, Escalating Commitment to a Course of Action: A
Reinterpretation, 11 Acad. Mgmt. Rev. 311321 (1986).
153. See, e.g., Barry M.Staw, Knee-Deep in the Big Muddy: A Study of Escalating
Commitment to a Chosen Course of Action, 16 Org. Behav. & Human
Performance 2744 (1976); Barry M.Staw, The Escalation of Commitment
to a Course of Action, 6 Acad. Mgmt. Rev. 577587 (1981); Barry
M.Staw, The Escalation of Commitment: An Update and Appraisal, in
Organizational Decision Making 191215 (Zur Shapira ed., 1997); Karl
E.Weick, Reduction of Cognitive Dissonance Through Task Enhancement
and Effort Expenditure, 56J.Abnormal & Soc. Psych. 152155 (1964).
154. Cf. Glen Whyte, Diffusion of Responsibility: Effects on the Escalation
Tendency, 76J.Applied Psych. 408415, 411413 (1991) (finding less
support for the self-justification account of escalation of commitment in
collective decision-making settings). See generally Joel Brockner, The
Escalation of Commitment to a Failing Course of Action: Toward
Theoretical Progress, 17 Acad. Mgmt. Rev. 3961 (1992) (reviewing the
relationship between self-justification explanations for the escalation of
commitment and alternative accounts, including prospect theory).
155. See Donald C.Hambrick & Richard A.DAveni, Large Corporate Failures
as Downward Spirals, 33 Admin. Sci. Q. 123 (1988).
156. See generally, e.g., Will Steffen, Jacques Grinewald, Paul Crutzen & John
McNeil, The Anthropocene: Conceptual and Historical Perspectives, 369
Phil. Trans. Royal Socy A 842867 (2011); Jan Zalasiewicz, Mark
Williams, Will Steffen & Paul Crutzen, The New World of the Anthropocene,
44 Envtl. Sci. & Tech. 22282231 (2010).
157. See generally Keith E.Stanovich & Richard F.West, Individual Differences
in Reasoning: Implications for the Rationality Debate, 23 Behav. & Brain
Sci. 645726 (2000).
317.
159. Id.
CHAPTER 9
Specific Applications ofProspect Theory

toBehavioral Finance
9.1 SKEWNESS PREFERENCE AND THE LONGING

FORLOTTERIES
Prospect theory illuminates multiple problems in behavioral finance. In

economics generally and in finance in particular, the real challenge lies in
know[ing] exactly how to apply prospect theorys many remarkable
insights.1 Whatever its limitations, prospect theory represents a valu-
able refinement to the maximization assumption of rational choice and
expected utility theory, and accordingly should inform policymaking.2
Among other possibilities, prospect theory may shed light on the dis-
position effect, momentum in returns, and the equity premium puzzle.3
One of the most useful applications of this theory begins with an exami-
nation of the cluster of preferences aptly predicted by the lottery effect
within prospect theorys fourfold pattern. It has long been known that
many investors fail to diversify their portfolios. Many investors hold just a
few stocks, or even just one.4 The failure to diversifyperhaps more per-
suasively characterized as the tendency to concentrate holdings within a
portfoliois the true manifestation of optimism and risk seeking. Positive
sentiment leads to optimistic judgments and choices; negative sentiment,
conversely, triggers pessimistic decisions.5 Individual investors, whether
optimistic or realistic, tend to allocate similar amounts to equity, relative
to safer asset classes such as bonds or cash. The difference is that optimists
eschew diversification, to their detriment.6

DOI10.1007/978-3-319-32711-2_9
214 J.M. CHEN
Optimism does appear to have positive effects on entrepreneurship,

self-employment, and raw pleasure in ones work.7 But its impact on
financial prudence is decidedly mixed at best, and arguably quite delete-
rious to those investors who are unfortunate enough to be more confi-
dent and optimistic than their counterparts. (Gentlemen: Im speaking
to you.)8 Optimism leads investors to overestimate returns;9 overconfi-
dence prompts investors to underestimate risk.10 Optimism may also be
highlighted by the contrast with its emotional opposite, pessimism.11 The
ultimate economic embodiment of pessimism, right through layers of fear
to sheer terror, may be the fear of death. On that understanding, the pres-
ence of money serves as a talisman against death anxiety.12 One especially
vivid example of pessimism consists of the outsized fear of dying right
after buying an annuity, which suppresses demand for an otherwise useful
financial tool for retirement planning.13
The mathematical expression of a sunny outlook is a preference for
skewness.14 In markets around the world, stock investors appear willing
to accept lower returns in exchange for positive skewness on holdings
within their portfolio.15 Positive skewness suggests the possibility that cer-
tain holdings in the portfolio will offer disproportionately large payouts, as
though they were winning lottery tickets.16 On average, stocks perceived
to be positively skewed will earn lower returns.17 This preference for posi-
tively skewed, or lottery-like, wealth distributions may be expressed by
adjusting the shape parameters of cumulative prospect theorys weight-
ing functions.18 By shifting Tversky and Kahnemans original gain and loss
weighting parameters, respectively, from 0.61 and 0.6919 to 0.4
and 0.65, we can observe an even more pronounced overweighting of
low probabilities in the domain of gains (which is again depicted in blue,
relative to red for the domain of losses and gold for perfect risk neutrality)
(Fig. 9.1).20 Once again, humans overestimate low-probability events and
underestimate high-probability events.21
Although Postmodern Portfolio Theory and this book (especially Chapter
6s discussion of risk aversion) both emphasize the behavioral pitfalls
accompanying downside risk, considerable financial risk lurks when inves-
tors look toward the higher end of their expectations. Finance offers several
variations on the theme that Tennessee Williams described as the catastro-
phe of success.22 We have already caught a glimpse of how investors and
institutions shred their investment plans in the presence (or even the mere
anticipation) of upside gain.23 There is credible evidence that risk seek-
ing rather than risk aversion is the true behavioral impetus behind stock
SPECIFIC APPLICATIONS OFPROSPECT THEORY TOBEHAVIORAL FINANCE 215
Fig. 9.1 Shifting prospect theorys weighting function so that it reflects skew-
ness preference.
prices.24 Research into the risky choices made by pigeons, long a staple
of studies into animal psychology paralleling work on prospect theory in
humans,25 has revealed a similar pattern in avian behavior. As between fixed
payouts offering a higher amount of food and variable payouts offering
occasional bonanzas but a lower expected amount, pigeons prefer variable
payouts.26 This evidently innate preference for gambling may reveal a bio-
logical mechanism for gambling in birds and in humans.27
The relationship between low returns, positive skewness, and investor
affinity toward investments offering lottery-like payoffs appears across a
wide variety of financial settings:
Prize-linked savings accounts. Americans spent $70.2 billion on lottery
tickets in fiscal year 2014.28 That amount, confined to the jurisdictions
where lotteries are legal (44 states and the District of Columbia), exceeds
the amount spent in the entire country on sports tickets, books, video
games, movie tickets, and recorded music sales.29 Of the $70.2 billion in
lottery sales, $19.9 billion was remitted to lottery-sponsoring jurisdictions
as revenue.30 Worldwide lottery sales totaled $284 billion in 2014.31 The
profoundly negative economic impact of the lottery falls disproportion-
ately on the poor,32 many of whom treat the regular purchase of lottery
tickets as a form of savings.33 Even the economic impact of winning the
lottery falls far short of being unambiguously positive.34 The most that
might be said in defense of lottery participation is that syndicate play
buying tickets as part of a pool of friends, relatives, and/or coworkers
conveys social value independent of economic losses.35
The yearning for lotteries is sufficiently powerful that banks and credit
unions, with the strong backing of advocates for higher levels of savings
216 J.M. CHEN
and financial literacy among the poor, have resorted to depositor lotter-
ies to induce lower- to middle-income customers to open and fund sav-
ings accounts.36 Although prize-linked savings accounts are illegal in states
that treat them as nonsanctioned lotteries, they are prevalent around the
world.37 They operate on the same principle as lottery bonds, typically
issued by a sovereign that promises to redeem a select number of bonds
(substantially) above face value.38 The practice is actually quite old, with
American39 and European40 antecedents predating the twentieth century.
The oldest lottery bond may be the UKs Million Adventure of 1694,
a lottery intended to retire debt from the then-ongoing Nine Years War
(168997).41
Private equity. At the opposite end of the socioeconomic spectrum
from state-sponsored lotteries, wealthier households indulge in private
equity. Skewness preference may explain why anyone willingly invests
large amounts in a single privately held firm,42 despite a wretched trade-
off between risk and return. Returns on private equity are typically no bet-
ter than those on publicly traded stocks, and high correlation with public
equity reduces any diversification value that private equity might otherwise
offer.43 Investment in private equity, to say the least, is extremely concen-
trated: About 75 percent of all private equity is owned by households
for whom it constitutes at least half of their total net worth.44 Households
holding entrepreneurial equity invest on average more than 70 percent
of their private holdings in a single private company in which they have an
active management interest.45 In many households, income from entre-
preneurial venturesmore generally designated as proprietary business
incomerepresents a large source of undiversifiable risk that is more
highly correlated with common stock returns.46
In exchange for this dramatic lack of diversification, to say noth-
ing of the risk inherent in aligning returns from investment with highly
compensated but easily terminated personal labor, private equity investors
realize rather unimpressive returns that are no higher than the mar-
ket return on all publicly traded equity.47 Ownership of entrepreneur-
ial equity therefore constitutes not so much an exercise in staking out a
personalized corner along the efficient frontier, but rather a plunge off
that frontier in search of extreme skewness. Older households, at least, do
seem to understand the risk inherent in relying too heavily on proprietary
business income. As households age, they tend to move to safer assets,
by substituting riskier proprietary business ownership with stocks, bonds,
and especially cash.48
Equity crowdfunding. Among species of private equity, perhaps none

captures academic and popular attention as vividly as equity crowdfund-
ing.49 Before 2012, Regulation D of the SEC,50 the traditional exemption
that facilitated the offering of securities without SEC registration, made
private placement exemptions from federal securities laws generally
unavailable for crowdfunding transactions.51 In particular, the SECs def-
inition of an accredited investor52 put Regulation D beyond the reach
of most crowdfunding efforts, which are not intended to be limited to
investors that meet specific qualifications.53 Natural persons (as opposed
to corporations and other business or nonprofit entities) seeking to qualify
as accredited investors must either (1) have a net worth, either alone or
with a spouse, exceeding $1 million without the benefit of equity in a
primary residence,54 or (2) report individual income exceeding $200,000
or joint income with a spouse exceeding $300,000 within each of the
two most recent years and ha[ve] a reasonable expectation of reach-
ing the same income level in the current year.55 One journalistic source
estimated that there were 8.5 million accredited investors in the USA
in 2014.56 Given approximately 345.2 million adults in the USA as of
2014,57 the accredited cohort represented about 3.5% of the prospective
pool of American investors.
Congress passed the Jump Our Business Startups Act of 2012 (JOBS
Act)58 in order to establish a regulatory structure for startups and small
businesses to raise capital using Internet-based equity crowdfunding.59
The JOBS Act was intended to reduce the cost of making relatively low
dollar offerings of securities featuring relatively low dollar investments by
the crowd.60 At the same time, the JOBS Act also sought to protect
investors who engage in equity crowdfunding transactions by imposing
investment limits and requiring intermediaries to make disclosures to
investors and potential investors.61
In concert with the SECs final crowdfunding rule (promulgated on
November 16, 2015),62 the JOBS Act exempts qualifying equity crowd-
funding transactions63 from registration requirements that would oth-
erwise be applicable under section 5 of the Securities Act of 1933.64
Exempted crowdfunding securities must raise no more than $1 million in
any 12-month period.65 Individual investments in all crowdfunding issuers
in a 12-month period are limited to [t]he greater of $2,000 or 5 per-
cent of the investors annual income or net worth if either the investors
annual income or net worth is less than $100,000.66 A higher limit of
10 percent of the lesser of the investors annual income or net worth, not
218 J.M. CHEN
to exceed an amount sold of $100,000, applies if both the investors

annual income and net worth are equal to or more than $100,000.67
In addition to imposing income- and wealth-based restrictions on
investors, the JOBS Act also placed limits on crowdfunding transac-
tions. Qualifying equity crowdfunding transactions must be conducted
through either a traditionally registered broker-dealer or a new type of
entity called a funding portal.68 The JOBS Act created funding por-
tals as online entities that may serve as equity crowdfunding intermediar-
ies69 without registering as broker-dealers under the Securities Exchange
Act of 1934.70 All intermediaries, whether registered as broker-dealers or
qualified as exempted funding portals, must provide regulatory disclosures
to investors and potential investors in crowdfunding transactions.71
As a brand-new regulatory mechanism for private stock placements, the
JOBS Act and the SECs crowdfunding regulations have no track record.
The chief executive officer of an equity crowdfunding platform predicts
that crowdfunding of all types will raise $34 billion in 2016, an amount
that would surpass venture capital.72 What is all but certain is that the
companies covered by this new mechanism will be offering the lowest-
capitalized, least liquid, and most starkly skewed securities in the universe
of investment opportunities. In some respects, the income restrictions
under the crowdfunding rules are even more stringent than those adopted
by Regulation Ds definition of an accredited investor, in that $100,000in
annual income rather than $200,000 constitutes a meaningful regulatory
boundary. It takes very little prognosticatory confidence to predict that
crowdfunded securities, after a few years of interaction, will rank among
the most highly skewed investment opportunities.
Conglomerate underpricing and spin-off overpricing. Single-segment
firms are more highly skewed than highly diversified conglomerates oper-
ating in the same industry.73 Consequently, conglomerates trade at a dis-
count relative to their single-segment counterparts and offer higher average
returns in the long run.74
The flip side of the underpricing of conglomerates is the correlative
overpricing of corporate spin-offs. Meta-analyses of the deep literature on
the pricing of spin-offs have suggested that the issuance of stock in a spun-
off former subsidiary as a dividend to shareholders of the parent corpora-
tion generates immediate gains in shareholder wealth.75 Both spun-off
subsidiaries and their parents fare well, but the spin-offs outperform their
parents by a large margin.76 On average, the announcement of a spin-off
generates abnormal returns of roughly 3%.77 These gains are enough to
fuel a trading strategy focusing on spin-offs. A portfolio consisting solely

of spun-off subsidiaries outperforms a hypothetical Spin-Off Exchange-
Traded Fund, which in turn outperforms a broad value-weighted market
index.78
Over the long run, however, abnormal returns from spin-offs dissipate.79
One explanation is that the short-term gain from a spin-off announce-
ment is more a result of the anticipation of the spin-off than the actual
event.80 Another explanation, not necessarily exclusive of the behavioral
explanation, is that arbitrage eventually erases the temporary mispricing of
spin-offs.81 Notably, however, option prices on spin-offs raise the costs of
shorting to very high levels, all but eliminating opportunities for arbitrage
based on the overpricing of spin-offs.82
Stock options. The allure of a large potential payoff explains the prevalence
of stock options rather than stock as such in executive83 and nonexecutive84
compensation packages.85 This is merely a variation on the more com-
mon tendency of employees to eschew diversification of their retirement
accounts, to the point that some 401(k) accounts consist entirely of a stock
whose correlation with the investors labor income approaches 100%.86
Death and jackpot. Firms teetering on the verge of default, unsur-
prisingly, offer abnormally low returns.87 Controlling for size and value,
studies have shown that bankruptcy risk is not rewarded by higher subse-
quent returns.88 Firms suffering from extreme financial distress face starkly
asymmetrical incentives to take outsized risks. If an outlandish gamble
succeeds, the firm reaps enormous gains. If it fails, however, the firm has
the option to declare bankruptcy and abandon some or all of its debt.89
The heads-I-win, tails-you-lose game that bankruptcy law effectively
invites such firms to play affects investor welfare. Because highly distressed
firms also offer a higher probability of extremely large payoffs, relative
to financially healthier firms, firms threatening default excel in attracting
individual investors (predominantly nave noise traders), but not more
sophisticated institutional investors that presumably trade on information,
to a financial game of death and jackpot.90
Jackpots, indeed. Certain investors approach the stock market as if
finance resembled the lottery and other forms of overt gambling.91 One
study found that large lottery jackpots in Taiwan decreased trading volume
among stocks preferred by individual investors, especially in stocks with
highly skewed, lottery-like returns.92 Over-the-counter stocks are known
to have extremely negative average returns.93 Skewness preference provides
the only plausible explanation for the willingness of some investors to sink
220 J.M. CHEN
money in the pink sheets.94 At least in the movies, [p]lot lines within
the genre of gambling stories are not only simplethe gambler falls into
debt; the gambler goes after one big score to get even; the gambler either
does or does not get evenbut also offer some chance of redemption as
long as the filmmaker sees gambling as a charming hustle rather than
a catastrophic vice.95 By contrast, gambling in any number of real-life
venuescasinos, racetracks, state-sponsored lotteries, the pink sheets, one-
week fantasy sportshas an even simpler narrative: the house always wins.
The inescapable implication is that trading in stocks, at least for some
investors, is a thrilling form of gambling, interchangeable with other
forms of risk-seeking entertainment.96 Such thrill-seeking may even be
detectable through neural scans (functional magnetic resonance imaging)
of brain activity during trading.97 At the same time, cognitive bias evi-
dently keeps investors and gamblers from recognizing the rather inconve-
nient truth that the preference for positively skewed assets systematically
lowers returns on those assets, from stocks in general to initial public
offerings (IPOs).98 Financial assets are subject to the same psychology that
affects gambling assets such as lottery tickets and long-shot bets in horse
racing.99 At its most destructive, the thrill of gambling mixes with the fear
of regret to form an all-consuming fear of missing out on some fantastic
financial opportunity.100
Intriguingly, religious differences appear to affect the propensity to
gamble.101 Although Americans are presumably drawn to positively skewed
assets, as people around the world appear to be, the Protestant churches
harsh stance toward gambling suppresses that preference in certain regions
of the USA, while the Catholic Churchs more lenient view of gambling
makes it easier for people in Catholic regions to act on their prefer-
ence for skewness.102 It was a devout Catholic, after all, who framed the
choice to believe in God as a wager.103 In harmony with these conclusions,
a study of health and retirement choices found noticeable differences in
risk tolerance by religion.104 Protestants are the least risk tolerant,
and Jews the most.105 Blaise Pascals Catholics come about halfway [in]
between.106
9.2 INITIAL PUBLIC OFFERINGS

In any consideration of financial instruments with abnormally low but
highly skewed returns, the pricing of IPOs warrants especially close scru-
tiny.107 In the short run, IPOs appear to be underpriced.108 A 2002 study
concluded that issuers of IPOs on average leave $9.1 million in wealth on

the table after the first day of trading, as measured by the number of shares
available, times the difference between the first-day closing price and the
offer price.109 The typically sanguine reaction of issuers to so much forgone
wealth provides keen insight into prospect theory and, more generally,
into behavioral economics. Contentment at selling an article at one-third
of its subsequent value is a rare quality.110 Such tranquility appears to be
a psychological artifact. Prospect theory predicts that in most situations
occurring in the IPO market, issuers will sum the wealth loss from leaving
money on the table with the larger wealth gain on the retained shares from
a price jump.111 In technical terms, issuers are influenced by the covari-
ance of the amount of money and unanticipated wealth changes.112 In
simpler terms, it takes extraordinary emotional energy to get upset over
millions left on the table when an IPO has made you a billionaire.
The underpricing of IPOs also affects underwriters. The actions of
underwriters who leave money on the table can also be explained in
behavioral terms. Underpriced IPOs indirectly compensate underwrit-
ers in two ways. First, they allow underwriters to reduc[e] marketing
costs in finding buyers for new offerings.113 Second, underwriters recoup
some of their losses by steering investors to affiliated brokerages, who
extract frothy commissions from IPO purchasers, who rarely object to
paying inflated fees in exchange for access to the offering.114 In short,
the opportunity cost of leaving money on the table is less important to
underwriters than the direct fees.115
Not all IPOs are created alike. The distribution of IPOs is lopsided.
IPOs undergoing downward revisions average 4% returns on their first
day, while IPOs experiencing upward revisions return an average of 32%
on the first day.116 Underpriced IPOs tend to run together in hot issue
cycles lasting many months at a time, in which the average initial return
is much higher than at other times.117 [I]n the long run, however,
initial public offerings appear to be overpriced.118 The reason subsists in
investor psychology: IPO investors are irrationally over-optimistic about
the future potential of certain industries.119 Ultimately, notwithstanding
the initial underpricing of IPOs (at least relative to the offer price that
would have left none of the issuers money on the table), investors as a
class overpay for IPOs in the long run.120
As the nearest cousins to lotteries in highly liquid, publicly regulated stock
markets, initial public offerings exhibit long-run pricing anomalies consistent
with prospect theory.121 An economy containing investors reacting according
222 J.M. CHEN
to the predictions of prospect theory allows IPOs to be overpriced to earn

low average returns, because those investors value[] highly the chance,
even if it is very small, of a very large return.122 The higher the predicted
skewness of an IPO, the lower its expected long-run return.123 There are also
intriguing relationships between the pricing of IPOs, on the one hand, and
the even-numbered moments on the distribution of returns on IPOs, on the
other. Volatility is negatively related to subsequent returns on IPOs; kurtosis,
by contrast, is positively related.124 Of its own force, this observation violates
the basic psychophysics of the higher-moment CAPM, which presumes, in
simple terms, that investors like odd-numbered moments of the distribution
of returns and dislike even-numbered moments.125
There is some disagreement over the plausibility of this account of
IPOs underperformance over the long run. One source mildly disputes
the ability of skewness preference to provide a complete explanation for
the long-run underperformance of IPOs.126 Five-year returns to IPOs are
higher for venture-backed firms than for firms not backed by venture capi-
tal.127 Because small nonventure-backed IPOs are more likely to be held
by individuals, bouts of investor sentiment are a possible explanation for
these firms severe underperformance.128 Individuals are arguably more
likely to be influenced by fads or lack complete information.129 It is true
that returns on small IPOs covary with the discount on closed-end funds,
which is considered to be another benchmark of investor sentiment.130
Ultimately, however, this critique concludes that [u]nderperformance is
a characteristic of small, low book-to-market firms regardless of whether
they are IPO firms.131
On balance, IPO returns follow the fault lines of heterogeneity among
investors and financial intermediaries. IPOs with more informed investors
require higher returns to attract the interest of investors trading on the
basis of information rather than faddishness, emotional affinity, or unfil-
tered speculation.132 Prestigious underwriters are associated with lower
risk offerings, which leave less money on the table (to the benefit and
delight of issuers as well as underwriters) and therefore lower the returns
to IPO investors in the long run.133 These traits of IPOs demonstrate
that the investor preferences predicted by cumulative prospect theory
may be reconciled with the conventional CAPM134 and with first- and
second-order stochastic dominance.135 Notably, a broader study of finan-
cial professionals outside the IPO context has found that private bankers
and fund managers, though less risk averse than other agents studied in
the literature on behavioral finance, behave according to the predictions of
prospect theory: risk aversion for gains, risk seeking for losses.136
9.3 PROSPECT THEORY ANDBOWMANS PARADOX

Section 4.1 of this book discussed Bowmans paradox, a variation on
the theme of the low-volatility anomaly, viewed from the perspective of
accounting and strategic management. In the early 1980s, Edward Bowman
observed that high levels of managerial risk actually corresponded to low
rather than high returns.137 Two elements of prospect theory may explain
this reversal of the usual positive relationship between risk and return.
First, prospect theory and related work in behavioral decision theory have
suggested that individuals use target, or reference, points in evaluating
risky choices.138 Moreover, contrary to expected utility theory, individu-
als are not uniformly risk averse but adopt a mixture of risk-seeking and
risk-averse behaviors.139 Accordingly, when returns have been below tar-
get, most individuals are risk seeking.140 Conversely, when returns have
been above target, most [individuals] are risk averse.141
Proponents of prospect theory interpret Bowmans work as
confirm[ing] the hypothesis that companies behaviors may be similar
to those of individuals.142 Companies in distress, like troubled individu-
als, take larger risks.143 This comparison of corporate decision-making
with its individual equivalent support[s] the adaptation of risky choice
research to corporate decision making.144 Corporate managersthe
individuals who actually carry out decisions on behalf of companies
are not risk averters.145 Instead, under certain conditions, they are risk
seekers.146
According to prospect theory, the choice between risk-averse and risk-
seeking behavior pivots on a reference point.147 The reference point is a
critical element in the application of prospect theory to Bowmans para-
dox since [prospect] theory predicts that most individuals exhibit a mix-
ture of risk-seeking and risk-averting behavior when the outcome is either
below or above the reference point respectively.148 In finance and in cor-
porate strategy, the appropriate reference point is a target level of return.
Generally speaking, the target return level at which prospect theorys
fourfold pattern would predict a turn from risk-averse to risk-seeking
behavior is some sort of industry-wide average. Long before Bowman
identified his paradoxical relationship between risk and return, research
in accounting had already established the managerial desirability of
adjusting the firms financial ratios to predetermined targets which are
usually based on industry wide averages.149 Whether active attempts
by management or passive industry-wide effects ultimately motivate
firms to adjust financial ratios in a dynamic fashion in response to those
224 J.M. CHEN
industry-wide averages,150 some sort of industry-wide benchmark for

performance or return may serve as an appropriate proxy for a given
firms target level.151
Upon examining returns on equity across a range of industries, Avi
Fiegenbaum and Howard Thomas found that risk and return are neg-
atively correlated for below-target means and positively correlated for
above-target firms.152 Fiegenbaum and Thomas chose the median
return on equity (ROE) rather than the mean because the median
represents the middle point of the return distribution and is unaffected
by extreme outliers.153 (Given evidence that humans are more sensitive
to rankings and to relative rather than absolute outcomes,154 it is certainly
plausible to surmise that the median may better indicate managerial per-
ceptions of target performance than the mean.155) Studies of Bowmans
paradox in banking have likewise used median returns.156
The introduction of a target ROE value at either the firm or
industry level disrupts the riskreturn relationship, regardless of the
period or the underlying environmental conditions.157 In accordance
with the behavioral assumptions of prospect theory, Fiegenbaum and
Thomas found that most firms may be risk seeking when they are suf-
fering losses or are below targeted aspiration levels.158 Conversely, they
will tend to be risk averse following achievement of aspirations and tar-
gets.159 Working alone, Avi Fiegenbaum later quantified the extent to
which prospect theory correctly predicted that the tradeoff between risk
and return for the loss domain [would] be much steeper than for the gain
domain.160 He determined that the ratio of the slopes indicating these
differences on either side of target return was approximately 3:1.161 The
rough congruence between this ratio, the value of 2.25 for the coef-
ficient in cumulative prospect theorys value function,162 and the 4.8:1
ratio found in Peter Fishburn and Gary Kochenbergers two-piece utility
functions163 lends credence to Fiegenbaums claim that his ratio of 3 to
1 connects his work on Bowmans paradox to prospect theorys predic-
tion of a steeper association for risk-return relationships when returns
fail to meet their targets.164
Subsequent work has found a similar divide between above- and below-
target levels of risk among Belgian companies.165 Another study, however,
found mixed results in Australia.166 Although evidence on the connec-
tion between risk and return was less conclusive above the median than
below, the Australian study did find that decision-makers sharply compart-
mentalized their actions according to their relationship to a firm-specific
or industry-wide target.167 A more nuanced view of the reference point

suggests that the relevant benchmark in managerial decision-making may
rest above the industry median.168 That benchmark almost certainly varies
across time and across industries.169
The application of prospect theory to Bowmans paradox generates
many lessons for mathematical and behavioral finance. At a minimum, this
line of inquiry invites deeper probing of the relationships between mana-
gerial decision-making and corporate performance, between investment
risk and financial return, and between mathematical finance and strate-
gic management as competing scientific approaches. Among the simplest
and strongest findings concerns the nature of risk-taking that reverses the
expected relationship between risk and return. One of the most common
ways to elevate risk, and simultaneously to reduce expected return, is to
concentrate corporate capital.170 Concentration is tantamount to disre-
garding diversification and ignoring correlation among assets, so com-
monplace that it might be rightly regarded as humanitys modal financial
preference.171
Applications of prospect theory to Bowmans paradox have suggested
that predictions regarding the relationship between risk and return will
break down for firms with returns close to a target level.172 Fiegenbaum
and Thomas found support [for] the hypothesis that there is no sig-
nificant correlation between risk and return around the target level.173
To the extent that markets come close to normality, many firms should
find themselves near industry-wide medians for target returns, in absolute
terms if not in terms of rank. Absent extreme departures at either end
of the distribution, perhaps something more characteristic of a mature
industry free from technological disruption, mean return may even be
modal. Moreover, reversion to the mean, something to be expected in the
presence of robust if not perfect competition, should realign managerial
incentives into the vast middle ground where neither the conventional
riskreturn relationship nor Bowmans paradoxical reversal strongly deter-
mines risk-taking incentives.
The most provocative inferences to be drawn from the absence of clear
evidence on riskreturn relationships in ordinary managerial situations
involve not just Bowmans paradox or the closely related low-volatility
anomaly. If the behavioral account of strategic management holds true
at the edges of the distribution and is inconclusive in circumstances
approaching the mean, median, or mode, then Bowmans paradox at least
partially confirms the foundational finding of contemporary financethat
226 J.M. CHEN
beta, the standard measure of systematic risk in the conventional CAPM,

has no explanatory power in the cross-section of stock market returns.174
The opening chapters of Postmodern Portfolio Theory175 and 12.2 of
this book provide a more comprehensive introduction to the conven-
tional CAPM and its eventual displacement.176 For current purposes, it
suffices to identify the three factors that have supplemented (or even sup-
planted) beta within the FamaFrenchCarhart four-factor model: value
(high book-to-market ratio), small firm size, and momentum in returns.
Efforts to explain Bowmans paradox as a manifestation of prospect theory
have identified at least one mechanism by which each of these additional
factorsvalue, size, momentummight meaningfully disrupt the tradi-
tional relationship between risk and return, especially in circumstances
where the broad psychophysics of economic decision-making do not pre-
dispose corporate managers either to avoid or to seek risk.
Value. The very existence of a subcategory of value stocks gave rise
to both the low-volatility anomaly and Bowmans paradox. As noted in
4.1, the tendency of value stocks [to] earn higher expected returns
than growth stocks presents a troublesome anomaly that undermines
the rational expectations underlying neoclassical economics, modern
portfolio theory, and conventional managerial theories.177
Size. In their original article linking Bowmans paradox to prospect
theory, Fiegenbaum and Thomas speculated that X-inefficiency may
impede large firms from achieving high-return-low risk profiles.178
X-inefficiency, defined as the shortfall in efficiency actually achieved by
a firm, relative to its fully efficient potential,179 may be regarded as the
downside corollary to the usual benefits of large firm size: economies of
scale, economies of scope, and positive, intrafirm spillover effects.180 These
traits, so essential in industries as diverse as pharmaceuticals,181 banking,182
and higher education,183 arguably impede managerial acumen or initiative.
Mathematical finance traditionally associates large firm size with abnor-
mally low returns. A glimpse of the reason for financial performance by
large firms, relative to the market portfolio, emerges from studies of small
firm success. Small firms, as it turns out, may succeed either by exploiting
efficiency or by exploiting flexibility, but not both.184
Momentum. The most contested of the four factors in the FamaFrench
Carhart model, momentum warrants a closer look in 12.2 and 12.3 of
this book. Fiegenbaum and Thomas hinted at the presence of momentum
effects in their study of Bowmans paradox, by suggesting that certain firms
may succumb to organizational inertia despite changed circumstances.185
Studies of the escalation of commitment to losing strategies, especially

those that connect this phenomenon to prospect theory,186 reinforce the
suggestion that high-volatility, low-return strategies, once adopted, may
continue to hold sway even in the face of demonstrated failure.
In concert, these limitations on the power of prospect theory to explain
Bowmans paradox highlight the role of idiosyncratic, firm-specific factors
distinct from either systematic risk or the presumptive heuristics of mana-
gerial behavior. Edward Bowman himself suggested that the paradoxical
reversal of the riskreturn relationship might be best explained through
industry-specific characteristics, whether economic (service versus man-
ufacturing businesses) or legal (the degree and nature of regulation).187
Industry-specific studies promise more detailed, sophisticated views of the
relationship between risk and return.188 For instance, one study of family-
controlled olive oil mills in Spain has revealed an intricate balance between
the business security promised by membership in a cooperative and the
emotional and entrepreneurial freedom inherent in remaining indepen-
dent.189 While family firms may avoid venturing risks, they may be willing
to incur the risk of greater performance hazard in order to preserve [the]
socioemotional wealth arising from independent ownership of a mill.190
Closer examination of Bowmans paradox as a behaviorally mediated
phenomenon demonstrates that the relationship of return to risk is not
merely an intractable artifact of ill-behaved markets. The strategic man-
agement literature more quickly and cohesively recognized the breakdown
in the conventional riskreturn relationship, relative to economics and
financial economics studies that assumed risk averse behavior and
consequently overlooked the mixture of risk averse/risk-taker behavior
in corporate practice.191 Indeed, evidence supporting Bowmans paradox
is stronger than evidence supporting the low-volatility anomaly in the
sense that negative-risk-return associations are more common when
measures are accounting-based rather than market-based.192
The strategic management literatures preference for accounting-based
over market-based data, arising in part from the assumption that account-
ing data more directly reflects managerial control,193 raises the question of
whose risk is important.194 Specifically, [i]s a manager acting as an agent
for a firms principal owners, or is a managers appropriate role to enhance
the strategic management capability of the organization?195 If managers
are the only relevant decision makers in a firm,196 then Bowmans para-
dox may stand wholly apart from the low-volatility anomaly. On the other
hand, if shareholders are regarded as making at least some contribution
228 J.M. CHEN
to corporate decision-making, then examinations of Bowmans paradox

should also consider the implications of capital market theory and the
other aspects of mathematical finance underlying the low-volatility anom-
aly.197 Prospect theory bridges not only Bowmans paradox and the low-vol-
atility anomaly, but also strategic management and mathematical finance.
9.4 PROSPECT THEORY ANDTHEEQUITY PREMIUM

PUZZLE: MYOPIC LOSS AVERSION
Prospect theory has also inspired multiple efforts to solve the equity pre-
mium puzzle. Perhaps the most celebrated of these is Shlomo Benartzi
and Richard Thalers argument that the equity risk premium arises from
myopic loss aversion, a combination of loss aversion and a short eval-
uation period that contribute[s] to an investor being unwilling to bear
the risks associated with holding equities.198 Beginning with the baseline
of a choice between a risky asset that pays an expected 7 percent per
year with a standard deviation of 20 percent (like stocks) and a safe asset
that pays a sure 1 percent, Benartzi and Thaler ask how often an inves-
tor with preferences consistent with cumulative prospect theory would
have to evaluate his portfolio in order to be indifferent between the his-
torical distribution of returns on stocks and bonds.199 Their answers to
that question fall in the neighborhood of one year, clearly a plausible
result.200 And given a one-year evaluation period, they conclude that
close to a 5050 split between stocks and bonds represents the opti-
mal asset allocation for such an investor.201
The intuition underlying this explanation of the equity risk premium as
a product of myopic risk aversion may be expressed in the terms that an
ordinary, individual investor uses to evaluate stock market risk: Well
stocks could go up over the next year ; but they could also go down,
and since Im more sensitive to losses than to gains, that would be really
painful.202 The aspect[] of prospect theory that serves as the main
determinant of the equity risk premium that such an investor would
demand is loss aversion.203 The specific functional forms of the value
function and weighting functions are not critical.204 The equity risk pre-
mium is equivalent to the psychic cost stemming from an investors
use of an evaluation period that is shorter than his or her true investment
horizon (such as the length of time before retirement).205 At a 20-year
horizon, 5.1 percent is the price of excessive vigilanceand an estimate
of the premium demanded to bear the risk inhering in equities.206
One interesting implication of the myopic loss aversion hypothesis is

this models guidance on retirement savings. Conventional financial advice
embraces the strong intuition that a rational risk-averse investor would
decrease the proportion of his assets in stocks as he nears retirement.207
The intuition stems from the assumption that a longer time horizon
pushes the probability that the return on stocks will exceed the return on
bonds toward 1, whereas stock investments may experience substantial
shortfalls over short[er] horizons.208 This intuition holds [i]f stock
returns are mean reverting, but not if the returns on stocks and bonds
[follow] a random walk.209 Adopting the latter, more realistic assumption
leads the myopic loss aversion model to a result that most investors find
extremely counterintuitive: [A]n investor who wants mostly stocks in
his portfolio at age 35 should still want the same allocation at age 64.210
Such advice is not only counterintuitive; it squarely contradicts the logic
underlying target-age retirement funds and other mechanisms for reduc-
ing portfolio volatility as an individual investor ages.211
9.5 ANOTHER EQUITY PREMIUM SOLUTION: PROSPECT

