A Study On Using Business Intelligence F

Business Intelligence Journal
July, 2010 Vol.3 No.2

Volume 3 - Number 2 - July 2010 - Semiannual Publication
Published by the IIU Press and Research Centre, A.C., Brussels EU Commission Building, Rond Point, Schuman
6, Box 5, 1040 Brussels, Belgium, for the Department of Business Management and Economics (BME) of the
School of Doctoral Studies (European Union) at the Isles Internationale Universit (IIU-EU), Brussels, Belgium
in collaboration with the Business Intelligence Service of London, UK (Sayco UK).
Editorial Note 1
Profile of authors included in this number 2
Information for Contributors 4
Articles REVAAM Model to determine a company's value by multiple valuation and linear regression
analysis
9
Luis G. Acosta-Calzado, Carlos Acosta-Calzado, Humberto Murrieta Romo

A comparison between Hybrid Approaches of ANN and ARIMA for Indian Stock Trend Forecasting 23
Nitin Merh, Vinod P. Saxena, Kamal Raj Pardasani

Cash Flow Valuations Using Capital Cash Flow Method, Comparing It with Free Cash Flow Method 45
and Adjusted Present Value Method in Companies Listed on Theran Stock Exchange
Mohsen Dastgir, Vali Khodadadi, Maryam Ghayed
Effects of Value Added Tax on Mexican Economy. A Multisectorial Analysis by Developing an 59
Applied General Equilibrium Model (MEGA)
Eduardo Marquez Pea
Application of Ants Colony System for Bankrupcy Prediction of Companies Listed in Theran Stock 89
Exchange
Vali Khodadadi, Abolfazl (Parviz) Zandimia, Marzieh Nouri
The Value Relevance of Earnings Components in Two Different GAAPs 101
Zahra Nikbakht, Abbas Aflatooni

A Word on Game Theory Application to Corporate Finance 113
Jos Carlos Arias

The Impact of International Business in Global Economy 119
Niki Geiersbach
Reasoning Disparities Between HK and US Managers 131
Isola Oluwabusuyi
A Study on Using Business Intelligence for Improving Marketing Efforts 141
Arpan Kumar Kar, Ashis Kumar Pani, Supriya Kumar De

The Effect of Corporate Governance Structure on the Extent of Voluntary Disclosure in Iran 151
Vali Khodadadi, Soheila Khazami, Abbas Aflatooni
Business Intelligence Journal - July, 2010 Vol.3 No.2

Copyright 2010 IIU Press and Reserach Centre, A.C.
All Rights reserverd. No part of this publication may be reproduced, stored in a retreival system, or transmitted,
in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, whitout the written
prior permission of the author.
ISBN: 978-1-4251-8179-6
ISSN: 1918-2325
2010 Business Intelligence Journal 1
EDITORIAL NOTE Business Intelligence Journal

In collaboration with the Business Intelligence Service of London, UK and with the European
Business School of Cambridge, UK, the Business Intelligence Journal (BIJ), produced by
the Department of Business Management and Economics (BME) at the School of Doctoral
Studies of the European Union, hosted at the Isles Internationale Universit (IIU-EU) in
SPARC Europe Brussels, Belgium, publishes research, analysis and inquiries into issues of importance to
the business community. Articles in BIJ examine emerging trends and concerns in the areas
Seal of the Scholarly Publishing
of general management, business law, public responsibility and ethics, marketing theory
and Academic Resources Coalition
(Granted to the Business Intelligence and applications, business finance and investment, general business research, business
Journal on the 20th day of August, and economics education, production/operations management, organizational behavior
2008) and theory, strategic management policy, social issues and public policy, management
organization, statistics and econometrics, personnel and industrial relations, technology
and innovation, case studies, and management information systems. The goal of BIJ is
to broaden the knowledge of business professionals and academicians by promoting free
access and provide valuable insight to business-related information, research and ideas.
All articles included in the BIJ are peer-reviewed. The Business Intelligence Journal is
published semiannually (one volume per year) by the Business Intelligence Service of
Secured Assets Yield Corporation Limited based in London, UK.
School of Doctoral Studies
(European Union)
Isles Internationale Universit Department of Business Management and Economics (BME)
Brussels, Belgium.
School of Doctoral Studies (European Union)
Isles Internationale Universit (IIU-EU)
Brussels EU Parliament Building:
Square de Meeus 37 4th Floor
1000 Brussels, Belgium
edit.bij@saycocorporativo.com
Head of Department (BME): Dr. Jnger Albinger (PhD)
European Business School,
School of Graduate Studies Published by
Cambridge, UK. IIU Press and Research Centre, A.C.
Brussels EU Commission Building:
Sayco Rond Point, Schuman 6, Box 5
Business Intelligence Service 1040 Brussels, Belgium
London, UK
Research Centres Director: Professor Michael Rockwell (PhD)
DIRECTORY OF
Periodical Publications Editorial Unit Director: Dr. Anne D. Surrey (PhD)
OPEN ACCESS
JOURNALS Business Intelligence Journal Editor: Robert B. Stacey
Directory of Open Access Journals Associate Editors: Michael Summers
Lund University Libraries Susan G. Boots
Sweden.
Martin A. Miller
Kenneth C. Michaels
Business Intelligence
Journal by Business Reviewers Coordinators: Anita Peters
Intelligence Service Roger Puig
is licensed under a Creative Commons Robert Miller
Attribution 2.0 UK: England & Wales
License. Further tips for using the Editorial Design: Pablo Gmez-Olivo
supplied HTML and RDF are here:
http://creativecommons.org/learn/
technology/usingmarkup ISSN 1918-2325
http://www.saycocorporativo.com/saycouk/BIJ/journals.html
Copyright: IIU Press and Research Centre A.C.

2 Business Intelligence Journal July

Profile of authors included in this number
Article 1: REVAAM Model to determine a companys value by multiple valuation and linear regression
analysis
Author: 1 - Luis G. Acosta-Calzado BA in Actuarial Science from Universidad Marista, and
currently studying a Masters in Finance. His undergraduate thesis is on Relative Valuation.
He has worked in important financial institutions, and currently at Monex in Mexico City.
email: act_gacosta@yahoo.com.mx
2 - Carlos Acosta-Calzado BS in Industrial Engineer from ITESM, MBA with focus
on Finance and Economics from NYU Stern. He is currently partner at Ibkan Consultants,
Investment Banking.
email: cacosta@ibkan.mx
2 - Humberto Murrieta-Romo BA in Accounting from Escuela Bancaria y Comercial,
MBA from NYU Stern. He is currently partner at Ibkan Consultants, Investment Banking.
email: hmurrieta@ibkan.mx
Article 2: A Comparison between Hybrid Approaches of ANN and ARIMA for Indian Stock Trend Forecasting
Author: 1 - Nitin Merh Faculty IT & Systems, IBS- Indore, Jaisingh Palace, 7, Jaisingh Nagar,
Rajiv Gandhi Circle, Indore, M.P, INDIA
email: n_merh@rediffmail.com
2 - Vinod P. Saxena Ex- Vice Chancellor, Jiwaji University and Ex- Professor, School of
Mathematics and Allied Sciences (SOMAAS), Jiwaji University Gwalior, M. P. INDIA
email: vinodpsaxena@gmail.com
3 - Kamal Raj Pardasani Professor and Head, Department of Mathematics, Bioinformatics
and Computer Applications, Maulana Azad National Institute of Technology (MANIT),
Bhopal, M.P. INDIA
email: kamalrajp@rediffmail.com
Article 3: Cash Flows Valuation Using Capital Cash Flow Method Comparing it with Free Cash Flow
Method and Adjusted PresentValue Method in Companies Listed on Tehran Stock Exchange
Author: 1 - Mohsen Dastgir Shahid Chamran University of Ahwaz.
2 - Vali Khodadadi Shahid Chamran University of Ahwaz.
3 - Maryam Ghayed MA in accounting of Khuzestan Science and Research Branch/
Islamic Azad University / Ahwaz/ Iran.
email: m_ghayed@yahoo.com
Article 4: Effects ofValue Added Tax on Mexican Economy A multisectorial analysis by developing an
Applied General Equilibrium Model. MEGA
Author: Eduardo Mrquez Pea Masters in Economic Theory Student. Instituto Tecnolgico
Autnomo de Mxico (ITAM), Mxico, D.F.
email: lalomarquez_mx@yahoo.com

Article 5: Application of Ants Colony System for Bankruptcy Prediction of Companies listed in Tehran Stock
Exchange
Author: 1 - Vali Khodadadi Shahid Chamran University of Ahwaz.
2 - Abolfazl (Parviz) Zandinia Allameh Tabatabaei University.
3 - Marzieh Nouri MA in accounting.
e-mail: m_n592007@yahoo.com.
Article 6: TheValue Relevance of Earnings Components in Two Different GAAPs
Author: 1 - Zahra Nikbakht Payam Noor University (PNU), Kohpayeh, Esfahan, Iran.
email: Zahra.Nikbakht77@gmail.com
2 - Abbas Aflatooni Department of Accounting, College of Economics and Social
Sciences, Shahid Chamran University of Ahvaz, Ahvaz, Iran.
email: abbasaflatooni@gmail.com
Article 7: AWord on Game Theory Application to Corporate Finance
Author: Jos Carlos Arias (PhD, DBA) Vice-Chancellor (Rector) of the Isles Internationale
Universit, School of Doctoral Studies of the European Union (Brussels, Belgium) and Senior
Partner of Secured Assets Yield Corporation Limited (London, UK)
email: josecarlosarias@saycocorporativo.com
Article 8: The Impact of International Business on the Global Economy
Author: Niki Geiersbach Candidate to PhD in Management Science at the School of Doctoral
Studies of the European Union.
Article 9: Reasoning Disparities Between HK and Us Managers
Author: Dr. Isola Oluwabusuyi (DBA - Argosy University) Leading instructor at the School of
Business and Accounting, Brown Mackie College, Atlanta.
Article 10: A Study on Using Business Intelligence for Improving Marketing Efforts
Author: 1 - Arpan Kumar Kar Doctorate student Information systems area XLRI School of
business and Human Resources.
email: arpan.kar@astra.xlri.ac.in
2 - Ashis Kumar Pani Faculty Information systems area XLRI School of business and
Human Resources.
email: akpani@xlri.ac.in
3 - Supriya Kumar De Faculty Information systems area XLRI School of business and
Human Resources.
email: skde@xlri.ac.in
Article 11: The Effect of Corporate Governance Structure on the Extent ofVoluntary Disclosure in Iran
Author: 1 - Vali Khodadadi Department of Accounting, College of Economics and Social Sciences,
Shahid Chamran University of Ahwaz, Ahwaz, Iran.
2 - Soheila Khazami M.A in Accounting, Islamic Azad University, Shushtar branch,
Khuzestan, Iran.
3 - Abbas Aflatooni Department of Accounting, College of Economics and Social
Sciences, Shahid Chamran University of Ahwaz, Ahwaz, Iran.
In order to make contact with any of the Authors referred to above, please forward your request to: edit.bij@saycocorporativo.
com, including BIJs edition (BIJ Volume 3, Number 2, July 2010), articles and authors names with your requirement. BIJs
Editor will be glad to submit your requests or inquiries before authors.