THEORY ANDASSET PRICING.
Nicholas Barberis, Ming Huang, and Tano Santos use prospect theory to
devise a more elaborate solution to the equity premium puzzle.212 They
start with the premise that the investor derives direct utility not only
from consumption but also from changes in the value of his wealth.213
This assumption stands in contrast with the traditional approach to asset
pricing, which focuses exclusively on the future consumption utility
of investment-based wealth.214 This bifurcated notion of investor utility is
drawn from a more general optimal expectations framework in which
the agent derives utility not only from [immediate] consumption,
but also from anticipating future consumption.215 The investor must
trade[] off the benefit of holding pleasurable beliefs about a speculative
future against the cost that she or he would incur if indulging those beliefs
would spur unwise actions.216
The twofold nature of investor utility within this optimal expectations
framework captures two intuitions.217 First, the investor is much more
sensitive to reductions in financial wealth than to increases.218 This
observation is consistent with prospect theory in general and with the
Benartzi and Thalers model of myopic loss aversion in particular. Second,
and of greater importance, loss aversion depends on prior investment
230 J.M. CHEN
performance.219 On the one hand, prior gains reduce loss aversion; they
cushion any subsequent loss, making it more bearable both financially
and psychologically. But prior losses increase loss aversion: after being
burned by the initial loss, an investor becomes more sensitive to addi-
tional setbacks.220 These reactions combine the predictions of prospect
theory with those of the house money effect, which holds that people
faced with sequential gambles are more willing to take risk if they
made money on prior gambles, than if they lost.221 Put simply, gamblers
are more willing to bet when they are ahead because losses are less pain-
ful if they [follow] prior gains, and more painful if they follow prior
losses.222
Quite evidently, this models consideration of past investment out-
comes is congruent with time series analysis and with the intertemporal
CAPM, both of which reject a static, single-period approach to returns
and volatility.223 In addition, this model is entirely in line with prospect
theory. In their original article on prospect theory, Daniel Kahneman and
Amos Tversky invoked the well known observation that the tendency to
bet on long shots increases in the course of the betting day.224 They used
this insight to defend their hypothesis that a failure to adapt to losses or
to attain an expected gain induces risk seeking.225
Barberis, Huang, and Santos address Kahneman and Tverskys con-
cern that prospect theory was originally developed only for elemen-
tary, one-shot gambles and that any application of prospect theory to
a dynamic context demanded further evidence on how people think
about sequences of gains and losses.226 Ignoring the psychological reac-
tion to sequences of investment outcomes is tantamount to assuming that
investors aggregate the totality of their stock market gains and losses
over vast periods of time, in contradiction of prospect theorys emphasis
on changes in wealth over absolute levels of wealth.227
The value of Barberis, Huang, and Santoss second factor becomes evi-
dent as their model confronts otherwise perplexing features of the tra-
ditional asset pricing framework, including the equity risk premium, the
otherwise bafflingly low risk-free rate, volatility clustering, return momen-
tum, and the low-volatility anomaly.228 Allowing risk aversion to change as
a dynamic function of investment outcomes allows returns to be much
more volatile than the underlying dividends.229 Although an unusually
good dividend raises prices, the resulting price increase also makes the
investor less risk averse and drives prices even higher.230 As a result, sub-
sequent returns are on average lower. At the same time, this model
produces a substantial equity premium: high volatility often gener-

ates poor performance and inflicts considerable discomfort on loss-
averse investors, who in turn demand a large premium to hold stocks.231
The inclusion of behavioral responses to prior investment outcomes
propels prospect theorys response to the equity premium puzzle beyond
the limitations of an equilibrium model based exclusively on myopic
loss aversion.232 [L]oss aversion cannot by itself explain the equity pre-
mium; incorporating the effect of prior outcomes is a critical ingredient
as well.233 Ignoring prior outcomes renders risk aversion constant over
time and deprives stock prices [of] an important source of volatility.234
Lowering volatility and the consequent reduction of risk make it impos-
sible to produce a substantial equity premium.235 The contribution of
loss aversion to cracking the equity premium puzzle must coincide with
a mechanism that makes stock returns more volatile than underlying cash
flows.236 Observed stock market fluctuations are not enough to scare
even a loss-averse investor into demanding a high equity premium.237
Consistent with this analysis of the equity risk premium through loss
aversion and prior investment outcomes, investors appear to be more
loss averse over fluctuations in individual stocks that they own.238 What
Nicholas Barberis and Ming Huang have called individual stock account-
ing describes loss aversion over individual stock fluctuations.239 The
painfulness of a loss on a particular stock depends on that stocks prior
performance.240 Historically, the impact of the performance of individual
stocks on loss aversion was greater because individuals with holdings in
the stock market typically owned only a very small number of stocks
before the advent of mutual funds.241 Despite the more recent ascen-
dancy of exchange-traded funds and the correspondingly greater ease with
which individual investors can build diversified portfolios, home bias242
and the more general failure to diversify suggest that individual investors
portfolios remain disproportionately concentrated.
Indeed, the failure to diversify is a manifestation of skewness prefer-
ence and the innate longing for lottery-like payouts.243 Consequently,
individual stock accounting may be primarily an artifact of the highly
individualized portfolios held by many investors, rather than the result
of excessively narrow framing.244 By contrast, narrow framing would
explain individual stock accounting if it can be shown that investors evalu-
ate investing opportunities in isolation, separately from other risks245
and consequently become unreasonably averse to a small, independent
gamble, even when the gamble is actuarially favorable.246
232 J.M. CHEN
To the extent that frames of mind do elevate the premium demanded

by equity investors, such framing can be reset. In their original presenta-
tion of prospect theory, Kahneman and Tversky recognized that reformu-
lating a problem in terms of final assets rather than in terms of gains
and losses, contrary to the instinctive human framing of risky or uncer-
tain choice, effectively resets the reference point to zero.247 Such
reframing is likely to render the value function concave everywhere and
essentially eliminates risk seeking except for gambling with low prob-
abilities.248 If the goal of private financial planning or public policy is to
eliminat[e] risk seeking in the domain of losses, there may be no more
effective procedure than to formulate decision problems [explicitly] in
terms of final assets.249 Against the backdrop of behaviorally influenced
decision-making, of course, such rationality comes at an emotional price.
Resetting the decisional baseline by forcing an investor to confront and
accept [a] loss after a market decline is the part that is painful.250
NOTES
1. Barberis, Thirty Years of Prospect Theory, supra note 1 (Chapter 8), at 173
174 (emphasis in original).
2. Guthrie, supra note 45 (Chapter 8), at 1163.
3. See Yan Li & Liyan Yang, Prospect Theory, the Disposition Effect, and Asset
Prices, 107J.Fin. Econ. 715739 (2013).
4. See Wilbur G. Lewellen, Gary E. Schlarbaum & Ronald C. Lease, The
Individual Investor: Attributes and Attitudes, 29 J. Fin. 413433, 425
(1974) (table 4). See generally Wilbur G.Lewellen, Ronald C.Lease & Gary
E. Schlarbaum, Patterns of Investment Strategy and Behavior Among
Individual Investors, 50J.Bus. 296333 (1977).
5. See, e.g., Hal R. Arkes, Lisa Tandy Herren & Alice M. Isen, The Role of
Potential Loss in the Influence of Affect on Risk-Taking Behavior, 42 Org.
Behav. & Human Decision Processes 181193 (1988); Gordon H.Bower,
Mood and Memory, 36 Am. Psychologist 129148 (1981); William
F. Wright & Gordon H. Bower, Mood Effects on Subjective Probability
Assessment, 52 Org. Behav. & Human Decision Processes 276291 (1992).
6. See Goetzmann & Kumar, supra note 130 (Chapter 1); Statman,
Diversification Puzzle, supra note 191 (Chapter 7).
7. See Manju Puri & David T. Robinson, Optimism and Economic Choice,
86J.Fin. Econ. 7199 (2007).
8. See supra 1.6, at 1417.
9. See Werner F.M. de Bondt & Richard H. Thaler, Does the Stock Market
Overreact?, 40J.Fin. 793805 (1985).
10. See Brad M. Barber & Terrance Odean, Boys Will Be Boys: Gender,
Overconfidence, and Common Stock Investment, 116 Q.J. Econ. 261292
(2001); Terrance Odean, Volume, Volatility, Price, and Profit When All
Traders Are Above Average, 53J.Fin. 18871934 (1998).
11. See Andrew B.Abel, An Exploration of the Effects of Pessimism and Doubt on
Asset Returns, 26J.Econ. Dynamics & Control 10751092 (2002).
12. See Tomasz Zaleskiewicz, Agata Gasiorowska, Pelin Kesebir, Aleksandra
Luszczynska & Tom Pyszczynski, Money and the Fear of Death: The Symbolic
Power of Money as an Existential Anxiety Buffer, 36J.Econ. Psych. 5567
(2013).
13. See Wei-Yin Hu & Jason S. Scott, Behavioral Obstacles in the Annuity
Market, 63:6 Fin. Analysts J. 7182 (Nov./Dec. 2007).
14. See Daniel R. Smith, Conditional Coskewness and Asset Pricing,
14J.Empirical Fin. 91119, 103 (2007) ([I]nvestors are more interested
in positive skewness than they are averse to negative skewness).
15. See Tang & Shum, Unsystematic Risk, Skewness and Stock Returns, supra
note 34 (Chapter 7).
16. See Bali, Cakici & Whitelaw, supra note 66 (Chapter 3); Barberis & Huang,
Stocks as Lotteries, supra note 117 (Chapter 8), at 20802082.
17. See Brian Boyer, Todd Mitton & Keith Vorkink, Expected Idiosyncratic
Skewness, 23 Rev. Fin. Stud. 169202 (2010).
18. Barberis & Huang, Stocks as Lotteries, supra note 117 (Chapter 8), at 2071.
19. Cf. Tversky & Kahneman, supra note 58 (Chapter 7), at 312.
20. Barberis & Huang, Stocks as Lotteries, supra note 117 (Chapter 8), at 2071.
The following image was generated at http://www.wolframalpha.com/inp
ut/?i=plot+x%5E.4%2F%28x%5E.4%2B%281-x%29%5E.4%29%5E%281%2
F.4%29+and+x%5E.65%2F%28x%5E.65%2B%281-x%29%5E.65%29%5E%2
81%2F.65%29+and+x+for+x%3D0+to+1.
21. See, e.g., Camerer & Ho, supra note 153 (Chapter 7); Tversky & Kahneman,
Advances in Prospect Theory, supra note 58 (Chapter 8); Wu & Gonzalez,
supra note 58 (Chapter 8).
22. See Tennessee Williams, The Catastrophe of Success, in N.Y.Times, Nov. 30,
1947, reprinted in Tennessee Williams, The Glass Menagerie 99103
(Robert Bray introd., 1999) (1st ed. 1945).
23. See, e.g., Statman, Diversification Puzzle, supra note 191 (Chapter 7).
24. See Thierry Post & Haim Levy, Does Risk Seeking Drive Stock Prices? A
Stochastic Dominance Analysis of Aggregate Investor Preferences and Beliefs,
18 Rev. Fin. Stud. 925853 (2005).
25. See Hamm & Shettleworth, supra note 29 (Chapter 8).
26. See Carla H.Lagorio & Timothy D.Hackenberg, Risky Choice in Pigeons:
Preference for Amount Variability Using a Token-Reinforcement System,
98J.Experimental Analysis Behav. 139154 (2012).
234 J.M. CHEN
27. See Thomas R.Zentall & Jessica P.Stagner, Maladaptive Choice Behavior by
Pigeons: An Animal Analogue and Possible Mechanism for Gambling (Sub-
Optimal Human Decision-Making Behavior), 278 Proc. Royal Socy Biol.
Scis. 12031208 (2011).
28. See North American Association of State and Provincial Lotteries (data com-
piled at http://www.naspl.org/index.cfm?fuseaction=content&menuid=1
7&pageid=1025) (hereinafter National Lottery Data).
29. Derek Thompson, Lotteries: Americas $70 Billion Shame, The Atlantic,
May 11, 2015 (available at http://www.theatlantic.com/business/archive/
2015/05/lotteries-americas-70-billion-shame/392870).
30. See National Lottery Data, supra note 28.
31. See La Fleurs 2015 World Lottery Almanac (23d ed. 2015).
32. See Charles T.Clotfelter, Philip J.Cook, Julie A.Edell & Marian Moore,
State Lotteries at the Turn of the Century: Report to the National Gambling
Impact Study Commission 13 (April 23, 1999) (reporting that lottery
expenditures represent a much larger burden on the household budget for
those with low incomes than for those with high incomes) (available at
http://govinfo.library.unt.edu/ngisc/reports/lotfinal.pdf).
33. See Tina Rosenberg, Playing the Odds on Savings, N.Y.Times, Jan. 15, 2014
(available at http://opinionator.blogs.nytimes.com/2014/01/15/playing-
the-odds-on-saving).
34. See, e.g., Scott Hankings, Mark Hoekstra & Paige Marta Skiba, The Ticket to
Easy Street? The Financial Consequences of Winning the Lottery, 93 Rev. Econ.
& Stat. 961969 (2011) (concluding that lottery prizes merely postpone
rather than prevent bankruptcy); Guido W. Imbens, Donald B. Rubin &
Bruce I. Sacerdote, Estimating the Effect of Unearned Income on Labor
Earnings, Savings, and Consumption: Evidence from a Survey of Lottery Players,
91 Am. Econ. Rev. 778794 (2001) (finding that unearned income from lot-
teries reduces labor earnings, especially for players between 55 and 65 years
old, and that lottery winners saved approximately 16% of their proceeds).
35. See Mauro F.Guilln, Roberto Garva & Andrs Santana, Embedded Play:
Economic and Social Motivations for Sharing Lottery Tickets, 28 Eur. Sociol.
Rev. 344354, 346347 (2012).
36. See generally Melissa Schettini Kearney, Peter Tufano, Janathan Guryan &
Erik Hurst, Making Savers Winners: An Overview of Prize-Linked Savings
Products, in Financial Literacy: Implications for Retirement Security and the
Financial Marketplace 218240 (Olivia S. Mitchell & Annamaria Lusardi
eds., 2011) (originally published as National Bureau of Economic Research
[NBER] Working Paper 16433 [Oct. 2010] and available at http://www.
nber.org/papers/w16433).
37. See Mauro F.Guilln & Adrian E.Tschoegl, Banking on Gambling: Banks and
Lottery-Linked Deposit Accounts, 21J.Fin. Servs. Research 219231 (2002).
38. See, e.g., Bjarne Florentsen & Kristian Rydqvist, Ex-Day Behavior When
Investors and Professional Traders Assume Reverse Roles: The Case of Danish
Lottery Bonds, 11J.Fin. Intermediation 152175 (2002); Richard C.Green
& Kristian Rydqvist, Ex-Day Behavior with Dividend Preference and
Limitations to Short-Term Arbitrage: The Case of Swedish Lottery Bonds,
53J.Fin. Econ. 145187 (1999).
39. See Charles W. Calomiris, The Motives of U.S. Debt-Management Policy,
17901880: Efficient Discrimination and Time Consistency, 13 Research in
Econ. Hist. 67105 (1991); Robert M. Jennings, Donald F. Swanson &
Andrew P.Trout, Alexander Hamiltons Tontine Proposal, 45 Wm. & Mary
Q. 107115 (1988).
40. See Henri Lvy-Ullmann, Lottery Bonds in France and in the Principal
Countries of Europe, 9 Harv. L.Rev. 386405 (1896).
41. Anne L.Murphy, Lotteries in the 1690s: Investment or Gamble?, 12 Fin. Hist.
Rev. 227245 (2005) (describing the Million Adventure as an offering of
100,000 tickets at 10 each, with 2.5 % (or 2500 tickets) paying prizes
ranging from 10 per year to 1,000 per year for 16 years); accord Kearney,
Tufano, Guryan & Hurst, supra note 36, at 223.
42. See Todd Mitton & Keith Vorkink, Equilibrium Underdiversification and
Preference for Skewness, 20 Rev. Fin. Stud. 12551288 (2007); cf. Campbell,
Household Finance, supra note 68 (Chapter 6), at 1564 (observing that
private business assets take the place of public equity holdings in many
wealthy households); William M. Gentry & Glenn R. Hubbard,
Entrepreneurship and Household Saving, 4:1 Advances in Econ. Analysis &
Poly, art. 8 (2004) (electronic copy at 8) (available at https://www0.gsb.
columbia.edu/faculty/ghubbard/Articles%20for%20Web%20Site/
Entrepreneurship%20and%20Household%20Saving.pdf).
43. See Tobias J. Moskowitz & Annette Vissing-Jrgensen, The Returns to
Entrepreneurial Investment: A Private Equity Premium Puzzle?, 92 Am.
Econ. Rev. 745778, 745746 (2002).
44. Id. at 745.
45. Id.
46. John Heaton & Deborah J. Lucas, Portfolio Choice and Asset Prices: The
Importance of Entrepreneurial Risk, 55J.Fin. 11631199, 1163 (2000).
47. Moskowitz & Vissing-Jrgensen, supra note 43, at 745.
48. Heaton & Lucas, supra note 46, at 1168.
49. See generally, e.g., C. Steven Bradford, Crowdfunding and the Federal
Securities Laws, 2012 Colum Bus. L.Rev. 1150; Paul Belleflamme, Thomas
Lambert & Armin Schwienbacher, Crowdfunding: Tapping the Right
Crowd, 29J.Bus. Venturing 589609 (2014); Ajay K.Agrawal, Christian
Catalani & Avi Goldfarb, Some Simple Economics of Crowdfunding (June
2013) (NBER Working Paper No. 19133) (available at http://www.nber.
236 J.M. CHEN
org/papers/w19133). The popular literature on crowdfunding is deep.

The title of Gary Spirer, Crowdfunding: The Next Big Thing (2014), con-
veys the sense of anticipation that characterizes that body of work.
50. Regulation D, Limited Offer and Sale of Securities Without Registration
Under the Securities Act of 1933, 17 C.F.R. 230.500230.508.
51. Securities and Exchange Commission, Crowdfunding, 80 Fed. Reg.
71,38871,551, 71,389 (Nov. 16, 2015) (final rule to be codified at 17
C.F.R. parts 200, 227, 232, 239, 240, 249, 269, and 274) (hereinafter
Final Crowdfunding Rule).
52. 17 C.F.R. 230.501(a); see also Securities and Exchange Commission,
Office of Investor Education and Advocacy, Investor Bulletin: Accredited
Investors (available at http://www.investor.gov/sites/default/files/ib_
accreditedinvestors.pdf).
53. Final Crowdfunding Rule, supra note 51, at 71,389.
54. 17 C.F.R. 230.501(a)(5).
55. Id. 230.501(a)(6).
56. See Devin Thorpe, SEC Mulls Changes to Accredited Investor Standards, 18
Crowdfunders React, Forbes, July 15, 2014 (available at http://www.forbes.
com/sites/devinthorpe/2014/07/15/sec-mulls-changes-to-accredited-
investor-standards-18-crowdfunders-react) (crediting SeedInvest, an equity
crowdfunding site, with that information).
57. Official estimates of the US population are available at http://quickfacts.
census.gov.
58. Pub. L.No. 112106, 126 Stat. 306327.
59. Final Crowdfunding Rule, supra note 51, at 71,388.
60. Id.
61. Id. at 71,38871,389.
62. See id. at 71,53771.538 (outlining a new regulatory provision to be codi-
fied at 17 C.F.R. 27.100).
63. See 15 U.S.C. 77d(a)(6) (codifying a new 4(a)(6) of the Securities Act of
1933).
64. See 15 U.S.C. 77e.
65. See 17 C.F.R. 227.100(a)(1).
66. Id. 227.100(a)(2)(i).
67. Id. 227.100(a)(2)(ii).
68. Final Crowdfunding Rule, supra note 51, at 71,389; accord 17 C.F.R.
227.100(a)(3).
69. See 15 U.S.C. 78c(h) (codifying a new 3(h) of the Exchange Act of
1934).
70. See 15 U.S.C. 78o(a)(1) (also known as 15(a)(1) of the Exchange Act).
71. See 15 U.S.C. 77d-1 (codifying a new 4A of the Securities Act); 17
C.F.R. 227.201.
72. See Chance Barnett, SEC Approves Title III of JOBS Act, Equity Crowdfunding
with Non-Accredited, Forbes, Oct. 30, 2015 (available at http://onforb.
es/1WlWtM4).
73. See Philip G. Berger & Eli Ofek, Diversifications Effect on Firm Value,
37J.Fin. Econ. 3965 (1995); Larry H.P.Lang & Ren M.Stulz, Tobins
Q, Corporate Diversification, and Firm Performance, 102 J. Pol. Econ.
12481280 (1994).
74. See Todd Mitton & Keith Vorkink, Why Do Firms with Diversification
Discounts Have Higher Expected Returns?, 45 J. Fin. & Quant. Analysis
13671390 (2010).
75. See Thomas M.Fischer, Christian W.Mller & Annette Ober, Wertsteigerung
Durch Sell-Offs: Eine Meta-Analyse, 63 J. fr Betriebswirtschaft 157205
(2013); John J. McConnell, Steven E. Sibley & Wei Xu, The Stock Price
Performance of Spin-Off Subsidiaries, Their Parents, and the Spin-Off ETF,
20012013, 42:1J.Portfolio Mgmt. 143152 (Fall 2015).
76. See McConnell, Sibley & Wu, supra note 75.
77. See Chris Veld & Yulia V.Veld-Merkoulova, Value Creation Through Spin-
Offs: A Review of the Empirical Evidence, 11 Intl J.Mgmt. Revs. 407420
(2009).
78. See McConnell, Sibley & Wu, supra note 75.
79. See Veld & Veld-Merkoulova, supra note 77.
80. Stephen Davidoff Solomon, A Test to See If the Parts Are Worth More Than
the Whole, N.Y.Times, Jan. 18, 2011.
81. See generally, e.g., Mark B.Garman & James A.Ohlson, Valuation of Risky
Assets in Arbitrage Free Economies with Transaction Costs, 9 J. Fin. Econ.
271280 (1981); Andrei Shleifer & Robert W. Vishny, The Limits of
Arbitrage, 52 J. Fin. 3556 (1997); Bruce Tuckman & Jean-Luc Vila,
Arbitrage with Holding Costs: A Utility Based Approach, 47J.Fin. 1283
1302 (1992).
82. See Owen A. Lamont & Richard H. Thaler, Can the Market Add and
Subtract? Mispricing in Tech Stock Carve-Outs, NBER Working Paper No.
8302 (May 2011) (available at http://www.nber.org/papers/w8302.pdf).
83. See Ingolf Dittman, Ernst Maug & Oliver G. Spalt, Sticks or Carrots?
Optimal CEO Compensation When Managers Are Risk Averse, 65 J. Fin.
20152050 (2010).
84. See Oliver G. Spalt, Probability Weighting and Employee Stock Options,
85. See generally Brian H. Boyer & Keith Vorkink, Stock Options as Lotteries,
69J.Fin. 14851527 (2014).
86. See supra 7.8, at 155156.
87. See John Y.Campbell, Jens Hischer & Jan Szilagyi, In Search of Distress Risk,
63J.Fin. 28992939 (2008).
238 J.M. CHEN
88. See Ilia D.Dichev, Is the Risk of Bankruptcy a Systematic Risk?, 53J.Fin.
11311147 (1998).
89. See, e.g., George A. Akerlof & Paul M. Romer, Looting: The Economic
Underworld of Bankruptcy for Profit, 2 Brookings Papers on Econ. Activity
160 (1993); Michelle J. White, The Corporate Bankruptcy Decision,
3J.Econ. Persp. 129151 (1989).
90. See Jennifer Conrad, Nishad Kapadia & Yuhang Xing, Death and Jackpot:
Why Do Individual Investors Hold Overpriced Stocks?, 113J.Fin. Econ. 455
475 (2014).
91. See Meir Statman, Lottery Players/Stock Traders, 58:1 Fin. Analysts J. 1421
(Jan./Feb. 2002).
92. See Xiaohui Gao Bakshi & Tse-Chun Lin, Do Individual Investors Treat
Trading as a Fun and Exciting Gambling Activity? Evidence from Repeated
Natural Experiments, Rev. Fin. Stud. (forthcoming 2015) (available at
http://ssrn.com/abstract=1622184).
93. See Bjrn Eraker & Mark J. Ready, Do Investors Overpay for Stocks with
Lottery-Like Payoffs? An Examination of the Returns on OTC Stocks,
115J.Fin. Econ. 486504 (2011).
94. See id.
95. Jay Caspian Kang, The Perfect Predictability of Gambling Movies, N.Y.Times
Mag. (Oct. 5, 2015) (available at http://nyti.ms/1j109Ro).
96. See generally Nicholas Barberis, A Model of Casino Gambling, 58 Mgmt. Sci.
3551 (2012); Daniel Dorn & Paul Sengmueller, Trading as Entertainment?,
55 Mgmt. Sci. 591603 (2009); Mark Grinblatt & Matti Keloharju,
Sensation Seeking, Overconfidence, and Trading Activity, 64J.Fin. 549578
(2009); Hartley & Farrell, supra note 66 (Chapter 6); Alok Kumar, Who
Gambles in the Stock Market?, 64J.Fin. 18891933 (2009).
97. See Cary Frydman, Nicholas Barberis, Colin Camerer, Peter Bossaerts &
Antonio Rangel, Using Neural Data to Test a Theory of Investor Behavior:
An Application to Realization Utility, 69J.Fin. 907946 (2014).
98. Markus K.Brunnermeier, Christian Gollier & Jonathan A.Parker, Optimal
Beliefs, Asset Prices, and the Preference for Skewed Returns, 97 Am. Econ.
Rev. 159165, 159 & n.2 (2007).
99. See id.; cf. Eric Snowberg & Justin Wolfers, Explaining the Favorite-Long Shot
Bias: Is It Risk-Love or Misperception?, 118J.Pol. Econ. 723746 (2010).
100. Compare Andrew K. Przybylski, Kou, Murayama, Cody R. DeHaan &
Valerie Gladwell, Motivational, Emotional, and Behavioral Correlates of
Fear of Missing Out, 29 Computers in Human Behav. 18441848 (2013)
with John Grable, Ruth Lytton & Barbara ONeill, Projection Bias and
Financial Risk Tolerance, 5J.Behav. Fin. 142147 (2004) (exploring how
projection bias and regret theory may shape individuals tolerance and taste
for financial risk).
101. See Alok Kumar, Jeremy K. Page & Oliver G. Spalt, Religious Beliefs,
Gambling Attitudes, and Financial Market Outcomes, 102 J. Fin. Econ.
671708 (2011); cf. Walid Mansour & Mouma Jlassi, The Effect of Religion
on Financial and Investing Decisions, in Investor Behavior: The Psychology
of Financial Planning and Investing, supra note 158 (Chapter 1), at
135152.
102. Nicholas Barberis, The Psychology of Tail Events: Progress and Challenges, 103
Am. Econ. Rev. 611616, 615 (2013).
103. See generally Nicholas Rescher, Pascals Wager: A Study of Practical
Reasoning in Philosophical Theology (1985).
104. Robert B.Barsky, F.Thomas Juster, Miles S.Kimball & Matthew D.Shapiro,
Preference Parameters and Behavioral Heterogeneity: An Experimental
Approach in the Health and Retirement Survey, 112 Q.J.Econ. 537579,
550 (1997).
105. Id.
106. Id.
107. See, e.g., Jay R.Ritter, The Long-Run Performance of Initial Public Offerings,
46J.Fin. 327 (1991).
108. See id. at 3.
109. See Tim Loughran & Jay R. Ritter, Why Dont Issuers Get Upset About
Leaving Money on the Table in IPOs?, 15 Rev. Fin. Stud. 413443, 414
(2002).
110. Richard A. Brealey & Stewart C. Myers, Principles of Corporate Finance
389 (5th ed. 1996); accord Loughran & Ritter, supra note 109, at 414.
111. Loughran & Ritter, supra note 109, at 414.
112. Id.
113. Id. at 416; see also Alexander Ljungqvist & William J. Wilhelm, Jr., Does
Prospect Theory Explain IPO Market Behavior?, 60J.Fin. 17591790, 1782,
1789 (2005) (concluding that underwriters defer their compensation by
extract higher fees from satisfied IPO participantsissuers and purchasers
alikein future transactions).
114. See Loughran & Ritter, supra note 109, at 416.
115. Id. (citing Richard H.Thaler, Toward a Positive Theory of Consumer Choice,
1J.Econ. Behav. & Org. 3960 (1980)).
116. See Kathleen Weiss Hanley, Underpricing of Initial Public Offerings and the
Partial Adjustment Phenomenon, 34J.Fin. Econ. 231250 (1993); see also
Michel A.Habib & Alexander Ljungqvist, Underpricing and Entrepreneurial
Wealth Losses: Theory and Evidence, 14 Rev. Fin Stud. 433458 (2001).
117. Ritter, supra note 107, at 3. See generally Roger G. Ibbotson, Jody
L.Sindelar & Jay R.Ritter, Initial Public Offerings, 1J.Applied Corp. Fin.
3745 (1988).
118. Ritter, supra note 107, at 3 (emphasis added).
240 J.M. CHEN
119. Id. at 4. See generally Edward M.Miller, Risk, Uncertainty, and Divergence
of Opinion, 32J.Fin. 11511168 (1977); cf. Robert J.Shiller, Speculative
Prices and Population Models, 4J.Econ. Persp. 5565 (1990) (describing
the vulnerability of nave investors to fads).
120. See Reena Aggarwal & Pietra Rivoli, Fads in the Initial Public Offering
Market, 19:4 Fin. Mgmt. 4557 (Winter 1990) (finding negative returns to
investors who buy stock in an IPO and hold it for one year and even more
negative returns to investors who buy stock in early aftermarket trading
immediately after an IPO and hold that stock for one year).
121. See, e.g., Nicholas Barberis, Ming Huang & Tano Santos, Prospect Theory
and Asset Prices, 116 Q.J.Econ. 153 (2001).
122. Barberis & Huang, Stocks as Lotteries, supra note 57, at 2091.
123. See T.Clifton Green & Byoung-Hyoun Hwang, Initial Public Offerings as
Lotteries: Skewness Preference and First-Day Returns, 58 Mgmt. Sci. 43244
(2012).
124. See Jennifer Conrad, Robert F.Dittmar & Eric Ghysels, Ex Ante Skewness
and Expected Stock Returns, 68J.Fin. 85124, 87, 104106 (2013).
125. See Estrada, An Alternative Behavioural Model, supra note 33 (Chapter 3),
at 241; supra 3.2, at 6066.
126. See Alon Brav & Paul A. Gompers, Myth or Reality? The Long-Run
Underperformance of Initial Public Offerings: Evidence from Venture and
Nonventure Capital-Backed Companies, 52J.Fin. 17911821 (1997).
127. See id. at 1791.
128. Id. at 1792.
129. Id.
130. See id. See generally Charles Lee, Andrei Shleifer & Richard H. Thaler,
Investor Sentiment and the Closed-End Fund Puzzle, 46 J. Fin. 2948
(1991).
131. Brav & Gompers, supra note 126, at 1792.
132. See Richard Carter & Steven Manaster, Initial Public Offerings and
Underwriter Reputation, 45J.Fin. 10451067 (1990).
133. See id. at 1046; Randolph P Beatty & Jay R.Ritter, Investment Banking,
Reputation, and the Underpricing of IPOs, 15 J. Fin. Econ. 213232
(1986); Richard B.Carter, Frederick H.Dark & Ajai H.Singh, Underwriter
Reputation, Initial Returns, and the Long-Run Performance of IPO Stocks,
53J.Fin. 285311 (1998); cf. Randolph P.Beatty, Auditor Reputation and
the Pricing of Initial Public Offerings, 64 Accounting Rev. 693709 (1989).
134. See Barberis & Huang, Stocks as Lotteries, supra note 57, at 2068, 2073
2074; Enrico G. de Giorgi, Thorsten Hens & Haim Levy, CAPM Equilibria
with Prospect Theory Preferences (Aug. 30, 2011) (available at http://ssrn.
com/abstract=420184).
135. See Barberis & Huang, Stocks as Lotteries, supra note 57, at 2074, 2096.
136. See Mohammed Abdellaoui, Han Bleichrodt & Hilda Kammour, Do

Financial Professionals Behave According to Prospect Theory? An Experimental
Study, 74 Theory & Decision 411429 (2013).
137. See Bowman, Risk/Return Paradox, supra note 20 (Chapter 4); Bowman,
Risk Seeking by Troubled Firms, supra note 21 (Chapter 4); Bowman,
Content Analysis of Annual Reports, supra note 21 (Chapter 4).
138. Avi Fiegenbaum & Howard Thomas, Attitudes Toward Risk and the Risk-
Return Paradox: Prospect Theory Explanations, 31 Acad. Mgmt. J. 85106,
85 (1988).
139. Id.
140. Id. at 8586.
141. Id. at 86; accord Avi Fiegenbaum, Prospect Theory and the Risk-Return
Association: An Empirical Examination in 85 Industries, 14J.Econ. Behav.
& Org. 187203, 188 (1990).
142. Fiegenbaum & Thomas, supra note 138, at 86.
143. Id.
144. Id.
145. Fiegenbaum, supra note 141, at 188.
146. Id.
147. See Kahneman, Thinking, Fast and Slow, supra note 11 (Chapter 1), at 281;
Kahneman & Tversky, Prospect Theory, supra note 90 (Chapter 6), at 274.
149. Baruch Lev, Industry Averages as Targets for Financial Ratios, 7J.Accounting
Research 290299, 290 (1969); accord Fiegenbaum & Thomas, supra note
138, at 91; Fiegenbaum, supra note 141, at 193.
150. Thomas J.Frecka & Cheng F.Lee, Generalized Financial Ratio Adjustment
Processes and Their Implications, 21J.Accounting Research 308316, 308
(1983). See generally Cheng F.Lee & Chunchi Wu, Expectation Formation
and Financial Ratio Adjustment Processes, 63 Accounting Rev. 292306
(1988); Samuel Tung & John Crowe, Expectation Formation and Financial
Ratio Adjustment Processes: A Comment and an Extension, 68 Accounting
Rev. 942952 (1993); Cheng F.Lee & Chunchi Wu, Expectation Formation
and Financial Ratio Adjustment Processes: A Reply, 68 Accounting Rev.
953955 (1993).
152. Id. at 97.
153. Id. at 97 n.1.
154. See, e.g., Hartzmark, supra note 28 (Chapter 2) (finding a propensity among
investors to focus on their first- and worst-ranked holdings, which is tanta-
mount to ignoring the rest of their portfolios); sources cited supra note 47
(Chapter 4)(documenting how humans measure welfare in relative rather
than absolute terms).
242 J.M. CHEN
155. Fiegenbaum & Thomas, supra note 138, at 97 n.1.

156. See Hazel J.Johnson, Prospect Theory in the Commercial Banking Industry,
7J.Fin. & Strat. Decisions 7389, 77 (1994); cf. Hazel J.Johnson, The
Relationship Between Variability, Distance from Target, and Firm Size: A
Test of Prospect Theory in the Commercial Banking Industry, 21 J. Socio-
Econ. 153171 (1992) (contrasting median returns with other factors
potentially affecting risky behavior in banking).
158. Id.
159. Id.
161. Id.
162. See Tversky & Kahneman, Advances in Prospect Theory, supra note 58
(Chapter 8), at 311; supra 8.4, at 189.
163. See Fishburn & Kochenberger, supra note 24 (Chapter 8).
165. See Marc Jegers, Prospect Theory and the Risk-Return Relation: Some Belgian
Evidence, 34 Acad. Mgmt. J. 215225 (1991).
166. See Tapen Sinha, Prospect Theory and the Risk Return Association: Another
Look, 24J.Econ. Behav. & Org. 225231 (1994).
167. See id. at 229; cf. Michele Cohen, Jean-Yves Jaffray & Tanios Said,
Experimental Comparison of Individual Behavior Under Risk and Under
Uncertainty for Gains and for Losses, 39 Org. Behav. & Human Decision
Processes 122, 1213 (1987) (finding similar compartmentalization effects
in individual decision-making).
168. See Richard Z.Gooding, Sanjay Goel & Robert M.Wiseman, Fixed Versus
Variable Reference Points in the Risk-Return Relationship, 29 J. Econ.
Behav. & Org. 331350 (1996).
169. See id.
170. See, e.g., Richard A. Bettis & William K. Hall, Diversification Strategy,
Accounting-Determined Risk, and Accounting-Determined Return, 25
Acad. Mgmt. J. 254264 (1982); Richard A. Bettis & Vijay Mahajan,
Risk/Return Performance of Diversified Firms, 31 Mgmt. Sci. 785799
(1985); W.Chan Kim, Peter Hwang & Willem P.Burgers, Multinationals
Diversification and the Risk-Return Trade-Off, 14 Strat. Mgmt. J. 275286
(1993).
171. See supra 7.8, at 155156.
173. Id.
174. See Fama & French, Cross-Section of Expected Stock Returns, supra note 5
(Chapter 3).
1), 2.13.4, at 538.
176. Cf., e.g., Javier Estrada & Ana Paula Serra, Risk and Return in Emerging
Markets: Family Matters, 15J.Multinatl Fin. Mgmt. 257272, 259 (2004)
(distinguishing between the traditional family of measures based on the
conventional CAPM and the factor family of measures emphasizing value,
size, and momentum).
177. Zhang, The Value Premium, supra note 4 (Chapter 4), at 67.
179. See generally, e.g., Harvey Leibenstein, Allocative Efficiency vs. X-Efficiency,
56 Am. Econ. Rev. 392415 (1966); George J. Stigler, The Xistence of
X-Efficiency, 66 Am. Econ. Rev. 213216 (1976).
180. Knowledge spillovers within a large firm may be regarded as a special case of
MarshallArrowRomer, or MAR, spillovers among geographically concen-
trated firms within a common industry. See generally, e.g., Edward L.Glaeser,
Hedi D. Kallal, Jos A. Scheinkman & Andrei Shleifer, Growth in Cities,
100J.Pol. Econ. 11261152 (1992); Adam B.Jaffe, Manuel Trajtenberg
& Michael S.Fogarty, Knowledge Spillovers and Patent Citations: Evidence
from a Survey of Inventors, 90 Am. Econ. Rev. 215218 (2000); Alfred
Marshall, Principles of Economics (1890); Kenneth J.Arrow, The Economic
Implications of Learning by Doing, 29 Rev. Econ. Stud. 155173 (1962);
Paul M.Romer, Increasing Returns and Long-Run Growth, 94J.Pol. Econ.
10021037 (1986).
181. See, e.g., Rebecca Henderson & Iain Cockburn, Scale, Scope, and Spillovers:
The Determinants of Research Productivity in Drug Discovery, 27 Rand
J.Econ. 3259 (1996).
182. See, e.g., Allen N. Berger, Gerald A. Hanweck & David B. Humphrey,
Competitive Viability in Banking: Scale, Scope, and Product Mix Economies,
20J.Monet. Econ. 501520 (1987); Jeffrey A.Clark, Economies of Scale
and Scope at Depository Financial Institutions: A Review of the Literature,
Fed. Reserve Bank Kansas City Econ. Rev., Sept./Oct. 1988, at 1633.
183. See, e.g., Elchanan Cohn, Sherrie L.W.Rhine & Maria C.Santos, Institutions
of Higher Education as Multi-Product Firms: Economies of Scale and Scope,
71 Rev. Econ. & Stat. 284290 (1989); Rajinder K.Koshal & Manjulika
Koshal, Economies of Scale and Scope in Higher Education: A Case of
Comprehensive Universities, 18 Econ. Educ. Rev. 269277 (1999).
184. See Jay J.Ebben & Alec C.Johnson, Efficiency, Flexibility, or Both? Evidence
Linking Strategy to Performance in Small Firms, 26 Strat. Mgmt. J. 1249
1259, 12561257 (2005).
185. Fiegenbaum & Thomas, supra note 138, at 99 (citing Ian C.MacMillan &
M.L. McCaffery, Strategy for Low Entry Barrier Markets, 2 J. Bus. Strat.
115119 (1982)).
187. Bowman, Risk/Return Paradox, supra note 20 (Chapter 4), at 27; accord
Fiegenbaum & Thomas, supra note 138, at 99.
244 J.M. CHEN
188. See, e.g., David A.Aaker & Robert Jackobson, The Role of Risk in Explaining
Differences in Profitability, 30 Acad. Mgmt. J. 277296 (1987); David
B. Jemison, Risk and the Relationship Among Strategy, Organizational
Processes, and Performance, 9 Mgmt. Sci. 10871101 (1987); Carolyn
Y. Woo, Path Analysis of the Relationship Between Market Share, Business-
Level Conduct and Risk, 8 Strat. Mgmt. J. 149168 (1987). On method-
ologies for studying managerial strategy within a specific industry, see
generally Inga Skromme Baird & Howard Thomas, Toward a Contingency
Model of Strategic Risk Taking, 10 Acad. Mgmt. Rev. 230243 (1985);
Kathryn R.Harrigan, Research Methodologies for Contingency Approaches to
Business Strategy, 8 Acad. Mgmt. Rev. 398405 (1983).
189. See Luis R.Gmez-Meja, Katalin Takcs Haynes, Manuel Nez-Nickel,
Kathryn J.L.Jacobson & Jos Moyano-Fuentes, Socioemotional Wealth and
Business Risks in Family-Controlled Firms: Evidence from Spanish Olive Oil
Mills, 52 Admin. Sci. Q. 106137 (2007).
190. Id. at 107.
192. Fiegenbaum & Thomas, supra note 138, at 86. See generally Avi Fiegenbaum
& Howard Thomas, Dynamic and Risk Measurement Perspectives on
Bowmans Risk-Return Paradox for Strategic Management: An Empirical
Study, 7 Strat. Mgmt. J. 395407 (1986).
193. See Bettis & Mahajan, supra note 170; Fiegenbaum, supra note 141, at 188.
In addition, whereas financial data exists solely by virtue of regulatory obli-
gations borne by publicly traded companies, accounting data is available to
and used by regulators, privately held firms, and state-owned enterprises. See
Fiegenbaum, supra note 141, at 188.
194. Fiegenbaum & Thomas, supra note 138, at 100 (emphasis added).
195. Id.
196. Jegers, supra note 165, at 223.
197. Id.; cf. Michael C. Jensen & William H. Meckling, Theory of the Firm:
Managerial Behavior, Agency Costs and Ownership Structure, 3J.Fin. Econ.
305360, 311 (1976) (The firm is not an individual. It is a legal fiction
which serves as a focus for a complex process in which the conflicting objec-
tives of individualsincluding managers and shareholdersare brought
into equilibrium within a framework of contractual relations. [emphasis in
original]).
199. Id.
200. Id. at 76.
201. Id.
202. Nicholas Barberis & Ming Huang, The Loss Aversion/Narrow Framing
Approach to the Equity Premium Puzzle, in Handbook of the Equity Risk
Premium 199229, 201 (Rajnish Mehra ed., 2008).
204. Id.
205. Id. at 86.
206. Id. at 87.
207. Id. at 81.
208. Id.
209. Id. at 81 & n.7; see Robert C. Merton, Lifetime Portfolio Selection and
Uncertainty: The Continuous Time Case, 51 Rev. Econ. & Stat. 247257
(1969); Paul A.Samuelson, Lifetime Portfolio Selection by Dynamic Stochastic
Programming, 51 Rev. Econ. & Stat. 238246 (1969).
211. See generally, e.g., Turan G.Bali, K.Ozgur Demirtas, Haim Levy & Avner
Wolf, Bonds Versus Stocks: Investors Age and Risk Taking, 56J.Monetary
Econ. 817830 (2009); Anup K.Basu, Alistair Byrne & Michael E.Drew,
Dynamic Lifecycle Strategies for Target Date Retirement Funds,
37:2J.Portfolio Mgmt. 8396 (Winter 2011); John J.Spitzer & Sandeep
Singh, Shortfall Risk of Target-Date Funds During Retirement, 17 Fin.
Servs. Rev. 143153 (2008).
212. See Barberis, Huang & Santos, supra note 121.
213. Id. at 2.
214. Id.
215. Barberis & Huang, Stocks as Lotteries, supra note 117, at 2090 (emphasis in
original). See generally Markus K.Brunnermeier, Christian Gollier & Jonathan
A.Parker, Optimal Beliefs Asset Prices, and the Preference for Skewed Returns,
97 Am. Econ. Rev. 159165 (2007); Markus K.Brunnermeier & Jonathan
A.Parker, Optimal Expectations, 95 Am. Econ. Rev. 10921118 (2005).
216. Barberis & Huang, Stocks as Lotteries, supra note 117, at 2090.
217. See Barberis, Huang & Santos, supra note 121, at 2.
218. Id.
219. Id.
220. Id.
221. Id. at 4; see also id. at 1719 (discussing Thaler & Johnson, supra note 15
(Chapter 2)).
222. Barberis, Huang & Santos, supra note 121, at 4.
8.18.5, at 155172 (time series); supra 5.1, at 9398 (intertemporal
CAPM).
224. Kahneman & Tversky, Prospect Theory, supra note 90 (Chapter 6), at 287
(citing William H. McGlothlin, Stability of Choices Among Uncertain
Alternatives, 69 Am. J.Psych. 604615 (1956)).
226. Barberis, Huang & Santos, supra note 121, at 19 (emphases added) (citing
Tversky & Kahneman, The Framing of Decisions and the Psychology of Choice,
supra note 21 (Chapter 2)).
246 J.M. CHEN
227. See Barberis, Huang & Santos, supra note 121, at 19 n.16.
228. Id. at 2.
229. Id.
230. Id. at 23.
231. Id. at 3.
232. Id. at 44.
234. Id.
235. Id.
236. Id. at 48.
237. Id. at 47; accord Michaela Pagel, Expectations Based Reference-Dependent
Preferences and Asset Pricing, J.Eur. Econ. Assn (forthcoming 2015).
238. See Nicholas Barberis & Ming Huang, Mental Accounting, Loss Aversion,
and Individual Stock Returns, 56J.Fin. 12471292, 1255 (2001).
239. Id.
240. Id.
241. Id.
242. See supra 7.8, at 155156.
243. See supra 9.1, at 215216.
244. Barberis & Huang, Individual Stock Returns, supra note 238, at 1255;
accord Gurevich, Kliger & Levy, supra note 124 (Chapter 8), at 1222 & n.5
(confirming narrow framing at the level of individual stocks); see also Klaus
Abbink & Bettina Rockenbach, Option Pricing by Students and Professional
Traders: A Behavioural Investigation, 27 Managerial & Decision Econ.
497510 (2006).
245. Nicholas Barberis, Ming Huang & Richard H. Thaler, Individual
Preferences, Monetary Gambles, and Stock Market Participation: A Case for
Narrow Framing, 96 Am. Econ. Rev. 10691090, 1069 (2006).
246. Id. at 1085. See generally Kahneman, Maps of Bounded Rationality:
Psychology for Behavioral Economics, supra note 2 (Chapter 8); Daniel
Kahneman & Dan Lovallo, Timid Choices and Bold Forecasts: A Cognitive
Perspective on Risk Taking, 39 Mgmt. Sci. 1731 (1993).
248. Id.
249. Id.
250. Barberis, Huang & Santos, supra note 121, at 17 n.14.
CHAPTER 10
Beyond Hope andFear:

Behavioral Portfolio Theory
10.1 Prospects Progress:

Beyond Theories ofEveryman
This book and its companion volume, Postmodern Portfolio Theory,1 have
devoted most of their attention to two models of finance. Each of these
two models is sensitive to human behavior. Postmodern Portfolio Theory
treated mathematical finance as a pattern of timeless moments, a deeply
quantitative puzzle whose answer lies in statistical distributions and their
properties. The presentation of a higher-moment CAPM in Chapter 3 of
this book enables the overtly behavioral interpretation of moment-based
theories of finance, which associate different statistical moments (mean,
variance, skewness, and kurtosis) with different emotions.2 Other chapters
in this book, so far, have presented financial models whose primary or
even exclusive purpose is to describe economic behavior as undertaken by
actual humans, as opposed to hypothetical economic reason dictated by
quantitative logic. Prospect theory, in particular, reflects the psychophys-
ics of chances.3
Even though these theories aspire to express the full range of consid-
erations, rational and emotional, that affect human decisions, they ulti-
mately fall victim to the excessive abstraction that has discredited modern
portfolio theory, the strong form of the efficient markets hypothesis, and
expected utility theory. Psychophysical theories and moment theo-
ries are theories of Everyman because they are based mechanistically

DOI10.1007/978-3-319-32711-2_10
248 J.M. CHEN
on principles that are presumed to bind all of humanity.4 As its very

name suggests, the idea of psychophysics invites precise quantitative speci-
fication, as though human emotion and behavior, if scrutinized closely
enough, could be contained [by] the axioms and corollaries of a book of
mathematics.5
In particular, despite being the most fully developed of psychophysical
theories, prospect theory is a rather blunt tool of analysis and cannot
explain the way all actors make decisions in all contexts.6 Prospect theory
has difficulty rebutting this statement: Everyman is risk averse for gains
even though every man (or woman) is not.7 Risk-averse people empha-
size the likelihood of meeting or exceeding a goal and are mostly
concerned about doing badly.8 Risk-seeking people, by contrast, have
a clear focus on large prizes.9 It is quite possible that risk-seeking
investor behavior is the true driver of stock prices.10 Consequently, a con-
vincing approach to behavioral finance must account for this obvious and
pervasive temperamental difference among individuals.11
All models are wrong; some models are useful.12 For example, the
ideal gas law is not exactly true for any real gas but nevertheless fre-
quently provides a useful approximation based upon a physical view of
the behavior of gas molecules.13 As the ideal gas law predicts the real-
world behavior of a hypothetical gas, prospect theory depicts the psycho-
physics of chances in broad, helpful generalities. To extend the descriptive
power of prospect theory beyond its limitations, this chapter will outline
an alternative theory called SP/A theory, named for its three components:
security, potential, and aspiration. SP/A theory has a palpable claim to
being the alpha and omega of behavioral theories of finance. It bridges
Roys safety-first principle, arguably the first behaviorally sophisticated cri-
tique of modern portfolio theory, with contemporary efforts to specify an
unapologetically behavioral approach to portfolio theory. Behavioral port-
folio theory, in turn, relies on mathematical techniques that are equivalent
to value-at-risk (VaR) analysis.
10.2 SP/A Theory

Psychologist Lola Lopes propounded SP/A theory in a 1987 arti-
cle, Beyond Hope and Fear: The Psychology of Risk. Her opening gambit
lamented that the psychology of risk had spawned so much theory
for so little substance.14 Writers trained in economics and in psychology
had spent countless hours trying to explain theoretically why people buy
BEYOND HOPE ANDFEAR: BEHAVIORAL PORTFOLIO THEORY 249
both lottery tickets and insurance.15 Lopes found it disconcerting that

experts on risk were struggling with the obvious.16 Lottery tickets cost a
dollar. One. We buy insurance so that we can sleep better. Is it really so
strange that we should want to buy both?17
Her playful introduction to the subject notwithstanding, Lopes out-
lined an ambitious plan to offset what was then the social sciences over-
whelming preference for weighted-value models of risk, the best
known of which [was] expected utility theory.18 She sought to defend the
less popular but more intuitively appealing body of work treating risky
choice as a process of trading off potential return with risk, defined as
variability in the outcome distribution or as variability among potential
outcomes.19 Within this literature, Lopes included approaches to port-
folio theory based on the premise that choices among risks reflect a
compromise between maximizing expected value and achieving an indi-
vidually determined ideal level of risk.20
Lopes likewise credited portfolio theory as one approach that has rec-
ognized, contrary to expected utility theorys dogmatic assumption of uni-
versally concave utility curves, that people are not always risk averse.21
Lopez recognized that people buy lottery tickets even though they know,
or should know, that, on average, lotteries are mathematically unfair in
the sense that the price of the ticket is more than the expected value.22
True to the Allais paradox and to prospect theory, Lopes also acknowl-
edged that many people would prefer to face an 80% chance of losing
$4000 than to lose $3000 for sure even though the expected value of the
risky loss ($3200) is greater than the sure loss.23 What Daniel Kahneman
and Amos Tversky said of prospect theorys value function applied with
equal force to Lopess view of choice under risk and uncertainty: Our
perceptual apparatus is attuned to the evaluation of changes or differ-
ences rather than to the evaluation of absolute magnitudes.24 The same
sensory psychophysics that make the difference between 1 pound and
2 pounds seem greater than the difference between 10 pounds and 11
pounds25 are not to be confined to attributes such as brightness, loud-
ness, or temperature.26
This capacity for incremental perception applies in particular to
the evaluation of monetary changes.27 In particular, humans seem to
respond more to the shapes of lotteriesthe physical representation of
the statistical distribution of financial returnsthan to the amounts
and probabilities of individual outcomes.28 Treating risk [as] a func-
tion of shape connected theories of risk perception and theories of
250 J.M. CHEN
risk preference with strictly statistical financial models.29 Moment-based

theories associate the shape of a lottery with the statistical moments of
the distribution, particularly mean, variance, and skewness.30 Specifically,
[v]ariance is generally considered to be bad (i.e., risky), whereas positive
skewness has been identified with hope and negative skewness has
been identified with fear.31 These are precisely the interpretations of odd
and even moments provided by four-moment CAPM.32 And transforming
the cumulative distribution function of a typically asymmetrical financial
distribution into prospect theorys value function brings to life the very
notion that fear, hope, and greed can take physical form.33
Moments models have several virtues.34 Chief among them, perhaps,
is that any distribution can be described, in principle, to any desired
level of precision by a sufficiently large set of its moments.35 That was
the instinct underlying Chapter 3s presentation of the progression of
the CAPM from its conventional two-moment formulation to the four-
moment variant based on a Taylor series expansion of logarithmic returns.
But moments-based models of behavioral finance also face major difficul-
ties.36 Perhaps the most serious is the fact that such theories implicitly
assume that moments have independent psychological reality.37 In truth,
they do not. Variance in particular, despite its prominence as the proxy for
risk in modern portfolio theory and the conventional CAPM, offers little
intuitive meaning or guidance.38 It is not all that easy to intuit the relative
variance of lotteries that differ in skewness except when the differences
are very great.39
10.3 The Human Heart inConflict withItself
SP/A theory represents a two-factor theory for risky choice.40 Behavior

has sources, both inner and outer.41 Although humans are disposed
by [their] unique constitutions and histories to behave in certain ways,
ultimately, it is situational stimuli that evoke [our responses], and it is
changes in conditions that alter them.42 SP/A theory accordingly com-
bines both a dispositional factor and a situational factor to explain risky
choice.43 The name SP/A theory refers to this models balance of its
dispositional security-potential factor (which reflect[s] the way individuals
typically respond to rules) against a situational factor called aspiration.44
SP/A = security-potential/aspiration.
The dispositional factor describes the underlying motives that dis-
pose people to be generally oriented to achieving security or to exploiting
potential.45 The former disposition is risk averse; the latter, risk seeking.
Risk averse people appear to be motivated by a desire for security, whereas
risk seeking people appear to be motivated by a desire for potential. The
former motivation values safety and the latter, opportunity.46 Notably,
these traits represent the two poles, refuge and prospect, in Jay Appletons
pathbreaking analysis of landscape painting as a projection of human aes-
thetics onto the environment.47 Every picture tells a story, either of risk-
averse refuge or of risk-seeking prospect.
Lopess distinction between the risk-averse desire for security and risk-
seeking yearning for potential invites the deployment of a two-parameter
probability weighting function that distinguishes between [d]iscrim-
inability (how people discriminate probabilities in an interval bounded
away from 0 and 1) and [a]ttractiveness (the degree of over/under
weighting).48 If the parameter represents curvature (discriminability)
(in the sense of an individuals responsiveness to changes in the probability
of success) and represents elevation, we can define probability weights
as a function of these parameters:49
d pg
w ( p) =
d pg + (1 - p )
g

The and parameters have a neat psychological interpretation.50

, the curvature parameter, reflects a subjects responsiveness to changes
in probability: the smaller , the less responsive the subject is.51 By con-
trast, the interpretation of , the elevation parameter, depends on the
domain considered.52 For winning gambles can be viewed as a gam-
bles attractiveness. The more elevated the probability weighting curve,
the greater are the weights placed on the probability, and the more
attractive that gamble appears.53 For losing gambles, however, greater
elevation as expressed by the parameter makes the gamble less attrac-
tive and the decision-maker more pessimistic.54
The one-parameter probability weighting function used in cumulative
prospect theory is a special case of this two-parameter function in which
, the elevation or attractiveness parameter, is equal to 1.55 One limitation
on the definition of weighting function w(p) as a continuous approxi-
mation of actual, subjective decision weights is that this function does
not allow investigation of discontinuities near p=0 and p=1, another
property suggested for prospect theorys decision-weighting function.56
252 J.M. CHEN
Because weighting function w(p) is fixed at both ends in the sense that
w(0)=0 and w(1)=1, the [e]levation and slope of this function can-
not be varied totally independently from each other.57
As between the two dispositions, security motivation (or risk aversion)
is the far more common patternso common that models such as the
ArrowPratt equations treat risk aversion as the pattern for Everyman.58
A pair of maxims about gambling illustrates the contrast between the
risk-averse security motivation and the risk-seeking potential motivation.
American journalist Damon Runyon (18801946) once advised the risk
averse, the race is not always to the swift nor the battle to the strong, but
thats the way to bet.59 For his part, Nkos Dndolos (18831966)bet-
ter known as Nick the Greekcelebrated the [r]isk seekers who dog
the long slots, waiting for that one streak of luck, properly ridden and
encouraged to compensate them for all the bad times.60 For riches
do not fall to the intelligent, nor favor to [people] of skill; but time and
chance happen to them all.61
By contrast, the situational factor describes peoples responses to
immediate needs and opportunities.62 People occasionally act contrary
to their own dispositions: [R]isk seekers may play it safe from time to
time, and even the most risk averse person will take chanceseven big
chanceswhen necessary.63 The aspiration level is a situational variable
that reflects the opportunities at hand (What can I get) as well as the
constraints imposed by the environment (What do I need?).64 In formal
terms, the aspiration criterion assumes that individuals assess the attrac-
tiveness of lotteries by the probability that a given lottery will yield an
outcome at or above the aspiration level, :65
A = p (v a )

Satisfying the aspirational level, in simpler language, means that even if

you do not get what you want, you can still get what you need.
The aspiration level arises from three potential sources. The first is
the direct assessment of what is reasonable or safe to hope for.66 A sec-
ond source is the direct contextual influence of other alternatives in the
choice set.67 The seemingly special status of certainty in risky choice,
particularly if it is defined by the sort of subjective framing that prospect
theory calls categorical or boundary effects at a reference point, may be
considered a special case of this factor.68
Third and finally, aspiration levels get set by outside influence.69

For dramatical examples, consider a competitive game in which players
attempt[] to take or defend territory.70 When players find themselves in
a bad position near the end of a round, they prefer riskier moves.71
Generally speaking, when people are in economic difficulty, they tend
to take risks that they would avoid under better circumstances.72 Agents
in real-life situations, from football to bond trading to guerrilla warfare,
behave precisely this way. In a celebrated, dramatic example of a bet the
company gamble, the founder of FedEx reputedly took the companys
last $5000 and won $27,000in Las Vegas to keep the struggling startups
planes aloft for another week.73 This paradoxical relationship between risk
and return, known as Bowmans paradox, commands detailed attention in
4.1 and 9.3 of this book, respectively, because of its connection to the
low-volatility anomaly in mathematical finance and to prospect theory.74
A heightened propensity to take deep chances is a prediction that SP/A
theory shares with the risk-seeking corner of prospect theorys fourfold
pattern.75
Critically, risky choice is not conflict free.76 The dispositional and
situational factors are sometimes in conflict and sometimes in concert;
their interaction produc[es] complex patterns of behavior in which
risk averse choices and risk seeking choices exist side by side in the same
individual[].77 [S]upport for the conceptual distinction between the
security-potential and aspiration factors exists in the fact that the
factors often act in opposition to each other.78 SP/A theorys internal
conflict thus appears to arise from the more general wellspring of intraper-
sonal conflict between emotional and rational evaluations of risk.79 William
Faulkner expressed a similar sentiment when he accepted his Nobel Prize
for a decidedly noneconomic subject: Economics, psychology, and litera-
ture share a concern with the problems of the human heart in conflict
with itself which alone can make good writing because only that is worth
writing about, worth the agony and the sweat.80
The most interesting conflicts within SP/A theory are less obvious
than straightforward tensions between security and potential (as in the
trade-off between risk and return or between yield and default risk).81
Truly unusual patterns of agreement and disagreement can arise
between dispositional motives toward security and potential, on the one
hand, and the immediate needs and opportunities affecting aspirational
level, on the other.82 In particular, losses, unlike prospective gains, pres-
ent a conflict between security and aspiration for the typically risk-averse
person.83
254 J.M. CHEN
A person disposed to be risk seeking faces the same conflict, merely

in reverse: For losses, potential and aspiration level are positively cor-
related, but for gains they are quite likely to be negatively correlated.84
Risk-averse people exhibit straightforward, monotonically increasing pref-
erences for potential gains, but an inverse U pattern of preferences with
respect to losses that are low at the extremes but higher in the middle.85
For risk seekers, the simple pattern occurs for losses: preferences tend
to increase from the sure thing to the long shot.86 Risk seekers exhibit a
more complex pattern of preferences toward gains. For them, the prom-
ise of riskless payouts tug[] against a general preference for bimodal out-
comes or even more extremely skewed long shots.87
The true value of SP/A theorywhat distinguishes SP/A theory from
prospect theory, to say nothing of expected utility theorylies in the abil-
ity of conflict between the disposition toward either security or poten-
tial and the situational opportunity to reach a certain aspirational level
to explain[] how [one] person can be risk averse in the economic sense
but sometimes make[s] the same choices as a person who is predisposed
to seek risk.88 SP/A theorys two-factor approach thus puts risk seekers
and risk averse people on equal footing.89 Despite profound differences
in their choices, the choice processes that motivate the risk averse and
the risk affine have more similarities than differences.90 In other words,
[t]heir goals may differ, but they have the same conceptual equipment.91
10.4 Roys Safety-First Criterion

In some senses, prospect theory and SP/A theory are conceptual antip-
odes.92 Although SP/A theory shares many important features with
cumulative prospect theory, such as an inverse S-shaped weighting func-
tion and a reference point, the two theories rest on quite different sets
of psychological underpinnings.93 For prospect theory, psychophysics
[serves] as a unifying principle underlying the shapes of the utility (value)
function and weighting functions.94 By contrast, SP/A theory empha-
sizes the impact of emotions such as fear and hope.95 These are the ways
in which prospect theory rest[s] on perceptual concepts, while SP/A
theory rest[s] on attentional and motivational concepts.96
As leading theories of decision-making under uncertainty, however,
prospect theory and SP/A theory are surprisingly similar when stud-
ied in an appropriate mathematical framework.97 Both prospect theory
and SP/A theory provide a psychological grounding that appeal[s]
directly to intuitions via easily understood and compellingly named

components.98 Because both of SP/A theorys basic criteria, security-
potential and aspiration, seem necessary to capture human choices
and risk,99 SP/A theory is arguably better suited than prospect theory
to serve as a theoretical basis for developing both behavioral portfolio
theory and behavioral asset pricing theory.100
Hersh Shefrin and Meir Statman have acted upon this suggestion in
devising behavioral portfolio theory as a positive portfolio theory on the
foundation of SP/A theory and prospect theory, two theories of choice
under uncertainty.101 Fear and hope operate on the willingness to take
riskby altering the relative weights attached to decumulative prob-
ability.102 Fear prompts investors to overweight[] the probabilities
attached to the worst outcomes relative to the best outcomes.103 Fearful
investors act is if they were unduly pessimistic when computing expected
wealth.104 By contrast, hope leads individuals to act as if they were unduly
optimistic.105
The validation of any economic model, including behavioral models
drawing insights from prospect theory and SP/A theory, begins with the
[m]athematical analysis necessary to specify the quantitative mecha-
nisms from which theoretical predictions flow.106 The mathematical
mechanics underlying behavioral portfolio theory begin with a closer look
at Roys safety-first criterion.107 The safety-first criterion consciously seeks
to minimize the probability that an investor would realize actual returns
(Ra) below some minimally acceptable baseline (Rb).
Formally, compliance with the safety-first criterion directs a portfolio
manager to minimize P(Ra < Rb). Roys safety-first criterion very mod-
estly substitutes the minimally acceptable baseline for the risk-free rate
E-d
of the Sharpe ratio. It directs a manager to maximize , where E
s
is the expected return and d represents a disastrous, unacceptable rate of
return.108 Roy eventually generalized his safety-first criterion by recog-
nizing that the disastrous level, d, may itself be uncertain and therefore
should not be regarded as a fixed value. Rather, d should be represented
by a variable, continuous utility function.109
In recognizing that an investor who seeks advice about his actions will
not be grateful for the suggestion that he maximises expected utility, Roy
departed from neoclassical economics assumption of pure investor ratio-
nality.110 Indeed, Harry Markowitz has described Roys safety-first crite-
rion as one of the two papers published in 1952 that opened the era
256 J.M. CHEN
of modern portfolio theory.111 Roy was among the first financial w riters
to recognize a difference in financial decision-making that arose from
varying behavioral sensitivities according to the magnitude of a potential
loss: whereas small-scale financial decisions may seek some target rate
of return, larger commitments involve[] the danger of insolvency.112
Most investors perceive a low probability of a large loss to be far more
risky than a high probability of a small loss, even when the expected losses
are the same.113 Although this perception does reject perfect rationality
and other economic assumptions underlying the efficient capital markets
hypothesis, it is more realistic to model financial markets with awareness
of this nearly universal human heuristic.
Properly extracting the foundation of behavioral finance from Roys
safety-first criterion would liberate Roys study [from] the shadow of
Markowitzs contemporaneous work on modern portfolio theory.114 In
the logic of behavioral finance, Roys safety-first criterion may be char-
acterized as the goal of minimizing the probability of ruin.115 Ruin
occurs when w, a variable representing terminal wealth, falls short of
subsistence level s, the probability of which we may formally indicate
as prob{w<s}.116
Roys original safety-first criterion may be treated as a special case based
on certain narrow assumptions, such as the absence of a risk-free security,
a low value for s (achieved by defining s to be less than the expected value
of all possible portfolios), and normally distributed returns.117 Dispensing
with the requirement that s meet any predetermined subsistence level,
let us redefine the goal of the safety-first criterion as that of ensuring that
the probability of remaining above subsistence, prob{w<s}, does not
exceed some probability .118 Optimizing a behaviorally sensitive inves-
tors portfolio would therefore require the maximization of w subject to
the constraint, prob{w<s}.119
One final elaboration of this framework consists of allowing , the
probability of ruin, to vary. This modification takes advantage of Arthur
Roys own modification of d, his disastrous rate of return, as a continu-
ously variable function rather than a constant.120 Should wealth fall below
the subsistence level, utility would fall by some fixed quantity:121
u ( w ) = w prob {w < s} a
u ( w ) = w - c prob {w < s} a

All of these models define financial danger as the possibility that wealth
might fall below a particular minimum level.122
10.5 Behavioral Portfolio Theory

To this day, many behaviorally sensitive financial models continue to
employ some variant of Roys safety-first criterion.123 Variations of par-
ticular interest include the intertemporal, multiperiod application of the
safety-first criterion,124 the performance of safety-first analysis with extreme
value distributions,125 and the application of safety-first principles to port-
folios containing unusually risky assets.126
Defining safety as the mirror image of ruinnamely, prob{ws}
recasts behavioral portfolio theory as a variation on Roys theme.127 By
working with the decumulative distribution function taking the form,
D(x)=prob{ws},128 SP/A theory replaces subsistence level s with the
more general aspiration level, A.129 Whereas earlier elaborations of the
safety-first criterion had treated the probability of ruin, , as a cumulative
probability, SP/A theory evaluates ruin in decumulative terms as 1.130
Although this description treat[s] the aspiration level as [a] crisp, dis-
crete value that either is or is not satisfied, the aspiration level could as
easily be fuzzy in the sense that some outcomes may satisfy the aspi-
ration level completely, others to a partial degree, and still others not at
all.131 Alternatively, the function [need] not be a step function, but
rather a smooth function that starts at zero and reaches one on some
interval around .132
Among the differences between prospect theorys reference point and
SP/A theorys aspiration level, two loom especially large. First, each of
these benchmarks differ[s] in how [it] exerts its impact.133 Prospect
theorys reference point is incorporated into the utility function and
mark[s] an inflation point dividing outcomes into gains and losses and
then scaling them nonlinearly in accord with a principle of diminish-
ing sensitivity.134 By contrast, SP/A theorys aspiration level reflect[s]
a principle of stochastic control that is distinct from individuals assess-
ment of security and potential.135 On occasion, SP/A theory predicts con-
flict between the security-potential and aspiration criteria, something not
contemplated by [prospect theory] and other single-criterion models.136
The second difference between prospect theorys reference point and
SP/A theorys aspiration level flows directly from this potential for c onflict
within SP/A theory. Prospect theory predicts a four-fold pattern across
gain preferences and loss preferences, one of smooth and symmetrical
258 J.M. CHEN
reflection between gains and losses.137 The resulting prediction[] of

(mostly) risk-seeking behavior for decisions about losses has posed a
major challenge to the application of prospect theory to problems in
behavioral economics and finance.138 To overcome seemingly anomalous
departures from assumed risk seeking for losses,139 models applying pros-
pect theory often incorporated parameters of risk aversion drawn from
expected utility theory.140
By contrast, SP/A theory allows considerable asymmetry between
gains and losses.141 In the most commonly observed case, people avoid
risks strongly for gains but [are] more-or-less risk neutral for losses.142
This pattern arises because security-minded or cautiously hopeful [indi-
viduals] set modest aspiration levels for gains, allowing the SP and A cri-
teria to reinforce each other.143 If the chances of attaining a desirable
outcome exceed the acceptable probability of success demanded by an
investor, that investor may well shed the defensive posture of a safety-first
approach in favor of a higher-yielding expected value criterion.144
On the other side of returns, a different relationship emerges. Mutual
reinforcement between security and aspiration levels does not occur for
losses. For losses, many individuals set high aspiration levels, hop-
ing to lose little or nothing.145 The high aspiration in the domain of
losses set[s] up a conflict between SP/A theorys aspiration level and its
security-potential criterion.146
SP/A does not evaluate terminal wealth value w directly, but works
instead with the Eh(w), expected value of w under the transformed decu-
mulative function of the probability of achieving a safe financial out-
come.147 For investors strongly driven by fear, Eh(w)<E(w).148 Indeed,
[t]he greater the fear, the lower the value of Eh(w).149 Hope in potential-
driven investors operates by increasing Eh(w) relative to E(w).150 SP/A
also evaluates D(A), the probability that the payoff from a particular
financial decision will be A or higher.151 Since Eh(w) and D(A) are vir-
tual analogues of the arguments used in the safety-first model, E(w) and
prob{w<s}, it is possible to compare the utility functions that each model
seeks to maximize:152
Safetyfirst:u(E (w),)
SP/A: u[Eh (w),D (A)]
Indeed, the standard safety-first model with a variable probability of ruin

is a special case of SP/A theory in which the strength of fear (the need for
security) and the strength of hope (the need for potential) are equal and
therefore neutralize each other.153
In practical terms, the contrast between conventional portfolio theory
and SP/A theory parallels the contrast between the two leading schools
of retirement planning.154 Conventional, probability-based approaches to
retirement income management aspire to determine a safe withdrawal
rate, in the sense of a rate of consumption that gives retirees a comfort-
ably high probability of not exhausting their resources before death.155
Probability-based retirement planning supplies the rule of thumb that
retirees may safely withdraw 4% of their portfolios per year.156 That bench-
mark has come under attack.157 SP/A theory and related work in behav-
ioral finance (especially life cycle theories of savings and consumption)
inform a competing safety-first school of retirement planning.158
At a (much) higher level of theoretical abstraction, portfolio optimi-
zation under SP/A theory closely resembles conventional mean-variance
optimization. Just as investors prefer higher mean return with lower vari-
ance, they prefer a higher decumulative expected return and a lower prob-
ability of failure to attain their aspiration level.159 SP/A theorys analogue
to modern portfolio theorys {, } space may therefore be defined as
{Eh(w), prob(wA)}, the space within which Eh(w) is maximized for a
given value of prob(wA).160 In other respects, however, SP/A theory
transcends modern portfolio theory. SP/A theorys multidimensional
treatment of risk stands in stark contrast with the reliance of modern port-
folio theory and expected value theory on risk tolerance as a single, free-
standing risk parameter.161 SP/A theory uses five parameters to describe
different dimensions of behavioral portfolio theory:
qs measures the strength of fear (need for security);

qp measures the strength of hope (need for potential);
A is the aspiration level;
measures the strength of fear relative to hope; and
determines the strength of the desire to reach the aspiration level
relative to fear and hope.162
What SP/A theory evidently sacrifices under the Akaike information

criterion,163 it gains through a highly nuanced model of fear, hope, and
conflicting desires. A behavioral portfolio theory built upon SP/A the-
ory therefore predicts that investors will choose portfolios by consider-
ing expected wealth, desire for security and potential, aspiration levels,
and probabilities of achieving aspiration levels.164 The application of
260 J.M. CHEN
SP/A theory to problems of asset allocation and pricing leads directly

to portfolio optimization over two-moment distributions where wealth
is maximized subject to reaching a threshold of return within a given
probability.165 The process that emerges from this generalization of
optimization under the safety-first criterion is mathematically equiva-
lent to mean-variance optimization.166 Portfolio maximization under
this framework can be generalized to other sources of wealth, including
housing and cash flow from labor.167 For all of the foregoing reasons, I
will henceforth treat the term SP/A theory, with the clarifications and
extensions outlined in this section, as if it were synonymous with behav-
ioral portfolio theory.
10.6 The Practical Consequences

ofBehaviorally Sensitive Portfolio Optimization
The application of behavioral portfolio theory produces outcomes that

differ even dramatically from those of modern portfolio theory. Portfolios
generated under behavioral influence do not reside on the efficient fron-
tier of mean-variance optimized portfolios.168 Behavioral portfolios violate
the two-fund separation theorem.169 Behavioral portfolio theory predicts
spectacular violations of the separation theorem, which contemplates an
equity subportfolio representing the entire securities market, offset for
purposes of liquidity by a cash position, of conventional mean-variance
optimization based on the CAPM, which counsels combinations of the
market portfolio and the risk free security.170 Instead, behavioral portfo-
lios resemble combinations of bonds and lottery tickets.171 Relative to
investors who adhere strictly to mean-variance optimization, behaviorally
motivated investors are much likelier to consider casino-type securities,
securities with low expected returns and high variance, and, even more
critically, high skewness.172
This effect becomes even more pronounced when portfolios are seg-
regated, either by mental accounting or by force of tax code provisions
encouraging investors to isolate accounts intended for retirement (401(k),
403(b), IRA, Roth IRA, and the like) and for education (529, Coverdell)
from other investment accounts.173 The layering or pyramiding of these
separate accounts, an outcome predicted by Maslowian psychology,174
magnifies even more dramatically the behavioral tendency to combine
bonds (or very conservative equity positions) with lottery-like invest-
ments. The simultaneous presence of low and high aspiration accounts
within a behaviorally motivated investors hierarchy of needs and desires