INFORMATION FOR CONTRIBUTORS

Electronic submission of manuscripts is Reviews: Submissions of reviews and
strongly encouraged, provided that the text, perspectives covering topics of current
tables, and figures are included in a single interest are welcome and encouraged.
Microsoft Word file (preferably in Times Reviews should be concise and no longer
New Roman, 12 size font) than 4-6 printed pages (about 12 to 18
Submit manuscript as e-mail attachment manuscript pages). Reviews manuscripts
to the BIJ Editorial Office at: edit.bij@ are also peer-reviewed.
saycocorporativo.com. A manuscript
number will be mailed to the corresponding Review Process
author within the following 7 days.
The cover letter should include the All manuscripts are reviewed by an
corresponding authors full address and editor and members of the Editorial Board
telephone/fax numbers and should be in or qualified outside reviewers. Decisions
an e-mail message sent to the Editor, with will be made as rapidly as possible, and
the file, whose name should begin with the the journal strives to return reviewers
first authors surname, as an attachment. comments to authors within 3 weeks. The
The authors may also suggest two to four editorial board will re-review manuscripts
reviewers for the manuscript (BIJ may that are accepted pending revision. It is the
designate other reviewers). goal of the BIJ to publish manuscripts within
BIJ will only accept manuscripts the following BIJ edition after submission.
submitted as e-mail attachments.
Regular Articles
Article Types
All portions of the manuscript must be
Three types of manuscripts may be typed double-spaced and all pages numbered
submitted: starting from the title page.
Regular Articles: These should describe The Title should be a brief phrase
new and carefully confirmed findings, describing the contents of the paper. The
and research methods should be given in Title Page should include the authors full
sufficient detail for others to verify the work. names and affiliations, the name of the
The length of a full paper should be the corresponding author along with phone, fax
minimum required to describe and interpret and e-mail information. Present addresses of
the work clearly. authors should appear as a footnote.
Short Communications: A Short The Abstract should be informative and
Communication is suitable for recording completely self-explanatory, briefly present
the results of complete small investigations the topic, state the scope of the work, indicate
or giving details of new models, innovative significant data, and point out major findings
methods or techniques. The style of main and conclusions. The Abstract should be 100
sections need not conform to that of full- to 200 words in length. Complete sentences,
length papers. Short communications are 2 active verbs, and the third person should be
to 4 printed pages (about 6 to 12 manuscript used, and the abstract should be written in
pages) in length. the past tense. Standard nomenclature

should be used and abbreviations should sections can include subheadings, and when
be avoided. No literature should be cited. appropriate, both sections can be combined.
Following the abstract, about 3 to 10 key The Acknowledgments of people,
words that will provide indexing references grants, funds, etc should be brief.
to should be listed. Tables should be kept to a minimum
A list of non-standard Abbreviations and be designed to be as simple as
should be added. In general, non-standard possible. Tables are to be typed double-
abbreviations should be used only when spaced throughout, including headings and
the full term is very long and used often. footnotes. Each table should be on a separate
Each abbreviation should be spelled out and page, numbered consecutively in Arabic
introduced in parentheses the first time it is numerals and supplied with a heading and
used in the text. a legend. Tables should be self-explanatory
The Introduction should provide without reference to the text. The details
a clear statement of the problem, the of the research methods should preferably
relevant literature on the subject, and the be described in the legend instead of in the
proposed approach or solution. It should be text. The same data should not be presented
understandable to colleagues from a broad in both table and graph form or repeated in
range of disciplines. the text.
Materials and methods should Figure legends should be typed in
be complete enough to allow possible numerical order on a separate sheet. Graphics
replication of the research. However, only should be prepared using applications
truly new research methods should be capable of generating high resolution GIF,
described in detail; previously published TIFF, JPEG or PowerPoint before pasting in
methods should be cited, and important the Microsoft Word manuscript file. Tables
modifications of published methods should should be prepared in Microsoft Word. Use
be mentioned briefly. Capitalize trade Arabic numerals to designate figures and
names and include the manufacturers upper case letters for their parts (Figure 1).
name and address. Subheadings should be Begin each legend with a title and include
used. Methods in general use need not be sufficient description so that the figure is
described in detail. understandable without reading the text
Results should be presented with clarity of the manuscript. Information given in
and precision. The results should be written legends should not be repeated in the text.
in the past tense when describing authors References: In the text, a reference
findings. Previously published findings identified by means of an authors name
should be written in the present tense. Results should be followed by the date of the
should be explained, but largely without reference in parentheses. When there are
referring to the literature. Discussion, more than two authors, only the first authors
speculation and detailed interpretation of name should be mentioned, followed by et
data should not be included in the Results al. In the event that an author cited has had
but should be put into the Discussion section. two or more works published during the
The Discussion should interpret the same year, the reference, both in the text and
findings in view of the results obtained in in the reference list, should be identified by
this and in past studies on the topic. State a lower case letter like a and b after the
the conclusions in a few sentences at the end date to distinguish the works.
of the paper. The Results and Discussion

Examples: Short Communications
Smith (2000), Wang et al. (2003), (Kelebeni, Short Communications are limited to a
1983), (Usman and Smith, 1992), (Chege, maximum of two figures and one table. They
1998; Chukwura, 1987a,b; Tijani, 1993, should present a complete study that is more
1995), (Kumasi et al., 2001) limited in scope than is found in full-length
papers. The items of manuscript preparation
References should be listed at the end of listed above apply to Short Communications
the paper in alphabetical order. Articles with the following differences: (1) Abstracts
in preparation or articles submitted for are limited to 100 words; (2) instead of a
publication, unpublished observations, separate Materials and Methods section,
personal communications, etc. should not research methods may be incorporated into
be included in the reference list but should Figure Legends and Table footnotes; (3)
only be mentioned in the article text (e.g., Results and Discussion should be combined
A. Kingori, University of Nairobi, Kenya, into a single section.
personal communication). Journal names
are abbreviated according to Chemical Ab- Proofs and Reprints
stracts. Authors are fully responsible for the
accuracy of the references. Electronic proofs will be sent (e-mail
attachment) to the corresponding author as
Examples: a PDF file. Page proofs are considered to
be the final version of the manuscript. With
Papadogonas TA (2007). The financial the exception of typographical or minor
performance of large and small firms: clerical errors, no changes will be made in
evidence from Greece. Int. J. Financ. the manuscript at the proof stage. Because
Serv. Manage. 2(1/2): 14 20. BIJ will be published online without access
restrictions, authors will have electronic
Mihiotis AN, Konidaris NF (2007). Internal access to the full text (PDF) of the article.
auditing: an essential tool for adding Authors can download the PDF file from
value and improving the operations of which they can print unlimited copies of
financial institutions and organizations. their articles.
Int. J. Financ. Serv. Manage. 2(1/2): 75
81. Copyright
Gurau C (2006). Multi-channel banking Submission of a manuscript implies: that

in Romania: a comparative study of the the work described has not been published
strategic approach adopted by domestic before (except in the form of an abstract or
and foreign banks Afr. J. Financ. Servic. as part of a published lecture, or thesis) that
Manage. 1(4): 381 399. it is not under consideration for publication
elsewhere; that if and when the manuscript
Yoon CY,Leem CS (2004).Development is accepted for publication, the authors agree
of an evaluation system of personal to automatic transfer of the copyright to the
e-business competency and maturity publisher.
levels Int. J. Electron. Bus. 2(4): 404
437.

Costs for Authors Manuscripts submitted will be judged

primarily on their substantive content,
Revision, edition and publishing costs though writing style, structure and length
will be totally paid by the IIU Press & will also be considered. Poor presentation
Research Centre A.C. and authors sole is sufficient reason for the rejection of a
contribution will be providing BIS with manuscript. Manuscripts should also be
their invaluable work. written as concisely and simply as possible,
without sacrificing clarity or meaningfulness
Publication Decisions of exposition. Manuscripts will be evaluated
by the editor when first received, on their
Decisions by the editor of all submitted contribution-to-length-ratio, meaning that
manuscripts reflect the recommendations of manuscripts with strong contributions will
members of the Editorial Board and other be assigned more pages than those making
qualified reviewers using a blind review narrower contributions. Papers intended to
process. Reviewers comments are made make very extensive contributions (over 35
available to authors. Manuscripts that are double-space pages, using one inch margins
inappropriate or insufficiently developed and Times New Roman twelve-pitch font)
may be returned to the authors without will, at discretion of the editor, be allotted
formal review for submission to a more additional space. Authors are expected to get
suitable journal or for resubmission to BIJ and use feedback from colleagues prior to
following further development. submitting a manuscript for formal review.


2010 Luis G. Acosta-Calzado, Carlos Acosta-Calzado and Humberto Murrieta-Romo 9
REVAAM MODEL TO DETERMINE A COMPANYS

VALUE BY MULTIPLE VALUATION AND LINEAR
REGRESSION ANALYSIS
Luis G. Acosta-Calzado, Carlos Acosta-Calzado and

Humberto Murrieta-Romo
Abstract
This paper shows an alternative model to the widely used method of multiple valuation (or relative
valuation) in order to calculate the value of a company by using either the Price Earnings (PE) and/or
the Enterprise Value to Earnings Before Interest, Taxes, Depreciation and Amortization (EV/EBITDA).
When calculating multiples, analysts tend to consider average multiples within an industry and apply them
directly to the target company; however, we believe that this practice is not considering differences among
the companies being compared, although they belong to the same sector or industry. REVAAM Model
uses linear regression to calculate adjusted PE and EV/EBITDA multiples by taking into consideration
profitability factors for each multiple in order to differentiate companies in the samples. Calculations
are based on public data for US companies, but could be further expanded to other markets. Not only
REVAAM Model provides a better estimate to relative valuation analysis than simply using average
multiples, but it could be used to compare under/overvalued companies or sectors, and also analyze
multiple value changes over time as the intrinsic fundamentals change.
Acosta-Calzado Luis G., Acosta Calzado C., Nurrieta-Romo H. - REVAAM Model to determine a companys value by multiple valuation and linear regression analysis
Introduction priced equally. Despite these arguments,

relative valuation can be a powerful tool
In financial analysis and company if we introduce some adjustment for the
valuation, it is common to use multiple differences in those companies.
valuation (or relative valuation) as a There are several multiples used in
method to determine the value or a range relative valuation: Price Earnings1 (PE),
of values for a specific company. Relative Enterprise Value to Earnings Before Interest,
valuation calculates the value of an asset Taxes, Depreciation and Amortization2 (EV/
(in our case companies) from the value of EBITDA), Price to Sales3 (PS), Price to Book
other comparable assets (for example, Value of Equity4 (PBV), and variations to
companies in the same sector). Although these. Each multiple is measured at different
this method gives us a quick reference of the levels of the income statement; multiples
companys possible fair value, based on using sales ratios are based on the top line,
the industry or on the comparable companies, whereas multiples using net earnings are
we believe that the best way to determine bottom line ratios.
a companys true value is by using the Sectors such as the Pharmaceutical tend
Discount Free Cash Flow (DCF) method to use multiples based on sales, such as
(either using Firm or Equity free cash flows), the PS, however PE and EV/EBITDA are
despite there exists controversy in choosing the most widely used multiples. The main
the adequate discount rates. The reason reason is that sales multiples do not consider
behind this is that there are big differences margins and costs, misleading the final
among companies even in the same sector; a value of comparable companies. Then we
company with sales of 100 whose operating would prefer using the PE ratio, as it is based
earnings have been growing at 10% for the on the bottom line of the Income Statement,
last five years and is expected to continue however, a) not all companies have the same
growing at that rate for the following five amount of financial income and expenses
years, should be worth more than a company and b) the value for this ratio is the equity
with sales of 100, ceteris paribus, but with value of our company which may be also
an operating earnings growth of 3%. misleading due to the debt amount for our
When comparing companies in the company. Therefore, we prefer using the EV/
same industry, we do obtain a reference of EBITDA ratio because the value we obtain
how the sector is behaving and what the is an Enterprise Value or the value of the
overall risk could be or to obtain parameters operating assets, which is also obtained
for that specific sector, but companies using the DCF from the free cash flows to
within the sample do not present the same the firm (FCFF). The EBITDA is a proxy
operating parameters, and therefore are not to obtain the FCFF as it adds depreciation
1
Price Earnings ratio is calculated as the market price per share
and amortization, non-cash expenses, to the
divided by net earnings per share operating earnings of the companies before
2
Enterprise Value to Earnings Before Interest, Taxes, Depreciation any financial expenses and taxes.
and Amortization ratio is calculated as the value of operating
assets divided by operating earnings before interests and taxes
Then, if the use of multiples lacks
plus depreciation and amortization accuracy and therefore confidence in
3
Price to Sales ratio is calculated as the market price per share determining a value for our company, we
divided by sales per share
need to establish a method that takes into
4
Price to Book Value of Equity is calculated as the market price
per share divided by the book value per share consideration the differences among the
companies in our sample. For our analyses,

we will consider the use of the PE and EV/ aPE = the value of the intercept calculated
EBITDA multiples. by the regression analysis
bROE = the value of the coefficient for
Design ROE
ROE = the return on equity, calculated
As we said before, the best way to as the net income divided by the companys
determine a companys value is by using beginning equity value
the DCF method, and in order to do so we bNM = the value of the coefficient for Net
need to forecast the free cash flow, either Margin
to the firm or to the equity shareholders. Net M arg in = calculated as the net
This process has to be detailed and requires income of year t divided by the sales of year
knowledge of the companys financials and f = the sum of errors derived from the
the expected growth. In consequence, we regression analysis
propose to adjust the PE multiple by using
Equation 2
the Return on Equity5 (ROE) and the Net
Margin6 as the indicators of profitability for EV/EBITDA = aEVEBITDA + bROC * ROC + bOM *
equity shareholders of the companies being Pre - tax Operating M arg in + f
analyzed. In contrast, for the EV/EBITDA
multiple we propose adjusting by using Where
the Return on Capital7 (ROC) and the Pre-
tax Operating Margin8. These adjustments EV/EBITDA =the EBITDA multiple
should be made using multiple regression calculated as the Enterprise Value or
analysis, in order to minimize errors, where Value of the Net Operating Assets divided
the independent variables are the multiples by the Earnings before Interests, Taxes,
and the other factors are the dependent Depreciation and Amortization
variables. These relationships are shown in aEVEBITDA = is the value of the intercept
the equations below: calculated by the regression analysis
bROC = the value of the coefficient for
Equation 1
ROC
P/E = aPE + bROE * ROE + bNM * Net M arg in + f ROC = the return on capital, calculated
as the after-tax operating income (EBIT*(1-
Where, tax)) divided by the companys beginning
equity value plus the debt balance
P/E = the price earnings ratio calculated bOM = the value of the coefficient for Net
as the price per share divided by net earnings Margin
per share Pre - tax Operating M arg in = calculated
5
Return on Equity measures the rate of return on shareholders
as the pre-tax operating income of year t
equity and is calculated as the Net Income divided by the book divided by the sales of year
value of equity f = the sum of errors derived from the
6
Net Margin measures the profitability and is calculated as the net
income divided by revenues
regression analysis
7
Return on Capital measures the rate of return on the companys
capital and is calculated as operating earnings after taxes divided
by the book value of equity plus the book value of debt
8
Pre-tax Operating Margin or EBIT measures profitability before
interests and taxes and is calculated as operating revenues less
operating expenses plus non-operating income
We extracted information for public table shows the number of companies,

companies in the United States for the ending divided by sector, that were analyzed each
of the years 2000 to 2009. The following year.
Table 1. Number of companies by sector
Industries 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Advertising 30 33 37 32 34 32 36 39 28 36
Aerospace/Defense 39 64 76 72 66 70 73 68 65 67
Air Transport 36 44 42 43 45 45 56 48 43 44
Auto & Truck 19 20 22 21 24 29 31 27 19 22
Auto Parts 55 64 61 62 59 58 64 55 53 54
Bank 211 489 505 558 546 537 598 551 521 527
Beverage 32 41 43 41 37 41 48 43 40 41
Biotechnology 0 75 90 84 89 87 105 102 107 121
Building Materials 50 65 68 62 60 57 60 59 64 67
Chemical 125 141 149 144 136 143 154 143 137 145
Coal 0 5 7 8 10 12 16 17 17 21
Computers 399 467 630 535 530 533 572 518 445 462
Diversified Co. 92 102 105 103 117 118 134 106 117 126
Drug 286 285 288 276 304 306 334 367 341 337
Educational Services 27 35 36 34 37 38 37 38 33 38
Electric Utility 88 82 75 77 74 74 75 71 67 66
Electrical Equipment 81 91 94 86 91 91 94 85 82 87
Electronics 150 208 211 194 189 186 196 187 181 192
Entertainment 91 106 126 116 117 118 132 129 115 130
Environmental 50 79 85 77 84 91 96 88 78 91
Financial Svcs. 184 223 238 231 232 244 269 294 295 296
Food Processing 86 115 114 104 103 110 123 122 108 121
Furniture and Furnishings 33 32 35 36 37 36 38 38 33 35
Healthcare Information 32 32 35 35 31 35 34 37 28 33
Homebuilding 54 56 48 44 33 34 41 35 31 28
Hotel/Gaming 52 84 94 80 76 76 84 74 67 74
Household Products 29 32 32 32 29 26 31 27 25 23
Human Resources 0 21 28 27 27 30 35 34 30 30
Industrial Services 173 211 205 190 199 207 230 194 166 168
Information Services 0 18 30 29 32 36 41 37 33 29
Insurance 86 84 96 115 120 127 137 135 121 124
Internet 315 446 422 330 346 365 389 320 260 295
Investment Co. 45 41 39 39 36 36 35 31 41 44
Machinery 134 167 172 153 147 149 153 135 123 130
Manuf. Housing/RV 20 19 20 18 18 16 19 17 17 15
Maritime 14 19 19 23 27 39 46 51 55 53

Industries 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Medical Services 156 207 226 197 194 184 186 177 159 162
Medical Supplies 182 224 244 236 261 261 279 273 251 264
Metal Fabricating 38 45 45 37 37 41 37 36 34 36
Metals & Mining 36 44 55 54 75 77 82 77 77 79
Natural Gas 71 72 72 70 66 65 64 69 75 75
Newspaper 18 20 19 20 19 19 18 17 15 15
Office Equip/Supplies 28 33 33 31 27 27 26 24 25 25
Oilfield Svcs/Equip. 71 87 87 88 92 98 110 112 111 113
Packaging & Container 36 42 41 34 34 37 36 34 32 31
Paper/Forest Products 48 47 43 40 38 40 42 38 37 39
Petroleum 134 175 170 162 177 178 208 210 211 222
Pharmacy Services 10 14 15 15 13 15 20 18 18 21
Power 15 30 33 29 33 36 55 69 76 87
Precious Metals 30 37 42 48 60 62 67 83 74 78
Precision Instrument 85 108 114 102 103 104 104 102 89 98
Property Management 0 0 0 0 0 0 0 11 16 20
Publishing 43 44 42 40 42 47 50 39 26 30
R.E.I.T. 143 152 151 144 134 122 143 146 143 143
Railroad 16 15 15 16 17 18 20 15 14 15
Recreation 81 86 88 73 77 74 84 72 63 65
Restaurant 90 93 94 85 83 82 81 74 67 68
Retail (Special Lines) 232 240 255 255 238 237 227 219 206 213
Retail Automotive 10 18 17 12 26 27 25 15 15 15
Retail Building Supply 11 10 10 8 8 10 9 8 7 7
Retail Store 80 117 119 90 89 90 91 72 67 75
Securities Brokerage 27 27 24 27 25 31 32 30 31 30
Semiconductor 113 126 123 128 138 135 138 152 136 139
Shoe 26 26 27 24 23 22 24 19 18 19
Steel 48 48 45 41 36 40 46 38 32 35
Telecom. Equipment 112 140 146 124 119 122 136 123 109 115
Telecom. Services 200 252 234 180 176 188 196 173 163 164
Thrift 124 116 146 242 221 221 248 233 233 227
Tobacco 11 12 12 11 12 11 11 10 11 12
Toiletries/Cosmetics 17 20 24 21 22 20 21 20 22 19
Trucking 45 42 40 39 35 37 38 31 32 33
Water Utility 13 15 16 16 16 16 16 15 15 15
Wireless Networking 69 79 76 63 65 66 73 73 56 60
Total 5587 6959 7320 6913 6973 7092 7659 7249 6752 7036
The following graphs show the relationships 2009 of all the companies analyzed. At
between P/E and ROE and Net Margin first glance, it appears to be no linear
and between EV/EVITDA and ROC and relationship at all.
Pre-tax Operating Margin for the year
Graph 1. P/E vs ROE Graph 2. P/E vs Net Margin
P/E multiple P/E multiple

1800 1800
1600 1600
1400 1400
1200 1200
1000 1000
800 800
600 600
400 400
200 200
0 0
0 1000 2000 3000 4000 5000 6000 7000 0 1000 2000 3000 4000 5000 6000 7000
ROE Net Margin
Graph 3. EV/EBITDA vs ROC Graph 4. EV/EBITDA vs Pre-tax Operating Margin
EV/EBITDA EV/EBITDA
multiple multiple
1800 1800
1600 1600
1400 1400
1200 1200
1000 1000
800 800
600 600
400 400
200 200
0 0
0 1000 2000 3000 4000 5000 6000 7000 0 1000 2000 3000 4000 5000 6000 7000
ROC Pre-tax Operating Margin
The basic linear regression model assumes From equations 1 and 2, we apply natural
that the contributions of the different logarithms to obtain:
Equation 3
independent variables to the prediction of
the dependent variable are additive and Ln (P/E) = aPE + bROE * Ln (ROE) + bNM *
they tend to follow normal distributions. Ln (Net M arg in) + f
Equation 4
In our case, the relationships between our
variables may be multiplicative and also Ln (EV/EBITDA) = aEVEBITDA + bROC * Ln (ROC) +
they have highly skewed distributions bOM * Ln (Pre - tax Operating Margin) + f
(positive values). Hence it may be
possible to make their distributions Now, scatter plots for the relationships
more normal-looking by applying the between the variables transformed
logarithm transformation, as shown in to natural logarithms show a linear
the following equations: relationship.

Graph 5. Ln(P/E) vs Ln(ROE) Graph 6. Ln(P/E) vs Ln(Net Margin)
Ln(P/E) Ln(P/E)
8 8
6 6
4 4
2 2
0 0
-2 -2
-4 -4
-6 -6
-8 -6 -4 -2 0 2 4 -8 -6 -4 -2 0 2 4
Ln(ROE) Ln(Net Margin)
Graph 7. Ln(EV/EBITDA) vs Ln(ROC) Graph 8. Ln(EV/EBITDA) vs Ln(Pre-tax Operating

Margin)
Ln(EV/EBITDA) Ln(EV/EBITDA)
8 7
6
6
5
4 4
2 3
2
0
1
-2 0
-1
-4
-2
-6 -3
-8 -6 -4 -2 0 2 4 -8 -6 -4 -2 0 2 4
Ln(ROC) Ln(Pre-tax Operatin Margin)
It has to be noted, that the number of P/E ratio. Although the R2 coefficients do
companies that can be used is reduced by not present high values, the F-test shows that
transforming variables to natural logarithms we can use the linear model. For all years,
due to the fact that negative values for our ROE is a significant variable (not equal to
variables cannot be transformed to natural zero), but that is not the case for Net Margin
logarithms. This is a limitation for the model. in years 2004, 2007 and 2008. For purposes
of comparison we will use both factors in
Results all years.
We calculated equation (3) for years 2000

to 2009 to determine the relationship for the
Table 2. Regression results for equation (3) - P/E
Concept 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Multiple correlation factor 0.2715 0.4975 0.5090 0.5882 0.6216 0.5984 0.5673 0.5344 0.4400 0.5324
R-square coefficient 0.0737 0.2475 0.2591 0.3459 0.3864 0.3581 0.3218 0.2856 0.1936 0.2834
F test value 799.74
111.064 545.417 503.213 800.975 964.306 890.183 622.593 367.977 543.500
Critical F value 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
Intercept 3.2672 1.7281 1.6491 2.0136 1.9820 1.9316 1.8731 1.9007 1.2959 1.5055
Prob Intercepcin 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
ln(ROE) -0.2856 -0.5912 -0.5763 -0.5552 -0.5558 -0.6222 -0.5305 -0.5242 -0.4720 -0.4952
Prob ln(ROE) 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
ln(Net Margin) 0.3490 0.0744 0.0406 0.0297 -0.0212 0.0436 -0.0652 -0.0242 0.0066 -0.0841
Prob (Net Margin) 0.0000 0.0000 0.0444 0.0449 0.1424 0.0051 0.0000 0.1398 0.7257 0.0000
In the case of the EV/EBITDA relationship other factors that need to be considered in
established in equation (4), we obtained the our analysis. As with the Net Margin for
results in the table below. The F-test also equation (3), there are some years (2003,
indicates that the linear regression model 2005 and 2007) in which Pre-tax Operating
is adequate. R2 values are smaller than for Margin is not significant.
equation (3); this may mean that there are
Table 3. Regression results for equation (4) EV/EBITDA
Concept 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Multiple correlation factor 0.1125 0.2354 0.2566 0.2196 0.2201 0.1790 0.1844 0.2308 0.2535 0.2192
R-square coefficient 0.0127 0.0554 0.0658 0.0482 0.0485 0.0321 0.0340 0.0533 0.0643 0.0481
F test value 22.458 113.001 127.863 93.823 95.170 61.088 68.194 98.852 116.734 86.831
Critical F value 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
Intercept 1.9749 1.4948 1.3906 1.6710 1.7411 1.8215 1.8645 1.7209 1.0705 1.5180
Prob Intercepcin 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
ln(ROC) -0.1282 -0.2349 -0.2316 -0.1768 -0.1377 -0.1462 -0.1260 -0.1995 -0.2607 -0.1388
Prob ln(ROC) 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
ln(Pre-tax Op Margin) 0.0795 0.0649 0.0566 -0.0042 -0.0682 -0.0023 -0.0334 -0.0016 0.0597 -0.0141
Prob (Pre-tax Op Margin) 0.0001 0.0001 0.0001 0.7744 0.0000 0.8795 0.0143 0.9147 0.0004 0.2998
Now that we have established our the P/E and EV/EBITDA and compare them
regression parameters for equations (3) and to arithmetic averages.
(4) by year, we now calculate for each year
Table 4. Average multiples vs REVAAM multiples
Net P/E - P/E - Pre-tax In(EV/ EV/EVITDA- EV/EVITDA-