magnifies the extremes within the overall portfolio so that the low aspi-
ration account will look more like [a] risk free bond, while the high
aspiration account looks more like a lottery ticket.175
Although treating wealth as a proxy for risk aversion carries serious pit-
falls,176 the level of wealth may affect an investors choice as between low-
and high-aspiration accounts or investment strategies. Because more
wealth implies a riskier portfolio with a higher rate of return (to say noth-
ing of additional resources for weathering higher risk), wealth may be
expected to grow faster the larger it is.177 In turn, greater wealth lowers
risk aversion, thus enabling nearly all investors to choose[] a riskier port-
folio and entrepreneurially minded investors to go into risky ventures
even with small amounts of capital.178
Consistent with Amos Tversky and Daniel Kahnemans observation that
behaviorally motivated human choices are orderly rather than chaotic
and intractable, though not always rational in the traditional sense,179
behavioral portfolio theorys all-or-nothing pattern is the product of a
reasoning process that is systematic and rules-based, even if it is not finan-
cially optimal. A high-aspiration investor chooses a lottery-like or casino-
like stock because it provides the highest probability of reaching the
aspiration level.180 Investors with high aspirations are not inherently risk
seeking.181 Instead, they choose stocks with lottery-like returns because
such investments give them the quantifiably best chance of reaching the
aspiration level.182 There is an efficient frontier in sight, even if it is not
the same one projected by modern portfolio theory. The desire to reach
the right tradeoff between expected returns and the probability of failing
to reach an investors subjectively desired threshold level explains why
risk seeking can be optimal under behavioral portfolio theory, contrary
to models that predict risk aversion and concave utility functions through-
out all domains.183
Behavioral portfolio theory has an answer for the putative paradox
that inspired the consideration of risk aversion in financial behavior. As
behavioral portfolio theory sees finance, it is far from paradoxical that
people simultaneously purchase insurance and lottery tickets.184 People
often accept[] a lottery, a game with a high positive skewness and a nega-
tive expectation, while simultaneously reject[ing] noninsurance, a game
with a strong negative skewness and a positive expectation.185 There is
no contradiction in this sort of conduct. Rather, such simultaneity is the
hallmark of efficient [behavioral] portfolios.186
262 J.M. CHEN
This phenomenon grows directly from SP/A theorys conflict between

its security-potential criterion and the individual decision-makers aspira-
tion level.187 On the one hand, a security-minded weighting function
pays more attention to the worst outcomes.188 With respect to the A
side, however, the model operates on the probability of achieving the
aspiration level.189 The resulting pattern is consistent with observations
of investors and mutual fund managers who focus entirely on the extremes
of their portfolioin the domain of gains as well as in the domain of
lossesat the price of ignoring the rest of their holdings.190
Affluent investors in developed economies are not alone in blending
defensive, risk-averse strategies with all-or-nothing bets.191 Similar behav-
ior has been observed among subsistence farmers, for whom risk is such
a stark fact of [the] physical and social environments that ones liveli-
hood can literally be threatened from all sides (by floods, by pests, by
invading armies).192 The portfolio optimization problem in subsistence
agriculture is one of allocating extremely scarce resources between two
assets with radically different risk profiles. Food crops provide food for
the table and low variance of return, but their expected return is also
low.193 In other words, food crops provide the closest thing to a guar-
antee of survival, but at a level of abject poverty. Cash crops, by con-
trast, are more variable but have higher expected return.194 Saving seed
corn while simultaneously shooting for the moon appears to be the innate,
modal financial strategy of humankind.
The myriad solutions to this problem all boil down to a simple rule:
first take care of subsistence needs (food for the larder and seed for the
coming season) and then plant cash crops.195 Subsistence farmers port-
folio strategy consists of planting low-return food crops to the point
where subsistence needs are met but remaining willing to allocate
the remainder of [available] land to cash crops.196 It is tempting to restate
this strategy in simple fashion: [F]armers gamble on cash crops because
they aspire to escape poverty.197 Critically, however, subsistence farmers
do not view themselves as gambling.198 Paraphrasing a subsistence farm-
ers preference for subsistence over starvation as a preference for X
chance of Y income over X totally misrepresents his options.199 It is
more honest and accurate to visualize food crop and cash crop allocations
in subsistence agriculture as the simultaneous satisfaction of the same emo-
tions of fear, hope, and aspiration that motivate wealthier actors. Although
fear of falling below subsistence motivates the allocation to food crops
in what must be considered a truly compelling application of the safety-first

criterion, the enduring aspiration of escaping poverty motivates the
allocation of the remainder to cash crops.200 Indeed, under sufficiently
dire conditions, it is quite rational for subsistence farmers to devote
relatively greater acreage to cash crops, since such a strategy represents
their sole hope to maximize their chances of survival.201
Similar incentives motivate diamond miners in desperately poor coun-
tries such as Sierra Leone.202 Lacking ex ante capital of their own, African
artisanal diamond miners typically work on credit and share [any]
profits from mining with their creditors.203 Because these m iners
do not refund any part of the credit that cannot be recovered from
their earnings in the event of a loss from a mines failure to recover
diamonds, miners engage in a form of gambling for resurrection akin
to the excessive risk taking that characterizes financially distressed firms
on the verge of default in developed economies.204 The lopsided allure
of a huge payoff for the luckiest miners contributes to the resource
curse that stunts the economic development of some of the worlds
countries holding immense amounts of mineral wealth and other natural
resources.205
Wealthier investors combine low and high aspirations similar to those
of subsistence farmers and diamond miners in the developing world: they
want to avoid poverty, understandably, but they also want a shot at
riches.206 True to the Maslowian vision of portfolio management, such
investors build layered pyramids that divide current wealth between
a bottom layer, designed to avoid poverty, and a top layer, designed for
a shot at riches.207 In the special case of cautious optimism, fear pre-
dominates asset allocation except for the upper end of the range.208
By overweight[ing] the probabilities attached to both the worst and
best outcomes, a cautious optimist produces a weighting scheme with
thicker tails than the underlying distribution.209 In the language and
logic of statistical moments, this is kurtosis preference. Thus arises the all-
or-nothing structure, or at least the bonds-plus-lottery-tickets allocation,
associated with behavioral portfolio theory. A cautious optimist, being
inclined to extremes in the two layers of his portfolio, chooses a risk-
free bond for his low aspiration account and then searches on the basis
of [i]ncreased hope for the maximum possible payoffs in [the] high
aspirationaccounts.210
264 J.M. CHEN
Let us return to the general observation that behavioral portfolio

theory does not satisfy mean-variance optimization.211 In principle, there
will be a loss of efficiency and of welfare in the classic utilitarian sense.
The normative question is whether investors who maximize their welfare
under behavioral optimization should be willing to bear that sacrifice.
As long as short-selling is permitted, mental accounting and behavioral
optimization can still generate combinations of [mean-variance] efficient
portfolios result[ing] in an efficient aggregate portfolio.212 Even without
short-selling, efficiency losses relative to strict mean-variance optimization
have been characterized as a few basis points.213 Such losses decline as
investors levels of risk aversion increase and are arguably small compared
to the loss that arises when investors, confused by the nonintuitive
abstractions of modern portfolio theory, inaccurately specify their risk
aversions.214 And this is to say nothing of losses incurred by uninformed
investors pursuing their own instincts or trading with undue confidence.
It is also worth asking how low- and high-aspiration accounts generated
by behavioral portfolio theory will fare within a competitive investment
environment, given their lack of mean-variance efficiency.215 In strictly
evolutionary terms, the highest probability of survival actually belongs to
log-utility portfolios (i.e., portfolios produced by investors with constant
relative risk aversion), which themselves do not typically lie on the mean-
variance frontier.216 A subpopulation of traders optimizing according to
conventional mean-variance criteria and the CAPM, subsisting within a
broader, heterogeneous population of traders following other algorithms,
will vanish in the long run.217
Indeed, one model predicts that the entry of investors with constant
relative risk aversion (logarithmic utility) is a sufficient condition for the
long-run extinction of mean-variance optimizers.218 The reasoning is that
very low and very high levels of risk aversion both lead to portfolios
that are low in fitness.219 High levels of risk aversionessentially, those
exceeding constant relative risk aversionlead to low expected returns,
while levels of risk aversion close to zero, essentially the risk neutrality
associated with the conventional CAPM, lead to low long-run wealth
because wealth declines precipitously during disasters.220 Accounting for
gamblers ruin and sequence-of-returns risk leads to the conclusion that
behavioral portfolio theory, relative to the platonic ideal of mean-variance
optimization in the absence of behavioral considerations, has mixed
evolutionary consequences.221
10.7 Behavioral Portfolio Theory asaForm

ofValue-at-Risk (VaR) Analysis
Behavioral portfolio optimization according to SP/A theory is closely

related to VaR analysis,222 a species of risk analysis introduced in
Chapter 12 of Postmodern Portfolio Theory.223 In SP/A and VaR analy-
sis alike, optimization involves tradeoffs between expected wealth and
probabilities of falling short of an aspiration level.224 The only differ-
ence is the tail of the distribution at issue. The usual aspiration level in
VaR analyses is a poverty level; the goal is to combine a low probabil-
ity of falling below a poverty level with the highest possible expected
wealth.225 A behavioral portfolio theory based on SP/A theory, by
contrast, consists of a VaR framework that corresponds to the high
aspiration framework of SP/A theory, where the aspiration level per-
tains to riches rather than to poverty.226
Closer analytic-mapping between the different probability formu-
lations of the trade-off between risk and return in VaR analysis and
behavioral portfolio theory shows the mathematical equivalence of these
two methodologies.227 The earliest generalizations of Roys safety-first
criterion228 sought to honor the investors desire that the return on a
portfolio should not fall below a level H with more than probability.229
At least for normal distributions, this inquiry is connected to VaR and
is the same as saying, VaR = H.230 The full implementation of SP/A
theory under behavioral portfolio theory enables investors to maximiz[e]
expected returns subject to a constraint that the probability of failing to
reach a threshold level H not exceed a specified maximum probability
.231 This inquiry is the same as expected wealth maximization with a
VaR constraint.232
[M]aximizing expected wealth subject to a maximum probability of
failing to reach a threshold level of return under behavioral portfolio
theory is directly analogous to conventional mean-variance optimizations
goal of minimizing variance, subject to a level of return.233 Under nar-
row conditionsnamely, when investors are quadratic utility maximizers
or returns are multivariate normalbehavioral portfolio optimization
maps directly into the Markowitz mean-variance problem and gener-
ates optimal portfolios that are mean-variance efficient.234 Under other
circumstances, however, the parameters that an investor specifies may
elude a feasible solution to behavioral portfolio optimization using VaR
analysis.235
266 J.M. CHEN
Without loss of generality, we may use the simplest form of parametric

VaR analysis according to a Gaussian distribution to illustrate how VaR can
help an investor achieve high-aspiration level under SP/A and behavioral
portfolio theory.236 I draw upon an example elaborated in 13.313.4 of
Postmodern Portfolio Theory. Suppose that an investor stakes $1 million on
an index fund tracking the S&Ps 500.237 She asks her financial advisor, If
capital markets go down to an extent witnessed only once in a hundred
trading days, what can I lose by tomorrows market close? Adjusted to
the appropriate scale, this problem represents virtually every question of
financial market risk.
To answer our investors question, an advisor using conventional
parametric VaR may assume a mean daily return of 0, with a standard devi-
ation over that interval of 100 basis points (bps) (equal to 1%). On those
assumptions, that advisor will report a one-day value of VaR0.01 as $23,264
for a $1 million portfolio. VaR0.01=$23,264 is a fancy, technocratic way
of telling this investor that she faces a 1% chance of losing at least $23,264
from her S&P 500 index fund on any given trading day. Equivalently, the
advisor could tell the investor client that her portfolio has a 99% chance
tomorrow (after a single trading day) of being worth at least $976,736
($1,000,000$23,264).
In formal terms, VaR for a certain risk or confidence level is the quantile
that satisfies the confidence level, , in the following equation:238
-VaR
e= f ( x ) dx
-
For the investor with a $1 million position in an S&P 500 Index fund,
= 10.01, or 0.99. f(x) refers to the probability density functionin this
case, of the distribution of returns on the S&P 500 Index fund. VaR may
also be defined as the greatest lower bound (infimum) on the cumulative
distribution function F of any financial position Y, expressed as a real-
valued, random variable:239
VaRa (Y ) = - inf { x |FY ( x ) a }

Parametric VaR analysis requires the computation of statistical quan-

tiles.240 The quantile function of a distribution is the inverse of its
cumulative distribution function, which in turn is the definite integral
of the probability density function. For normally distributed returns, the

quantile function is designated by the inverse of the capital phi symbol
that designates the cumulative distribution function of the Gaussian distri-
bution: 1(p). The quantile function of the standard normal distribution,
also known as the probit function,241 is expressed as a transformation of
the inverse error function:242
z p = F -1 ( p ) = 2 erf -1 ( 2 p - 1)

Conventional VaR notation designates the quantile function as zp.

There are four alternate ways for referring to the same mathematical con-
ceptquantile function, inverse cumulative distribution function, probit,
and zp. Formally, [t]he quantile zp represents such a value that a standard
normal random variable X has the probability of exactly p to fall inside the
(, zp] interval.243 In effect, we are asking what standard score, or z,244
corresponds to the value of the cumulative distribution function repre-
senting a certain percentage of the total under the curve that defines the
probability density function of the returns on an investment.
We can now complete our simple VaR analysis of the investor who has
staked $1 million on an S&P 500 index fund, where mean daily return ()
is 0 and the standard deviation of that mean return () is 100bps (0.01).
The variable VaRp expresses the VaR given a particular probability of a
loss as the product of zp, standard deviation , and the total value of the
portfolio (v):245
VaR = - z p s v

This definition of VaR for a single time period enables us to recast

behavioral portfolio theory in terms of VaR analysis. Defining p as , the
lower bound on the probability of success for attaining aspiration level H,
and substituting the quantile function of a Gaussian distribution yields:
VaR = -F -1 (a ) s v

Aspiration level H can be expressed as the expected return on a port-

folio, plus the VaR value associated with the variance-based risk profile of
that portfolio:
268 J.M. CHEN
H = m v - F -1 (a ) s v
H = v m - F -1 (a ) s

It is quite evident that the value of H, given a choice of investments, depends

solely on the second factor in the final equation above, 1(). The
choice between investments therefore hinges on the value of each invest-
ments expected rate of return and the variance of that rate.
A simple worked example illustrates how this criterion may favor an
investment with lower expected return but higher variance. Consider
an investor whose desired probability of achieving aspiration level H is
0.25. Since =0.25, the investment that produces the highest value of
1(0.25) is the investment that provides the best chance of attaining
H. 1(0.25)0.67449. Assume further that our investor has exactly
two choices: an exchange-traded fund (ETF) tracking the S&P 500 Index
or common stock in a biotech company with a hotly controverted p atent
case that has reached the Supreme Court.246 The S&P 500 ETF is expected
to return 6%, with a standard deviation of 12%. Given the vagaries of the
drug approval process, the biotech stock is actually expected to report a
negative 1% return. But the corresponding standard deviation of the bio-
tech stocks return is 25% (Table 10.1). If our investors goal is to maxi-
mize her chances of reaching her aspiration level within a year, she will
choose the investment that generates the higher value for 1(0.25).
Despite a lower expected return, the higher variance of the biotech stock
gives our investor a better chance to reach her aspiration level. Indeed, on
the foregoing assumptions, our hypothetical investor would choose the
more speculative biotechnology stock for any value of , the lower bound
on the probability of attaining aspiration level H, up to nearly 0.30:
Table 10.1 The VaR-style evaluation of alternative equity investments under

SP/A theory
Investment 1(0.25)
S&P 500 index 6% 12% 0.1409

Biotech stock 1% 25% 0.1586
0.06 - F -1 (a ) 0.12 = -0.01 - F -1 (a ) 0.25

0.07 = F -1 (a ) 0.13
7
F -1 (a ) = -
13
7
2 erf -1 ( 2a - 1) = -
13
1 7
a = 1 - erf
2 13 2
a 0.2951
The higher probability of reaching the aspiration level, projected from

the level of individual decision-making to its firm-level equivalent, read-
ily explains Edward Bowmans observation that troubled firms are more
willing to take risk.247 This framing of behavioral prospect theory as a
variation on the theme of VaR analysis likewise confirms the entire class of
risk-seeking behaviors predicted by prospect theorys fourfold pattern and
observed in practice in settings as diverse as state-sponsored lotteries and
initial public offerings.248
This sections upshot is as noteworthy as it is succinct. The mathematics
of behavioral portfolio theory bears more than a passing resemblance to
the mathematics of conventional risk management. VaR analysis, which is
traditionally enlisted for tempering unreasonably risky trading behavior by
banks, supplies in this instance the tool that guides an investor to choose
a much riskier stock, with lower, even negative expected returns, over the
market portfolio. Thanks to these methodological similarities, behavioral
portfolio theory proves vulnerable to the pitfalls of VaR analysis, especially
its perverse propensity to forgo diversified portfolios in favor of the bonds-
and-bullets allocations that typify portfolios following the behavioral pre-
scriptions of SP/A theory.249
Notes
1. See generally Chen, Postmodern Portfolio Theory, supra note 1 (Chapter 1).
2. See, e.g., Tore Ellingsen, Cardinal Utility: A History of Hedonimetry, in
Cardinalism: A Fundamental Approach 105165 (Maurice Allais &
Ole Hagen eds., 1994); Hagen, supra note 24 (Chapter 3); Ole
270 J.M. CHEN
Hagen, Neo-Cardinalism, in Progress in Utility and Risk Theory

145164 (Ole Hagen & Fred Wenstp eds., 1984).
3. Daniel Kahneman & Amos Tversky, Choices, Values, and Frames, 39 Am.
Psychologist 344350, 344 (1984).
4. Lola L.Lopes, Between Hope and Fear: The Psychology of Risk, 20 Advances
Experimental Soc. Psych. 255295, 283 (1987).
5. Oliver Wendell Holmes, Jr., The Common Law 1 (Sheldon M.Novick
intro., 1991) (1st ed. 1881).
6. Guthrie, supra note 45 (Chapter 8), at 1163.
7. Lopes, Between Hope and Fear, supra note 4, at 268.
8. Id.
9. Id.
10. See Thierry Post & Haim Levy, Does Risk Seeking Drive Stock Prices? A
Stochastic Dominance Analysis of Aggregate Investor Preferences and
Beliefs, 18 Rev. Fin. Stud. 925953 (2005).
11. Cf. Charles F.Manski, Measuring Expectations, 72 Econometrica 1329
1376 (2004) (advocating the direct measurement of subjective probabili-
ties in place of the traditional economic practice of inferring decision
processes from data on observed choices).
12. G.E.P. Box, J. Stuart Hunter & William G. Hunter, Statistics for
Experimenters: Design, Innovation, and Discovery 440 (2d ed. 2005);
see also id. at 208, 384; G.E.P. Box & Norman R. Draper, Empirical
Model-Building and Response Surfaces 74 (1987) (Remember that all
models are wrong; the practical question is how wrong do they have to
be to not be useful.); id. at 424 (Essentially, all models are wrong, but
some are useful.).
13. G.E.P. Box, Robustness in the Strategy of Scientific Model Building, in
Robustness in Statistics 201236, 203 (Robert L. Launer & Graham
N.Wilkinson eds., 1979). The ideal gas law states that PV=nRT for an
ideal gas, where P, V, and T designate pressure, volume, and tempera-
ture, respectively; n indicates the amount of gas in moles; and R is the
universal gas constant. See generally J. Patrick Abulencia & Louis
Theodore, Fluid Flow for the Practicing Chemical Engineer 109119
(2011).
15. Id.
16. Id.
17. Id.
18. Id. at 256.
19. Id.
20. Id. (citing Clyde H. Coombs, Portfolio Theory and the Measurement of
Risk, in Human Judgment and Decision Processes 6385 (Martin
F.Kaplan & Steven Schwartz eds., 1975)).

22. Id. But see Charles Clotfelter & Philip Cook, Selling Hope 76 (1989)
(reporting survey data showing that lottery players tend to underestimate
rather than overestimate their probability of winning).
24. Kahneman & Tversky, Prospect Theory, supra note 90 (Chapter 6), at
278279; accord Lopes, Between Hope and Fear, supra note 4, at 262.
26. Id. at 262
27. Id.
29. Id. (citing, inter alia, Harry M.Markowitz, Portfolio Selection: Efficient
Diversification of Investments (1959); Coombs, supra note 20; Hagen,
supra note 2).
31. Id. (citing Hagen, supra note 2).
32. See, e.g., Scott & Howath, supra note 52 (Chapter 3); Jondeau &
Rockinger, supra note 26 (Chapter 3); Athayde & Flres, supra note 25
(Chapter 3). See generally supra 3.2, at 6066.
33. See supra 8.48.5, at 189198; cf. Lopes, Between Hope and Fear,
supra note 4, at 269275 (illustrating emotional responses to lotteries
through Lorenz curves, a distinct method for visually depicting risk and
human reaction to it).
35. Id.
36. Id.
37. Id.
38. See Das, Markowitz, Scheid & Statman, supra note 42 (Chapter 2), at
325.
40. Lola L. Lopes, Risk and Distributional Inequality, 10 J. Experimental
Psych. 465485 (1985); accord Lopes, Between Hope and Fear, supra
note 4, at 275.
42. Id. (quoting Walter Mischel, Personality and Assessment 296 (1968)).
44. Id. at 276277.
45. Id. at 275.
46. Id.
47. See generally Jay Appleton, The Experience of Landscape (rev. ed. 1996).
48. Richard Gonzalez & George Wu, On the Shape of the Probability Weighting
Function, 38 Cognitive Psych. 119166, 139 & n.3 (1999).
272 J.M. CHEN
49. See id. at 139.

50. Fehr-Duda, de Gennaro & Schubert, supra note 140 (Chapter 1), at
294; see also id. ([P]arameter largely governs the elevation of the
curve, while largely governs its slope: the smaller the value of , the
more curved (flatter in the range of medium probabilities and steeper
near the ends) the w(p) curve). See generally Pamela K. Lattimore,
Joanna R. Baker & Ann Dryden Witte, The Influence of Probability on
Risky Choice: A Parametric Examination, 17 J. Econ. Behav. & Org.
377400, 380382 (1992).
294.
52. Id.
53. Id. at 294295.
54. Id. at 295.
55. See Fehr-Duda, de Gennaro & Schubert, supra note 140 (Chapter 1), at
294; Gonzalez & Wu, supra note 48, at 139.
56. Lattimore, Baker & Witte, supra note 50, at 382; see also id. at 382 n.3
(defining subadditivity and subproportionality as other properties that a
probability weighting function should satisfy). See generally Uzi Segal,
Some Remarks on Quiggins Anticipated Utility, 8J.Econ. Behav. & Org.
145154 (1987).
294.
58. Lopes, Between Hope and Fear, supra note 140 (Chapter 1), at 276.
59. Ellsberg, Savage Axioms, supra note 75 (Chapter 6), at 644; accord
Lopes, Between Hope and Fear, supra note 4, at 276.
60. Lopes, Between Hope and Fear, supra note 4, at 276 (quoting Ted
Thackrey, Jr., Gambling Secrets of Nick the Greek 67 (1968)).
61. Ecclesiastes 9:11 (Revised Standard Version).
62. Lopes, Between Hope and Fear, supra note 4 (Chapter 1), at 275. See
generally Sidney Siegel, Level of Aspiration and Decision Making, 64
Psych. Rev. 253262 (1957); Herbert A.Simon, A Behavioral Model of
Rational Choice, 69 Q.J.Econ. 99118 (1955).
64. Id.
65. Lopes & Oden, supra note 55 (Chapter 2), at 291.
67. Id.
68. Id. at 278.
69. Id.
70. Id.
71. Id.
73. See Robert Frock, Changing How the World Does Business: FedExs
Incredible Journey to Success The Inside Story 101 (2006) (quoting
founder Fred Smith: What difference does it make? Without the funds
for the fuel companies, we could not have flown anyway.).
74. See, e.g., Bowman, Risk/Return Paradox, supra note 20 (Chapter 4);
Bowman, Risk Seeking by Troubled Firms, supra note 21 (Chapter 4);
Andersen, Denrell & Bettis, supra note 35 (Chapter 4).
75. See supra 8.8, at 203207. See generally Lopes & Oden, supra note 55
(Chapter 2).
76. Lopes, Between Hope and Fear, supra note 4 (Chapter 1), at 284; see also
Sandra L. Schneider & Lola L. Lopes, Reflection in Preferences Under
Risk: Who and When May Suggest Why, 12J.Experimental Psych. 533
548 (1986).
78. Id. at 279.
79. See Thomas C.Schelling, Self-Command in Practice, in Policy, and in a
Theory of Rational Choice, 74 Am. Econ. Rev. 111 (1984).
80. William Faulkner, Acceptance, in Nobel Lectures, Literature 19011967,
at 444445, 444 (Horst Frenz ed., 1969).
82. Id.
83. Id. at 280.
84. Id.
85. Id. at 281; see also id. at 285.
86. Id. at 282.
87. Id.
88. Id.
89. Id. at 283.
90. Id. at 283284.
91. Id. at 284; see also id. at 286 (observing that risk-averse and risk-seeking
people are basically doing the same thing, but their values differ at least
on the security/potential factor, and probably on aspiration level as
well).
92. Marc Oliver Rieger, SP/A and Cumulative Prospect Theory: A
Reconciliation of Two Behavioral Decision Theories, 108 Econ. Letters
327329, 327 (2010).
93. Hersh Shefrin, A Behavioral Approach to Asset Pricing 429 (2d ed.
2008).
94. Id.
95. Id.
97. Rieger, supra note 92, at 327.
274 J.M. CHEN
99. Id. at 311. See generally Charles F. Manski, Ordinal Utility Models of
Decision Making Under Uncertainty, 25 Theory & Decision 79104
(1988).
100. Shefrin, supra note 93 (Chapter 1), at 429.
101. Shefrin & Statman, Behavioral Portfolio Theory, supra note 24 (Chapter
2), at 127.
102. Id. at 131.
103. Id.
104. Id.
105. Id.
107. See Arthur D.Roy, Safety First and the Holding of Assets, 20 Econometrica
431449 (1952). See generally Chen, Postmodern Portfolio Theory,
supra note 1 (Chapter 1), 4.2.
108. See Roy, Safety First, supra note 107 (Chapter 1), at 433.
109. See Arthur D. Roy, Risk and Rank in Safety-First Generalized, 23
Economica 214228 (1956). See generally Haim Levy & Marshall Sarnat,
Safety FirstAn Expected Utility Principle, 7J.Fin. & Quant. Analysis
18291834 (1972); David H.Pyle & Stephen J.Turnowsky, Safety-First
and Expected Utility Maximization in Mean-Standard Deviation Portfolio
Analysis, 52 Rev. Econ. & Stat. 7581 (1970).
110. Roy, Safety First, supra note 107 (Chapter 1), at 433; accord Levy, The
CAPM Is Alive and Well, supra note 10 (Chapter 3), at 45 n.1.
111. Harry M. Markowitz, Mean-Variance Analysis in Portfolio Choice and
Capital Markets 37 (1987). The other paper was Harry M.Markowitz,
Foundations of Portfolio Theory, 46 J. Fin. 469477 (1991). Other
sources trace the origins of portfolio selection to Helen Makower &
Jacob Marschak, Assets, Prices and Monetary Theory, 5 Economica 261
288 (1938) and Jacob Marschak, Money and the Theory of Assets, 6
Econometrica 311325 (1938). See Justin Fox, The Myth of the Rational
Market: A History of Risk, Reward, and Delusion on Wall Street 347
(2009).
112. James C.T. Mao, Survey of Capital Budgeting: Theory and Practice,
25J.Fin. 349360, 354 (1970).
113. Balzer, supra note 47 (Chapter 3), at 115.
114. V.I.Norkin & S.V.Boyko, Safety-First Portfolio Selection, 48 Cybernetics
& Sys. Analysis 180191, 180 (2012).
2), at 130.
116. Id.
117. See id.
118. See Shinji Kataoka, A Stochastic Programming Model, 31 Econometrica
181196 (1963).
119. See Leslie G.Telser, Safety-First and Hedging, 23 Rev. Fin. Stud. 116
(1955).
120. See Roy, Risk and Rank, supra note 109 (Chapter 1).
121. See Enrique R.Arzac & Vijay S.Bawa, Portfolio Choice and Equilibrium
in Capital Markets with Safety-First Investors, 4J.Fin. Econ. 277288
(1977).
2), at 131.
123. See, e.g., Yuanyao Ding & Bo Zhang, Optimal Portfolio of Safety-First
Models, 139 J. Stat. Planning & Inference 29522962 (2009); Jos
Encarnacin, Jr., Portfolio Choice and Risk, 16J.Econ. Behav. & Org.
347353 (1991); Haim Levy & Moshe Levy, The Safety First Expected
Utility Model: Experimental Evidence and Economic Implications,
33 J. Banking & Fin. 14941506 (2009) (proposing an optimization
approach that uses the weighted average of expected utility and a safety-
first maximization of utility).
124. See D.Li, T.F.Chan & W.L.Ng, Safety-First Dynamic Portfolio Selection,
4 Dynamics Continuous, Discrete & Impulsive Sys. 585600 (1998);
Wei Yan, Continuous-Time Safety-First Portfolio Selection with Jump-
Diffusion Processes, 43 IntlJ.Sys. Sci. 622628 (2012).
125. See Dennis W.Jansen, Kees G.Koedijk & Casper G. de Vries, Portfolio
Selection with Limited Downside Risk, 7 J. Empirical Fin. 247269
(2000).
126. See Mahfuzul Haque, M. Kabir Hassan & Oscar Varela, Safety-First
Portfolio Optimization for U.S.Investors in Emerging Global, Asian and
Latin American Markets, 12 Pac. Basin Fin. J. 91116 (2004); Mahfuzul
Haque, Oscar Varela & M.Kabir Hassan, Safety-First and Extreme Value
Bilateral U.S.-Mexican Portfolio Optimization Around the Peso Crisis and
NAFTA in 1994, 47 Q.Rev. Econ. & Fin. 449469 (2007).
127. See Shefrin & Statman, Behavioral Portfolio Theory, supra note 24
(Chapter 2), at 131132.
128. Id. at 131.
129. See id. at 132.
130. Id.
131. Lopes & Oden, supra note 55 (Chapter 2), at 291. See generally Gregg
C.Oden & Lola L.Lopes, Risky Choice with Fuzzy Criteria, in Qualitative
Aspects of Decision Making 5682 (Roland W.Scholz & Alf C.Zimmer
eds., 1997).
132. Rieger, supra note 92 (Chapter 1), at 328.
134. Id.
135. Id.
136. Id.
276 J.M. CHEN
137. Id.
138. Rieger, supra note 92 (Chapter 1), at 327.
139. See, e.g., Antoni Bosch-Domnech & Joaquim Silvestre, Reflection on
Gains and Losses: A 227 Experiment, 33J.Risk & Uncertainty 217
235 (2006).
140. See generally Thorsten Hens & Marc Oliver Rieger, Financial
Economics A Concise Introduction to Classical and Behavioral
Finance 7580 (2010).
142. Id.
143. Id.
144. Jos Encarnacin, Portfolio Choice and Risk, 16J.Econ. Behav. & Org.
347353, 352 (1991).
146. Id. See generally Lola Lopes, Algebra and Process in the Modeling of Risky
Choice, in 32 Decision Making from a Cognitive Perspective 177220
(Jerome Busemeyer, Reid Hastie & Douglas L.Medin eds., 1995).
2), at 132.
148. Id.
149. Id.
150. Id.
151. Id.
152. Id.
153. See id.
154. See generally Wade D. Pfau & Jeremy Cooper, The Yin and Yang of
Retirement Income Philosophies (Nov. 10, 2014) (available at http://ssrn.
155. See Philip L. Cooley, Carl M. Hubbard & Daniel T. Walz, Retirement
Savings: Choosing a Withdrawal Rate That Is Sustainable, 20:2 Am. Assn
Indiv. Investors J. 1621 (Feb. 1998). Because these authors taught
finance at Trinity University of Texas, this article is known as the Trinity
study. See also Philip L.Cooley, Carl M.Hubbard & Daniel T.Walz,
Portfolio Success Rates: Where to Draw the Line, 24:4 J. Fin. Planning
4860 (April 2011).
156. See William P.Bengen, Determining Withdrawal Rates Using Historical
Data, 7:4J.Fin. Planning 171180 (Oct. 1994); cf. sources cited supra
notes 3031 (Chapter 5) (describing sequence-of-returns risk, which
may undermine assumptions about safe withdrawal rates).
157. See, e.g., Wade D.Pfau, An International Perspective on Safe Withdrawal
Rates from Retirement Savings: The Demise of the 4 Percent Rule,
23:12 J. Fin. Planning 5261 (Dec. 2010); Jason S. Scott, William
F.Sharpe & John G.Watson, The 4% Rule At What Price? (April 2008)
(available at http://www.stanford.edu/~wfsharpe/retecon/4percent.
pdf and http://ssrn.com/abstract=1115023) (decrying the 4% rule, the
advice most often given to retirees for managing spending and investing,
for prescribing a constant, non-volatile spending plan according to a
risky, volatile investment strategy that leads retirees to accumulate
unspent surpluses when markets outperform and to face spending
shortfalls when markets underperform).
158. See, e.g., Zvi Bodie, On the Risks of Stocks in the Long Run, 51:3 Fin.
Analysts J. 1822 (May/June 1995); Jason K. Branning & M. Ray
Grubbs, Using a Hierarchy of Funds to Reach Client Goals, 23:12J.Fin.
Planning 3133 (Dec. 2010).
160. See id.
161. See id. at 137.
162. Id.
163. See generally, e.g., Hirotugu Akaike, Information Theory and an Extension
of the Maximum Likelihood Principle, in Second International Symposium
on Information Theory 267281 (B.N.Petrov & F.Cski eds., 1973);
Hirotugu Akaike, A New Look at the Statistical Model Identification, 19
IEEE Transactions on Auto. Control 716723 (1974); Wei Pan, Akaikes
Information Criterion in Generalized Estimating Equations, 57
Biometrics 120125 (2001).
2), at 128.
165. See Das, Markowitz, Scheid & Statman, supra note 42 (Chapter 2), at
314.
166. Id.
167. See id. at 332 n.10; Alexandre M.Baptista, Optimal Delegated Portfolio
Management with Background Risk, 32 J. Banking & Fin. 977985
(2008).
(Chapter 2), at 128, 135.
169. See id. at 135. See generally sources cited supra notes 1518 (Chapter 5)
(describing the separation theorem).
2), at 128.
171. Id.
172. Id. at 139.
173. See id. at 142. See generally, e.g., John B.Shoven & Clemens Sialm, Asset
Location in Tax-Deferred and Conventional Savings Accounts, 88J.Pub.
Econ. 2338 (2004); Edward A. Zelinsky, The Defined Contribution
Paradigm, 114 Yale L.J. 451534 (2004).
278 J.M. CHEN
174. See supra 2.2, at 3135.