Year ROE ln(P/E) ROC
Margin regression average OM EVITDA) regression average
2000 23.0% 8.9% 2.8417 17.14 44.08 14.4% 12.8% 2.0596 7.84 14.64
2001 17.2% 10.5% 2.5995 13.46 48.35 14.8% 9.6% 1.7916 6.00 9.71

Net P/E - P/E - Pre-tax In(EV/ EV/EVITDA- EV/EVITDA-

Year ROE ln(P/E) ROC
Margin regression average OM EVITDA) regression average
2002 19.8% 9.3% 2.4848 12.00 36.46 13.9% 9.2% 1.7120 5.54 7.74
2003 18.9% 11.6% 2.8738 17.71 45.02 14.3% 13.0% 2.0239 7.57 11.96
2004 17.6% 9.6% 2.9976 20.04 50.85 14.5% 13.0% 2.1458 8.55 13.62
2005 25.8% 11.0% 2.6783 14.56 45.20 15.8% 13.5% 2.0955 8.13 15.07
2006 18.5% 15.7% 2.8883 17.96 44.05 16.1% 13.8% 2.1612 8.68 12.88
2007 22.9% 15.4% 2.7195 15.17 49.59 16.4% 14.1% 2.0852 8.05 11.94
2008 22.2% 11.7% 1.9913 7.33 18.64 15.7% 14.7% 1.4383 4.21 6.10
2009 21.6% 11.9% 2.4441 11.52 30.04 15.9% 14.1% 1.8008 6.05 8.61
As can be seen in the numbers and in the P/E and EV/EBITDA in each year are above
following graphs simple average values for those calculated from the regressions.
Graph 9. P/E regression vs P/E average Graph 10. EV/EBITDA regression vs EV/EBITDA
average
2009 P/E - average 2009 EV/EBITDA - average

2008 P/E - regression 2008 EV/EVITDA - regression
2007 2007
2006 2006
2005 2005
2004 2004
2003 2003
2002 2002
2001 2001
2000 2000
- 10.00 20.00 30.00 40.00 50.00 60.00 - 5.00 10.00 15.00 20.00
In the case of P/E ratios, average values to obtain an equity value for the company
are about 2.2 times the values obtained from being analyzed.
regression analysis, and for EV/EBITDA
average values are 1.57 times the regression Graph 11. Regression Multiples over time
values.
Another important aspect can be derived 30.00
P/E
from the year to year comparison of our 25.00 EV/EBITDA
multiples. As shown on Graph 11, P/E 20.00
multiple tends to be more volatile over 15.00

time than the EV/EBITDA multiple. For 10.00
the former, the average from 2000 to 2009
has been 19.04 with a standard deviation of 5.00
5.02, whereas for the latter the average has -

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
been 7.06 with a standard deviation of 1.5.
Crises of 2001 and 2008 contributed to these
deviations. Then, at first glance, it would be
a better idea to use an EV/EBITDA multiple
and then subtract all financial debt in order
Further Applications discounting future cash flows, then both

companies should have the same resulting
REVAAM Model provides adjusted multiples since their margins are the same
multiples that should be more accurate than (not the same value because cash flows
using simple average multiples. So far we amount depend on the company size).
have applied the model to the entire market, Common use suggests taking multiples by
without distinguishing among different industry, and the two important supporting
sectors. arguments for this are differences in risk and
There are several analysts that support potential growth (technological sectors tend
the idea that companies in different sectors to have bigger growths than utility sectors,
should not be priced equally even they have for example).
the same characteristics. This would imply We used REVAAM model to calculate
that company A and company B should the corresponding EV/EBITDA and P/E
have a different EV/EBITDA multiple multiples for all sectors for year end 2009
even though they have exactly the same (each year should be calculated depending
ROC and Pre-tax Operating Margin. Others on prevailing factor values), as shown in
suggest that since DCF valuation involves Graphs 12 and 13
Graph 12. EV/EBITDA calculation by sector for 2009
25.00
EV/EBITDA regression
EV/EBITDA average
20.00
15.00
10.00
5.00 4.7
-
Information Services
Medical Supplies
Machinery
Advertising
Oilfield Svcs/Equip.
Telecom. Services
Pharmacy Services
Computers
Medical Services
Metal Fabricating
Trucking
Petroleum
Securities Brokerage
Diversified Co.
Building Materials
Tobacco
Hotel/Gaming
Chemical
Telecom. Equipment
Aerospace/Defense
Toiletries/Cosmetics
Metals & Mining (Div.)
Maritime
Property Management
Biotechnology
Investment Co.
Thrift
Office Equip/Supplies
Household Products
Precious Metals
Insurance
Environmental
Beverage
Coal
Power
Auto & Truck
The solid line in Graph 12 is the average multiple that analysts use when
calculated EV/EBITDA for each sector doing relative valuation.
using the corresponding average ROC and Sectors whose average EV/EBITDAs
average Pre-tax Operating Margin in 2009. are above the REVAAMs EV/EBITDAs
Dots represent the simple EV/EBITDA are considered to be overvalued, whereas

those below are considered undervalued. Metals, Homebuilding, Healthcare

Looking at Graph 12, we can observe that Information and Educational Services are
25 sectors are undervalued, 10 sectors are the most overvalued sectors, while Steel,
within a 5% range of fair price, and 37 Petroleum, Air Transport and Trucking are
are overvalued. Thrift/Banking, Precious the most undervalued.
Graph 13. P/E calculation by sector for 2009
80.00
P/E regression
70.00 P/E average
60.00
50.00
40.00
30.00
20.00
10.00
-
Healthcare Information
Precision Instrument
Financial Svcs. (Div.)
Entertainment
Recreation
Machinery
Water Utility
Food Processing
Advertising
Electrical Equipment
Telecom. Services
Computers
Pharmacy Services
Petroleum
Diversified Co.
Internet
Hotel/Gaming
Publishing
Telecom. Equipment
Aerospace/Defense
Electric Utility
Retail Automotive
Retail Building Supply
Metals & Mining (Div.)

Property Management
Industrial Services
Thrift
Office Equip/Supplies
Household Products
Semiconductor
Insurance
Auto Parts
Beverage
Steel
Coal
Power
Similar to Graph 12, the solid line in Taking into consideration results from
Graph 13 is the calculated P/E for each both graphics 12 and 13, we find common
sector using the corresponding average sectors that have the same results. However,
ROE and average Net Margin in 2009, and there are sectors that show opposite results
dots represent the arithmetic average of P/E in both analysis, such as the Electric Utility,
multiples in each sector. Retail Automotive, REIT, Manufacturing
Comparing average P/Es to REVAAMs Housing/RV, Trucking and Telecom Services
P/Es, we can observe in Graph 13 that 27 sectors.
sectors are undervalued, 13 sectors are We have used the REVAAM Model
within a 5% range of fair price, and 32 in comparing market-wide multiples in
are overvalued. Thrift/Banking, Precious specific years and through time, and also by
Metals, Tobacco, Healthcare Information, comparing different sectors. Now we will
Insurance and Educational Services are the apply the model to a specific company that
most overvalued sectors, while Maritime, belongs to the Drug sector (Pharmaceutical
Steel, Manufacturing Housing/RV, Office Laboratories), and we will use a US based
Equipment and Supplies and Metal company called Perrigo Co. (Ticker PRGO).
Fabricating are the most undervalued. The table below shows the parameters
for Perrigo Co. at the end of each year and REVAAM Model in the specific Drug
the actual P/E and EV/EBITDA multiples sector. The same is done to the multiples of
that the company is trading at, and then the whole Drug sector.
we calculate these multiples by using the
Table 5. Perrigos and Drug sector multiples (REVAAM and averages)

Perrigo Co. 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
ROC 10.33% 13.80% 17.11% 18.58% 22.35% 4.78% 9.62% 9.84% 9.84% 16.40%
Pre-tax OM 7.99% 7.10% 7.20% 13.10% 14.62% 9.52% 12.94% 11.39% 11.39% 16.79%
ROE 5.63% 10.68% 10.69% 11.57% 12.59% 6.42% 11.57% 10.42% 10.42% 19.05%
Net Margin 2.74% 5.47% 5.39% 6.28% 7.52% 3.70% 5.42% 5.43% 5.43% 8.75%
Current P/E 35.11 22.58 20.05 21.74 18.41 37.10 21.70 42.96 37.15 20.40
Current EV/EBITDA 11.57 8.70 6.76 7.77 6.76 15.52 9.44 17.84 15.79 10.22
REVAAM P/E 21.86 52.42 46.29 63.25 61.39 69.50 71.20 58.42 19.11 36.61
REVAAM EV/EBITDA 17.13 19.76 11.14 15.72 16.97 18.83 18.09 15.28 6.19 8.98
Av. Sector ROC 18.20% 15.80% 20.10% 21.06% 19.96% 19.17% 22.46% 19.10% 22.93% 22.10%
Av. Sector Pre-tax 16.05% 12.90% 18.02% 25.35% 22.90% 21.38% 23.02% 22.93% 23.49% 23.80%
OM
Av. Sector ROE 18.25% 11.43% 13.49% 14.54% 12.06% 12.42% 13.01% 13.44% 15.46% 16.64%
Av. Sector Net 15.68% 14.59% 15.51% 14.40% 12.27% 12.03% 11.16% 12.84% 15.14% 16.62%
Margin
Av. Sector P/E 34.1 35.5 23.5 30.3 28.9 27.0 25.2 23.9 12.3 14.7
Av. Sector EV/EBITDA 19.4 15.1 10.2 13.6 10.9 12.9 9.9 9.8 6.2 7.5
REVAAM Sector P/E 34.85 29.09 21.16 28.04 24.23 24.16 23.57 21.82 10.53 14.22
REVAAM Sector EV/ 17.50 17.49 9.96 13.71 13.14 13.24 12.43 10.72 5.74 7.49
EBITDA
The following graphs (14 and 15) show Perrigo was trading at the end of year 2009,
the REVAAM calculations for the P/E and and simple average multiples for the Drug
EV/EBITDA multiples respectively and the sector at the same date.
comparison against the current multiples
Graph 14. Perrigos and Drug sector P/E Graph 15. Perrigos and Drug sector EV/EBITDA
calculation calculation
Current P/E REVAAM P/E Av. Sector P/E REVAAM Sector P/E Current EV/EBITDA REVAAM EV/EBITDA
80.00 25.00
Av. Sector EV/EBITDA REVAAM Sector EV/EBITDA
70.00
20.00
60.00
50.00 15.00
40.00
30.00 10.00
20.00
5.00
10.00
- -
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

As it can be seen, in both graphs the individual companies, such as the one used
REVAAM multiples (both P/E and EV/ as example, we may be understating its value
EBITDA) for the Drug sector are very close if we use only the average sector multiple.
to the simple averages of those multiples. In our experience most (over 60%) Mergers
However, when looking at the multiples and Acquisitions are valued using simple
at which Perrigo was trading compared average multiples.
to calculated REVAAM multiples for the As stated initially, the preferred
company, it seems that Perrigo was trading valuation method for the authors is the
at lower multiples than it should had been use of DCF analysis, however relative
or it was undervalued. In the case of the P/E valuation may give a quick insight of the
multiple this is true for all years except for intrinsic value. Just considering one, two
2000 and 2008, and for the EV/EBITDA or more financial factors that differentiate
multiple Perrigo was undervalued for all companies valuations, is a better proxy to
years except for 2007, 2008 and barely 2009. what a DCF analysis will show for different
Another important conclusion is that, companies. The model also lets us compare
in general, Perrigo is a company whose market valuations to those derived from
REVAAM multiples are over the sector the REVAAM Model. As seen in the recent
REVAAM multiples, suggesting that it crises the market sometimes is not rational,
outperforms other companies in the sector, and this model let us take into consideration
and may be a good investment in the long objective financial and operational factors
run. without the inertia of bull or bear markets.
Additionally, the REVAAM Model provides
Conclusions different depth levels of analysis, from
the whole market to sector and company
REVAAM Model does present several specific calculations, and allows for time
limitations: such as a) it considers few comparison based on the prevailing factors
factors to define the multiples, b) due at each period.
to logarithms transformation it cannot Further work in this model may include
consider companies with negative values determining other significant factors that
for the factors, c) it is based on historical may affect valuations to be included in
data and does not take into consideration the regression analyses, also developing a
future growth, d) it requires at least 30 version for mature markets and markets in
companies for the linear regressions to be development, that take into consideration
significant, and some sectors do not include risk differences among countries.
that number of companies, e) is applied to
public companies because of information References
availability and f) may not apply to markets
where there are few public companies in Schreiner, Andreas. 2007. Equity
order to be statically significant. Valuation Using Multiples: An
However, we believe that it provides a Empirical Investigation Dissertation der
better estimate to relative valuation analysis Universitt St. Gallen (HSG)
than simply using average multiples. We
have seen that even for sector analysis some Damodaran, Aswath. 2002. Investment
valuations could be 50% percent off (both Valuation: Tools and Techniques for
positive and negative), and when valuing
Determining the Value of Any Asset Myers, Brealey. 2002. Principles of