2), at 144.
176. See Marc F.Bellemare & Zachary S.Brown, On the (Mis)use of Wealth as
a Proxy for Risk Aversion, 92 Am. J.Agric. Econ. 273282 (2010).
177. Encarnacin, supra note 144 (Chapter 1), at 352.
178. Id.
179. Tversky & Kahneman, Advances in Prospect Theory, supra note 58
(Chapter 8), at 317 (emphasis in original).
2), at 139.
181. See id. at 140.
182. See id.
183. Das, Markowitz, Scheid & Statman, supra note 42 (Chapter 2), at 312.
2), at 137.
185. Hagen, Separation of Cardinal Utility, supra note 24 (Chapter 3), at
100.
2) at 137; cf. Hagen, Separation of Cardinal Utility, supra note 24
(Chapter 3), at 101 (observing that many peoplein all brackets of
wealth may be willing to pay administrative costs and profit for lotter-
ies as well as insurance companies).
187. See Lopes & Oden, supra note 215 (Chapter 2), at 307.
188. Id.
189. Id.
190. See Hartzmark, supra note 28 (Chapter 2).
191. See generally John Kenneth Galbraith, The Affluent Society (4th ed.
1984) (1st ed. 1958).
193. Id.
194. Id.
195. Id. See generally Jack R. Anderson, Perspective on Models of Uncertain
Decisions, in Risk, Uncertainty, and Agricultural Development 3962
(James A.Roumasset, Jean-Marc Boussard & Inderjit Singh eds., 1979).
2), at 137.
197. Id.
198. See Howard Kunreuther & Gavin Wright, Safety First, Gambling, and the
Subsistence Farmer, in Risk, Uncertainty, and Agricultural Development,
supra note 35 (Chapter 2), at 213230; Lopes, Between Hope and Fear,
supra note 4 (Chapter 1), at 287.
199. Sutti Ortiz, The Effect of Risk Aversion Strategies on Subsistence and Cash
Crop Decisions, in Risk, Uncertainty, and Agricultural Development,
supra note 35 (Chapter 2), at 231246, 235; accord Lopes, Between Hope
and Fear, supra note 4 (Chapter 1), at 287.
2), at 137 (emphasis in original).
201. Quazi Shahabuddin, Farmers Crop Growing Decisions Under
Uncertainty A Safety-First Approach, 10 Bangladesh Dev. Stud.
95100, 95 (1982).
202. See generally Victor A.B.Davies, Sierra Leone: Ironic Tragedy, 9J.African
Econs. 349369 (2000); Victor A.B. Davies, Sierra Leones Economic
Growth Performance, 19612000, in 2 The Political Economy of
Economic Growth in Africa, 19602000: Country Case Studies 660
696 (Benno J.Ndulu etal. eds., 2008).
203. Victor A.B.Davies, Alluvial Diamonds: A New Resource Curse Theory 3
(Feb. 2009) (available at http://www.csae.ox.ac.uk/conferences/2009-
EDiA/papers/128-Davies.pdf).
204. Id. See generally Akerlof & Romer, supra note 89 (Chapter 9); White,
205. See, e.g., Roy Maconachie & Tony Binns, Beyond the Resource Curse?
Diamond Mining, Development and Post-Conflict Reconstruction in
Sierra Leone, 32 Resources Poly 104115 (2007).
2), at 141.
207. Id.
208. Id. at 145.
209. Id.
210. Id.
211. See id. at 128, 135.
212. Das, Markowitz, Scheid & Statman, supra note 42 (Chapter 2), at 315
(citing William F. Sharpe, Gordon J. Alexander & Jeffrey V. Bailey,
Investments 194 (6th ed. 1999)).
213. Das, Markowitz, Scheid & Statman, supra note 42 (Chapter 2), at 315
(citing Jean L.P.Brunel, How Suboptimal If at All Is Goal-Based
Asset Allocation?, 9:2 J. Wealth Mgmt. 1934 (Fall 2006)); cf.
E.Ballestero, M.Gnther, D.Pla-Santamaria & C.Stummer, Portfolio
Selection Under Strict Uncertainty: A Multi-Criteria Methodology and Its
Application to the Frankfurt and Vienna Stock Exchanges, 181 Eur.
J.Oper. Research 14761487, 1477 (2007) (applying compromise pro-
gramming model[s] such as those described by E. Ballestero &
C. Romero, Portfolio Selection: A Compromise Programming Solution,
280 J.M. CHEN
47J.Oper. Research Socy 13771386 (1996), to determin[e] the util-

ity function for an individual investor in order to select the most attrac-
tive point on that frontier for that investor).
(Chapter 2), at 146; cf. Lawrence E.Blume & David Easley, Evolution
and Market Behavior, 58J.Econ. Theory 940 (1992) (establishing an
evolutionary framework for predicting the long-term fitness and survival
of competing portfolios).
217. Emanuela Sciubba, The Evolution of Portfolio Rules and the Capital Asset
Pricing, 29 Econ. Theory 123150 (2006); accord Shefrin & Statman,
Behavioral Portfolio Theory, supra note 24 (Chapter 2), at 146.
219. Id.
220. Id. at 147.
221. See id. at 148.
222. See id. at 140.
1), 13.113.4, at 247259.
2), at 140.
225. Id.
226. Id.
On portfolio optimization within VaR constraints, id. at 315 n.3, see
generally Gordon J.Alexander & Alexandre M.Baptista, Active Portfolio
Management with Benchmarking: Adding a Value-at-Risk Constraint,
32J.Econ. Dynamics & Control 779820 (2008); Gordon J.Alexander,
Alexandre M.Baptista & Shu Yan, Mean-Variance Portfolio Selection with
At-Risk Constraints and Discrete Distributions, 31J.Banking & Fin.
37613781 (2007); Suleyman Basak & Alex Shapiro, Value-at Risk
Based Risk Management: Optimal Policies and Asset Prices, 14 Rev. Fin.
Stud. 371405 (2001).
228. See Telser, supra note 119 (Chapter 1).
230. Id.
231. Id.
232. Id.
233. Id. at 321.
234. Id. at 318.
235. See id. at 315, 324325.

236. See id. at 319.
237. This example is drawn from Linda Allen, Jacob Boudoukh & Anthony
Saunders, Understanding Market, Credit, and Operational Risk: The
Value at Risk Approach 67 (2004) and more generally from the same
source at 120. The ensuing discussion in text also draws upon (and
considerably simplifies) the analysis outlined in Philippe Jorion, Value at
Risk: The New Benchmark for Managing Financial Risk 106113 (3d ed.
2006). I develop a similar model in greater detail in Chen, Postmodern
Portfolio Theory, supra note 1 (Chapter 1), 13.313.4, at 251254.
See generally Simon Benninga & Zvi Wiener, Value at Risk (1998);
Philippe J.S. de Brouwer, Understanding and Calculating Value at Risk,
6 J. Derivatives Use, Trading & Reg. 306322 (2001); Winfried
G.Hallerbach, Decomposing Portfolio Value-at-Risk: A General Analysis,
5:2 Risk 1 (Winter 2002).
238. See Jn Danelsson & Jean-Pierre Zigrand, On Time Scaling of Risk and
the Square-Root-of-Time Rule, 30J.Banking & Fin. 27012713, 2702
n.1 (2006).
239. See Johanna F.Ziegel, Coherence and Elicitability, at 1 (March 8, 2013)
(available at http://arxiv.org/pdf/1303.1690v2) (forthcoming in
Mathematical Finance).
240. See https://en.wikipedia.org/wiki/Quantile_function
241. See https://en.wikipedia.org/wiki/Probit
242. https://en.wikipedia.org/wiki/Normal_distribution.
243. Id.
244. See https://en.wikipedia.org/wiki/Standard_score
245. See Allen, Boudoukh & Saunders, supra note 77 (Chapter 2), at 7.
246. Cf. supra 1.3, at 89 (discussing Association for Molecular Pathology v.
Myriad Genetics Inc., 133S.Ct. 2107 (2013)).
247. See Bowman, Risk Seeking by Troubled Firms, supra note 21 (Chapter 4).
248. See generally supra 8.8, at 203207 (prospect theorys fourfold pattern);
9.19.2, at 215224 (lotteries and lottery-like financial instruments).
249. Compare Jn Danelsson, The Emperor Has No Clothes: Limits to Risk
Modelling, 26J.Banking & Fin. 12731296, 1289 (2002) with Shefrin
& Statman, Behavioral Portfolio Theory, supra note 24 (Chapter 2), at
128 (describing behavioral portfolios as resembl[ing] combinations of
bonds and lottery tickets). See generally Chen, Postmodern Portfolio
Theory, supra note 1 (Chapter 1), 16.2, at 292296 (describing VaRs
failure to satisfy subadditivity and the other conditions required of coher-
ent risk measures); James Ming Chen, Measuring Market Risk Under the
Basel Accords: VaR, Stressed VaR, and Expected Shortfall, 8 Aestimatio
184201, 192195 (2014) (same).
CHAPTER 11
Behavioral Gaps Between Hypothetical

Investment Returns andActual Investor
Returns
11.1 Hypothetical Investment Returns Versus

Actual Investor Returns
Individual investors behave badly. [I]nvestors with strong behavioral
biases or lack of attention to meaningful financial news are more likely to
forgo equity ownership or to participate in capital markets for the wrong
reasons.1 Individual investors, often the product of that lethal combina-
tion of ignorance and bias, trade frequently, tend to time their buys
and sells badly, and prefer high expense [mutual] funds and active funds
rather than index funds.2 Behavioral biases cost investors. Narrow fram-
ing extracts a 2.16% premium from the most heavily affected quintile of
individual investors relative to the least affected quintile.3 The disposition
effect costs the worst quintile 0.89% in annual returns relative to best
quintile.4
This chapter begins with a closer look at losses that biased investors
incur when they chase fund performance, in defiance of financial prin-
ciples guiding rational fund selection decisions.5 Actual investor returns
from mutual funds lag behind those same funds hypothetical investment
returns based on a fixed initial investment and reinvestment of all distribu-
tions. This gap between actual investor returns and hypothetical invest-
ment returns arises from behaviorally driven errors in timing. By chasing
recent returns and abandoning funds at any sign of weakness, actual inves-
tors buy high and sell low.

DOI10.1007/978-3-319-32711-2_11
284 J.M. CHEN
Economic losses from trading by individual investors are not trivial. One
study, focusing exclusively on Taiwan from 1995 to 1999, estimated that
individual investors aggregate losses in the worlds 12th largest financial
market stripped an annual 3.8% in annual return, roughly equivalent to
2.8% of total personal income and a staggering 2.2% of the countrys gross
domestic product.6
The benchmark for investor returns depends on the financial instru-
ment being traded. For trading in individual stocks, one plausible bench-
mark is whether investor returns on brokerage accounts, net of trading
costs, can match a passive index.7 For mutual funds, where differences in
investor preferences are reflected in choices within a galaxy of passively
and actively managed funds,8 close examination of net inflows and out-
flows should reveal whether fund investors are realizing the same return
as a hypothetical buy-and-hold investment in the fund, or whether behav-
iorally motivated decisions to buy and redeem fund shares affect realized
investor returns. This chapter will focus on the latter question. In focusing
on mutual funds, I remain mindful of evidence that individual investors are
likelier to treat winning and losing stocks, as a class, differently from the
way they treat winning and losing mutual funds.9 The difference between
asymmetrical treatment of stocks and asymmetrical treatment of mutual
funds suggest[s] that households motive for mutual funds investment is
different from individual stocks investment.10
Persistent gaps between actual investor returns and hypothetical invest-
ment returns expose a deep flaw in portfolio theorys mechanical reliance
on simple mean-variance optimization. The financial services industry, or
at least the financial academy, should devise and deploy sophisticated mea-
sures of the impact of investor behavior on portfolio-wide performance.
This chapter will formalize the definition of the behaviorally driven inves-
tor gap. After surveying the literature exploring gaps between hypotheti-
cal investment returns and the actual financial performance attained by
real-life investors, this chapter defines a statistic that it will call , which
quantifies this performance gap.
11.2 The Disposition Effect

The financial literatures focus on the impact of investor behavior on
investment returns began no later than the identification by Hersh Shefrin
and Meir Statman of the so-called disposition effect, or the tendency of
investors to ditch winning stocks too soon and to ride losers too long.11
BEHAVIORAL GAPS BETWEEN HYPOTHETICAL INVESTMENT RETURNS... 285
The disposition effect is quite subtle and may not be dismissed merely as
the artifact of an irrational belief that security prices will eventually, inevi-
tably revert to the mean.12
The disposition effect appears to have some relationship to prospect
theory, even if the connection is rather complex. It is crucial to define
prospect theorys contribution to the disposition effect according to an
assumption that investors distinguish between paper and realized gains.13
The more straightforward application of portfolio theory to annual gains
and losses does not predict the disposition effect.14 Notably, even though
the disposition effect is consistent with prospect theory ex postin other
words, differential treatment of winning and losing stocks by an investor,
once it has occurred, would be predicted by prospect theoryit must be
acknowledged that prospect theory cannot explain how investors who sell
winners and hold losers came to buy stocks in the first place.15
The disposition effect does arise to some degree from prospect theorys
S-shaped utility function, insofar as an investor realizing gains is risk
averse and therefore unwilling to gamble about future uncertain gains,
while risk seeking in the domain of losses will spur an investor to gamble
and postpone realizing losses.16 This purely static argument, however,
consider[s] only a single sale and takes no account of any effects of
reinvestment.17 In the dynamic context of sequential purchases and
sales of stock, prospect theorys sigmoid value function actually serves
to reduce the disposition effect.18 The very act of realizing a loss resets
the reference point for future possible gains.19 The potential to offset
the direct disutility of [a] loss transforms the realization of a loss into
some sort of valuable option.20 Such intertemporal realization util-
ity therefore explains why investors realize any voluntary losses at all,
rather than none.21
Among other consequences of the disposition effect, the tendency to
sell winners and hold losers biases the results of any analysis of the inter-
nal rate of return on round-trip trades toward positive performance.22
The work of Brad Barber and Terrance Odean sharpened the focus on the
mostly negative contribution of investor behavior to actual investment
returns.23 In their review of 66,465 accounts at a discount retail broker-
age from 1991 through 1996, Barber and Odean found a significant
underperformance of 15 to 31 basis points per month, equivalent to 1.8
percent to 3.7 percent per year.24 The disposition effect also emerges
in real estate markets, where powerful emotions associated with homes
make owners extremely reluctant to sell at a loss, relative to their original
purchase price.25
286 J.M. CHEN
Professional investors may temper the impact of the disposition effect,

relative to ordinary individuals,26 but they do not appear to exercise any-
thing approaching omniscient control over the timing of their investment
decisions. One study has found, for instance, statistically significant evi-
dence of market-timing ability by some actively managed equity funds,
but only over three- and six-month horizons and not over 1- or 12-month
horizons.27 Moreover, gains from market timing come with a caveat: funds
exhibiting positive timing activity tend to be more highly concentrated in
their industry-specific allocations.28 Ceteris paribus, higher concentration
suggests that gains from market-timing extract a sacrifice in diversifica-
tion and elevate downside volatility, precisely when other sectors and asset
classes feel downward pressure.
More generally, the extent to which strategic asset allocation determines
portfolio performance is the subject of deep, contentious debate among finan-
cial professionals.29 Whether the contribution of tactical asset allocation is rel-
atively modest (less than 10% of total portfolio return)30 or on equal footing
with strategic allocation, 31 few active managers are able to enhance portfolio
value by accurately timing changes in the relative weight of asset classes.32 A
new, multinational study of active fund managers indicate[s] overall that
timing skill is rare and is found among a small minority of funds.33
Although a few authors have suggested that professional advisors
and/or active fund managers can profitably steer their choices in capital
markets, insofar as stocks held by mutual funds experiencing net inflows
tended to rise, even works finding that certain funds enjoyed superior
performance34 and that certain some managers actually exhibit selection
ability35 ultimately support the importance of identifying and quantifying
gaps in investor performance. The stock selection ability of mutual fund
managers, even if it does beat a purely passive index, must still cover the
fees that their funds charge,36 lest their claims to alpha fail to overcome the
arithmetic of active management.37
What has been said of putatively inefficient market for small-cap stocks
applies, a fortiori, to more liquid segments of the market where alpha is even
more elusive. Abnormal returns, even if they are theoretical available, will
most likely in practice be offset by the costs of active managementnamely
the search costs of identifying inefficiencies in the market and the transaction
costs of trading to exploit those opportunities.38 Moreover, the zero-sum
nature of financial markets means that the presence of exceptional inves-
tors requires the presence of offsetting subpar investors.39 By and large,
the evidence indicates that individual investors are subpar investors.40
11.3 The Behavioral Origins

oftheInvestmentGap
Investment professionals have joined academics in taking interest in mea-

suring gaps in investor performance. Since 1994, Dalbar has published an
annual Quantitative Analysis of Investor Behavior.41 Morningstar pro-
vides investor return information not only to subscribers, but also on
a limited basis to the public.42 By 2007, Morningstar and two academic
works had firmly established the use of net cash flows as the most read-
ily implemented methodology for distinguishing between endogenous
growth in fund assets and changes in net asset value (NAV) more properly
attributed to purchases and redemptions by investors. Almost simultane-
ously, Geoff Friesen and Travis Sapp (on the one hand) and Ilia Dichev
(on the other hand) applied nearly identical net cash flow methodolo-
gies to detect annualized investment performance gaps of roughly 1.5%,
respectively, among equity mutual fund investors43 and stock investors in
the New York Stock Exchange (NYSE) and other markets around the
world.44
Specifically, Friesen and Sapp found that investors in the average equity
mutual fund earn[ed] 0.13% less per month (1.56% annually) than the
fund itself.45 Dichev reported that dollar-weighted returns systematically
underperformed buy-and-hold strategies by 1.3% for NYSE and American
Stock Exchange (AMEX) trades from 1926 to 2002, 5.3% for NASDAQ
from 1973 to 2002, and 1.5% for 19 major stock markets around the
world from 1973 to 2004. An application of Dichevs methodology, as
posted in its prepublication form in 2004, to American equity mutual
funds reported 0.1% and 0.21% monthly degradation in investor perfor-
mance, respectively, for 198490 and 19912003.46 These results bridged
Dichevs work on stocks with Friesen and Sapps later work on mutual
funds.
Extensions of these studies to other markets have found comparable
results for Canada,47 the UK,48 and hedge funds in the USA.49 Writing
with Mercer Bullard, Friesen and Sapp found striking differences in inves-
tor performance among the major retail classes of load-bearing mutual
funds.50 In particular, Class B investors underperformed buy-and-hold
strategies by 2.28% a year.51 Two authors have gone so far as to propose
a trading strategy based on the dumb money effect of mutual fund
investors:
288 J.M. CHEN
[O]n average, retail investors direct their money to funds which invest in
stocks that have low future returns. To achieve high returns, it is best to do
the oppose of those investors. [M]utual fund investors experience total
returns that are significantly lower due to their reallocations. Therefore,
mutual fund investors are dumb in the sense that their reallocations
reduce their wealth on average. We call this predictability the dumb
money effect.52
Given the presence of noisy speculation under the guise of informed

trading53 and the very real, economically significant difference between
smart money and noise trading,54 an informed investors optimal strategy
may well consist of taking large, aggressive positions contrary to those of
nave speculators.55 In the absence of transaction costs (a difficult proposi-
tion, given that transaction costs are what give rise to arbitrage opportu-
nities in the first place),56 informed traders may use a contrarian strategy
to exploit uninformed investors given sufficient time or, if equipped with
appropriate derivatives, in a single round of trading.57 On one hand, a
dumb money strategy may backfire over time horizons as short as a
week; individual investors enjoy strong returns during the next 20 trad-
ing days58 and in the immediate neighborhood of trading announce-
ments.59 On the other hand, stocks heavily favored by individuals beyond
a monthly horizon subsequently underperform relative to stocks that indi-
vidual investors tend heavily to sell.60
The literature on investor-driven gaps in financial performance has
established a clear methodology for identifying the net cash flows, in and
out, of mutual funds and similar securities. This methodology enables aca-
demic researchers and investment professionals to duplicate, enhance, or
supplant outright the proprietary models of investor behavior devised by
Morningstar and Dalbar, which those companies will grudgingly reveal
to nonsubscribers, if at all, in vague (Morningstar) or completely opaque
(Dalbar) terms. Indeed, in light of the contribution of mutual fund rat-
ings by companies such as Morningstar to the aggressive marketing prac-
tices and sufficiently volatile strategies that enable some one-star and
two-star funds to rise phoenix-like from the ashes and join the exalted
ranks of four- and five-star funds, independent economists bear some
obligation to devise their own methodologies for measuring the extent
of investor losses from perverse decisions to sell[] low after having
bought high.61
I therefore turn to the task of formally specifying the dollar-weighted
method of detecting net cash flows into and out of mutual funds and other
investment portfolios.
11.4 Measuring Behavioral Gaps

inInvestment Performance
Generally speaking, investor performance is measured by the periodic

change in the market value of all securities held within an investors
account.62 The investor-specific performance gap may be formally defined
as the net reduction (or increase) in a publicly traded, open-ended mutual
funds internal rate of return based on net cash flows in and out of that
fund, relative to the hypothetical rate of return that would have been real-
ized by an investor who left undisturbed an initial investment in that fund
and reinvested all dividends and capital gains distributions. I shall name
the resulting investor gap (psi), after the Greek letter representing psy-
chology ( + ).
Although this methodology can be and has been applied to other
investment portfolios, I will focus on publicly traded, open-ended mutual
funds in the USA.The nature of the regulatory scheme governing these
securities makes them ideal subjects for a study on the impact of investor
behavior on actual returns. Under Securities and Exchange Commission
(SEC) rules63 implementing the Investment Company Act of 1940,64
mutual fund managers must honor all requests for new investments and for
redemptions at a price based on the current net asset value of the fund
which is next computed after receipt of a tender for redemption or of
an order to purchase or sell.65 Moreover, the NAV of that mutual fund
must be computed at least once daily.66 Mutual funds must issue redemp-
tion payments within seven days, excluding periods during which the
NewYork Stock Exchange is closed other than customary week-end and
holiday closings and other exceptional circumstances.67 Compliance with
the Investment Company Act and the SECs implementing regulations
generates the net cash flow data that academic researchers, investment
professionals, and advisory firms such as Morningstar and Dalbar may use
to contrast the actual experience of investors with buy-and-hold invest-
ment history typically provided by mutual funds to prospective investors.
In the interest of thoroughness, I restate 3.2s definition of the con-
tinuously compounded return or log return of an asset as the natural
logarithm of [the assets] gross return (1+Rt): rt=ln(1+Rt).68 If returns
are computed in logarithmic rather than arithmetic returns, we can express
continuously compounded multiperiod return [as] simply the sum of
continuously compounded single-period returns.69
In light of the foregoing, the rate of return for a mutual fund over k
periods should be expressed thus:
290 J.M. CHEN
rt ( k ) = ln (1 + Rt ) + ln (1 + Rt -1 ) + + ln (1 + Rt - k +1 )
rt ( k ) = rt + rt -1 + + rt - k +1

The foregoing expression is a much simpler and more tractable version of

the periodic rate of return that many funds report on the basis of the total
change in v, the value of the fund over those k periods:
vk
rt ( k ) = k -1
v0

The actual experience of investors in the fund during these k periods

is better reflected by net cash flows in and out of the fund. For any given
period t in the range of 0 to k, net cash flow ct may be expressed as the
change in vt, the NAV of the fund for that period, that cannot be explained
by periodic investment returns on the assets held by the fund:
ct = vt - vt -1 (1 + rt )

The computation of rt(k), the buy-and-hold return for the fund over k
periods, and the sequence of periodic net cash flows enables the computa-
tion of an internal rate of return based on the initial NAV of the fund and
the cash flow sequence for periods 1 to k. Investor return is given by the
internal rate of return, ri, in the following formula for zero net present
value (NPV):70
k cj
NPV = =0
(1 + ri )
n
j =0

Although this calculation may not have an analytical solution, it does

lend itself to computation, especially if we select rt(k) as the initial guess or
seeding value for a numerical computation of the internal rate of return.
Perhaps the simplest and most familiar method for computing an internal
rate of return is the secant method:71
rj - rj -1
rj +1 = rj - NPV j
NPVn - NPVn -1
There are more elaborate and accurate methods.72 In most real-world

applications, software will compute the internal rate of return to an arbi-
trary degree of accuracy. For example, Excel computes ri, the investor-
experienced internal rate of return (IRR), according to the following
specification:
ri = IRR -v0 , c1 , c2 , , cn -1 , cn - vn ; rt ( k )

Armed with ri, the computationally determined internal rate of return,

we may now express the NAV of the fund at any given interval T in the
temporal range from 0 to k as function of the initial net asset value, this
series of net cash flows, and ri itself:73
T
vT = v0 (1 + ri ) + ct (1 + ri )
T T -t
t =1
Alternatively and equivalently:74
T
vT ct
= v0 +
(1 + ri ) (1 + ri )
T T
t =1

One significant extension of this basic model involves accounting for

the effect of distributions such as dividends and capital gains distributions.
Unless we are prepared to ignore the impact of these events on net cash
flow, or we are alternatively willing to adopt an unrealistic assumption
such as 100% or 0% reinvestment of dividends and capital gains, some
modest adjustment of this elegant basic model is warranted. Let dt be the
distributions in any period and bt be the reinvestment rate. Cash flow for
period t may now be stated as:
ct = vt - vt -1 (1 + rt ) - dt bt

292 J.M. CHEN
The final NAV may be expressed as the sum of the compounded value
of the initial NAV, the sum of all net cash flows over the relevant inter-
val, and the sum of all distributions, net of reinvestment, over the same
interval:
k k
vk = v0 (1 + ri ) + ct (1 + ri ) + dt bt
k k -t
t =1 t =1
Most importantly, however, we have a value for . The gap between

hypothetical investment return and actual investor return may be expressed
as the difference between the funds overall rate of return and the IRR
computed from dollar-weighted net cash flows:
y = rt ( k ) - ri

11.5 The Effect ofCapital Gains Taxation

One parting note on the impact of capital gains taxation is warranted.
Studies of investor gaps in performance tend to downplay or ignore the
tax effects associated with high levels of portfolio turnover and with other
mistakes in timing. [B]ecause no existing dataset contains the account-
level tax liabilities incurred on dividends and realized capital gains, exist-
ing studies focus exclusively on pre-tax returns.75 Given that the average
holding period for stocks in American discount brokerages is roughly 16
monthsa figure implied by the finding that average annual turnover is
75% and by the mathematical definition of the holding period as the recip-
rocal of the turnover rate76it is likely that taxes will only exacerbate the
performance penalty from trading.77
Although the legal basis for asymmetrical treatment of capital gains
and losses is rather convoluted, the crucial point is that capital gains are
taxable income, while capital losses, within limits, may be deducted from
other taxable income.78 The disposition effect squarely contradicts the
conventional tax-sensitive strategy of harvesting losses more aggressively
(since they may be deducted from ordinary income) than realizing prof-
its (since these are subject to taxation as short-term or long-term capital
gains).79 The taxation of capital gains has two effects: capital gains tax not
only blunts the utility from the immediate realization of gains, but also
postpones the harvesting of losses by reducing the value of gains relative
to the reference point that would be reset by a realization event.80
In fairness, there is evidence of tax-minimizing loss harvesting,81 con-
sistent with standard financial advice to recoup some investment losses
through the tax codes favorable treatment of capital losses.82 Even the
January effect, an otherwise annoying calendar anomaly, appears to
be an artifact of different countries tax systems and conscious efforts
by investors to minimize their tax exposure.83 Every so often, rationality
does claim a victory, however modest, over cognitive bias and behavioral
heuristics.
Notes
1. Warren Bailey, Alok Kumar & David Ng, Behavioral Biases of Mutual Fund
Investors, 102J.Fin. Econ. 127, 3 (2011).
2. Id.
3. See id.
4. See id.
5. Id.
6. See Brad M.Barber, Yi-Tsung Lee, Yu-Jane Liu & Terrance Odean, Just
How Much Do Individual Investors Lose by Trading?, 53 Rev. Fin. Stud.
609632, 610611 (2009).
7. See Gary E.Schlarbaum, Wilbur G.Lewellen & Ronald C.Lease, Realized
Returns on Common Stock Investments: The Experience of Individual
Investors, 51J.Bus. 299325, 300 (1978).
8. See Ronald T. Wilcox, Bargain Hunting or Star Gazing? Investors
Preferences for Stock Mutual Funds, 76J.Bus. 645663 (2003).
9. See Laurent E.Calvet, John Y.Campbell & Paolo Sodini, Fight or Flight?
Portfolio Rebalancing by Individual Investors, 124 Q.J. Econ. 301348,
334, 342, 346 (2009).
10. Ingersoll & Jin, supra note 51 (Chapter 8), at 753 n.42.
11. See Hersh Shefrin & Meir Statman, The Disposition to Sell Winners Too
Early and Ride Losers Too Long: Theory and Evidence, 40J.Fin. 777782
(1985); see also, e.g., Calvet, Campbell & Sodini, supra note 9; Andrea
Frazzini, The Disposition Effect and Underreaction to News, 61 J. Fin.
21172144 (2006); Hyun-Jung Lee, Jongwon Park, Jin- Yong Lee &
Robert S. Wyer, Disposition Effects and Underlying Mechanisms in
E-Trading of Stocks, 45J.Marketing Research 362378 (2008).
12. See Barberis & Xiong, Realization Utility, supra note 51 (Chapter 8), at
263; Markku Kaustia, Prospect Theory and the Disposition Effect, 45J.Fin.
294 J.M. CHEN
& Quant. Analysis 791812 (2010); Martin Weber & Colin F.Camerer,
The Disposition Effect in Securities Trading: An Experimental Analysis,
33J.Econ. Behav. & Org. 167184 (1998).
13. Nicholas Barberis & Wei Xiong, What Drives the Disposition Effect? An
Analysis of a Long-Standing Preference-Based Explanation, 64J.Fin. 751
784, 753 (2009). See generally Barberis & Xiong, Realization Utility,
14. Barberis & Xiong, Disposition Effect, supra note 13, at 753 (emphasis
added). See generally id. at 76570.
15. See Thorsten Hens & Martin Vlcek, Does Prospect Theory Explain the
Disposition Effect?, 12J.Behav. Fin. 141157 (2006).
16. Ingersoll & Jin, supra note 51 (Chapter 8), at 731.
17. Id.
18. Id. at 732.
19. Id.
20. Id.
21. Id. at 733.
22. Barber & Odean, The Behavior of Individual Investors, supra note 197
(Chapter 7), at 1535 n.1.
23. See Terrance Odean, Are Investors Reluctant to Realize Their Losses?,
53 J. Fin. 17751798 (1998); Terrance Odean, Do Investors Trade Too
Much? 89 Am. Econ. Rev. 12791298 (1999).
24. Brad M.Barber & Terrance Odean, Trading Is Hazardous to Your Wealth:
The Common Stock Investment Performance of Individual Investors,
40J.Fin. 773806, 789 (2000).
25. See David Genesove & Christopher Mayer, Loss Aversion and Seller
Behavior: Evidence from the Housing Market, 116 Q.J.Econ. 12331260
(2001).
26. Cf. Abdellaoui, Bleichrodt & Kammoun, supra note 136 (Chapter 9);
Ravi Dhar & Ning Zhu, Up Close and Personal: Investor Sophistication and
the Disposition Effect, 52 Mgmt. Sci. 726740 (2006).
27. See George J. Jiang, Tong Yao & Tong Yu, Do Mutual Funds Time the
Market? Evidence from Portfolio Holdings, 86 J. Fin. Econ. 724758
(2007).
28. See id. at 727, 744, 750. See generally Marcin Kacperczyk, Clemens Sialm
& Lu Zheng, On the Industry Concentration of Actively Managed Equity
Mutual Funds, 60J.Fin. 19832011 (2005) (devising an industry con-
centration index).
29. See generally, e.g., Bruno Solnik & Dennis McLeavey, Global Performance
Evaluation, in International Investments 469526 (6th ed. 2008); Jeffrey
V.Bailey, Thomas M.Richards & David E.Tierney, Evaluating Portfolio
Performance, in Managing Investment Portfolios: A Dynamic Process
717782 (John L. Maginn, Donald L. Tuttle, Dennis W. McLeavey &

Jerald E.Pinto eds., 3d ed. 2007).
30. See Gary P. Brinson, L. Randolph Hood & Gilbert L. Beebower,
Determinants of Portfolio Performance, 42:4 Fin. Analysts J. 3944 (July/
Aug. 1986); Gary P. Brinson, Determinants of Portfolio Performance II:
An Update, 47:3 Fin. Analysts J. 4048 (May/June 1991); L.Randolph
Hood, Determinants of Portfolio Performance 20 Years Later, 61:5 Fin.
Analysts J. 68 (Sept./Oct. 2005); cf. Meir Statman, The 93.6% Question
of Financial Advisors, 9:1 J. Investing 1620 (Spring 2000) (describing
strategic asset allocation as movement along the efficient frontier and tacti-
cal asset allocation as movement of the efficient frontier).
31. See Roger G. Ibbotson, The Importance of Asset Allocation, 66:2 Fin.
Analysts J. 1820 (March/April 2010); Roger G. Ibbotson & Paul
D.Kaplan, Does Asset Allocation Policy Explain 40, 90, or 100 Percent of
Performance?, 56:1 Fin. Analysts J. 2633 (Jan./Feb. 2000); James
X. Xiong, Roger G. Ibbotson, Thomas M. Idzorek & Peng Chen, The
Importance of Asset Allocation, 66:2 Fin. Analysts J. 1820 (March/April
2010).
32. See generally Roy D. Hendriksson & Robert C. Merton, On Market
Timing and Investment Performance. II. Statistical Procedures for
Evaluating Forecasting Skills, 54J.Bus. 513533 (1981).
33. Andrew Clare, Niall OSullivan, Meadhbh Sherman & Steve Thomas,
Multi-Asset Class Mutual Funds: Can They Time the Market? Evidence from
the US, UK and Canada, Research in Intl Bus. & Fin. (forthcoming
2015) (prepublication draft available at http://ssrn.com/
abstract=2589043).
34. Mark Grinblatt & Sheridan Titman, Mutual Fund Performance: An
Analysis of Quarterly Portfolio Holdings, 62J.Bus. 393416, 415 (1989).
35. Kent Daniel, Mark Grinblatt, Sheridan Titman & Russ Wermers,
Measuring Mutual Fund Performance with Characteristic-Based
Benchmarks, 52J.Fin. 10351058, 1037 (1997).
36. See Eugene F.Fama & Kenneth R.French, Luck Versus Skill in the Cross-
Section of Mutual Fund Returns, 55 J. Fin. 19151947 (2010); Robert
Kosowski, Allan Timmermann, Russ Wermers & Hal White, Can Mutual
Fund Stars Really Pick Stocks? New Evidence from a Bootstrap Analysis,
51J.Fin. 25512595 (2006).
37. William F.Sharpe, The Arithmetic of Active Management, 47 Fin. Analysts
J. 7 (1991); see also Kenneth R. French, The Cost of Active Investing,
63J.Fin. 15371573 (2008).
38. See Richard M.Ennis & Michael D.Sebastian, The Small-Cap Alpha Myth,
28:3J.Portfolio Mgmt. 1116 (Spring 2002).
296 J.M. CHEN
40. Id.
41. See DALBAR, Inc., Research & Communications Division, Quantitative
Analysis of Investor Behavior: Volatility Unchecked (March 2012) (described
at http://www.qaib.com/public/about.aspx; previewed at http://www.
qaib.com/public/downloadfile.aspx?filePath=freelook&fileName=fulledit
ionfreelook.pdf).
42. See Morningstar Investor Return: Morningstar Methodology Paper (Aug.
31, 2010) (available at http://corporate.morningstar.com/us/docu-
ments/MethodologyDocuments/MethodologyPapers/
InvestorReturnMethodology.pdf); Fact Sheet: Morningstar Investor
Return (2006) (available at http://corporate.morningstar.com/US/
documents/MethodologyDocuments/FactSheets/InvestorReturns.pdf).
Morningstars most recent Investor Return report, Russel Kinnel, Mind
the Gap 2015 (Aug. 11, 2015), is available at http://news.morningstar.
com/articlenet/article.aspx?id=710248.
43. See Geoffrey C. Friesen & Travis R.A. Sapp, Mutual Fund Flows and
Investor Returns: An Empirical Examination of Fund Investor Timing
Ability, 31J.Banking & Fin. 27962816 (2007).
44. See Ilia D. Dichev, What Are Stock Investors Actual Historical Returns?
Evidence from Dollar-Weighted Returns, 97 Am. Econ. Rev. 386401
(2007).
45. Friesen & Sapp, supra note 43, at 2802.
46. See Oded Braverman, Shmuel Kandel & Avi Wohl, The (Bad?) Timing of
Mutual Fund Investors? 8 (Aug. 2005) (available at http://www.econ-pol.
unisi.it/labsi/papers2006/Kandel.pdf).
47. See Rajeeva Sinha & Vijay Jog, Fund Flows and Performance: A Study of
Canadian Equity Funds (2004) (available at http://economics.ca/2005/
papers/0387.pdf and http://www.centerforpbbefr.rutgers.edu/2005/
Paper%202005/PBFEA85.pdf).
48. See Andrew Clare & Nick Motson, Do UK Retail Investors Buy at the Top
and Sell at the Bottom? (Sept. 2010) (available at http://www.cass.city.
ac.uk/__data/assets/pdf_file/0003/69933/Do-UK-retail-investors-
buy-at-the-top-and-sell-at-the-bottom.pdf).
49. See Ilia D.Dichev & Gwen Yu, Higher Risk, Lower Returns: What Hedge
Fund Investors Really Earn, 100J.Fin. Econ. 248263 (2011).
50. See Mercer Bullard, Geoff Friesen & Travis Sapp, Investor Timing and
Fund Distribution Channels (2007) (available at http://funddemocracy.
com/Investor%20Timing%20final%20final&2012.4.07.pdf).
51. Id. See generally Erik R.Sirri & Peter Tufano, Costly Search and Mutual
Fund Flows, 53J.Fin. 15891622 (1998) (identifying search costs as the
factor that drives consumers to invest disproportionately in funds that per-