Second Edition. Corporate Finance.
McKinsey & Company Inc., et al. 2005. Pagano, Robert R. 1998. Estadstica para
Valuation: Measuring and Managing the las Ciencia del Comportamiento, 5
Value of Companies, Edition, University Edicin.
Edition.
Beaver, W. 1967. Financial ratios as
Damodaran, Aswath. 2010. The Dark Side predictors of failure, Empirical Research
of Valuation: Valuing Young, Distressed, in Accounting. (1966), Selected
and Complex Businesses, 2nd Edition, studies, supplement to the Journal of
Pearson Education, Inc. Accounting Research, January, 1966.
Medlen, Craig. 2007. The bubble machine: Greene W. 1993. Econometric Analysis,
relative capital valuation, distributive 2th Edition, Mc. Millan Publishing Co.
shares and capital gains An article from:
Journal of Economic Issues. Mandala G.S. 1983. Limited Dependent and
Qualitative Variables in Econometrics,
Iwanowski, Raymond J. 1995. An Econometric Society, Monographs
investors guide to floating-rate notes: number 3, Cambridge, University.
Conventions, mathematics and relative
valuation (United States fixed-income Bloomberg Professional monitoring service.
research. Portfolio strategues)
Yahoo Finance (http://finance.yahoo.com)
Hoover, Scott. 2005. Stock Valuation: An
Essential Guide to Wall Streets Most Capital IQ (http://www.capitaliq.com)
Popular Valuation Models McGraw-Hill
Library of Investment and Finance. Value Line (http://www.valueline.com)
Frykman, David and Tolleryd, Jakob. 2003.

Corporate Valuation: an easy guide to
measuring value.

2010 Nitin Merh, Vinod P. Saxena and Kamal Raj Pardasani 23
A COMPARISON BETWEEN HYBRID APPROACHES

OF ANN AND ARIMA FOR INDIAN STOCK TREND
FORECASTING
Nitin Merh, Vinod P. Saxena and Kamal Raj Pardasani
Abstract
In this paper an attempt is made to develop hybrid models of three layer feed forward back propagation
artificial neural network (ANN) and autoregressive integrated moving average (ARIMA) for forecasting
the future index value and trend of Indian stock market viz. SENSEX, BSE IT, BSE Oil & Gas, BSE 100 and
S& P CNX Nifty. Simulations have been done using prices of daily open, high, low and close of SENSEX,
BSE IT, BSE Oil & Gas, BSE 100 and S& P CNX Nifty. These are chosen as input data values and output
is the forecasted closing price of SENSEX, BSE IT, BSE Oil & Gas, BSE 100 and S& P CNX Nifty for the
next day and future trend. Simulation results of hybrid models are compared with results of ANN based
models and ARIMA based models. Convergence and performance of models have been evaluated on the
basis of the simulation results done on MATLAB6.1 and SPSS13.0.
Merh N., Saxena Vinod P., Raj Pardasani K. - A Comparison between Hybrid Approaches of ANN and ARIMA for Indian Stock Trend Forecasting
Introduction (ANN) and auto regressive integrated

moving average (ARIMA). Studies also
Many studies have established that non- show that using dissimilar models result in
linearity exits in financial data and artificial lower generalization error. Recent studies
neural networks can be successfully used show that the techniques of ANN and
to model the relationship among this ARIMA when combined, offer a competitive
data. Neural network can mine valuable edge over each of the individual model.
information from a mass of historical The benefits of such methods appear to be
information and can be efficiently used substantial especially when dealing with
in financial areas, so the applications of non-stationary series (Temizel and Ahemad,
neural networks to financial forecasting 2005). Results from study (Sallehudin, et al.,
have become very popular over the last few 2008) show that GRANN_ARIMA(Grey
years [Zhang, et al, 2004, Windrow, et al, Relational Artificial Neural Network_
1994, Refenes, 1995; Kate, et. al, 2000, J. T. Autoregressive Integrated Moving
Yao, et al, 1999; Abu- Mostafa, at el, 2001]. Averages) is better than the individual model
Artificial neural networks are a natural way and conventional hybrid model in terms of
for solving problems that involve learning accuracy and robustness since it produces
and pattern recognition. It can detect patterns small forecasting errors and can work well in
in data through learning and are much easier both small and large scale data. Studies also
to program since they elicit general rules show (Tseng, et al., 2002) that SARIMABP
from exemplars [Afolabi, et al, 2007]. (Seasonal Autoregressive Integrated
Though back-propagation (BP) neural Moving Averages Neural Network Back
networks are often used because of their Propagation) is superior to the SARIMA
better prediction ability than other models, (Seasonal Autoregressive Integrated Moving
they have certain limitations. Lawrence, Averages) model, the BP (Neural Network
Tsoi, and Giles (1996) pointed out that, Back Propagation) with depersonalized
when the training of a BP tends to be difficult data, and the BP with differenced data for
due to the noise of data, then the networks the test cases. A hybrid ARIMAANN
fall into a naive solution such as always model for time series prediction synergically
predicting the most common output. Miao et combines the advantages of the easy-to-use
al. (2007) indicates that the solutions of the and relatively easy-to-tune ARIMA models
BP usually are forced to the local minimum and the computational power of ANN.
due to the gradient descent algorithm used The process of predicting as accurately
to get weights of connection (Yudong and as possible the stock price of the day is
Lenan, 2009). very difficult if not impossible (Afolabi and
New researches suggest that seeing the Olude, 2007; Brabazon, 2000). Technical
nature of stock market, combining two or analysis involves the analysis of the
more computational models synergistically statistics generated by market activity like
is better approach for prediction problem. past prices and volume. In other words
The basic idea behind this is that each technical analysis uses only historical data
models unique capability can be used to like past prices, volume of trading, volatility
model different patterns of data. The hybrid etc., to determine the movements in the
models can have similar components like price of some financial assets. Technical
two neural networks or can have dissimilar approach is based on three premises that
components like artificial neural networks are (a) market action discounts everything

(b) prices move in trend (c) history repeats Next objective is to compare the results
itself. The use of technical analysis (Malkiel, and trends of actual and predicted values
1996; Skabar and Cloete, 2001) goes of above mentioned indices.
against the conservative academic opinion,
which regards this behavior as irrational To evaluate the performance of the
given the efficient markets hypothesis. models by calculating average absolute
According to Efficient Market Hypothesis error (AAE), root mean square error
(Samanta and Bordoloi, 2005; Fama, (RMSE), mean absolute percentage
1965 and Fama, 1970) all available market error (MAPE), mean percent square
information are factored immediately into error (MPSE) and compare their values.
the formation of stock price and therefore,
the best predictor of future stock price is To compare the results and trends
the latest stock price is the latest available between the traditional hybrid and new
price. Technical analysis, (Murphy, 1999) hybrid approach for forecasting stock
refers to the various methods that aim to market in Indian context.
predict future price movements using past
stock prices and volume information. It is To compare the trends between models
(Mills, 1990; Priestley, 1988) possible that based on single approach based of ANN/
non-linear models are able to explain this ARIMA and hybrid approach using
residual variance and produce more reliable techniques of ANN and ARIMA.
predictions of the stock price movements.
In the present study a hybrid models Financial Time Series
have been developed using techniques of
both artificial neural network (ANN) and A time series is a sequence of vectors
autoregressive integrated moving average Xt = (xt - n, ... ..., xt - i, ... ..., xt - 2, xt - 1) where xt - i
(ARIMA) and their performances are represents past value that varies with time.
compared to the ANN and ARIMA models. In stock prices also the data are typical
In the study MATLAB 6.1.0 and SPSS example of time series. Models in the study
13.0 for windows have been used for predict the future trends on the basis of the
simulations. past values. In the current study open, high,
low and close of BSE 30 (SENSEX), BSE
Purpose of the Research IT, BSE Oil & Gas, BSE 100 and S&P CNX
Nifty are used for the prediction of next
Main objective of the study is to develop stock value and future trend. The price of
models for forecasting the next days volume traded is not a significant predictor
close value of SENSEX, BSE IT, BSE Oil for the time series as during learning in the
& Gas, BSE 100 and S& P CNX Nifty. In ANN model most of the prices against traded
the study hybrid model using techniques volume are zero as the data is not available
of feed forward back propagation with respect to volume traded. Thus traded
supervised learning artificial neural volume is dropped as one of the variables or
network and autoregressive integrated predictor.
moving average are developed and
results are compared.
Table 1. Input and Output Time Series for the Study. Table 3. Control Parameters used in the Hybrid
Model (Error Forecast through ARIMA to ANN) in
Output the Study.
Index Name Input Variables
Variable
BSE 30 (SENSEX) ANN (input for ANN will be
from april Next day Model Index ARIMA
Open High Low Close residual from ARIMA)
16,2004 to april close price
16,2009 ANN(5-5-1)Linear-log
BSE30 ARIMA(p,d,q)= sigmoid-log sigmoid, a =0.5,
BSE IT (SENSEX) ARIMA(1,1,1) h=0.001, Epochs=1000, Error
from april Next day Tolerance=0.01
Open High Low Close
16,2009 ANN(5-5-1)Linear-log
ARIMA(p,d,q)= sigmoid-log sigmoid, a =0.9,
BSE Oil & Gas BSEIT
ARIMA_ANN (Hybrid Model)

ARIMA(0,1,2) h=0.001, Epochs=1000, Error
from august Next day Tolerance=0.01
Open High Low Close
16,2009 BSE Oli &
N/A N/A
Gas
BSE 100
from april Next day ANN(5-5-1)Linear-log
Open High Low Close ARIMA(p,d,q)= sigmoid-log sigmoid, a =0.5,
16,2004 to april close price BSE 100
16,2009 ARIMA(1,1,1) h=0.001, Epochs=1000, Error
Tolerance=0.01
S&P CNX Nifty
from april Next day ANN(5-5-1)Linear-log
Open High Low Close S&P CNX ARIMA(p,d,q)= sigmoid-log sigmoid, a =0.5,
16,2009 NIFTY ARIMA(1,1,1) h=0.001, Epochs=1000, Error
Tolerance=0.01
* In ANN 5-5-1 five input values are set containing samples

Table 2. Control Parameters used in the Hybrid of five values of residuals generated from ARIMA using
Model (Error Forecast through ANN to ARIMA) in suitable model.
the Study.
Use of ANN in Hybrid Modelling
ARIMA (input for
ARIMA will be
Model Index ANN
residual from ANN) Artificial neural network model has
ARIMA(p.d.q.)= been used in two different ways in the
ANN(4-4-1)Linear-log first case where input data set comprising
BSE30 sigmoid-log sigmoid, a =0.9,
SENSEX h=0.01, Epochs=1000, Error
ARIMA(1,0,0) of open, high, low and close price time
Tolerance=0.001 series of different indexes selected for the
ANN(4-4-1)Linear-log study is provided to the ANN model and
sigmoid-log sigmoid, a =0.9,
BSEIT
h=0.01, Epochs=1000, Error
ARIMA(0,0,1) in the second case the residual generated
from suitable ARIMA model is provided
ANN_ARIMA (Hybrid Model)
Tolerance=0.001
ANN(4-4-1)Linear-log to ANN model. Hence in the first case data
BSE Oli & sigmoid-log sigmoid, a =0.9,
Gas h=0.01, Epochs=1000, Error
ARIMA(1,0,1) set contains four different time series with
Tolerance=0.001 respect to open, high, low and close price
ANN(4-4-1)Linear-log whereas in the second case data set contains
BSE 100
h=0.01, Epochs=1000, Error
ARIMA(1,0,0) time series generated as the residual from
Tolerance=0.001 ARIMA model which are represented in the
ANN(4-4-1)Linear-log following diagrams (Figure 1 and Figure 2).
S&P CNX
NIFTY h=0.01, Epochs=1000, Error
ARIMA(1,1,1) Input data for all Indices consist of
Tolerance=0.001 data collected from http://www.cmie.com
(CMIE Prowess database release 3.1, Date
of Downloading: May 15, 2009).

[a] Procedure for Scaling of time series

(1)
ymax - ymin
F=
xmax - xmin
where xmin to xmax are scaled from ymin to ymax

(2)
F = 0.9 - 0.1
xmax - xmin
Figure 1. Three Layer Feed-forward Neural
such that ymax = 0.9 and ymin = 0.1
Network with One Hidden Layer for Hybrid Model
(3)
(Error Forecast through ANN to ARIMA).
Offset = (ymin - (F) * xmin)
(4)
When data is loaded in the ANN (Artificial
Neural Network), it must be preprocessed Offset = (0.1 - (F) * xmin)
from its numeric range into the numeric
range that the ANN can deal with efficiently. The input data is preprocessed by using
In this process, proper transformation of data the following scaling function and the output
simplifies the process of learning and may data is again processed using reverse scale.
improve the generalizability of the learned
(5)
results (Kim and Lee, 2004).
F * X + Offset
where X is the data which is to be processed
[b] Procedure for Descaling of time series
After processing the input value through

the artificial neural network model proposed
in the paper the data at the output layer is
again scaled in the original numeric range
using the following method:
(6)
Figure 2. Three Layer Feed-forward Neural ymax = 0.9 and ymin = 0.1
(7)
Network with One Hidden Layer for Hybrid Model
(Error Forecast through ARIMA to ANN). ymax - ymin
F=
xmax - xmin
Where A is the scaled value which to be
In artificial neural network model at the
hidden and output layer log sigmoid transfer converted in the original numeric value.
(8)
function has been used and its range lies in
[0, 1]. So before providing the data to the A - ymin
val =
F + xmin
network, the data is to be preprocessed or
scaled. Once the input scaled data is processed
through the artificial neural network and the Back propagation algorithm (supervised
output obtained, it is descaled. Methods for learning) is chosen for the training of the
preprocessing and post processing data are neural network, in which finite number of
as described as follows (Merh, Saxena and pairs of input patterns Xk and target output
Pardasani, 2008): patterns Yk in the form (X1, T1), (X2, T2), ...
..., (Xk, Tk) are initialized. Training dataset sigmoid as a neuron signal function is often
for artificial neural network module consists justified by the biological argument that it
of approximately 1218 samples. The dataset represents the behaviour of the firing rate of
used for prediction consist of 30 days closing action potential, averaged over a population
price value. In model I (Figure 1) for of neurons (Kumar, 2004; Koch, 1999).
training open, high, low and close price has More specifically, for the input layer the
of about 1218 sample been provided to the values are as follows:
(9)
model and next day close is provided as the
target. While for the prediction next thirty xik = /un = 0 wuik pik, i = 1, 2, ... ..., n
(10)
days forecasted values have been generated
from the model. In model-II (Figure 2) f (xik) = xik, i = 1, 2, ... ..., n
1218 residuals generated from ARIMA are where xik is i th component of the input vector Xk
provided for training and residual of next and k is the iteration index
day is the target for training. In the current
research work control parameters learning Input for the hidden layer is as follows:
rate, momentum, epochs, error tolerance,
(11)
target and number of hidden layers are
taken randomly. z =/
k
h
n
i=0
k k
w x , h = 1, 2, ... ..., m
ih i
Neural Network Architecture and Output generated by the hidden layer will
Design be as follows:
(12)
k
Network architecture consists of an input f (z hk) = 1/ (1 + e ), h = 1, 2, ... ..., m

- zh
layer, one hidden layer and an output layer;

index range for the above three layers are Input for the output layer of the neural
i=1,2,.n, h=1,2,..m and j=1,2, network will be:
(13)
p respectively and for input dataset
index range is u=1,2,n. In the model, y =/
k
j
m
h=0
k k
w f (z ), j = 1, 2, ... ..., p
hj h
selection of number of hidden layers is

random. Linear transfer function is used at Output derived at the output layer is as
the input layer which receives the external follows:
(14)
real input and produces the same output, k
while for the output layer and the hidden f (y kj ) = 1/ (1 + e ), j = 1, 2, ... ..., p
-y j
layer log sigmoid transfer function is used

as the transfer function. Results from the In the above architecture biases are
research (Merh, Saxena and Pardasani, included with weights. The weights having
2008) show that in the case of only three index 0 will be treated as bias for the given
k k k
layers an input layer, one hidden and an layer and neuron i.e., w0i, w0h, and w0j are the
output layer results where linear function at biases for the input layer, hidden layer and
the input layer, log sigmoid transfer function output layer respectively. The inputs datasets
at the hidden & output layer are used and and output dataset is defined in Table 1.
when linear function at the input layer, tan After deciding the architecture of the
sigmoid transfer function at the hidden layer neural network, next step is to evaluate the
and output layer are used, generate similar other control parameters as indicated in
patterns and errors. The selection of the Table 2 and Table 3 which are important

for the development of the overall network differentiable processing element function
structure. These factors include learning for the hidden layer PEs would allow chain
rate, momentum factor, initial weight, error rule of partial differentiation to be used to
function, bias function and target output. calculate weight changes for any weight in
For the estimation of above said parameters, the network. This was realized while solving
various strategies are required rather it credit assignment problem (Barto, 1984;
is an art to decide these parameters. The Minsky, 1961; Sinencio and Lau, 1992).
current work without emphasizing on these Multilayer learning (in this case three
strategies assumes learning rate, momentum, layer learning) for the Figure 3 can be
epochs, error tolerance and target randomly. explained as follows (Sinencio and Lau,
1992):
Learning Process The output error across the entire FZ
processing element is found by the following
Back propagation algorithm (supervised cost function:
(19)
learning) is chosen for the training of the
neural network, in which finite number of /= / 1
2
q
j=1 (bkj - z j) 2
pairs of input patterns Xk and target output Following equation computes the output
patterns Tk in the form (X1, T1), (X2, T2), ... of FZ processing element z j :
(20)
..., (Xk, Tk) are initialized.
Instantaneous error Ek for kth learning zj = / p
y Wij
1=1 i
pattern will be: For the hidden layer FY PE, yi is calculated

(15)
by the following equation:
Ek = TK - f (yk) (21)
(16) yi = f (/ n
akh vhi) = f (ri)
h=1
where (22)
TK = (e1k, e2k, ... ..., e kp) T = (t1k - f (y1k), ... ..., where ri = / n
h=1 akh vhi
t kp - f (y kp))
(17)
The weights adjustment is performed
Mk = 1 / pj = 1 (t kj - f (y kj )) 2 by moving along the cost function in the
2
opposite direction of the gradient to a
Mean Square Error on the entire training minimum (minimum is the input /output
set T on a particular neural network will be mapping producing the least amount of
as follows: the total error). The connection weights
(18)
between the FY and FZ processing elements
M = 1 / Qk = 1 (M) k are adjusted and for the two layers the error
Q
correction learning will be explained by the
where M is the mean square error and following equations:
(23)
k is the iteration index. We have a set of Q
training vector pairs. 2E/2wij = 2/2wij 6 21 / qj = 1 (bkj - z j) 2 @
(24)
Multilayer Error Correction Learning = (bkj - z j) yi

(25)
Processing element (PE) is a portion = d j yi

of the artificial neural network where all
computing is performed. A continuously
A positive constant valued learning rate range be 0.1 to .0001, depending upon the
(h) has been added to adjust the amount of application.
change with each move down the gradient.
Adjustments to the connection weights Use of ARIMA in Hybrid
between the FX and FY processing elements Modelling
are derived using the chain rule (Sinencio
and Lau, 1992) The ARIMA model is representative
of linear models and has achieved great
(26)
popularity since the publication of Box
2E/2vih = (2E/2yi) (2yi /2ri) (2ri /2xh) (2xh /2vhi) Jenkins classic book: Time-Series Analysis:
(27)
Forecasting and Control (Box and Jenkins,
= / (bkl - yl) yl whl f' (ri) akh
p
i=1
1976; Zhang, 2001). Different stages of
designing and implementing an ARIMA
A positive constant valued learning rate (p, d, q) based model are identification,
(b) has been added to adjust the amount of estimation, diagnosis and forecasting. If
change with each move down the gradient. seasonal component is also included then the
model becomes ARIMA (p, d, q) (P, D, Q)s
where s is number of periods per session and
P, D, Q are seasonal components. Financial
time series can be stationary or non-
stationary. For converting a time series to a
stationary series it is necessary to difference
the time series d times. ARIMA (p, d, q) is
applying ARMA (p, q) to a stationary series
where p is the order of auto regression,
q is the order of moving-average and d is
the order of differencing. It is also known
Figure 3. Three Layer Neural Network With set of as Box Jenkins (BJ) Method. Training
Inputs and Outputs. dataset ARIMA consists of more than 1218
samples.
From equations 25 and 27 weight
adjustment equation becomes
(28)
Hybrid Model of ANN and ARIMA
(ANN_ARIMA)
wijnew = wijold - h (2E/2wij)
and It may be reasonable to consider a
(29)
time series to be composed of a linear
w hinew = w hiold - b (2E/2vhi) autocorrelation structure and a non-linear
component. In the proposed model there are
Large learning rate can lead to oscillations mainly two stages. In the first stage ANN is
during learning. A refined method of used to forecast the future value of the close
increasing learning rate is to introduce price and then the residual generated is
a momentum term into weight update provided to the ARIMA which will forecast
procedure, where (learning rate) > 0. The the error forecast. In the second stage the
training process will continue until the error predicted close price by ANN is summed
tolerance reaches some specific value let the

with the error forecast generated by ARIMA

which produces the final forecasted value.
Let Xt be the time series for t = 1, 2, 3
which generates Nt as the forecasted series
by implementing ANN and the residual (et)
will be
(30)
et = Xt - Nt where t = 1, 2, 3, 4...
Later ARIMA is used to model et
(residual of the ANN forecasting), which
will generate a series of forecast, let it be
Pt . Studies (Zhang, 2003; Sallehuddin, et
al., 2008) indicate that, this step can be Figure 4. Proposed Hybrid Approach (ANN and
said as a process of error generation of time ARIMA) (Flow Chart of Error Forecast through
series prediction and ARIMA is used as ANN to ARIMA).
error correction of multivariate time series
forecasting for ANN model (Sallehuddin, et Merged ARIMA: Model Identification,
al., 2008). Estimation and Diagnosis
Final forecasting Xp will be as follows:
(31)
The input series is the residual or error
generated from the difference of actual and
X p = Nt + Pt
predicted closing price of the indices from
ANN model during the same time period.
Framework of the proposed model is First stage is to verify that whether the series
shown in figure 4. is stationary or not with the help of sequence
As seen in the Figure 4 the first step is graphs in SPSS 13.0. It can be seen from
the input time series selection. In the present the Figure 5 that series show almost constant
model five data sets BSE 30, BSE IT, BSE mean and variance. Hence it is clear that the
Oil & Gas, BSE 100 and S&P CNX Nifty time series can be said a stationary series
with four variables each has been used. and no differencing is required. ACF and
In the next step the time series is used for PACF on the input series will help to find
training then prediction through ANN. The out the tentative values of p, d and q three
ANN forecasts the next thirty days closing parameters of ARIMA (p, d, q)
price of SENSEX from February 26, 2009
to March 16, 2009. Residual generated by
prediction through ANN is provided to the
ARIMA which in turn generates its residual
series. Finally the residual series generated
by ARIMA is added to the predicted series
generated by ANN, thus giving the final
forecasted values.
Figure 5. Sequence Graph of Residual Generated from ANN Model Indicates that for Indices BSE 30, BSE IT,
BSE Oil& Gas, BSE 100 and S&P CNX Nifty Input Time Series are Stationary.
Values of ACF, PACF, P-Value 0 and best fit was found having minimum
(significance) and LB Statistics (Q*) for Akaike Information criteria (AIC) value.
the maximum lag 10 also indicates that the The best fit was selected after applying
residual series is stationary do not needs any various p and q values the model. Table 5
differencing or transformation using natural shows the best fit model for five cases. To
log (Table 4). It is also found that in case check the adequateness of the model the
of S&P CNX Nifty after taking first order ACF and PACF plots of the error between
differencing results are better. the input series and its fit were generated.
Table 4. ACF, PACF, Significance Value and LB Table 5. Best Fit ARIMA Model for the Given Data
-Statistics Q* Value for all the Indices Taken in The Set.
Study at Ten Degrees of Freedom.
ARIMA
Index p d q
(p, d, q)
LB - Statistic
Significance
(P-Value)
BSE 30(SENSEX) 1 0 0 ARIMA(1,0,0)