formed very well in the period immediately preceding the purchase).
52. Andrea Frazzini & Owen A.Lamont, Dumb Money: Mutual Fund Flows
and the Cross-Section of Stock Returns, 88 J. Fin. Econ. 299322, 300
(2008); see also F.Albert Wang, Informed Arbitrage with Speculative Noise
Trading, 34 J. Banking & Fin. 304313 (2010) (suggesting that an
informed investor may wish to trade aggressively on the basis of reliable
information and to take large positions contrary to those of nave specula-
tors). See generally Humphrey B. Neill, The Art of Contrary Thinking
(1992; 1st ed. 1954); William X. Scheinman, Why Most Investors Are
Mostly Wrong Most of the Time (1991; 1st ed. 1970).
53. See Brad M. Barber, Terrance Odean & Ning Zhu, Systematic Noise,
12J.Fin. Mkts. 469547 (2009); Fischer Black, Noise. 41J.Fin. 529543
(1986); Albert S. Kyle, Continuous Auctions and Insider Trading. 53
Econometrica 13151335 (1985).
54. See, e.g., Maik Schmeling, Institutional and Individual Sentiment: Smart
Money and Noise Trader Risk?, 23 Intl J. Forecasting 127145 (2007)
(using institutional and individual sentiment, respectively, as proxies for
smart money and noise trading).
55. See Wang, Informed Arbitrage, supra note 52.
56. See Michael J.Brennan and Eduardo S.Schwartz, Arbitrage in Stock Index
Futures, 63J.Bus. S7S31, S7 (1990).
57. See Michael J. Brennan & H. Henry Cao, Information, Trade, and
Derivative Securities, 9 Rev. Fin. Stud. 163208 (1996).
58. See Ron Kaniel, Gideon Saar & Sheridan Titman, Individual Investor
Trading and Stock Returns, 63J.Fin. 273310 (2008).
59. See Ron Kaniel, Shuming Liu, Gideon Saar & Sheridan Titman, Individual
Investor Trading and Return Patterns Around Earnings Announcements,
67J.Fin. 639680 (2012).
60. See Brad M.Barber, Terrance Odean & Ning Zhu, Do Retail Trades Move
Markets?, 22 Rev. Fin. Stud. 151186 (2009); Soeren Hvidkjaer, Small
Trades and the Cross-Section of Stock Returns, 21 Rev. Fin. Stud. 1123
1151 (2008).
61. David F.Swensen, The Mutual Fund Merry-Go-Round, N.Y.Times, Aug.
14, 2011, at SR6.
62. See, e.g., Rob Bauer, Mathijs Cosemans & Piet Eichholtz, Option Trading
and Individual Investor Performance, 33J.Banking & Fin. 731746, 734
(2009) (defining investor performance as the monthly change in the mar-
ket value of all stocks and options in an investors account).
63. 17 C.F.R. part 270.
64. 15 U.S.C. 80a-1 to 80a-64. On the regulation of the capital structure
of mutual funds, including a proposal to allow mutual funds to issue debt
298 J.M. CHEN
as well as stock, see generally John Morley, The Regulation of Mutual Fund
Debt, 30 Yale J. on Reg. 343376 (2013). In 2008, the SEC proposed
rules for exchange-traded funds (ETFs), whose shares may be traded
throughout the trading day, but those rules have not been codified. See
Exchange-Traded Funds, 73 Fed. Reg. 14,618, 14,65558 (March 18,
2008). See generally David Herzig, Exchange Funds: A Proposal for
Regulations, Finally, 135 Tax Notes 865 (2012).
65. 17 C.F.R. 270.22c-1(a).
66. See 17 C.F.R. 270.22c-1(b)(1).
67. 15 U.S.C. 80a-22(e).
69. Id.
70. See https://en.wikipedia.org/wiki/Internal_rate_of_return.
71. See id.; https://en.wikipedia.org/wiki/Secant_method.
72. See James Matthew Moten, Jr. & Chris Thron, Improvements on Secant
Method for Estimating Internal Rate of Return, 42 Intl J.Applied Math.
& Stat. 8493 (2013).
73. See Friesen & Sapp, supra note 43, at 2801.
74. See Dichev & Yu, supra note 44, at 251.
76. See Barber & Odean, Trading Is Hazardous to Your Wealth, supra note 24,
at 781. Other estimates of the average holding period for stocks in the
USA range from one year, see Levy, CAPM in the 21st Century, supra note
41 (Chapter 1), at 224226; Benartzi & Thaler, Myopic Loss Aversion,
supra note 20 (Chapter 2), at 83 (reaching, more precisely, an equilib-
rium evaluation period of roughly thirteen months for nominal
returns and a period between ten and eleven months for real returns),
to two or three years, see Allen B.Atkins & Edward A.Dyl, Transaction
Costs and the Holding Periods for Common Stocks, 52 J. Fin. 309325
(1997) (concluding that holding periods vary in response to the size of
bid-ask spreads on a stock exchange).
78. See I.R.C. 1(h) (specifying maximum statutory federal income tax rates
on capital gains); id. 1001(c) (providing for recognition in full of all capi-
tal gains or losses unless otherwise provided); id. 1222 (defining terms
such as net short-term capital loss, net long-term capital gain, and
net capital gain); id. 1211 (generally limiting the deduction for capital
losses in any taxable year to $3000). The full mechanics of the income tax
treatment of capital gains and losses are specified at I.R.C. 10011288.
For discussions of the policies underlying preferential tax treatment of
capital gains and the distinct case for limitations on the deductibility of
capital losses, see Nol B.Cunningham & Deborah H.Schenk, The Case
for a Capital Gains Preference, 48 Tax L.Rev. 319380 (1993); Robert
H. Scarborough, Risk, Diversification and the Design of Loss Limitations
Under a Realization-Based Income Tax, 48 Tax L.Rev. 677717 (1993).
79. See Ingersoll & Jin, supra note 51 (Chapter 8), at 752.
80. See id.
81. See S.G. Badrinath & Wilbur G. Lewellen, Evidence on Tax-Motivated
Securities Trading Behavior, 46J.Fin. 369382 (1991) (reporting some
evidence of tax-loss harvesting); Brad M.Barber & Terrance Odean, Are
Individual Investors Tax Savvy? Evidence from Retail and Discount
Brokerage Accounts, 88J.Pub. Econ. 419442 (2004) (finding, consistent
with tax-based considerations, that investors do sell losing stocks in taxable
accounts in December, but not in tax-deferred accounts); Zoran Ivkovi,
James M. Poterba & Scott Weisbenner, Tax-Motivated Trading by
Individual Investors, 95 Am. Econ. Rev. 16051630 (2005) (finding
meaningful differences between taxable and tax-sheltered brokerage
accounts throughout the year); Josef Lakonishok & Seymour Smidt,
Volume for Winners and Losers: Taxation and Other Motives for Trading,
41 J. Fin. 951974 (1986) (finding limited evidence of tax-motivated
trading in December and January).
82. See George M.Constantinides, Optimal Stock Trading with Personal Taxes:
Implication for Prices and the Abnormal January Returns, 13J.Fin. Econ.
6589 (1984); Edward A. Dyl, Capital Gains Taxation and Year-End
Stock Market Behavior, 32J.Fin. 165175 (1977).
83. Compare Marc R. Reinganum, The Anomalous Stock Market Behavior of
Small Firms in January: Empirical Tests for Tax-Loss Selling Effects,
12J.Fin. Econ. 89104 (1983) (attributing the January effect, as docu-
mented by sources such as those cited supra note 38 (Chapter 1), to tax-
driven loss-harvesting activities) with Mustafa N. Gultekin & N. Bulent
Gultekin, Stock Market Seasonality: International Evidence, 12 J. Fin.
Econ. 469481 (1983) (confirming that stock market seasonality does
appear in developed countries around the world and, where it exists,
appears to coincide with the turn of the tax year in those countries).
CHAPTER 12
Irrational Exuberance: Momentum Crashes

andSpeculative Bubbles
12.1 Some Speculation About Speculative Bubbles

The statistic described in 11.4 is specific to a particular investment
vehicle, such as a mutual fund or a hedge fund, whose net inflows and out-
flows of cash enable us to quantify the gap between hypothetical invest-
ment returns and the actual returns that real-life investors realized in fact.
Unlike other anomalous departures from hypothetically efficient markets
and capital asset pricing, however, the effect that measures is almost
entirely an artifact of investor behavior. Indeed, among the many phe-
nomena documented in this book, the performance gap may be the purest
example of a behavioral departure from perfect efficiency, rational expecta-
tions, and platonically maximized utility.
But one suspects, if only as a matter of intuition rather than completely
theorized (let alone empirically validated) analysis,1 that there are other,
broader financial events reflecting similar cycles of fear and greed. This
chapter indulges some necessarily speculative thoughts about other finan-
cial phenomena attributable to mismatches between investor perceptions
and true market conditions, including momentum crashes and speculative
bubbles. The title of this chapter pays homage to the December 5, 1996,
speech in which Federal Reserve Board chairman Alan Greenspan gave
public voice to the kind of social phenomenon that perceptive people
[have seen] with their own eyes happening again and again in history,
when markets have been bid up to unusually high and unsustainable

DOI10.1007/978-3-319-32711-2_12
302 J.M. CHEN
levels under the influence of market psychology.2 Or, as John Kenneth

Galbraith said of the great crash of 1929: [T]he chances for a recurrence
of a speculative orgy are rather good.3 The investment gap measured by
is a direct outgrowth of strategic and tactical asset allocations decisions
made by investors, which in turn are highly significantly correlated with
investors subjective beliefs about the future returns offered by each asset
or asset class.4
Because there are severe limits to arbitrage as a corrective force for
behaviorally driven departures from rational asset pricing,5 it is entirely
conceivable that behaviorally driven decisions traceable to individual port-
folios (and whose effects are initially confined to that level) can cascade
until they reach the level of global contagion.6 It is not enough that the
very presence of noise trading imposes a limit on rationality, in the sense
that the sheer depth and duration of uninformed trading can disrupt
arbitrage.7
Worse still is the realization that arbitrageurs are not immune to cogni-
tive biases of their own. Arbitrageurs very limited trading horizons exac-
erbate their baseline preference as traders for short-run information to
long-run information, so much so that they will even ignore long-run
information.8 Trapped by their own cognitive biases and limitations,
arbitrageurs and other purveyors of smart money strive to gather the
same information.9 Institutional traders, who supposedly know better
than dumb money, trade in coordinated ways that are best understood
as the result of using other institutional traders transactions as a source of
financial information.10 Even the information-gathering phase in trading,
by the financial markets best-informed participants, constitutes a clinic in
herding behavior.11
Timing, as one might expect, also matters. Rational traders face syn-
chronization risk in the uncertainty in identifying the precise moment
when their peers will exploit a common arbitrage opportunity.12 As a
consequence of synchronization risk, putatively profit-driven arbitra-
geurs resort to timing the market, a rationally tenuous strategy that
injects a raw dose of cognitive bias of its own.13 In retrospect, it now
is clear that hedge funds rode the technology bubble of the 1990s and
were able to reduce their stock exposure before the bubble burst, but
exploited their informational and behavioral advantages over noise trad-
ers for personal gain without exerting any genuinely correcting force on
stock prices.14
IRRATIONAL EXUBERANCE: MOMENTUM CRASHES ANDSPECULATIVE BUBBLES 303
12.2 The Behavioral Origins

ofStock Market Momentum
Taking due note of what appears to be the greater financial impact of

investor behavior, relative to the constrained corrective power of arbi-
trage, I now offer some speculative thoughts on the relationship of
investor-specific performance gaps to the rise of speculative bubbles
some speculation on speculation, as it were. Speculative asset pricing,
after all, warranted the attention of Robert Shiller on the occasion of his
acceptance of the Nobel Prize in economics.15 If this very brief account of
investor exuberance fails to do justice to the topic of momentum crashes
and speculative bubbles,16 then I take refuge in Brad Barber and Terrance
Odeans terse but accurate observation: [A]ll humans err.17 As is the
case for individual investors, so is the case for those who study them.18
Financial interest in stock market momentum originates in academic
work of the 1970s and 1980s that identified significant departures from
the conventional capital asset pricing model.19 Firms with modest market
capitalization20 and firms with a high ratio of book value to market value21
offered returns exceeding those predicted solely by beta. In a comprehen-
sive synthesis of the literature, Eugene Fama and Kenneth French declared
in 1992 that the relation between and average return is weak, per-
haps nonexistent.22 Fama and French identified small firm size and high
book-to-market ratio as the most significant determinants of financial
returns.23 The resulting FamaFrench three-factor model, which explains
asset prices as a function of market risk, size, and value, is regarded as the
dominant paradigm in contemporary finance.24
A distinct body of scholarship, most closely associated with Mark
Carhart, has identified momentum in short-run stock prices.25 Because
it cannot be explained by beta or by Fama and Frenchs size and value
factors,26 momentum represents a distinct factor in asset pricing. In their
own scholarship, Fama and French have routinely evaluated momentum
alongside their original size and value factors.27 One may argue that the
three-factor model and Carharts momentum model have merged into a
comprehensive FamaFrenchCarhart four-factor model of risk, size, value,
and momentum, or simply the four-factor model.28 Subsequent scholar-
ship has reinforced the cogency of the four-factor model by considering
other factors, such as dividend yield, the term spread, the default spread,
and the yield on Treasury bills, and concluding that the primary drivers of
excess investment return are indeed size, value, and momentum.29
304 J.M. CHEN
Momentum transcends boundaries between geographic markets and

financial asset classes. Momentum appears in developed30 and emerging
markets31 around the world (with the curious exception of Japan).32 The
global prevalence of momentum warrants the reorientation of asset man-
agement strategies focusing on a single country (particularly the USA) in
order to capture this driver of returns.33 Likewise, momentum has been
detected in asset classes besides equities.34 As the likeliest source of diver-
sification for the stocks held in most individual and institutional portfo-
lios, bonds also exhibit momentum.35 The presence of positive returns
to momentum in commodities, coupled with those markets low correla-
tion with more conventional asset classes, may counsel the inclusion of
momentum-based commodity futures contracts in well-diversified portfo-
lios.36 Among equities, however, there is evidence that momentum inheres
in the movement of industries and sectors within the broader equity
market, rather than in the movement of individual stocks within a single
industry.37
Consistent with evidence of persistent investor overreaction to
market events over the long run,38 momentum in asset prices almost
certainly manifests the impact of investors behavior. At a minimum,
momentum accounts more of the behavioral impact on the Fama
FrenchCarhart model than any of the other factors. Although there
are admittedly nonbehavioral accounts of momentum,39 those typically
macroeconomic hypotheses are heavily contested.40 The prevailing
explanation for momentum in stock returns is behavioral.41 Of its own
force, the very fact that momentum correlates with trading volume
suggests that this phenomenon deviates from rational models of asset
pricing.42
More specifically, momentum originates in the basic behavioral frames
of the investing mind: mental accounting and prospect theory.43 A dif-
ference in risk attitudes that is driven entirely by whether [a] stock has
generated a paper capital gain or a paper capital loss gives investors a
greater tendency to sell stocks that have gone up in value since purchase
(and, correlatively, to cling longer to losers).44 The resulting disposition
effect conditions investors to favor their baseline holdings and to under-
react to news.45 At the same time, investors collectively brim with overcon-
fidence.46 Not only do they desire to do something; they have supreme
albeit misplaced confidence in their financial abilities.47 Overconfidence
in the sense of not searching for evidence that would disconfirm [ones]
beliefs is the fundamental error of noise traders.48
The remarkably slow rate at which certain types of information diffuse

throughout the market amplifies the behavioral drivers of momentum.49
Three findings suggest that market momentum reflects the gradual[]
response of the market to new information.50 First, if the market
is surprised by good or bad earnings news, then on average the market
continues to be surprised in the same direction at least over the next two
subsequent announcements.51 Second, inertia and prolonged adjust-
ment in analyst forecasts, especially in revising estimates in the
case of companies with the worst performance out of analysts possibl[e]
reluctance to alienate management, may not be helping the market to
assimilate new information in a timely fashion.52 Third, because cogni-
tive dissonance retards investors reaction to information that contradicts
their sentiment, bad news for losing stocks diffuse[s] slowly when senti-
ment is optimistic.53 When sentiment is pessimistic, cognitive dissonance
slows the diffusion of good news for winning stocks.54
The sluggish response of market participants, however, is neither an
entirely negative verdict on the information efficiency of the stock mar-
ket nor an opportunity for better reasoning (or at least more ruthlessly
efficient, informed traders) to restore order to noise.55 Whereas purely
rational markets are directed strictly by the minimal amount of new infor-
mation necessary to infer changes to the return distribution of the market
portfolio, the presence of noise traders disrupts this process and drives
prices away from efficiency.56 Over the long run, especially as salience bias
attributable to news coverage takes its toll on rational thought,57 investors
overreact to market events.58 The impact of momentum on asset prices
and on returns ebbs and flows over time: Although short-term and long-
term momentum strategies perform poorly (especially in time horizons
shorter than one month or longer than two years), momentum trading is
highly profitable at intermediate horizons up to 24 months, and espe-
cially in the 6- to 12-month range.59 These cognitive biases and behav-
ioral departures from rationality help build a unified theory of momentum
as a product of both under- and overreaction to news.60
12.3 Momentum Crashes

Momentum strategies, however, do have an Achilles heel. Their strong
positive average returns and Sharpe ratios are punctuated with occa-
sional crashes.61 Momentum crashes tend to occur in times of market
stress, when the market has fallen and ex-ante measures of volatility are
306 J.M. CHEN
high, coupled with an abrupt rise in contemporaneous market returns.62

True to the postmodern portfolio theory, prospect theory, and virtually
every other variation on the theme of behavioral economics, momentum
hinges on the state of the market. When the lagged three-year market
return is non-negative, a six-month momentum strategy generates a
significant mean monthly profit of 0.93%.63 The corresponding return
for down markets, defined as markets where the three-year lagged
return is negative, is an insignificant 0.37% profit.64 Moreover, con-
sistent with the broad understanding that falling markets are more vola-
tile than rising markets, returns from momentum reach their ebb when
volatility spikes.65 The conclusion that the state of the market is criti-
cally important to the profitability of momentum strategies reinforces
the behavioral account of momentum.66 Just as plausibly, although less
supportive of a purely behavioral explanation, asymmetry in momentum
arises from short-selling constraints and other limits on arbitrage that
prevent the timely correction of cognitively restrained reaction[s] to bad
news in good periods.67
There assuredly are cycles in financial markets,68 and momen-
tum represents but one manifestation of those cycles. Over time, as
momentum-driven strategies cull losers and increase their allocations to
recent winners, portfolios sorted on past returns will exhibit meaningful
changes in systematic risk.69 Stocks that fell with the market are likely to
have been, and to remain, high-beta stocks; recent winners are corre-
spondingly likely to have low beta.70 The resulting dramatic time varia-
tion in the betas of momentum portfolios renders them vulnerable
to sudden collapse: [W]hen the market rebounds quickly, momentum
strategies will crash because they have a conditionally large negative
beta.71
Specifically, because most of the up- versus down-beta asymmetry in
bear markets is driven by the past losers, it follows that momentum
strategies in bear markets behave like written call options on the market
when the market falls, they gain a little, but when the market rises they lose
a lot.72 We therefore have ample reason to suspect (and, perhaps eventu-
ally, empirical evidence to verify) that the presence of momentum crashes
and option-like features for momentum strategies across the spectrum
of marketsnot just equities, but also futures, bonds, currencies, and
commoditiesarises from the core behavioral insight that investors in
extreme situations tend to be fearful and to focus on losses, largely
ignoring [rational] probabilities.73
12.4 Liquidity Risk

Momentum crashes bode ill for other, more encompassing forms of
irrational exuberance. Especially intriguing (and disturbing) is evidence
that returns from value and returns from momentum are negatively cor-
related with each other.74 The difference appears to arise from liquidity
risk, both in the sense that an asset might lose market liquidity (i.e., the
ease with which the asset is traded) and in the sense that traders might
lose funding liquidity (i.e., the ease with which traders can secure fund-
ing).75 Although both of these factor loadings earn positive risk premia,
value and momentum have opposite signed exposure to liquidity risk:
momentum loads positively on liquidity shocks, while value loads neg-
atively on liquidity risk even as it earns positive return[s].76
The following account may provide a simple and intuitive explana-
tion for the opposite response of momentum and value to liquidity risk77:
[M]omentum captures the most popular trades, being long the assets whose
prices have recently appreciated as fickle investors flocked to these assets.
Value, on the other hand, expresses a contrarian view, where assets have
experienced price declines over several years. When a liquidity shock occurs,
investor liquidations (from cash needs, redemptions, risk management, run-
ning for the exit at the same time ) put[] more price pressure on the more
crowded trades. These liquidations may affect crowded high momentum
securities more than the less popular contrarian/value securities.78
Liquidity profoundly affects asset prices, most saliently by suppressing

the price of hard-to-trade securities relative to more liquid counterparts
with identical cash flows.79 Measures of liquidity are generalizable to all
asset classes through differences in bid-ask spreads.80 In all settings, reduc-
tions in liquidity reduce prices and raise expected returns.81 Liquidity risk
compounds the endemic inclination of the market toward speculation.
The mere presence of heterogeneous expectations within the community
of potential investors gives rise to the speculative buy[ing] of stock for
later sale at a higher price than the purchaser thinks [the stock] is actually
worth.82
This foundational assumption undergirds many heterogeneous agent
models that explain volatility clustering, high trading volume, specula-
tive bubbles and crashes, and fat tails in the distribution of returns.83
Heterogeneity may subsist in something as simple as the methodologi-
cal difference between fundamental analysis and technical analysis84
308 J.M. CHEN
or a difference in optimism about foreseeable market outcomes.85

Heterogeneity may be latent; it need not be immediately visible to the
naked eye.86
In the most obvious instance of heterogeneity within capital mar-
kets, investors beholden to reference points defined by their unrealized
capital gains and lossesa class that evidently includes university endow-
ments87generate demand distortions that are inversely related to
[their] unrealized profit[s].88 This source of [i]nvestor heterogeneity
between rationally informed traders and behaviorally constrained investors
generates not only diverse demand functions and trades of a type con-
sistent with the disposition effect, but also momentum through price
distortions translat[ing] into return distortions.89 Even small differences
between diverse subgroups within a heterogeneous population of traders
can lead to large speculative premiums.90 Speculation breeds more specu-
lation as these premiums generate fluctuations in wealth for competing
groups of investors with divergent expectations.91 And since liquidity is
persistent over time, any illiquidity that emerges as a result of these inter-
actions is more likely [than] not to cripple trading in the immediate
future.92 Correspondingly, the fact that high illiquidity today predicts
high expected illiquidity in the future impl[ies] a high required return
and a suppression of current prices.93
We must also account for heterogeneity between individual and insti-
tutional investors. Given the presence of rational decision-making mecha-
nisms in institutions and the putatively indefinite investment horizon that
institutions hold (and individuals cannot), it may well be that cumulative
prospect theory and other manifestations of behavioral economics do
not provide as good a description of institutional investor preferences
as they do for individual investor preferences.94 On the other hand,
because institutions are no better than their human agents, institutions
agency costs may produce their own sort of myopic loss aversion.95
Even if an institutional investor such as a pension fund expects to last for-
ever, or at least to endure, the funds manager does not expect to be in
[her or his] job forever.96 The human managers short horizon creates a
conflict of interest with the institutional investor.97
In practice, we may witness a blend of two effects. Agency costs may
dampen the differences between institutional and individual investors, to
the extent that human managers short horizons nudge institutional incen-
tives back in line with those of individuals. Unless institutional governance
mechanisms are utterly impotent, however, there should be meaningful (if
not entirely categorical and dispositive) differences between institutional

and individual investors. Those differences among investors, even if they
are incomplete, should propel the dynamics of a heterogeneous market.
There is, however, a contrary view casting doubt on the hypothesis that
nave investors create[] nondiversifiable risk that sophisticated investors
must address.98 Accepting this alternate account leaves little direct evi-
dence that trading by nave investors plays a substantial role in stock price
determination.99 If anything, [e]fficient prices protect particular noise
traders.100 Information traders refrain from trading during efficient
markets, electing instead to continue to hold the market portfolio.101
By contrast, nave noise traders, under behavioral influence, amass riskier
assets in proportion to bull sentiment and to their own assessments of
financial expertise (as a proxy for overconfidence).102 Ultimately, noise
traders trade only among themselves, and their sole impact on the mar-
ket is to generate excess volume.103
Sure enough, rising markets exhibit higher volume than falling mar-
kets.104 Market turnover, defined as the ratio of trading volume to the
number of shares listed, is regarded as a reliable measure of investor sen-
timent.105 Highly valued assetsperhaps to the point of being overval-
uedare the most heavily traded.106 As a rising market pushes investors
to the liquidation points where they will reap sufficient realization utility,
sales by these investors generate greater volume.107 A similar relationship
between investor optimism and trading volume has been observed in the
housing market.108 At some point, the logical implication of these linkages
between noise trading, investor sentiment, and trading volume is that vol-
ume alone, holding all other factors constant, may serve as an indicator of
irrational exuberance and speculative bubbles.
Perhaps most spectacularly of all, heterogeneity and liquidity risk
together generate leverage cycles within markets for emerging assets
that are not yet mature enough to be attractive to the general public,
such as high-yield bonds from emerging markets or subprime mortgages
anywhere.109 The resulting dynamic of the anxious economy in emerg-
ing assets, one of fluctuating uncertainty and disagreement, creates par-
allel cycles of leverage and tightening liquidity,110 and ultimately spasms
of contagion, flights to collateral, and the rationing of issuance of
good debt.111 The mechanics of momentum in ordinary markets assume
even greater prominence under conditions of crisis: negative momentum
may build as investors with realization utility sell their stakes, without
corresponding demand from the distinct population of conventionally
rational, informed investors.112
310 J.M. CHEN
What is said of the carry trade in currencies applies equally to many

other markets: security prices go up by the stairs and down by the ele-
vator.113 To paraphrase that master of the Southern gothic, Flannery
OConnor: Everything that crashes must converge.114 Noise trading,
speculative bubbles, liquidity shocks, leverage cycles, volatility spikes, price
crashes, flights to quality, and contagion are all highly interrelated, corre-
lated phenomena.115 The mutually procyclical nature of these phenomena
suggests that an intricate feedback mechanism connects abnormal markets
with cognitively biased investor behavior.
The discovery of new linkages between markets and risk factors prom-
ises to blur the conventional distinction between shift contagion and
pure contagion.116 Shift contagion supposedly occurs when high vola-
tility disrupts pre-existing market linkages, such as goods trade, financial
flows, or exposure to common shocks.117 By contrast, pure contagion
reflects excess not explained by market fundamentals or common
shocks.118 It purportedly arises from idiosyncratic shocks being trans-
mitted through channels that could not have been identified before-
hand.119 The identification of new linkages suggests that the distinction
between shift and pure contagion does not reflect a qualitative difference
in episodes of contagion, but rather exposes previous limits on economic
knowledge.
12.5 A Simple Model ofInformed andNave

Trading
Although there is no shortage of exhaustive models for extreme financial
events, including models that explicitly incorporate investor behavior and
cognitive bias,120 simple heterogeneity between nave and informed inves-
tors supplies a straightforward model for oscillations in markets known
to defy Gaussian normality and behavioral rationality. Robert Shillers
2013 Nobel Prize lecture revived a model that the laureate first devised in
1984.121 Because that model is not only elegant but also complete enough
to capture this chapters discussion of investor navet as the impetus for
irrational exuberance, I present it here.
Let us begin with a mathematical definition of stock price under the
simplest version of the efficient markets model.122 In perfectly effi-
cient markets, all stock price movements depend exclusively on publicly
available information about future payouts of dividends, which in turn

are discounted at a constant rate through time:123

Dt + k
Pt =
(1 + d )
k +1
k =0

Informed investors, or smart money, demand stock according to a

linear function of the expected return on the market over the next
time period:124
Et Rt - r
Qt =
j
where Qt is the demand for shares by smart money at time t expressed as a

portion of total shares outstanding, EtRt is the expected return starting
at time t, is the expected real return such that there is no demand for
shares by the smart money, and is the risk premium that would induce
smart money to hold all the shares.125
By contrast, all other investors are nave, ordinary investors who do[]
not respond to expected returns, but rather overreact to news or are
vulnerable to fads.126 The dumb money demand for stock per share is
defined as Yt.127 Since equilibrium demands that Qt+Yt/Pt=1 in order
to clear the stock market, [s]olving the resulting rational expectations
model gives us the model for stock price in this stylized economy:128

Dt + k + j Yt + k
Pt = ,0 <j <
(1 + r + j )
k +1
k =0

In other words, the real price is the present value, discounted at rate
+ , of both the expected future dividend payments and times the
expected future demand by ordinary investors.129 The two decisive vari-
ables in this model of stock price are (the risk premium) and Yt (ordi-
nary investors demand for stock per share).
The extreme limits on this definition of the stock price describe all-or-
nothing dominance of the stock market by either smart money or dumb
312 J.M. CHEN
money. As 0, smart money becomes more and more influential,

and the equation for stock price collapses to:130

Dt + k
Pt =
(1 + d )
k +1
k =0

This equation defines the ordinary efficient markets model that equates
price with the present value of expected dividends.131 On the other
hand, as , smart money drops out in favor of dumb money, and the
equation for stock price collapses to Pt=Yt in a market where ordinary
investors determine the price.132
The truly interesting situation arises in the intermediate case where
approaches neither zero nor infinity.133 In the intermediate case, price
may have low predictability from day to day or month to month, con-
sistent with efficient markets theory, even if animal spirits dominate the
broad movements in the stock price.134 Finite, nonzero risk premium
means that [l]ong, slow swings in Yt, dumb money stock demand, may
produce [correspondingly] long slow swings in stock prices, all the way
to the so-called multiyear bull and bear markets.135 The actual price
of stock now responds not solely to news about future dividends, as it
would in a perfectly efficient market, but also to news about animal spir-
its, or uninformed dumb money demand for stock.136 Given everything
we have discovered about the disposition effect, investors under- and
overreaction to news, the performance gap measure , and the mechan-
ics of momentum and liquidity, we should express no surprise whatsoever
that the interaction of the risk premium and the dumb money demand for
stock may generate immense market bubbles and even more spectacular
crashes.
Throughout it all, smart money exercises at best limited influence
on the pricing dynamics of the market. Inertia is the ultimate limit on
arbitrage; surrounded by an overwhelming miasma of noise, otherwise
informed traders may elect not to rebut nave investors mistaken evalu-
ation of companies and their securities.137 Meanwhile, competing, noisy
narratives may propagate along pathways resembling biological mecha-
nisms, such as those by which infections spread and ants recruit others
to food sources.138 All theories of herd behavior in financial markets are
ultimately theories of the failure of information about true fundamental
value to be disseminated and evaluated.139
Notes
1. See generally Cass R. Sunstein, Incompletely Theorized Agreements, 108
Harv. L.Rev. 17331772 (1995).
2. Shiller, Irrational Exuberance, supra note 89 (Chapter 1), at 12. Rather
obviously, this chapters allusion to irrational exuberance pays homage
to Robert Shillers book as well as Federal Reserve Chairman Alan
Greenspans December 1996 speech, The Challenge of Central Banking
in a Democratic Society (Dec. 5, 1996) (available at http://www.federal-
reserve.gov/boarddocs/speeches/1996/19961205.htm).
3. John Kenneth Galbraith, The Great Crash: 1929, at 194 (1954); accord
Shiller, Irrational Exuberance, supra note 89 (Chapter 1), at 243.
4. Annette Vissing-Jrgensen, Perspectives on Behavioral Finance: Does
Irrationality Disappear with Wealth? Evidence from Expectations and
Actions, in NBER Macroeconomics Annual 2003, at 137194, 162164
(Mark Gertler & Kenneth Rogoff eds., 2004); accord Markus
K. Brunnermeier, Christian Gollier & Jonathan A. Parker, Optimal
Beliefs, Asset Prices, and the Preference for Skewed Returns, 97 Am. Econ.
Rev. 159167, 159 & n.3 (2007).
5. Andrei Shleifer & Robert Vishny, The Limits of Arbitrage, 52 J. Fin.
3555 (1997); accord Vissing-Jrgensen, supra note 4, at 140, 150151.
See generally Chen, Postmodern Portfolio Theory, supra note 1 (Chapter
1), 9.2, at 174176.
6. See generally Markus K.Brunnermeier, Herding in Finance, Stock Market
Crashes, Frenzies, and Bank Runs, in Asset Pricing Under Asymmetrical
Information: Bubbles, Crashes, Technical Analysis, and Herding 165220
(2001).
7. See J. Bradford DeLong, Andrei Shleifer, Lawrence H. Summers &
Robert J.Waldman, Noise Trader Risk in Financial Markets, 98J.Pol.
Econ. 703738, 705 (1990).
8. Brunnermeier, supra note 6, at 191.
9. Id.
10. See Richard W.Sias, Institutional Herding, 17 Rev. Fin. Stud. 165206
(2004).
11. See generally Kenneth A.Froot, David B.Scharfstein & Jeremy C.Stein,
Herd on the Street: Informational Efficiencies in a Market with Short-Term
Speculation, 47 J. Fin. 14611484 (1992); David B. Scharfstein &
Jeremy C.Stein, Herd Behavior and Investment, 80 Am. Econ. Rev. 465
479 (1990).
12. See Dilip Abreu & Markus K.Brunnermeier, Synchronization Risk and
Delayed Arbitrage, 66J.Fin. Econ. 341360 (2002).
13. See Dilip Abreu & Markus K. Brunnermeier, Bubbles and Crashes, 71
Econometrica 173204 (2003).
314 J.M. CHEN
14. See Markus K. Brunnermeier & Stefan Nagel, Hedge Funds and the
Technology Bubble, 59 Econometrica 20132040 (2004).
15. Robert J. Shiller, Speculative Asset Prices, 104 Am. Econ. Rev. 1486
1517 (2014), reprinted in Shiller, Irrational Exuberance, supra note 89
(Chapter 1), at 239279. Subsequent references to Shillers Nobel Prize
lecture will be made to the reprinted version in Irrational Exuberance,
where the lecture serves as an appendix to the third edition of the book.
16. See also Chen, Postmodern Portfolio Theory, supra note 1 (Chapter 1),
4.2, at 4445 (discussing how the prospect of upside gain exposes
investors to the catastrophe of success identified by Williams, supra
note 22 (Chapter 9)).
18. Id.
19. See generally supra 3.1; Chen, Postmodern Portfolio Theory, supra
note 1 (Chapter 1), 2.22.8, at 617.
20. See, e.g., Rolf W. Banz, The Relationship Between Return and Market
Valuation of Common Stocks, 9J.Fin. Econ. 318 (1981).
21. See, e.g., S.Basu, Investment Performance of Common Stocks in Relation
to Their Price-Earning Ratios: A Test of the Efficient Market Hypothesis,
32J.Fin. 663682 (1977); cf. Marc R.Reinganum, Misspecification of
Capital Asset Pricing: Empirical Anomalies Based on Earnings Yield and
Market Values, 9J.Fin. Econ. 1946 (1981) (identifying both the size
and the value anomalies).
22. Eugene F. Fama & Kenneth R. French, The Cross-Section of Expected
Stock Returns, 47J.Fin. 427465, 464 (1992).
23. See, e.g., Kent Daniel & Sheridan Titman, Evidence on the Characteristics
of Cross-Sectional Variation in Stock Returns, 52 J. Fin. 133 (1997);
Eugene Fama & Kenneth R.French, Size and Book-to-Market Factors in
Earnings and Returns, 50J.Fin. 131155 (1995).
24. See generally, e.g., John M. Griffin, Are the Fama and French Factors
Global or Country Specific?, 15 Rev. Fin. Stud. 783803 (2002); Adam
Borchert, Lisa Ensz, Joep Knijn, Greg Pope & Aaron Smith,
Understanding Risk and Return, the CAPM, and the Fama-French Three-
Factor Model (Tuck School of Business Case No. 03111, supervised by
Kent Womack & Ying Zhang) (Dec. 2003) (available at http://ssrn.
25. See Mark M. Carhart, On Persistence in Mutual Fund Performance,
52 J. Fin. 5782 (1997); Mark Grinblatt, Sheridan Titman & Russ
Wermers, Momentum Investment Strategies, Portfolio Performance, and
Herding: A Study of Mutual Fund Behavior, 85 Am. Econ. Rev. 1088
1105 (1995); Narasimhan Jegadeesh & Sheridan Titman, Returns to
Buying Winners and Selling Losers: Implications for Market Efficiency,

48J.Fin. 6591 (1993).
26. See Louis K.C. Chan, Narasimhan Jegadeesh & Josef Lakonishok,
Momentum Strategies, 51J.Fin. 16811713 (1996).
27. See Eugene F. Fama & Kenneth R. French, Dissecting Anomalies,
63J.Fin. 16531678 (2008); Eugene F.Fama & Kenneth R.French,
Multifactor Explanation of Asset Pricing Anomalies, 51 J. Fin. 5585
(1996); Eugene F. Fama & Kenneth R. French, Size, Value, and
Momentum in International Stock Returns, 105J.Fin. Econ. 457472
(2012).
28. See, e.g., Subhrendu Rath & Robert B. Durand, Decomposing the Size,
Value and Momentum Premia of the FamaFrenchCarhart Four-Factor
Model, 132 Econ. Letters 139141 (2015); cf. Jimmy Liew & Maria
Vassalou, Can Book-to-Market, Size and Momentum Be Risk Factors That
Predict Economic Growth?, 57J.Fin. Econ. 221245 (2000).
29. See Doron Avramov & Tarun Chordia, Pricing Stock Returns, 82J.Fin.
Econ. 387415 (2006).
30. See John A. Doukas & Phillip J. McKnight, European Momentum
Strategies, Information Diffusion, and Investor Conservatism, 11 Eur.
Fin. Mgmt. 313338 (2005); K. Geert Rouwenhorst, International
Momentum Strategies, 53J.Fin. 267284 (1998).
31. See K. Geert Rouwenhorst, Local Return Factors and Turnover in
Emerging Stock Markets, 54J.Fin. 14391464 (1999).
32. See Cifford S.Asness, Momentum in Japan: The Exception That Proves the
Rule, 37:4 J. Portfolio Mgmt. 6775 (Summer 2011); Andy Chui,
Sheridan Titman & K.C. John Wei, Individualism and Momentum
Around the World, 65J.Fin. 361392 (2010).
33. See Clifford S.Asness, John M.Liew & Ross L.Stevens, Parallels Between
the Cross-Sectional Predictability of Stock and Country Returns,
23:3J.Portfolio Mgmt. 7987 (Spring 1997); Clifford S.Asness, Toby
J.Moskowitz & Lasse Heje Pedersen, Value and Momentum Everywhere,
58J.Fin. 929985 (2013).
34. See Claude B. Erb & Campbell R. Harvey, The Strategic and Tactical
Value of Commodity Futures, 62:2 Fin. Analysts J. 6997 (March/April
2006) (commodities); Tobias J.Moskowitz, Yao Hua Ooi & Lasse Heje
Pedersen, Time Series Momentum, 104 J. Fin. Econ. 228250 (2012)
(futures); John Okunev & Derek White, Do Momentum-Based Strategies
Still Work in Foreign Currency Markets? 38 J. Fin. & Quant. Analysis
425447 (2003) (currencies).
35. See Adams, Moskowitz & Pedersen, supra note 33, at 939951 (docu-
menting momentum and value effects in government bonds and value
effects in currencies and commodities).
316 J.M. CHEN
36. See Jolle Miffre & Georgios Rallis, Momentum Strategies in Commodity
Futures Markets, 31J.Banking & Fin. 18631886 (2007).
37. See Tobias J. Moskowitz & Mark Grinblatt, Do Industries Explain
Momentum?, 54J.Fin. 12491290 (1999).
38. See Navin Chopra, Josef Lakonishok & Jay R.Ritter, Measuring Abnormal
Performance: Do Stocks Overreact?, 31 J. Fin. Econ. 235268 (1985);
Werner F.M. De Bondt & Richard H. Thaler, Does the Stock Market
Overreact?, 40 J. Fin. 793805 (1985); Andrew W. Lo & A. Craig
MacKinlay, When Are Contrarian Profits Due to Stock Market
Overreaction?, 3 Rev. Fin. Stud. 175205 (1990).
39. See Dong H.Ahn, Jennifer Conrad & Robert F.Dittmar, Risk Adjustment
and Trading Strategies, 16 Rev. Fin. Stud. 459485 (2002); Tarun
Chordia & Lakshmanan Shivakumar, Momentum, Business Cycle, and
Time-Varying Expected Returns, 57J.Fin. 9851019 (2002); Timothy
Johnson, Rational Momentum Effects, 57J.Fin. 585605 (2002); Laura
Xiaolei Liu & Lu Zhang, Momentum Profits, Factor Pricing, and
Macroeconomic Risk, 21 Rev. Fin. Stud. 24172448 (2009); Jacob Sagi
& Mark Seasholes, Firm- Specific Attributes and the Cross-Section of
Momentum, 84 J. Fin. Econ. 389434 (2008); Tong Yao, Dynamic
Factors and the Source of Momentum Profits, 26 J. Bus. & Econ. Stat.
211226 (2008).
40. See Kent D.Daniel & Sheridan Titman, Testing Factor-Model Explanations
of Market Anomalies, 1 Critical Fin. Rev. 103139 (2012) (failing to find
correlation between momentum and the usual indicators of macroeco-
nomic risk, such as growth in consumption); John M.Griffin, Xiuqing Ji
& J. Spencer Martin, Momentum Investing and Business Cycle Risk:
Evidence from Pole to Pole, 58J.Fin. 25152547, 2515 (2003) (finding
no evidence that macroeconomic risk variables can explain momentum
and concluding instead that momentum profits around the world are
large and statistically reliable in both good and bad economic states).
41. See Nicholas Barberis, Andrei Shleifer & Robert Vishny, A Model of
Investor Sentiment, 49 J. Fin. Econ. 307343 (1998); cf. Narasimhan
Jegadeesh & Sheridan Titman, Profitability of Momentum Strategies: An
Evaluation of Alternative Explanations, 56J.Fin. 699720, 719 (2001)
(concluding that behavioral models provide at best a partial
explanation).
42. See Charles M.C.Lee & Bhaskaran Swaminathan, Price Momentum and
Trading Volume, 55J.Fin. 20172069 (2000); cf. Shefrin & Statman,
Behavioral Capital Asset Pricing Theory, supra note 79 (Chapter 1), at
334 (explaining how noise trading in rational markets generates volume
rather than price effects).
43. See Mark Grinblatt & Bing Han, Prospect Theory, Mental Accounting and
Momentum, 78J.Fin. Econ. 311339 (2005).
44. Id. at 312.
45. See Frazzini, supra note 11 (Chapter 11).
46. See Simon Gervais & Terrance Odean, Learning to Be Overconfident, 14
Rev. Fin. Stud. 127 (2001).
47. See Kent D. Daniel, David Hirshleifer & Avanidhar Subrahmanyam,
Overconfidence, Arbitrage, and Equilibrium Asset Pricing, 56 J. Fin.
921965 (2001).
79 (Chapter 1), at 331 n.21. See generally Hillel J. Einhorn & Robin
M. Hogarth, Confidence in Judgment: Persistence in the Illusion of
Validity, 85 Psych. Rev. 395416 (1978).
49. See Craig W. Holden & Avanidhar Subrahmanyam, News, Events,
Information Acquisition, and Serial Correlation, 75J.Bus. 132 (2002).
50. Louis K.C.Chan, Narasimhan Jegadeesh & Josef Lakonishok, Momentum
Strategies, 51J.Fin. 16811713, 1709 (1996).
51. Id. at 17091710.
52. Id. at 1710.
53. Constantinos Antoniou, John A.Doukas & Avanidhar Subrahmanyam,
Cognitive Dissonance, Sentiment, and Momentum, 48J.Fin. & Quant.
Analysis 245275, 246 (2013).
54. See id.
55. Chan, Jegadeesh & Lakonishok, supra note 50, at 1710.
79 (Chapter 1), at 324.
57. See Pedro Bordalo, Nicola Gennaioli & Andrei Shleifer, Salience and
Asset Prices, 103 Am. Econ. Rev. 623628 (2013); Kent Daniel, David
Hirshleifer & Siew Hong Teoh, Investor Psychology in Capital Markets:
Evidence and Policy Implications, 49J.Monetary Econ. 139209 (2002).
58. See Navin Chopra, Josef Lakonishok & Jay R.Ritter, Measuring Abnormal
Performance: Do Stocks Overreact?, 31 J. Fin. Econ. 235268 (1985);
Navin Chopra, Josef Lakonishok & Jay R. Ritter, Performance
Measurements Methodology and the Question of Whether Stocks Overreact,
31 J. Fin. Econ. 235268 (1992); de Bondt & Thaler, Does the Stock
Market Overreact?, supra note 9 (Chap 9); Andrew W.Lo & A.Craig
MacKinlay, When Are Contrarian Profits Due to Stock Market
Overreaction?, 3 Rev. Fin. Stud. 175205 (1990).
59. Moskowitz & Grinblatt, supra note 37, at 12491250; see also id. at
1250 n.1 (documenting sources that have revealed [a]nomalous strong
autocorrelation in stock market momentum).
318 J.M. CHEN
60. See Harrison Hong & Jeremy C.Stein, A Unified Theory of Underreaction,
Momentum Trading, and Overreaction in Asset Markets, 54J.Fin. 2143
2184 (1999); Kent D. Daniel, David Hirshleifer & Avanidhar
Subrahmanyam, Investor Psychology and Security Market Under- and
Over-Reactions, 53 J. Fin. 18391886 (1998). See generally Brad
M.Barber & Terrance Odean, All That Glitters: The Effect of Attention
and News on the Buying Behavior of Individual and Institutional Investors,
21 Rev. Fin. Stud. 785818 (2008).
61. Kent Daniel & Tobias J. Moskowitz, Momentum Crashes 1 (Aug. 8,
2014) (available at http://faculty.chicagobooth.edu/tobias.moskowitz/
research/mom11.pdf).
62. Id.
63. Michael J. Cooper, Roberto C. Gutierrez Jr. & Allaudeen Hameed,
Market States and Momentum, 59J.Fin. 13451365, 1347 (2004).
64. Id.; see also Dayong Huang, Market States and International Momentum
Strategies, 46 Q. Rev. Econ. & Fin. 437446 (2006) (confirming the
statistically significant presence of momentum profits in upside markets,
but also finding profits, albeit statistically insignificant ones, in downside
markets).
65. See Chris Stivers & Licheng Sun., Cross-Sectional Return Dispersion and
Time-Variation in Value and Momentum Premia, 45 J. Fin. & Quant.
Analysis 9871014 (2010).
66. Cooper, Gutierrez & Hameed, supra note 63, at 1347.
67. Antoniou, Dokas & Subrahmanyam, supra note 53, at 267.
68. See Clive W.J. Granger & Oskar Morgenstern, Spectral Analysis of
NewYork Stock Market Prices, 16 Kyklos 127 (1963).
69. See S.P. Kothari & Jay Shanken, Stock Return Variation and Expected
Dividends, 31J.Fin. Econ. 177210 (1992).
70. See Bruce Grundy & J.Spencer Martin, Understanding the Nature of the
Risks and the Source of the Rewards to Momentum Investing, 14 Rev. Fin.
Stud. 2978, 30 (2001); Jennifer Conrad & Gautam Kaul, An Anatomy
of Trading Strategies, 11 Rev. Fin. Stud. 489519 (1998). But cf.
Narasimhan Jegadeesh & Sheridan Titman, Cross-Sectional and Time-
Series Determinants of Momentum Returns, 15 Rev. Fin. Stud. 143157
(2002) (rejecting the suggestion that momentum profits are attributable
to cross-sectional differences in expected returns rather than time-series
dependence in returns and concluding instead that cross-sectional differ-
ences account for little or no momentum).
71. Daniel & Moskowitz, supra note 61, at 2.
72. Id. at 3.
73. Id. at 40.
74. See Clifford S. Asness, Tobias J. Moskowitz & Lasse Heje Pedersen,
Value and Momentum Everywhere, 68J.Fin. 929985, 930 (2013).
75. See generally Markus K.Brunnermeier & Lasse Heje Pedersen, Market
Liquidity and Funding Liquidity, 22 Rev. Fin. Stud. 22012238 (2009).
76. Asness, Moskowitz & Pedersen, supra note 74, at 962 (describing the
ability of value to earn a risk premium under these conditions an even
deeper puzzle).
77. Id.
78. Id. (citing Pedersen, When Everyone Runs for the Exit, supra note 56
(Chapter 4)).
79. See Yakov Amihud, Haim Mendelson & Lasse Heje Pedersen, Liquidity
and Asset Prices, 1 Foundations & Trends in Fin. 269364, 341350
(2006).
80. See id. at 322329.
81. See id. at 310311.
82. J.Michael Harrison & David M.Kreps, Speculative Investor Behavior in
a Stock Market with Heterogeneous Expectations, 92 Q.J.Econ. 323336,
323324 (1978).
83. See generally Cars H.Hommes, Heterogeneous Agent Models in Economics
and Finance, in 2 Handbook of Computational Economics 11091186
(Leigh Tesfatsion & Kenneth L.Judd eds., 2006).
84. See Carl Chiarella, Roberto Dieci & Laura Gardini, Asset Price and Wealth
Dynamics in a Financial Market with Heterogeneous Agents, 30J.Econ.
Dynamics & Control 17551786 (2006); Carl Chiarella & Xue-Zhong
He, Asset Price and Wealth Dynamics Under Heterogeneous Expectations,
1 Quant. Fin. 509526 (2001); Carl Chiarella, Xue-Zhong He & Duo
Wang, A Behavioral Asset Pricing Model with a Time-Varying Second
Moment, 29 Chaos, Solitons & Fractals 535555 (2006).
85. Compare Evan W. Anderson, Eric Ghysels & Jennifer L. Juergens, Do
Heterogeneous Beliefs Matter for Asset Pricing?, 18 Rev. Fin. Stud. 875924
(2005) (modeling asset prices on the basis of disagreement among analysts
about expected short-term and long-term earnings) with Simon Benninga
& Joram Mayshar, Heterogeneity and Option Pricing, 4 Rev. Derivatives
Research 727 (2000) (predicting option prices according to heterogene-
ity among traders based on the cohort holding out-of-the-money put
options, either as a function of relatively high risk aversion or as a function
of relatively high subjective expectations of low market outcomes).
86. See, e.g., James J. Heckman, Micro Data, Heterogeneity, and the
Evaluation of Public Policy: Nobel Lecture, 109 J. Pol. Econ. 673648
(2001); Joost M.E. Pennings & Philip Garcia, Risk and Hedging
Behavior: The Role and Determinants of Latent Heterogeneity, 33J.Fin.
Research 373401 (2010).
320 J.M. CHEN
87. See Jeffrey R. Brown, Stephen G. Dimmock, Jun-Koo Kang & Scott
J.Weisbenner, How University Endowments Respond to Financial Market
Shocks: Evidence and Implications, 104 Am. Econ. Rev. 931962 (2014).
88. Grinblatt & Hann, supra note 43, at 313.
89. Id. at 314.
90. See Stephen Morris, Speculative Investor Behavior and Learning, 111
Q.J.Econ. 11111133 (1996).
91. See Wei Xiong & Hongjun Yan, Heterogeneous Expectations and Bond
Markets, 23 Rev. Fin. Stud. 14331466 (2000) (arguing that this
dynamic explains several strange aspects of bond markets, including
excess volatility in yields and the ability of a tent-shaped linear combina-
tion of forward rates to predict bond returns).
92. Amihud, Mendelson & Pederson, supra note 79, at 290. Compare Viral
V. Acharya & Lasse Heje Pedersen, Asset Pricing with Liquidity Risk,
77J.Fin. Econ. 375410 (2005) (outlining the theoretical basis for per-
sistence in liquidity over time) with Yakov Amihud, Illiquidity and Stock
Returns: Cross-Section and Time Series Effects, 5 J. Fin. Mkts. 3156
(2002) (providing empirical evidence that illiquidity persists over time).
See generally Markus Brunnermeier & Lasse Heje Pedersen, Market
Liquidity and Funding Liquidity, 22 Rev. Fin. Stud. 22012238 (2009)
(explaining intertemporal variations in liquidity according to changes in
the funding conditions faced by market makers).
93. Amihud, Mendelson & Pederson, supra note 79, at 290.
94. Barberis & Huang, Stocks as Lotteries, supra note 117 (Chapter 8), at
2092.
88.
96. Id.
97. Id.
98. Kenneth D.West, Bubbles, Fads and Stock Price Volatility Tests: A Partial
Evaluation, 43J.Fin. 639656, 640 (1988).
99. Id.
79 (Chapter 1), at 334 (emphasis omitted).
101. Id.
102. See Bart Frijns, Ester Koellen & Thorsten Lehnert, On the Determinants
of Portfolio Choice, 66J.Econ. Behav. & Org. 373386, 385 (2008).
79 (Chapter 1), at 334. See generally Jonathan M.Karpoff, A Theory of
Trading Volume, 41 J. Fin. 10691098 (1986); Jonathan M. Karpoff,
The Relation Between Price Changes and Trading Volume: A Survey,
104. See John M. Griffin, Federico Nardari & Ren M. Stulz, Do Investors
Trade More When Stocks Have Performed Well? Evidence from 46
Countries, 20 Rev. Fin. Stud. 905951 (2007); Meir Statman, Steven
Thorley & Keith Vorkink, Investor Overconfidence and Trading Volume,
19 Rev. Fin. Stud. 15311565 (2006).
105. Malcolm Baker & Jeffrey Wurgler, Investor Sentiment in the Stock Market,
21J.Econ. Persp. 129151, 137 (2007); see also Jos Scheinkman & Wei
Xiong, Overconfidence and Speculative Bubbles, 111J.Pol. Econ. 1183
1219 (2003).
106. See Harrison Hong & Jeremy C. Stein, Disagreement and the Stock
Market, 12J.Econ. Persp. 109128 (2007).
107. See Barberis & Xiong, Realization Utility, supra note 51 (Chapter 8), at
265, 267.
108. See Jeremy C.Stein, Prices and Trading Volume in the Housing Market: A
Model with Down-Payment Effects, 110 Q.J.Econ. 379406 (1995).
109. Ana Fostel & John Geanakoplos, Leverage Cycles and the Anxious
Economy, 98 Am. Econ. Rev. 12111244, 1211 (2008) (emphasis
omitted).
110. Id. at 1214.
111. Id. at 1211.
112. See Barberis & Huang, Realization Utility, supra note 51 (Chapter 8), at
267.
113. Markus K. Brunnermeier, Stefan Nagel & Lasse H. Pedersen, Carry
Trades and Currency Crashes, in NBER Macroeconomics Annual 2008,
at 313347, 315 (Daron Acemoglu, Kenneth Rogoff & Michael
Woodford eds., 2009).
114. Cf. Flannery OConnor, Everything That Rises Must Converge (1965).
See generally Chen, Postmodern Portfolio Theory, supra note 1 (Chapter
1), 4.1, at 4144 (introducing the idea of dividing systematic risk into
its upside and downside components).
115. See, e.g., Acharya & Pedersen, supra note 92; Franklin Allen & Douglas
Gale, Financial Fragility, Liquidity, and Asset Prices, 2 J. Eur. Econ.
Assn 10151048 (2005); Sanford J. Grossman & Merton H. Miller,
Liquidity and Market Structure, 43 J. Fin. 617633 (1988); Lubo
Pstor & Robert F. Stambaugh, Liquidity Risk and Expected Stock
Returns, 111J.Pol. Econ. 642685 (2003).
116. See supra 4.2, at 78.
117. Flavin & Panopoulou, supra note 58 (Chapter 4), at 401. See generally
Gravelle, Kichian & Morley, supra note 59 (Chapter 4).
118. Flavin & Panopoulou, supra note 58 (Chapter 4), at 401402. See gener-
ally Pericoli & Sbracia, supra note 61 (Chapter 4).
119. Flavin & Panopoulou, supra note 58 (Chapter 4), at 402.
322 J.M. CHEN
120. See generally, e.g., Shefrin, A Behavioral Approach to Asset Pricing, supra
note 93 (Chapter 10); Didier Sornette, Critical Market Crashes, 378
Phys. Reports 198 (2003).
121. Robert J.Shiller, Stock Prices and Social Dynamics, 2 Brookings Papers on
Econ. Activity 457510, 477481 (1984); accord Shiller, Irrational
Exuberance, supra note 89 (Chapter 1), at 254255.
123. See id.; Shiller, Stock Prices and Social Dynamics, supra note 121, at 475
(equation 1).
124. Shiller, Stock Prices and Social Dynamics, supra note 121, at 477 (equa-
tion 2).
125. Id.
126. Id. On fads, see generally Shiller, Speculative Prices and Population
Models, supra note 119 (Chapter 9).
127. Shiller, Stock Prices and Social Dynamics, supra note 121, at 478; cf.
Shiller, Irrational Exuberance, supra note 89 (Chapter 1), at 254 (elect-
ing to designate this variable as At, in honor of animal spirits).
128. Shiller, Stock Prices and Social Dynamics, supra note 121, at 478 (equa-
tion 3)
129. Id.
130. Id.
131. Id.
132. Id.
133. See Shiller, Irrational Exuberance, supra note 89 (Chapter 1), at 255.
134. Id.
135. Id.
136. Id.
137. See Christopher Avery & Peter Zemsky, Multidimensional Uncertainty
and Herd Behavior in Financial Markets, 88 Am. Econ. Rev. 724748
(1998); In Ho Lee, Market Crashes and Informational Avalanches, 65
Rev. Econ. Stud. 741760 (1998); cf. Abhijit V. Banerjee, A Simple
Model of Herd Behavior, 107 Q.J. Econ. 787817 (1992) (outlining a
more general theory of herd behavior involving otherwise rational
agents); S.D.Bikhchandani, David Hirshleifer & Ivo Welch, A Theory of
Fashion, Social Custom and Cultural Change, 81J.Pol. Econ. 647654
(1992) (same).
138. See Shiller, Irrational Exuberance, supra note 89 (Chapter 1), at 182184
(citing Norman T. Bailey, The Mathematical Theory of Epidemics
(1957); Alan Kirman, Ants, Rationality and Recruitment, 108 Q.J.Econ.
137156 (1993)).
139. Shiller, Irrational Exuberance, supra note 89 (Chapter 1), at 178 (empha-
sis in original).
CHAPTER 13
The Monster andtheSleeping Queen
Postmodern portfolio theory, as depicted in this book, provides an over-

arching view of abnormal markets and less than fully rational investor
behavior. Whereas market abnormality could be described in terms of
econophysics, complete understanding of investor irrationality demands
knowledge of neuroscience, evolutionary biology, and epidemiology. A
federal court has described agriculture as a field so vast that fully to com-
prehend it would require an almost universal knowledge ranging from
geology, biology, chemistry and medicine to the niceties of the legisla-
tive, judicial and administrative processes of government.1 A comparable
claim befits finance, particularly as it embraces the perspectives and meth-
odologies of the behavioral sciences.
Economics has traditionally confined itself to the realm of the ideal,
to stylized facts and hypothetical decision-making mechanisms. Because
[m]ost economists view[ed] their discipline as one that deals with ideally
rational behavior, they generally attach[ed] little significance to discrep-
ancies between what the theory predicts and what people actually do.2
Neither the returns from actual markets nor behaviorally mediated human
responses to those markets sustain such a sanitized view of finance. Finance
does not live by rationality alone, but by every passion that animate[s]
peoples ideas and feelings, their animal spirit.3
Abnormality, in the sense of statistically significant departures
from Gaussian assumptions, is the modal condition of capital markets.

DOI10.1007/978-3-319-32711-2_13
324 J.M. CHEN
Confronted with volatility, spillover, and contagion, and perhaps even

touched by an awareness of profound limits on the human ability to
detect deep financial threats, human actors take refuge in time-worn albeit
imperfectly rational heuristics: mental accounting, subjective weighting of
probabilities, systematic disregard of correlation.
In behavioral portfolio theorys depiction of the embattled human
decision-maker at war with herself,4 the emotional basis of fast, system
1 thinking appears to have a built-in advantage over the cool, slow ratio-
nality of system 2. There are deep, evolutionary reasons for the preemi-
nence of emotion over cognitions: emotions can flood consciousness
because the wiring of the brain has evolved so that connections from
emotional systems to cognitive systems are stronger than vice versa.5
Even if rationality and cognition must concede the upper hand to the
emotional heuristics by which humans make most of their decisions, how-
ever, it is far from clear that the victory of fast, system 1 instinct over
cool, slow, system 2 reason represents a defeat. Resolutions of human
conflict between desire and reason appear to rest on ethical grounds, on
what people ultimately decide is right, and not simply on the basis of
what people want.6 And naturally enough, morality is perhaps the most
deeply emotional, least mechanistically rational projection of the human
mind at work.7
Moreover, as Amos Tversky and Daniel Kahneman have said of pros-
pect theory, even if humans make choices that are not always rational in
the traditional sense, behavioral decision-making is neither chaotic nor
intractable, but rather is orderly according to its own criteria.8 The
inconvenient fact that neoclassical economics and behavioral finance may
be incompatible with each other in some circumstances should not
entail relativism of [the] values with each framework, only the notion
of a plurality of values not structured hierarchically.9 Humanity, with all
its limitations, may well be the source of the stability we lack in our more
abstract conceptions in the universe of finance.10
The behavioral turn aligns finance with the mission of psychology,
whose practitioners have traditionally view[ed] their task as one of pre-
dicting behavior and describing its cognitive sources in psychologically
meaningful terms, whether or not that behavior is rational.11 From the
idealized, purely rational perspective of modern portfolio theory, staring
into the new behavioral universe might seem an unspeakable horror.12
But behavioral finance is neither madness nor Hell; it is Knowledge,
and we are obliged to open [our] eye[s] once again and [to] try to look
steadily.13
THE MONSTER ANDTHESLEEPING QUEEN 325
Nicht Gott sondern Mensch wrfelt mit dem Universum: It is not God,
but rather humanity, that plays dice with the universe.14 Behavioral finance
reflects the quantum mechanics of human judgment. The human animal
has its own elaborate mechanism for making decisions under uncertainty.
An analogy to linguistics may be helpful. Natural language is not just a
cultural artifact, but a species property arising from human biology, like
echolocation in bats or stereopsis in monkeys.15 Risk-taking is likewise
a species property, one whose underlying dispositions are susceptible to
formal description, mathematical specification, and empirical verification
or falsification. Behavioral finance may not be rational in a conventional
economic sense, but it arises from the same material as universal grammar
and the atoms of language.16
Cognitive biases, innate heuristics, and the entire apparatus of behav-
ioral finance coexist alongside the stylized rationality of neoclassic econom-
ics. As humans, we should strive to speak of our own condition exactly as
we find it. [N]othing extenuate / Nor set down aught in malice.17 What
we discover may give cause to shudder or adore / Before the flame of
financial behavior:
This is the monster and the sleeping queen

And both have roots struck deep in your own mind,
This is reality that you have seen,
This is reality that made you blind
If you at last must have a word to say,
Say neither, in their way,
It is a deadly magic and accursed,
Nor It is blest, but only It is here.18
NOTES
1. Queensboro Farms Prods., Inc. v. Wickard, 137 F.2d 969, 975 (2d Cir.
1943).
3. George A. Akerlof & Robert J. Shiller, Animal Spirits: How Human
Psychology Drives the Economy, and Why That Matters for Global
Capitalism 1 (2009).
4. See supra 10.3, at 252256.
5. Joseph LeDoux, The Emotional Brain 19 (1996).
6. See Daniel Read, Which Side Are You On? The Ethics of Self-Command,
27J.Econ. Psych. 681-693 (2006).
326 J.M. CHEN
7. See, e.g., Jesse Graham, Brian A.Nosek, Jonathan Haidt, Ravi Iyer, Spassena
Koleva & Peter H.Ditto, Mapping the Moral Domain, 101J.Personality
& Soc. Psych. 366385 (2011); Spassena P.Koleva, Jesse Graham, Ravi
Iyer, Peter H. Ditto & Jonathan Haidt, Tracing the Threads: How Five
Moral Concerns (Especially Purity) Help Explain Culture War Attitudes,
46J.Research in Personality 184194 (2012).
8. Tversky & Kahneman, Advances in Prospect Theory, supra note 58 (Chapter
8), at 317.
9. Isaiah Berlin, Alleged Relativism in Eighteenth-Century European Thought,
in The Crooked Timber of Humanity: Essays and Chapters in the History
of Ideas 7394, 83 (2d ed. 2013, John Banville foreword).
10. Greg B.Davies & Arnaud de Servigny, Behavioral Investment Management:
An Efficient Alternative to Modern Portfolio Theory, at viii (2012).
12. Edwin A.Abbott, Flatland: A Romance of Many Dimensions 64 (Dover
thrift ed., 1992) (1st ed. 1884).
13. Id.
14. God does not play dice with the universe, a saying attributed to Albert
Einstein, arises from a December 4, 1926, letter to Max Born:
Die Quantenmechanik ist sehr achtunggebietend. Aber eine innere Stimme
sagt mir, da das noch nicht der wahre Jakob ist. Die Theorie liefert viel,
aber dem Geheimnis des Alten bringt sie uns kaum nher. Jedenfalls bin
ich berzeugt, da der nicht wrfelt.
Max Born, Physik im Wandel Meiner Zeit 244 (3d ed. 1959); Albert Einstein,
Hedwig und Max Born: Briefwechsel 19161955, at 97 (Bertrand Russell
preface, Werner Heisenberg intro., 1972). In my English translation:
Quantum mechanics is certainly impressive. But a voice inside tells me that
its not yet the real McCoy. The theory says a lot, but it barely gets us
closer to the secrets of God. Anyway, I am convinced that He does not play
dice.
15. Steven Pinker & Paul Bloom, Natural Language and Natural Selection, in
The Adapted Mind: Evolutionary Psychology and the Generation of
Culture 451493, 451 (Jerome H.Barkow, Leda Cosmides & John Tooby
eds., 1992).
16. See generally Mark C.Baker, The Atoms of Language: The Minds Hidden
Rules of Grammar (2002).
17. Shakespeare, Othello, act V, sc. 2, ll. 351352, in The Oxford Shakespeare,
supra note 232 (Chapter 7), at 819853, 853; cf. Bernhard Schlink, The
Reader 217 (Carol Brown Janeway trans., 1997) (concluding that once we
know that a story is true, the question of whether it is sad or happy has
no meaning whatever).
18. Stephen Vincent Bent, John Browns Body 336 (1990) (1st ed. 1928).
INDEX
A annuities, 202
Abel, Andrew B., 145, 151, 152, anomalies. See also paradoxes and
166n73, 170n117, 171n131, puzzles
175n181, 233n11 calendar anomalies (e.g., the
adaptation level, 1846, 188, 189. See January effect and sell in May
also prospect theoryreference and go away), 7, 11, 19n39,
point 293, 299n83
adverse selection, 8, 143, 154 generally, 1, 2, 4, 7, 11, 16, 67, 73,
affect heuristic, 13, 39 93, 101, 182, 301
Affordable Care Act, 9 geomagnetic storms, 7
Africa, 263 The Ant and the Grasshopper (La
agriculture, 157, 262, 323 Cigale et la Fourmi), 40, 41
Akaike information criterion, 259 anthropology, 34
Allais, Maurice, 125, 126, 134n74, Appleton, Jay, 251, 271n47
134n76, 183, 269n2 arbitrage and limits to arbitrage, 4,
Allais paradox, 1226, 183, 198, 249 7, 219, 288, 302, 303,
alpha, 37, 74, 100, 208n7780, 306, 312
209n84, 209n85, 248, 286 Argentina, 79
alternative investments, 268 Arnott, Robert D., 139, 162n19,
American Stock Exchange, 287 162n20
Ang, Andrew, 81, 83, 85n10, 85n14, Arrow-Debreu model of complete
86n17, 91n86, 91n9294, markets, 94
91n96, 91n98, 92n105, 92n107, Arrow, Kenneth J., 104n11, 111,
106n26, 109n74, 169n104 129n7, 243n180

DOI10.1007/978-3-319-32711-2
328 INDEX
Arrow-Pratt measures of risk aversion, and beta, 2, 5, 11, 5760, 7380,

11113, 120, 182 82, 83, 93, 96, 98103, 226,
art, 39 303, 306
landscape painting, 251 endowment effect, 29, 149, 187
asset allocation and the equity premium puzzle, 1,
nave allocation (1/n), 31 3, 95, 111, 117, 13779, 181,
strategic allocation, 286 213, 2289, 231
tactical allocation, 286, 302 generally, 2, 30, 34, 66, 77, 154,
Australia, 7, 224 156, 185, 198, 200, 213, 216,
217, 221, 223, 250, 253, 286,
289, 323
B mental accounting, 1, 17, 2957,
Baker, Malcolm P., 53n123, 82, 85n8, 151, 202, 260, 264, 304, 324
88n52, 321n104 and risk aversion, 2, 3, 33, 44n15,
banking, 49n84, 224, 226 45n16, 48n66, 65, 67, 94,
bankruptcy risk, 219 11135, 138, 141, 1469,
Barber, Brad M., 22n74, 26n126, 1525, 167n78, 169n103,
44n16, 88n47, 176n197, 1825, 188, 198202, 208n72,
233n10, 285, 293n6, 294n22, 211n133, 214, 222, 228, 230,
294n24, 296n39, 297n53, 231, 252, 258, 261, 264,
297n60, 298n757, 299n81, 319n85
303, 314n17, 318n60 behavioral portfolio theory
Barberis, Nicholas C., 43n11, 162n30, compared to diamond mining, 263
205n1, 207n51, 210n117, compared to SP/A theory, 4, 67,
22931, 232n1, 233n16, 181, 24850, 2535, 25760,
233n18, 233n20, 238n96, 262, 265, 268, 269
238n97, 239n102, 240n121, compared to subsistence farming,
240n122, 240n134, 240n135, 256, 257, 262, 263
244n202, 245n212, Belgium, 224
245n21517, 245n222, Benartzi, Shlomo, 16n26, 45n20,
245n226, 246n227, 246n238, 45n27, 165n55, 165n58,
246n244, 246n245, 246n250, 170n123, 170n124, 175n185,
249n13, 293n12, 294n14, 175n186, 177n200, 177n210,
316n41, 320n94, 321n107, 177n212, 178n214, 178n220,
321n112 207n53, 207n64, 210n111,
behavioral economics and finance. See 210n115, 228, 229, 244n198,
also behavioral portfolio theory; 245n203, 245n210, 298n76,
cognitive biases; prospect theory; 320n95
SP/A theory Bernoulli, Daniel, 127, 135n83
behavioral environmental Bernstein, William J., 73, 85n1, 85n5
economics, 2, 3840 beta
behavioral portfolio theory, 4, 67, Bad Beta, Good Beta, 98101, 103
24781, 324 behavioral implications, 5760
INDEX 329
bifurcation, 75, 76, 103 performance relative to stocks, 288

and Bowmans paradox, 2, 735, 226 sovereign debt default, 153
cash-flow beta (bad beta), 98103 term structure, 11
correlated relative volatility, 77 treasury bills and notes, 125, 140
defined, 82 Born, Max, 326n14
discount-rate beta (good beta), Boston College, 142
98103 Bowie, David, 154
and emerging markets, 79, 80, 101 Bowman, Edward H., 74, 86n20,
generally, 2, 11, 5860, 76, 82, 83, 86n21, 269. See also Bowmans
93, 98, 303, 306 paradox
and growth stocks, 73, 98102, 226 Bowmans paradox
high- and low-beta portfolios, 74, 82 and prospect theory, 75, 181,
single-sided, 79 2238, 253
and value stocks, 73, 74, 98102, 226 and SP/A theory, 4, 67, 181,
Bhutan, 38 24850, 2535, 257, 258, 269
biology and medicine Bradley, Brendan, 82, 85n8
agriculture, 323 brain damage, 13
ants, 312 Brazil, 79
biotechnology, 37, 268 Buffett, Warren, 14, 27n132
blood glucose, 6
brain damage, 13
epidemiology, 323 C
evolution, 34, 35, 139, 323 calendar anomalies
infections, 312 generally, 7, 11,
magnetic resonance imaging, 220 January effect, 7, 19n39, 293,
neuroscience, 323 299n83
pigeons, 184, 215 sell in May and go away, 67, 11
polar bears, 41 and storytelling, 12
punctuated equilibria, 139 Campbell, John Y., 17n5, 17n7,
unicorns (fallacious medical 91n91, 91n92, 93, 98, 1003,
diagnoses), 30, 153, 173n155 105n25, 107n47, 108n48,
biotechnology, 37, 268 109n70, 109n71, 133n68, 151,
Black, Fischer, 22n79, 22n80, 67n2, 152, 165n57, 168n90, 168n103,
109n12, 158, 178n215, 297n53 172n145, 237n87, 293n9
Black-Scholes pricing model, 94 Canada, 7, 41, 287
bonds capital allocation line, 119, 131
generally, 3, 125, 138, 141, 213, capital asset pricing model (CAPM)
216, 228, 229, 260 and beta, 2, 11, 5760, 74, 76, 79,
liquidity risk, 309 93, 96, 98103, 226, 303
momentum, 304, 306, 309, 315n35 and consumption smoothing, 95
paired with lottery-like investments contrasted with models of behavioral
(bonds-and-bullets finance, 250
portfolios), 260, 269 criticism, 67, 260
330 INDEX
capital asset pricing model (CAPM) certainty effect, 126, 183, 202, 203
(cont.) cognitive dissonance, 305
and the equity premium puzzle, 1, discounting (exponential,
3, 95, 137, 143, 144, 148, hyperbolic, etc.), 99, 147
152, 153, 155 disposition effect, 213, 2836, 292,
and the equity risk premium, 3, 137, 304, 308, 312
143, 144, 148, 152, 153, 155 endowment effect, 29, 149, 187
four-moment (higher-moment) framing effects, 197, 205
CAPM, 57, 607, 76, 97, 116, generally, 11, 17, 41, 302, 305, 325
203, 222, 247, 250 house money effect, 230
generally, 1, 5760, 93, 95100, loss aversion, 1859, 198, 201,
137, 222, 226, 250 22831, 308
and initial public offerings (IPOs), myopic loss aversion, 2289, 231,
3, 203, 222 308
intertemporal CAPM, 67, 76, 938, narrow framing, 231, 283
102, 103, 148, 158, 230 projection bias, 97, 150
unobservability (Rolls critique), 96, rank effect, rankings, 31
155 realization utility, 185, 285, 309
capital gains, 95, 289, 2913, 298n78, reflection effect, 183, 184
308 status quo bias, 185
CAPM. See capital asset pricing model colleges and universities
Carhart, Mark M., 97, 303, 314n25. Boston College, 142
See also factor models, momentum Cornell University, 187
carry trade, 310 university endowments, 15760, 308
catastrophe of success, 214, 314n16 Yale University, 157
catching up with the Joneses, 151, commodities markets, 157, 304, 306
152, 156. See also Abel, Andrew concentration (of assets or portfolios),
B.; lagged consumption 5, 225, 286. See also
Catholic Church, 220 diversification
CBOE. See Chicago Board Options conglomerates, 218
Exchange (CBOE) conspicuous consumption, 151
Chen, Joseph, 81, 83, 85n10, 90n81, Constantinides, George M., 20n40,
91n86 133n68, 14952, 169n106,
Chen, Zhanhui, 83, 92n108, 102, 172n139, 174n168, 174n170,
103, 104n1 174n171, 299n82
Chicago Board Options Exchange constant relative risk aversion (CRRA),
(CBOE), 77 117, 118, 128, 144, 149, 264
Chile, 79 consumption smoothing, 95, 154, 158
China, 153 contagion, 78, 302, 309, 310, 324
cognitive biases Cornell University, 187
affect heuristic, 5, 11, 31, 41, corporate distress, 73
293, 325 corporate social responsibility (CSR), 36
INDEX 331
corporate spin-offs, 218 decision weights, 126, 197, 205, 251