(Q*)
Index ACF PACF

BSEIT 0 0 1 ARIMA(0,0,1)
BSE Oil & Gas 1 0 1 ARIMA(1,0,1)
BSE 30(SENSEX) -0.073 -0.230 0.190 13.641 BSE 100 1 0 0 ARIMA(1,0,0)
BSEIT -0.080 -0.178 0.345 11.166 S&P CNX Nifty 1 1 1 ARIMA(1,1,1)
BSE Oil & Gas -0.160 -0.199 0.242 12.684
BSE 100 -0.240 -0.285 0.708 7.188 In the process of diagnosis it is verified
S&P CNX Nifty -0.087 -0.161 0.311 11.625 whether the residual or error generated if
white noise or not. The autocorrelations
After analyzing ACF, PACF and specifying Q*-value, P-value (significance)
performing various iterations by taking at different degrees of freedom having
different p and q parameter values with d as a lag of maximum 10 are shown in the

table (Table 6). To determine whether the and ANN is used to model the residuals
time series is white noise or not, the Box- summation of the ARIMA forecasted stock
Ljung Q* statistic is compared with the price and the error forecast generated by
chi-square distribution with (h-m) degrees ANN produces the final forecasted value.
of freedom. Here h is the number of lags Let Xt be the time series for t = 1, 2, 3
and m is the number of parameters. Table which generates Nt as the forecasted series
6 clearly demonstrates that the Q* statistic by implementing ARIMA and the residual
has same distribution as chi-square with (h- (et) will be
(32)
m) degrees of freedom. The plots and the
autocorrelations generated indicate that the et = Xt - Nt where t = 1, 2, 3, 4...
model fits well.
Later ANN is used to model et (residual of
Table 6. LB Statistics and P-Value of the Error or the ANN forecasting), which will generate a
Residual Generated from ARIMA Model for the series of forecast, let it be Pt. Studies (Zhang,
Verification of White Noise. 2003; Sallehuddin, et al., 2008) indicate
that, this step can be said as a process of
error generation of time series prediction
LB - Statistics Q*
Is Series White
(h-m) Degrees
Significance
of Freedom
and ANN is used as error correction of

(P-Value)
Noise
Index
multivariate time series forecasting for
ARIMA model (Sallehuddin, et al., 2008).
Final forecasting Xp will be as follows:
BSE 30(SENSEX) (10-1)=9 10.209 0.334 Yes
BSEIT (10-1)=9 10.438 0.316 Yes (33)
BSE Oil & Gas (10-2)=8 5.234 0.732 Yes X p = Nt + Pt
BSE 100 (10-1)=9 6.240 0.716 Yes
S&P CNX Nifty (10-2)=8 2.930 0.939 Yes Framework of the proposed model is
shown in figure 6.
Forecasting Using Hybrid ANN As seen in the Figure 6 the first step is
and ARIMA the input time series selection. In the present
model five data sets BSE 30, BSE IT, BSE
Figure 4 outlines the forecasting Oil & Gas, BSE 100 and S&P CNX Nifty
procedure. The above mentioned ARIMA each has been used. In the next step the
models are not used for forecasting purpose. time series is used for model identification
The fit of the input (residual of ANN) series and the 30 samples of hold out data for
generated by the ARIMA models is summed forecasting through ARIMA. This forecasts
with the predicted close values generated the next thirty days closing price of SENSEX
by ANN. This results in the final predicted from February 26, 2009 to April 16, 2009.
value. Residual generated by ARIMA is provided
to ANN for residual analysis. Finally the
Hybrid Model of ARIMA and ANN residual predicted is added to the ARIMA
(ARIMA_ANN) forecasts thus giving the final forecasted
values.
The model has same components as the
ANN_ARIMA model. The difference is that
here ARIMA is used to forecast the values
taking the natural logarithm and differencing

of first order all input series ( BSE 30, BSE
IT, BSE 100 , S&P CNX Nifty) became
stationary (Figure 8) which is evident from
the values of ACF, PACF, LB statistics (Q*)
and P-value (Table 7). In case of BSE Oil
& Gas even after transformation through
natural logarithm and differencing of order
two the series does not become stationary
this is because of heteroscedasticity in the
data; hence it shows that ARIMA modelling
is not suitable technique for BSE Oil & Gas.
Table 7 ACF, PACF, Significance Value and LB

Figure 6. Conventional Hybrid Approach (ARIMA -Statistics Q* Value for all the Indices Taken in the
and ANN) (Flow Chart of Error Forecast through Study at Three Hundred (Maximum) Degrees of
ARIMA to ANN). Freedom.
It was found that without transforming Index ACF PACF

Significance
(P-Value)
LB -
Statistic
the series using natural logarithm and (Q*)
differencing it was non-stationary (Figure 7 BSE 30(SENSEX) 0.018 -0.007 0.191 321.255
(a), Figure 7 (b), Figure 7 (c), Figure 7 (d)). BSEIT 0.005 -0.011 0.353 308.625
If the series is transformed only by taking BSE 100 0.022 -0.004 0.114 329.811
differencing or natural logarithm then also S&P CNX Nifty 0.021 -0.011 0.222 318.425
series is not becoming stationary. After BSE Oil & Gas is not fitting in ARIMA modelling.
Figure 7. (a) Sequence Graph of BSE 30 Close Price from April 16, 2004 to April 16, 2009. Figure 7 (b)
Sequence Graph of BSE IT Close Price from April 16, 2004 to April 16, 2009.

Figure 7. (c) Sequence Graph of BSE 100 Close Price from April 16, 2004 to April 16, 2009. Figure 7 (d)
Sequence Graph of S&P CNX Nifty Close Price from April 16, 2004 to April 16, 2009.
Figure 8. Sequence Graph of Transformed Input Time Series for Indices BSE 30, BSE IT, BSE Oil& Gas, BSE
100 and S&P CNX Nifty. Series is Transformed Using Natural Logarithm and Differencing of First Order.
After analyzing ACF, PACF and AIC value. The best fit was selected after
performing various iterations by taking applying various p and q values the model.
different p and q parameter values with d as Table 8 shows the best fit model for five
0 and best fit was found having minimum cases. To check the adequateness of the
model the ACF and PACF plots of the error Residual derived from suitable ARIMA
between the input series and its fit were model is supplied to ANN 5-5-1 model for
generated. residual analysis and residual is forecasted
as shown in Figure 2. Control parameters
Table 8. Best Fit ARIMA Model for the Given Data for models are explained in Table 3. Various
Set. errors derived are explained in table 10.
ARIMA
Index p d q
(p, d, q)
Results and Simulation
BSE 30(SENSEX) 1 1 1 ARIMA(1,1,1)
Comparisons
BSEIT 0 1 2 ARIMA(0,1,2)
BSE 100 1 1 1 ARIMA(1,1,1)
Comparisons and performance of above
S&P CNX Nifty 1 1 1 ARIMA(1,1,1) models have been evaluated by calculating
BSE Oil & Gas is not fitting in ARIMA modelling.
errors between the actual close price and the
In the process of diagnosis it is verified predicted close price generated by all the
whether the residual or error generated if models (Hybrid ARIMA_ANN and Hybrid
white noise or not. The autocorrelations ANN_ARIMA). In the study average
specifying Q*-value, P-value (significance) absolute error (AAE), root mean square
at different degrees of freedom having a error (RMSE), mean absolute percentage
lag of maximum 300 are shown in the table error (MAPE)and mean square percent
(Table 9). To determine whether the time error MSPE) have been calculated. These
series is white noise or not, the Box-Ljung measures help in calculating and comparing
Q* statistic is compared with the chi-square accuracy of different techniques. It will
distribution with (h-m) degrees of freedom. not only help in measuring a particular
Here h is the number of lags and m is the techniques reliability, usefulness but also
number of parameters. Table 10 clearly help in searching the optimal technique
demonstrates that the Q* statistic has same (Hanke and Wichern, 2007).Figure 9
distribution as chi-square with (h-m) degrees displays the comparison between the actual
of freedom. Plots and the autocorrelations close of BSE 30 (SENSEX), BSE IT, BSE
generated indicated that the model fits well. Oil & Gas, BSE 100 and S&P CNX Nifty
forecasted close price generated by the
Table 9. LB Statistics and P-Value of the Error or models discussed above during Feb 26,
Residual Generated from ARIMA Model for the 2009 to April 16, 2009.
Verification of White Noise. Table 10. Comparison between the Errors of Hybrid
Model of ANN_ARIMA and Hybrid ARIMA_ANN
LB-Statistics Q*
Is Series White
(h-m) Degrees
Models.
Significance
of freedom
(P-Value)
Noise
Errors
Model
Index
Index
AAE RMSE MAPE MSPE
BSE30
(SENSEX)
77.8306 102.1076 0.8200 0.011496
BSE 30
(300-2)=298 289.956 0.620 Yes
(SENSEX) BSEIT 5.8395 7.1693 0.2591 0.001001
ANN_ARIMA
BSEIT (300-2)=298 279.365 0.774 Yes BSE Oil &

Gas
91.8356 117.5227 1.2688 0.023937
BSE 100 (300-2)=298 280.680 0.608 Yes
BSE 100 90.9771 108.8592 1.8602 0.04916
S&P CNX
(300-2)=298 295.677 0.527 Yes S&P CNX
Nifty 36.4592 44.1206 1.2430 0.022582
Nifty
BSE Oil & Gas is not fitting in ARIMA modelling.

Errors
Model
Index
AAE RMSE MAPE MSPE
BSE30
(SENSEX)
164.0544 198..9514 1.7619 0.0451531
ARIMA_ANN
BSEIT 40.4625 46.74907 1.8194 0.44264

BSE 100 87.1768 104.6332 1.8152 0.0469203
S&P CNX
Nifty
43.8553 56.8648 1.4678 0.034431
fBSE Oil & Gas is not fitting in ARIMA modelling.
Snapshot of the Output that Compares Actual and

Predicted Close Prices of SENSEX, BSE IT, BSE Oil
& Gas, BSE 100 and S&P CNX Nifty.
Figure 9. Snapshots of the Output that Compares

Actual and Predicted Close Prices of SENSEX (A),
BSE IT (B), BSE Oil & Gas (C), BSE 100 (D) and
S&P CNX Nifty (E) Generated by ANN, ARIMA,
ANN_ARIMA and ARIMA_ANN for the year 2009.
Conclusion References
In the current research an attempt was made Abu- Mostafa, Y. S., et al. (2001). Neural
to study whether a hybrid model achieves Networks in Financial Engineering,
better results than an individual mode? IEEE Transactions on Neural Networks,
Two hybrid models were developed using 12(4): 653-656.
techniques of ANN and ARIMA comparison
is also made between the conventional hybrid Afolabi, Mark O. and Olude, Olatoyosi
approach using ARIMA, ANN and proposed (2007). Predicting Stock Prices Using a
hybrid approach using ANN, ARIMA. A Hybrid Kohonen Self Organizing Map
hybrid model has been developed using (SOM), Proceeding of 40th International
techniques of back propagation artificial Conference on Systems Sciences, 1560-
neural networks (ANN) and ARIMA for 1605.
analyzing BSE 30 (SENSEX), BSE IT, BSE
Oil & Gas, BSE 100 and S&P CNX Nifty Barto, A., (1984). Simulation Experiments
for forecasting the next days closing price with Goal-Seeking Adaptive Elements,
and predicting the stock trend. In the current Air Force Wright Aeronautical
study, BSE 30 (SENSEX), BSE IT, BSE Oil Laboratory, Technical Report, AFWAL-
& Gas, BSE 100 and S&P CNX Nifty open, TR-84-1022.
high, low, and close prices have been taken
as the input data set and the output generated Box, G.E.P. and Jenkins, G.M., (1976). Time
is the predicted close price of the same. Series Analysis: Forecasting and Control.
Results show that prediction of San Francisco, CA: Holden-Day.
hybrid ANN_ARIMA model for BSE 30
(SENSEX), BSE IT and S&P CNX Nifty are Brabazon, T. (2000). A Connectivist
better than hybrid ARIMA_ANN whereas Approach to Index Modelling in
in the case of BSE 100 results of ARIMA_ Financing Markets, Proceedings of Coil
ANN are marginally better then hybrid / EvoNet Summer School. University of
ANN_ARIMA. Hybrid Model ARIMA_ Limerick, Briys, E., Bellalah, M., Mai, H.
ANN(Error Forecast through ARIMA to M. and Varenne, F. de (1998). Options,
ANN) exhibited over-fitting problem. Futures and Exotic Derivatives. England:
Various trials were done by modifying the John Wiley & Sons Ltd.
learning rate and momentum. Finally the
learning rate was reduced and momentum Fama, E.F. (1965). The Behaviour of Stock
changed accordingly to resolve this problem. Market Prices, Journal of Business, 38:
Results also demonstrate that for BSE Oil 34-105.
& Gas ANN and hybrid ANN_ARIMA are
able to manage the input data set and predict Fama, E.F. (1970). Efficient Capital Markets:
the future closing price, whereas ARIMA A Review of Theory and Empirical Work,
and ARIMA_ANN fails to forecast future Journal of Finance, 25: 1465-1468.
values.
Hanke, John E. and Wichern, Dean W.
(2007). Business Forecasting. New
Delhi: Pearson Education, Inc, 79-80.