correlation decumulative distribution function,
correlation risk, 78, 805, 257. See also cumulative
92n114, 102 distribution function
correlation tightening, 7780, de la Fontaine, Jean, 41, 55n147,
82, 83, 101, 102 56n149, 156
disregard, 30, 225, 324 developed markets, 7880, 90n68,
and emerging markets, 7880, 109n83
101, 304 Dichev, Ilia D., 238n88, 287, 296n44,
between financial and labor income, 296n49, 298n74
154, 219 disaster relief, 29, 54n131
in global markets, 157 discount brokerages, 292. See also
single-sided (downside), 79, 84, 120 individual investors
Corzine, Jon, 203 discounting (exponential, hyperbolic,
countercyclicality, 155 etc.), 43, 99, 147
covariance, 30, 145, 221 disposition effect, 213, 283, 2846,
crashes and crises 292, 304, 308, 312
financial crisis of 2008-09, 157, 158 distributions, statistical
stock market crash of 1929, 302 extreme value, 66, 257
stock market crash of 1987, 82 Fisk (log-logistic), 189
technology bubble of 2000, 158 Gaussian (normal), 64, 193, 266,
crowdfunding, 13, 217, 218 267, 310, 323
CRRA. See constant relative risk lognormal, 146, 148, 151, 166n75,
aversion (CRRA) 171n133, 1936, 208n81,
cumulative distribution function 209n83
(CDF), 120, 128, 18991, stable Paretian, 118, 131n36
1937, 250, 266, 267. See also diversification, 14, 31, 77, 78, 81, 82,
decumulative distribution 87n40, 102, 154, 156, 158, 213,
function 216, 219, 225, 286, 304. See also
cumulative prospect theory, 187, concentration (of assets or
1968, 200, 214, 222, 224, 228, portfolios); retirement and
251, 254, 308 retirement savings-failure to
cyclicality risk, 155 diversify
dividends, 14, 95, 99, 139, 140, 145,
218, 230, 289, 291, 292, 303,
D 311, 312
Dalbar (financial analytics firm), 2879 Driessen, Joost, 84, 91n88,
Dndolos, Nkos (Nick the Greek), 252 92n11315
death and jackpot (bankruptcy dumb money, 287, 288, 302,
risk), 219 311, 312. See also noise, noise
death anxiety, 214 trading
332 INDEX
E exchange-traded funds (ETFs), 231,

econophysics, 46n30, 323 268, 298n64
efficient frontier, 2, 31, 94, 157, 216, expected utility theory, 3, 111, 122,
260, 261, 295n30 123, 1257, 1824, 185, 197,
efficient markets hypothesis, 5, 7, 213, 223, 247, 249, 254, 258
11, 247
Einstein, Albert, 326n14
emerging markets F
and capital asset pricing model factor models
(CAPM), 79, 101, 109n83 Fama-French-Carhart four-factor
and correlation tightening, 7880 model (size, value and
Endangered Species Act, 42 momentum), 97, 100, 226,
endowment effect, 29, 149, 187 303
endowment hoarding, 159 Fama-French three factor model
endowment model of investing, (size and value), 97, 303
87n40, 157 fads, 222, 311
environmental economics, 2, 31, Fama, Eugene F., 1, 7, 19n38, 58,
3842 68n2, 68n5, 68n8, 90n81, 97,
environmental, social, and governance- 98, 107n42, 118, 131n33,
related (ESG) corporate 131n34, 295n36, 303, 314n22,
practices, 37 315n27. See also factor models;
Epstein, Larry G., 1467, 167n80 French, Kenneth R.
equity crowdfunding, 217, 218 family-owned firms, 227, 244n189
equity premium puzzle, 1, 3, 95, 111, fast vs. slow thinking, 5
117, 13779, 181, 207n66, 213, Faulkner, William, 253, 273n80
2289, 231 Federal Reserve System, 141, 142,
and prospect theory, 184, 255, 304 203, 301, 313n2
risk-free rate puzzle, 148 federal securities law. See also
equity risk premium, 3, 111, 13779, Securities and Exchange
228, 230, 231, 244n202 Commission (SEC)
escalation of commitment (to losing accredited investors, 217
strategies), 204, 212n154, 227 equity crowdfunding, 21718
Estrada, Javier, 63, 64, 70n33, 70n43, exchange-traded funds (ETFs),
70n45, 71n61, 79, 80, 83, 298n64
85n15, 90n75, 90n77, 90n81, fraud on the market, 8
1013, 243n176 Investment Company Act
ETFs. See exchange-traded funds of 1940, 289
(ETFs) Jump Our Business Startups Act
Europe, 23n96, 216, 235n40 of2012 (JOBS Act), 21718
evolutionary economics and finance, mutual funds, 288
34, 264 Regulation D, 217, 218, 236n50
Excel (spreadsheet software), 291 structured securities, 1067n33
INDEX 333
federal tax law gambler's ruin, 127, 264

capital gains and losses, 292, house money effect, 230
298n78, 308 Kelly criterion, 13, 25n116, 135n83
mortgage interest deduction, 163n41 and SP/A theory, 4, 67, 254, 255,
tax loss harvesting, 292, 293, 257, 259
299n81, 299n83 Gaussian mathematics. See
tax-sheltered retirement and distributions, statistical-Gaussian
educational savings (normal)
accounts, 150 GDP. See gross domestic product
Fiegenbaum, Avi, 86n30, 2246, (GDP)
241n138, 241n141, 244n193 general equilibrium models, 94, 149,
financial crisis of 2008-09, 82, 157, 158 169n103
financial (il)literacy, 142, 216, 234n36 Germany, 80, 139, 140, 174n164
Fishburn, Peter C., 128n4, 132n60, goal-oriented investing, 32, 33
206n24, 207n489, 224, 242 Graham, John, 138
foreign exchange, 90n68, 102, Great Britain, 80, 139, 140, 174n164,
134n74 216, 287
carry trade, 310 Greenspan, Alan, 301, 313n2
four-moment (higher-moment) Grizzard, Lewis, 207n55
CAPM, 57, 60, 67, 76, 97, 116, gross domestic product (GDP), 38,
203, 250 100, 284
France, 80, 139, 140 growth stocks, growth investment
fraud on the market, 8 styles, 73, 74, 98102, 152, 226
French, Kenneth R., 1, 19n35, 68n2,
68n5, 97, 174n164, 295n36,
295n37, 303, 314n223, H
315n27. See also factor models; habit formation, 14952, 155, 156,
Fama, Eugene F. 158, 160
FriedmanSavage utility function, 123, HARA. See hyperbolic absolute risk
124, 126 aversion (HARA)
Friesen, Geoffrey C., 296n43 Harvey, Campbell R., 69n25, 70n42,
fundamental vs. technical analysis, 89n57, 89n64, 90n68, 138,
3078 161n12, 176n190, 179n231,
future consumption utility, 229 315n34
Haugen, Robert A., 74, 85n13,
85n15
G hedge funds and hedge fund
Galbraith, John Kenneth, 24n109, managers, 287, 301, 302
278n191, 302, 313n3 Heins, A.James, 74, 85n13
gambling and gamblers. See also herding behavior, 302
lotteries, lottery tickets and heterogeneity, 4, 16, 222, 30710
lottery-like investments; risk higher-moment CAPM, 60, 222, 247
seeking behavior high-frequency trading, 8, 9, 11
334 INDEX
hoarding behavior (by institutional investing and investors,

endowments), 159 2, 302, 77, 100, 157, 158,
home bias (preference for domestic 219, 308
stock), 156, 157 insurance, 8, 47n56, 82, 95, 96,
household portfolios, 133n68, 151, 11415, 124, 143, 154, 166n66,
172n144, 175n178 182, 183, 196, 197, 202, 204,
housing. See real estate 249, 261, 278n186
Huang, Ming, 43n11, 162n30, internal rate of return (IRR), 285,
210n117, 22931, 240n121, 28992
244n202, 246n238, 246n245 secant method, 290
human capital and human capital risk, intertemporal capital asset pricing
38, 51n102, 96, 154 model, 2, 93109
hyperbolic absolute risk aversion intertemporal substitution, 147
(HARA), 2, 65, 11518, 128, investment gap (between hypothetical
132n58, 147, 182. See also risk and actual returns), 2878, 302
aversionquadratic utility investor return, 1, 4, 182, 28399
IPOs. See initial public offerings
(IPOs)
I irrational exuberance, 1, 30122
ideal gas law, 248, 270n13 Islamic finance, 36, 49n84
idiosyncratic risk, 81, 82, 84
Iksil, Bruno (London Whale), 203
implied volatility, 77 J
independence axiom, 126. Japan, 139, 140, 174n169, 304
See also FriedmanSavage utility
function
indexing, index funds, 199, 283 K
individual companies, by stock symbol Kahneman, Daniel, 5, 18n11, 18n27,
Aetna (AET), 8, 10 26n122, 27n128, 43n4, 45n21,
Anthem/WellPoint (WLP), 8, 10 52n115, 53n116, 127, 135n90,
Humana (HUM), 8, 10 169n108, 182, 183, 185, 187,
J.P.Morgan Chase (JPM), 203 188, 196, 197, 199204,
Myriad Genetics (MYGN), 8, 10 205n23, 207n52, 207n58,
individual investors, 9, 32, 97, 98, 207n68, 214, 230, 232, 249,
140, 153, 156, 157, 178n213, 261, 270n3, 324. See also fast
213, 219, 231, 283, 284, 286, and slow thinking; prospect
288, 303, 308, 309 theory; System 1 and System 2
information trading, informed trading, modes of thought; Tversky, Amos
11, 288, 302 Kaplan, Paul D., 120, 131n50,
initial public offerings (IPOs), 3, 13, 295n31
37, 51n97, 203, 2202, kappa, 11922
240n120, 269 Keating, Con, 120, 132n51
INDEX 335
Kelly criterion (Kelly, J.L.), 13, 271n43, 272n5860, 272n62,

25n116, 135n83 272n63, 272n65, 272n66,
Keynes, John Maynard, 13, 25n117, 272n72, 272n76, 273n75,
77, 88n49, 107n45 273n77, 273n81, 273n96,
Knowles, James A., 120, 131n50, 273n98, 274n106, 275n131,
132n52, 132n56 275n133, 276n141, 276n145,
Kochenberger, Gary A., 206, 224, 242 276n146, 278n187, 278n192,
Kocherlakota, Narayana R., 148, 279n198, 279n199, 325n2,
161n9, 167n77 326n11
Kronecker delta function, 187, loss aversion, 1859, 198, 201, 207n66,
208n69 210n115, 22831, 298n76
kurtosis, 1, 57, 61, 637, 222, 247 lotteries, lottery tickets and lottery-like
kurtosis preference, 263 investments, 35, 66, 124,
133n66, 196, 203, 21321,
234n34, 235n41, 249, 250, 252,
L 260, 261, 263, 269, 271n33,
lagged consumption, 152 278n186, 281n248, 281n249.
Las Vegas, 253 See also prospect theory
"law on the market," 711, 21n55 xskewness preference
Leeson, Nick, 203, 211n146 lottery bonds, 216
leverage, leverage effects, 151 low-volatility anomaly, 13, 37,
L'Hpital's rule (L'Hpital, Guillaume 7393, 98, 1013, 148, 223,
de), 117 2258, 230, 253
life-cycle model, 95, 133n72, 154
linguistics, 139, 325
Lintner, John, 58, 68n2, 130n27 M
liquidity, 8, 78, 79, 94, 109n83, 125, Maclaurin series (Maclaurin,
260, 30710, 312, 320n92 Colin), 70n31
funding vs. market liquidity, 307 macroeconomics
liquidity preference, 94 (see also and distorted savings, 151
(Tobin, James)) and financial risk aversion, 214
liquidity risk, 30710 generally, 8, 9, 78, 99, 100, 143,
logarithmic (log) returns, 2, 62, 63, 150, 151, 155, 304
250, 289 and the life cycle model, 150, 155
London whale (Iksil, Bruno), 203 and momentum, 100, 304,
Longin, Franois, 80, 91n82 316n40
Long-Term Capital Management, 203, Maenhout, Pascal J., 84, 91n88,
204, 211n147 92n11315
Lopes, Lola L., 47n55, 132n58, 248, magnetic resonance imaging
249, 270n4, 270n7, 270n14, (MRI), 220
271n21, 271n235, 271n30, Mandelbrot, Benoit B., 118, 130n30,
271n33, 271n34, 271n3941, 131n32, 131n33, 131n35
336 INDEX
marathons and marathon runners, 30, 295n32, 321n115. See also

43n12 intertemporal capital asset pricing
market timing, 286 model
Markham, Edwin, 42, 56n155 Mexico, 79
Markowitz, Harry M., 47n42, MF Global, 203
47n446, 47n50, 47n56, 124, modern portfolio theory, 1, 3, 4, 7,
127, 128, 130n28, 131n46, 11, 13, 16, 31, 32, 94, 98, 141,
132n64, 133n66, 169n104, 226, 247, 248, 250, 256,
169n105, 184, 206n30, 255, 25961, 264, 324, 326n10
256, 265, 271n29, 271n38, moments-based models, 250
274n111, 277n165, 278n183, momentum, 1, 4, 36, 79, 97, 100,
279n21214, 280n227, 280n229 108n65, 182, 213, 226, 230,
Marshall-Arrow-Romer (MAR) 243n176, 30122. See also factor
spillovers, 243n180 modelsFama-French-Carhart
martingale, 127, 135n85 four-factor model (size, value and
Maslow, Abraham H., 326, 46n34, momentum)
48n62, 48n70, 48n75 behavioral explanations, 306
Maslowian portfolio theory momentum crashes, 182, 30122
hierarchy of needs, 33, 34, 261 moral hazard, 154
self-actualization, 358 Morningstar (ratings firm), 2879,
mathematics, 42, 46n30, 70n31, 296n42
87n40, 129n12, 248, 269 mortgages. See also real estate
mean-variance optimization, 32, 60, continuous workout mortgages,
69n10, 78, 98, 152, 259, 260, 154, 174n167
264, 265, 284 impact on stock market
Mehra, Rajnish, 70n37, 107n33, participation, 141, 142
130n26, 13740, 1449, 151, reverse mortgages, 142
158, 160, 160n58, 161n13, and risk aversion, 154
162n21, 162n27, 166n68, subprime, 309
166n702, 166n747, 167n79, multiple discovery, 12, 24n100
167n869, 168n96, 169n109, mutual funds
170n11114, 171n12931, capital structure, 297n64
172n138, 172n146, 173n152, closed-end funds, 222
174n170, 174n174, 175n175, ethically conscious funds, 36
175n178, 177n199, 244n102 fund flows, 296n43, 296n51,
mental accounting, 1, 17, 2957, 151, 297n52
202, 260, 264, 304, 324 generally, 7, 36, 49n83, 94, 231,
Merton, Robert C., 22n80, 24n100, 283, 284, 2869, 297n64
24n101, 69n17, 93, 94, 97, 98, and investment gaps, 2878
104n1, 104n2, 104n9, 105n15, two mutual fund theorem, 2, 32 (see
105n21, 107n34, 129n18, also (separation theorem;
129n202, 133n72, 153, 158, Tobin, James))
173n156, 178n217, 245n209, and value investing, 36, 73, 289
INDEX 337
myopic loss aversion, 210n115, optimism and pessimism, 15, 34, 156,
2289, 231, 298n76, 308 213, 214, 263, 308, 309
mythology and storytelling, 12, options pricing, 22n80, 81, 96, 101,
40, 139 106n33, 185, 319n85
overconfidence
sex-based differences, 14
N over-the-counter (OTC) stocks, 219,
nave asset allocation (1/n), 31 238n93
narrow framing, 231
NASDAQ, 287
National Bureau of Economic P
Research (NBER), 160, 168n90, Panel Study of Income Dynamics, 140
179n235, 234n36, 235n49, paradoxes and puzzles. See also
321n113 anomalies
NBER. See National Bureau of Allais paradox, 1226
Economic Research contrasted with normal science, 4
net asset value (NAV), 287, 28992 equity premium puzzle, 3, 13779,
net present value (NPV), 43n5, 181, 2289
290, 291 low-volatility anomaly, 13, 37,
New York Stock Exchange (NYSE), 7392, 223, 2258
287, 289 risk-free rate puzzle, 148
Nobel Prizes St. Petersburg paradox, 122, 1268
economics, 182, 205n2, 253, 303 stock market participation puzzle, 140
literature, 253, 314n15 Pascal, Blaise, 220, 239n103
noise, noise trading, 288, 297n52, passive indexing and investment,
297n54, 302, 309, 310, 316n42 284, 286
noisy markets hypothesis, 22n79 patent law, 9
Norway, 156 pensions, pension funds, and pension
NYSE. See New York Stock Exchange managers, 28n150, 100, 140, 143,
(NYSE) 157, 158, 178n215, 204, 320
perfect competition, 225
Petkova, Railitsa, 83, 84, 92n108,
O 92n114, 92n117, 92n119,
Obamacare (Affordable Care Act), 92n120, 102, 103, 104n1,
8, 21n66 109n79, 109n82
O'Connor, Flannery, 310, 321n114 pharmaceutical industry, 226
Odean, Terrence, 22n74, 26n126, physical perception, 184
44n16, 88n48, 176n197, pigeons, 184, 215, 233n26, 234n27
233n10, 285, 293n6, 294n224, polar bears, 41, 56n153
296n39, 297n53, 297n60, postmodern portfolio theory, 17n1,
298n757, 299n81, 303, 18n11, 49n83, 131n48, 202,
314n17, 317n46, 317n60 281n237, 281n249, 306,
omega, 86n22, 120, 121, 132n55, 248 314n16, 321n114, 323
338 INDEX
Postmodern Portfolio Theory (book), 1, and the equity premium puzzle, 3,

5, 67, 75, 76, 80, 84, 97, 105n20, 13779, 181, 207n66,
101, 102, 120, 214, 226, 213, 2289, 231
247, 265, 266 fourfold pattern, 3, 181, 2015,
Pratt, John W., 111, 11415, 129n8, 211n133, 213, 223, 253, 269
129n13, 132n61 framing effects, 197, 205
precautionary saving, 54n127, 150, gain-loss asymmetry, 201
158, 172n143, 178n225 generally, 3, 4, 67, 75, 128, 156,
Prescott, Edward, 70n37, 130n26, 181255, 257, 258, 269, 285,
137, 138, 140, 1449, 151, 158, 304, 306, 308, 324
160n48, 162n27, 166n68, loss aversion, 1859, 198, 201,
166n702, 166n74, 166n757, 207n66, 210n115, 22831, 308
167n79, 167n86, 167n88, lottery (or possibility) effect, 202,
167n89, 168n96, 170n111, 203, 213
170n113, 170n114, and momentum, 4, 213, 226, 230,
171n12931, 172n138, 243n176, 30122
173n152, 174n174, 175n175, nonlinear probabilities, 201
175n178, 177n199 realization utility, 185, 285
price-to-book value ratio, 73. See also reference point, 3, 181, 1846,
value factor (book-to-market), 18891, 196, 223, 225, 232,
value stocks 252, 254, 257, 285
private equity, 44n14, 216, 217 reflection effect, 183, 184
prize-linked savings accounts, 215, 216 and skewness preference, 3, 203,
probability distribution function 204, 21320, 222, 231
(PDF), 61 source dependence, 198
procyclicality, 148, 310 and SP/A theory, 4, 67, 181,
projection bias, 97, 150, 238n100 24850, 2535, 257, 269
prospect theory status quo bias, 185
adaptation level, 1846, 188, 189, value function, 128, 1848, 190,
206n39 193, 196, 197, 199, 224, 228,
and Bowman's paradox, 3, 735, 232, 249, 250, 254, 285
181, 2238, 253 weighting function, 15, 197,
certainty effect, 183, 203 198202, 210n115, 214, 215,
cumulative prospect theory, 187, 228, 251, 252, 254
1968, 200, 210n115, psi (), 289. See also investment gap
210n124, 214, 222, 224, 228, psychology, 18n27, 20n48, 23n87,
251, 254, 308 24n110, 26n121, 26n122,
descriptive vs. normative theory, 204 28n158, 34, 35, 37, 38, 42,
diminishing sensitivity, 1869, 196, 44n12, 52n115, 53n116,
201, 203, 257 55n138, 105n2, 142, 149, 215,
and the disposition effect, 213, 220, 221, 248, 253, 260, 289,
2846, 304, 308 302, 324, 325n3, 326n15
INDEX 339
R constant relative (CRRA), 117, 128,

Rabin, Matthew, 107n43, 126, 149, 264
134n79, 170n16 and consumption smoothing, 149,
random walk, 229 1524
rational expectations hypothesis, 12 decreasing absolute, 65, 116, 123,
real estate. See also mortgages 132n61, 208n72
and the disposition effect, 285 hyperbolic absolute (HARA), 2, 65,
generally, 39, 96, 125, 133n68, 285 11518, 132n58, 147, 182
and risk aversion, 125 quadratic utility, 182
and speculative bubbles, 311 relative, 94, 111, 114, 117, 128,
and stock market participation, 140 146, 147, 149, 264
wealth effects, 39 risk-averse insurance premium,
recessions, 155, 157 114, 115
reference point, 3, 30, 77, 181, 1846, and the St. Petersburg paradox, 121,
18891, 196, 223, 225, 232, 252, 122, 126
254, 257, 285, 293, 308 risk-free rate
relative welfare, 77 risk-free rate puzzle, 148
religion, 12, 220 risk-free return, 60, 131n48, 145
retirement and retirement savings risk premium. See equity premium
defined contribution plans, 16, 140 puzzle; equity risk premium
failure to diversify, 156 risk-return relationship, 74, 139, 224,
and intertemporal CAPM, 97, 98 225, 227, 242. See also Bowman's
and life cycle theory, 150 paradox; low-volatility anomaly
and myopic loss aversion, 228, 229 risk-seeking behavior, 3, 34, 67, 122,
participant-directed retirement 181, 201, 223, 258, 269. See also
accounts, 15 gambling and gamblers; lotteries,
pensions, pension funds and pension lottery tickets, and lottery-like
managers, 106n31, 140, 143, investments; prospect theory
150, 158 skewness preference
retirement planning, 142, 214, 259 risk tolerance, 47n56, 115, 116, 119,
retirement preparedness, 142 122, 147, 220, 259
and risk aversion, 228 risk, types of
and SP/A theory, 259, 260 default, 253
target-age retirement funds, 229 downside, 13, 39, 82, 120, 121,
tax-sheltered retirement accounts 137, 202, 214
(e.g., 401(k)), 150 human capital, 154
return on equity, 1446, 224 idiosyncratic, 81, 82, 84
reversion to the mean, 162n20, 225 macroeconomic, 316n40
risk aversion procyclicality, 148
absolute, 2, 65, 11113, 11519, sequence-of-returns, 96, 155, 264,
123, 132n58, 132n61, 147, 276n156
182, 208n72 systematic, 36, 73, 76, 80, 96, 143,
and the Allais paradox, 122, 126, 226, 227, 306, 321n114
183, 198, 249 rogue trading, 203
340 INDEX
Roll, Richard, 19n38, 19n39, 69n20, Shadwick, William F., 120, 132n51
90n68, 108n57, 109n84 Sharpe ratio, 60, 11921, 305
Roll's critique, 96, 155 and Roy's safety-first criterion,
Roy, Arthur D., 274n10710, 2547
275n120 Sharpe, William F., 58, 68n2, 69n16,
Roy's safety-first criterion, 2547, 265 69n22, 69n23, 276n157,
Runyon, Damon, 252 279n212, 295n37. See also Sharpe
Russia, 153 ratio
Shefrin, Hersh, 22n79, 23n82, 23n84,
27n133, 45n24, 46n31, 105n19,
S 132n64, 165n54, 255, 273n93,
SAD. See seasonal affective 274n100, 274n101, 274n115,
disorder (SAD) 275n122, 275n127, 276n147,
safety-first criterion, 2547, 260, 277n159, 277n164, 277n168,
263, 265 277n170, 278n175, 278n180,
Sagoff, Mark, 41, 56n152 278n184, 278n186, 278n196,
Samuelson, Paul A, 42n3, 127, 128, 279n200, 279n206,
130n30, 133n72, 135n89, 280n21518, 280n224,
135n92, 245n209 281n249, 284, 293n11, 316n42,
Santos, Tano, 229, 230, 240n121, 317n48, 317n56, 320n100,
245n212, 245n217, 245n222, 320n103, 322n120
245n226, 246n227, 246n250 Shiller, Robert J., 23n87, 23n89,
Sapp, Travis R.A., 287, 296n43, 24n108, 24n110, 24n111,
296n45, 296n50, 298n73 88n49, 88n50, 105n25, 108n60,
Scandinavia, 156 109n71, 161n14, 170n125,
science, history and philosophy of 174n166, 174n167, 176n188,
anomalies, 2, 46 240n119, 303, 310, 313n2,
multiple discovery, 12 313n3, 314n15, 322n1214,
normal science, 4 322n1268, 322n133, 322n138,
seasonal affective disorder (SAD), 322n139, 325n3
20n48 short-termism, 143
Securities and Exchange Commission Sierra Leone, 263
(SEC), 106n33, 217, 218, 289. sigmoid curves, 186, 189
See also federal securities law skewness
security-potential/aspiration theory. and higher-moment CAPM, 60,
See SP/A theory 222, 247
semiconductors, 37 negative skewness preference, 204
separation theorem, 2, 32, 94, 95, skewness preference, 3, 35, 71n59,
118, 141, 260 116, 203, 204, 21320,
sequence of returns risk, 96, 155, 264, 222, 231
276n156 small firm factor, small-cap stocks, 99,
sex-based differences, 14, 16 100, 102, 226, 286, 303
INDEX 341
smart money, 288, 302, 311, 312. See Statman, Meir, 22n79, 23n82, 23n84,
also arbitrage; information 26n122, 45n23, 45n24, 46n31,
trading, informed trading 47n42, 47n446, 47n50, 47n56,
Social Security, 142, 143 52n113, 52n115, 53n124,
Solnik, Bruno, 80, 91n82, 174n164, 132n64, 163n39, 169n104,
294n29 176n191, 232n6, 233n23,
Spain, 227 238n91, 255, 271n38, 274n101,
SP/A (security, potential, aspiration) 274n115, 275n122, 275n127,
theory 276n147, 277n159, 277n164,
aspiration level, 2579, 262, 277n165, 277n168, 277n170,
265, 266 278n175, 278n180, 278n183,
and behavioral portfolio theory, 4, 278n184, 278n186, 278n196,
67, 181, 24850, 2535, 279n206, 279n212, 279n213,
25760, 262, 265, 268, 269 280m224, 280n21418,
conflicts within risky choice, 253 280n226, 280n229, 281n249,
curvature and elevation 284, 293n11, 295m30, 316n42,
parameters, 251 317n48, 317n56, 320n100,
dispositional vs. aspirational factors, 320n103, 321n104, 279n200
250, 253 stochastic dominance, 71n59, 92n104,
generally, 250, 253 197, 222
high- and low-aspiration accounts, stock exchanges
258, 261, 263 265, 266 American Stock Exchange
and modern portfolio theory, 3, 4, (AMEX), 287
248, 250, 259, 265, 266 Chicago Board Options Exchange
probability of ruin, 257, 258 (CBOE), 77
and prospect theory, 3, 4, 67, 181, NASDAQ, 287
24850, 2535, 257, 258, 269 New York Stock Exchange (NYSE),
and retirement planning, 259 287, 289
Roy's safety-first criterion, 254, 255, stock options (as part of executive
257, 265 compensation), 219
security-seeking vs. potential-seeking St. Petersburg paradox, 122, 1268
dispositions, 3, 4, 248, 251, strategic management, 67, 74, 75,
2535, 2579, 262 223, 225, 227, 228. See also
and value-at-risk (VaR) analysis, 4, Bowmans paradox
248, 265, 266, 269 structural breaks, 139
speculative bubbles, 1, 3, 4, 139, 182, structured financial products, 96
30122 sunk costs, 29, 42n1
spillover effects and knowledge Supreme Court of the United States,
spillovers, 226, 243n180 8, 9, 11, 186
Standard & Poor's (S&P) 500 Sweden, 156
(securities index), 77 Swensen, David F., 157, 177n203,
standard score, 267 177n209, 297n61
342 INDEX
sympathetic magic, 14 Their Eyes Were Watching God

synchronization, 302 (novel), 16
System 1 and System 2 modes of Thomas, Howard, 86n30, 2246,
thought, 5, 6, 18n11, 324 241n138, 241n142, 241n149,
systematic risk, 36, 73, 76, 80, 96, 241n151, 242n155, 242n157,
143, 226, 227, 306, 321n114 242n172, 243n178, 243n185,
243n187, 244n188, 244n192,
244n194
T thrill-seeking, 220
Taiwan, 219, 284 time-series analysis, 318n70
target return, 33, 120, 121, 185, 2235 Tobin, James, 94, 105n14, 141, 158,
disastrous rate of return, 256 (see 163n40, 178n216, 237n73
also (Roy, Arthur D.-Roy's Tolkein, J.R.R., 160, 177n207,
safety-first criterion)) 179n233
Taylor series expansion (Taylor, Brook) trading strategies
and absolute risk aversion, 2, 65, corporate spinoffs, 219
113, 115, 116 dumb money effect, 287, 288
generally, 2, 66, 250 Maslowian self-actualization, 357
and higher-moment CAPM, 60 Treynor ratio (Treynor, Jack L.), 59, 60
as a polynomial equation, 66 turnover, 292, 309
and the risk-averse insurance Tversky, Amos, 20n42, 44n13, 45n21,
premium, 114, 115 45n22, 127, 134n75, 135n90,
and skewness preference, 116, 203 182, 183, 185, 187, 188, 196,
technical analysis, 307 197, 199204, 205n2, 205n11,
technology bubble (of 2000), 158 205n14, 206n22, 206n23,
Thaler, Richard H., 19n38, 43n4, 206n25, 206n31, 207n56,
44n15, 45n19, 45n20, 46n27, 207n58, 207n63, 209n86,
54n133, 105n19, 162n25, 209n90, 209n94, 209n101,
162n26, 162n30, 165n54, 209n102, 210n105, 210n106,
165n55, 165n58, 169n108, 210n112, 210n115, 210n116,
170n23, 170n24, 171n28, 210n118, 210n120, 210n124,
175n185, 175n186, 177n200, 211n125, 211n133, 211n134,
177n210, 177n212, 178n214, 212n151, 212n158, 214, 230,
178n220, 207n52, 207n53, 232, 233n19, 233n21, 241n147,
207n64, 207n68, 210n111, 242n162, 245n2246, 246n247,
210n115, 228, 229, 232n9, 249, 260n40, 261, 270n3,
237n82, 239n115, 240n130, 271n24, 278n179, 324, 326n8.
244n198, 245n203, 245n210, See also Kahneman, Daniel;
245n221, 246n245, 298n76, prospect theory
316n38, 317n58, 320n95 two-piece utility functions, 224
INDEX 343
U low-volatility anomaly, 13, 37, 67,

underwriting and underwriters, 221, 7393, 98, 1013, 148, 223,
222, 239n113 2258, 230, 253
unicorns (companies valued at $1 volume, 1, 11, 203, 219, 247,
billion or more), 30, 173n155 270n13, 304, 307, 309, 316n42
unicorns (fallacious medical Vuolteenaho, Tuomo, 98, 100,
diagnoses), 153 108n48, 108n51, 108n55,
United Kingdom (UK). See Great 108n58, 108n61, 108n66, 108n67
Britain
United States of America (USA), 7, 8,
41, 51n104, 56n153, 80, W
133n68, 13841, 151, 165n61, Wile E.Coyote, 58
186, 200, 217, 220, 287, 289, Williams, Tennessee, 214, 233n22
298n76, 304 World War II, 153
universities and university Wurgler, Jeffrey, 53n123, 82, 85n8,
endowments, 15760 85n11, 85n14, 85n19, 86n16,
USA. See United States of 88n52, 91n95, 91n99, 92n103,
America (USA) 109n77, 168n98, 321n105
V X
value-at-risk (VaR) analysis, 4, 248, X-inefficiency, 226
2659 Xing, Yuhang, 83, 85n10, 85n14,
value factor (book-to-market), 86n17, 91n92, 91n93, 91n96,
value stocks, 73, 74, 83, 98102, 91n98, 92n105, 92n107,
226, 303 106n26, 109n74, 169n104,
vaunted Volatility Index (VIX), 77, 238n90
80, 81
venture capital, 13, 218, 222
Vilkov, Grigory, 84, 91n88, Y
92n11315 Yale University, 157
volatility
clustering, 230, 307
high-volatility portfolios, 82 Z
implied, 77 Zhang, Lu, 85n4, 85n9, 102, 109n82,
in initial public offerings (IPOs), 3, 175n180, 316n39
37, 222 Zin, Stanley E., 146, 147, 167n80

(Quantitative Perspectives On Behavioral Economics and Finance) James Ming Chen (Auth.) - Finance and The Behavioral Prospect - Risk, Exuberance, and Abnormal Markets-Palgrave Macmillan (2016)

Enviado por

Dados do documento

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

(Quantitative Perspectives On Behavioral Economics and Finance) James Ming Chen (Auth.) - Finance and The Behavioral Prospect - Risk, Exuberance, and Abnormal Markets-Palgrave Macmillan (2016)

Enviado por

Direitos autorais:

Formatos disponíveis

FINANCE AND

JAMES MING CHEN

More information about this series at

Finance and the

Quantitative Perspectives on Behavioral Economics and Finance

Library of Congress Control Number: 2016950218

The Editor(s) (if applicable) and The Author(s) 2016

Cover illustration TongRo Images / Alamy Stock Photo

Printed on acid-free paper

This Palgrave Macmillan imprint is published by Springer Nature

assistance. I am also grateful for contributions by several students at

1 The Structure ofaBehavioral Revolution 1

2 Mental Accounting, Emotional

3 Higher-Moment Capital Asset Pricing andIts

4 Tracking theLow-Volatility Anomaly Across

5 The Intertemporal Capital Asset Pricing Model:

6 Risk Aversion 111

7 The Equity Risk Premium andtheEquity

8 Prospect Theory 181

9 Specific Applications ofProspect Theory to

10 Beyond Hope andFear: Behavioral Portfolio Theory 247

11 Behavioral Gaps Between Hypothetical Investment

12 Irrational Exuberance: Momentum Crashes and

13 The Monster andtheSleeping Queen 323

The Structure ofaBehavioral Revolution

1.1 ABNORMAL MARKETS, IRRATIONAL INVESTORS

The Editor(s) (if applicable) and The Author(s) 2016 1

book synthesizes observations on abnormal markets and irrational inves-

paradox, suggest that conventional accounts of risk aversion do not

Chapter 10 presents a competing account of behavioral finance. SP/A

1.2 ANOMALIES, FAST ANDSLOW

of System 2 is their need for attention.21 Drawing attention away from a

1.3 SELL INMAY ANDGOAWAY?

Should a rational investor really sell in May and go away?36 Maybe, or

1.4 LAW ONTHEMARKET

Admittedly, the legal reasoning in the Sebelius and Myriad decisions

0.5 0.0 0.5 1.0 1.5

least under existing technology, still requires additional evaluation, both

1.5 RAW EMOTION

Gossip is theology translated into experience. In it we hear great stories of

removal of dot-com from a company name resulted in abnormally posi-

1.6 TRADE LIKE AGIRL

On average, female probability weighting functions differ from male ones

Most intriguingly of all, especially in light of other research into mens

An appropriately humble and nuanced approach to mathematical

explain. The descriptive abnormalities of financial markets cannot be

Regarding Market Efficiency, 6 J. Fin. Econ. 95101 (1978); Mark

68. See 133 S.Ct. 2107 (2013).

Implications of the Cox, Ingersoll, Ross Theory of the Term Structure of

where p is the probability of winning, q=1p is the probability of losing,

121. See Kent Daniel, David Hirshleifer & Avanidhar Subrahmanyan,

145. Fehr-Duda, de Gennaro & Schubert, supra note 140, at 305.

Mental Accounting, Emotional Hierarchies,

2.1 KEEPING EMOTIONAL SCORE

The Editor(s) (if applicable) and The Author(s) 2016 29

political economy of disaster assistance dictates perverse outcomes,7 which

from failing to integrate their assets in a single portfolio, institutional

2.2 MASLOWIAN PORTFOLIO THEORY

This purely rational approach to investing and risk management befits

within one need level.45 Quite realistically, an investor might be very

evolutionary biology, anthropology, and psychology, with the conscious

salience erodes generosity toward others, requiring active work to pro-

2.3 SHOTS AT GREATNESS: REHABILITATING SELF-