http//www.cmie.com (CMIE Prowess Mills, T. C. (1990). Non-linear Time

database release 3.1, Date of Series Models in Economics, Journal of
Downloading: May 15, 2009). Economic Surveys, 5: 215241.
Kate A., Simth, Jatinder, N. D, Gupta Minsky, M. (1961). Steps Towards AI,
(2000). Neural Networks in Business: Proceeding of the IRE, 49: 5-30.
Techniques and Applications for the
Operations Researcher, Computer and Merh, Nitin, V.P. Saxena and Kamal Raj
Operations Research, 27: 1023-1044. Pardasani (2008), Artificial Neural
Network for Stock Market Forecasting,
Kim, Kyoung-jae and Lee, Won Boo (2004). Nirma University Journal of Business
Stock Market prediction using Artificial and Management Studies (NUJBMS),
Neural Networks with Optimal Feature 2(3 & 4): 3-19.
Transformation, Neural Computing and
Application, 13: 255260. Murphy, J. J. (1999). Technical Analysis of
the Financial Markets: A Comprehensive
Koch, C., (1999), Biophysics of Guide to Trading Methods and
Computation: Information Processing in Applications. New York: Institute of
Single Neuron, Oxford university Press, Finance, 4-5.
New York.
Priestley, M. B. (1988). Non-linear and Non-
Kumar, Satish (2004), Neural Networks: stationary Time Series Analysis. London:
A Classroom Approach, Tata McGraw- Academic Press.
Hill Publishing Company Limited, New
Delhi. Refenes, A. P., (Ed) (1995). Neural Network
in the Capital Markets. England:John
Lawrence, S., Tsoi, A. C., and Giles, C. L. Wiley & Sons Ltd, 149-161.
(1996). Noisy Time Series Prediction
Using Symbolic Representation and Sallehuddin, Roselina., Shamsuddin, Siti
Recurrent Neural Network Grammatical Mariyam. and Hashim, Siti Zaiton Mohd
Inference. Technical report UMIACS- (2008). Hybridization Model of Linear
TR-96-27 and CS-TR-3625. Institute for and Non-linear Time Series Data for
Advanced Computer Studies, University Forecasting. Second Asia International
of Maryland. Conference on Modelling & Simulation,
IEEE Conference Proceedings, DOI
Malkiel, B.G. (1996). A Random Walk 10.1109/AMS.2008.142, 597 602.
Down Wall Street. New York: W. W.
Norton, 6th Edition. Samanta G.P. and Bordoloi, Sanjib (2005).
Predicting Stock Market- An Application
Miao, K., Chen, F., and Zhao, Z. G. (2007). of Artificial Neural Network Technique
Stock Price Forecast Based on Bacterial Through Genetic Algorithm, Finance
Colony RBF Neural Network, Journal of India, 19(1): 173-188.
QingDao University, 20 (2): 5054 (in
Chinese). Sinencio, Edgar Snchez and Lau, Clifford
(1992). Artificial Neural Networks:
Paradigms, Applications and Hardware Yao, J.T., and Poh, H. L. (1996). Equity
Implementations. New York: IEEE Press, Forecasting: A Case Study on the KLSE
16-17. Index. Neural Networks in Financial
Engineering, Proceedings of the Third
Skabar, Andrew and Cloete, Ian (2001). International Conference on Neural
Neural Networks, Financial Trading Networks in the Capital Markets. World
and The Efficient Markets Hypothesis, Scientific, Singapore, 341353.
Australian Computer Society, Inc. ,
Twenty-Fifth Australasian Computer Yudong, Zhang and Lenan, Wu (2009).
Science Conference (ACSC2002), Stock Market Prediction of S&P 500 via
Melbourne, Australia, Conferences in Combination of Improved BCO Approach
Research and Practice in Information and BP Neural Network, Expert Systems
Technology, 4: 241 249. with Applications, 36: 88498854.
Termizel, Taskaya T., and Casey, M.C. Zhang, Defu., Jiang, Qingshan and li, Xin
(2005). A Comparative Study of (2004). Application of Neural Networks
Autoregressive Neural Network Hybrids, in Financial Data Mining, International
Neural Networks, 18( 5-6): 781-789. Journal of Computational Intelligence,
1(2): 116-119.
Tseng, Fang-Mei., Yu, Hsiao-Cheng
and Tzeng, Gwo-Hsiung (2002). Zhang, G. Peter (2003). Time Series
Combining Neural Network Model with Forecasting Using a Hybrid ARIMA and
Seasonal Time Series ARIMA model, Neural Network Model, Neurocomputing,
Technological Forecasting & Social 50: 159 175.
Change, 69(1): 7187.
Zhang, Guoqiang Peter (2001). An
Windrow, B, D Rumelhart, E. and Lehr, M.A Investigation of Neural Networks
(1994). Neural Networks Applications for Linear Time-Series Forecasting,
in Industry, Business and Science, Computers and Operations Research, 28:
Communications of the ACM, 37(3): 93- 1183-1202.
105.

Annexure
Table 11. Comparison between the Actual and Table 12. Comparison between the Actual and
Predicted Values Generated by ANN, ARIMA, Predicted Values Generated by ANN, ARIMA,
Hybrid ANN_ARIMA and Hybrid ARIMA_ANN Hybrid ANN_ARIMA and Hybrid ARIMA_ANN
Models for BSE 30 (SENSEX). Models for BSE IT.
BSE 30 (SENSEX)
Predicted Close Value of BSE 30 Predicted Close Value of BSE IT

Actual Values of
Actual Values of
Using Following Model Using Following Models
BSE IT
Date
Date
ARIMA_
ARIMA_
Hybrid
Hybrid
Hybrid
Hybrid
ARIMA
ARIMA
ARIMA
ARIMA
ANN_
ANN_
ANN
ANN
ANN
ANN
26-FEB-2009 8954.90 9001.30 8875.57 8954.90 8940.26 26-FEB-2009 2110.70 2092.10 2082.62 2110.70 2132.37
27-FEB-2009 8891.60 9070.80 8908.93 8872.95 9003.59 27-FEB-2009 2096.20 2107.70 2094.06 2098.99 2136.73
02-MAR-2009 8607.10 9033.20 8875.15 8535.06 8873.20 02-MAR-2009 2057.60 2090.00 2088.89 2055.41 2099.85
03-MAR-2009 8427.30 8653.80 8558.21 8256.00 8305.54 03-MAR-2009 2019.20 2063.80 2062.57 2014.56 2046.14
04-MAR-2009 8446.50 8403.70 8441.90 8355.44 8342.52 04-MAR-2009 2032.80 2042.70 2028.77 2026.67 2000.29
05-MAR-2009 8197.90 8388.60 8338.91 8215.11 8358.90 05-MAR-2009 2011.00 2051.10 2023.89 2010.42 2042.54
06-MAR-2009 8325.80 8199.10 8304.31 8249.14 8178.69 06-MAR-2009 2072.30 2035.40 2039.78 2066.23 2041.72
09-MAR-2009 8160.40 8128.70 8270.04 8211.33 8306.92 09-MAR-2009 2025.30 2073.00 2050.24 2031.89 2084.09
12-MAR-2009 8343.80 8095.30 8228.51 8356.54 8134.24 12-MAR-2009 2073.50 2046.40 2045.77 2065.17 2035.71
13-MAR-2009 8756.60 8217.80 8415.51 8856.50 8546.12 13-MAR-2009 2190.30 2079.40 2090.56 2195.74 2122.25
16-MAR-2009 8943.50 8663.90 8769.14 9160.10 9125.61 16-MAR-2009 2199.20 2186.20 2166.94 2215.38 2281.13
17-MAR-2009 8863.80 9019.10 8938.47 8976.20 9128.24 17-MAR-2009 2161.30 2197.30 2181.95 2160.81 2228.77
18-MAR-2009 8976.70 8960.20 8922.82 8914.27 8863.54 18-MAR-2009 2184.80 2154.40 2176.88 2179.35 2165.62
19-MAR-2009 9001.70 9060.60 8926.35 9008.33 8995.96 19-MAR-2009 2219.20 2181.80 2191.26 2224.70 2213.85
20-MAR-2009 8966.70 9145.20 9018.28 8943.02 9109.10 20-MAR-2009 2217.40 2226.60 2209.45 2222.30 2252.05
23-MAR-2009 9424.00 9103.50 9070.86 9352.24 9034.63 23-MAR-2009 2271.60 2227.70 2232.91 2269.44 2255.47
24-MAR-2009 9471.00 9534.10 9420.80 9599.84 9789.33 24-MAR-2009 2263.00 2287.00 2250.89 2270.07 2300.15
25-MAR-2009 9667.90 9619.70 9536.24 9642.53 9601.84 25-MAR-2009 2286.00 2257.30 2280.31 2281.23 2307.30
26-MAR-2009 10003.10 9832.50 9742.85 10022.48 9889.88 26-MAR-2009 2361.20 2303.70 2297.32 2366.34 2317.39
27-MAR-2009 10048.50 10116.90 9963.24 10117.08 10238.85 27-MAR-2009 2337.80 2373.80 2325.16 2345.84 2403.54
30-MAR-2009 9568.10 10202.60 9891.09 9540.60 9991.71 30-MAR-2009 2234.30 2344.70 2314.72 2227.54 2342.29
31-MAR-2009 9708.50 9748.20 9663.57 9453.42 9356.06 31-MAR-2009 2285.70 2221.10 2270.98 2269.79 2205.26
01-APR-2009 9902.00 9870.90 9686.73 9886.04 9791.43 01-APR-2009 2358.10 2300.70 2290.72 2370.45 2319.84
02-APR-2009 10348.80 10103.00 10102.59 10361.30 10333.22 02-APR-2009 2454.70 2364.60 2389.35 2461.61 2470.73
06-APR-2009 10534.90 10369.70 10329.97 10633.71 10591.58 06-APR-2009 2454.70 2450.60 2416.18 2467.47 2482.51
08-APR-2009 10742.30 10513.10 10614.78 10808.71 10835.05 08-APR-2009 2498.40 2436.00 2474.98 2497.07 2516.52
09-APR-2009 10803.90 10612.20 10668.07 10896.04 10811.00 09-APR-2009 2499.30 2470.40 2469.76 2509.08 2508.05
13-APR-2009 10967.20 10669.70 10907.09 11044.27 11058.25 13-APR-2009 2487.00 2469.20 2488.09 2489.93 2532.51
15-APR-2009 11284.70 10728.30 11063.38 11404.30 11138.88 15-APR-2009 2444.60 2453.50 2486.54 2446.88 2500.39
16-APR-2009 10947.40 10825.60 11058.48 11171.08 11295.20 16-APR-2009 2391.80 2435.00 2450.64 2390.08 2423.57
Table 13. Comparison between the Actual and Table 14. Comparison between the Actual and
Predicted Values Generated by ANN and Hybrid Predicted Values Generated by ANN, ARIMA,
ANN_ARIMA Models for BSE Oil & Gas. Hybrid ANN_ARIMA and Hybrid ARIMA_ANN
Models for BSE 100.
Predicted Close Value of BSE

Actual Values of
Predicted Close Value of BSE 100
Actual Values of
BSE Oil & Gas
Oil & Gas Using Following

Using Following Models
BSE 100
Models
Date
Date
ARIMA_
Hybrid
Hybrid
ARIMA
ARIMA
ANN_
ANN
ANN
ANN Hybrid ANN_ARIMA
26-FEB-2009 6165.30 6053.20 6165.30 26-FEB-2009 4537.10 4390.20 4513.22 4537.10 4547.03
27-FEB-2009 6064.10 6137.10 6123.74 27-FEB-2009 4516.40 4385.90 4519.83 4612.21 4545.90
02-MAR-2009 5886.00 5979.70 5896.71 02-MAR-2009 4377.20 4374.40 4506.72 4462.31 4505.59
03-MAR-2009 5775.60 5867.90 5756.51 03-MAR-2009 4293.50 4297.20 4355.14 4295.33 4244.34
04-MAR-2009 5842.20 5781.00 5805.64 04-MAR-2009 4311.00 4252.50 4290.44 4308.59 4243.66
05-MAR-2009 5625.30 5842.50 5611.64 05-MAR-2009 4193.50 4244.80 4265.99 4231.65 4289.49
06-MAR-2009 5724.10 5694.90 5667.44 06-MAR-2009 4242.20 4222.50 4240.53 4208.74 4179.18
09-MAR-2009 5633.90 5806.60 5596.39 09-MAR-2009 4160.40 4198.80 4209.41 4173.25 4212.41
12-MAR-2009 5814.60 5719.40 5750.73 12-MAR-2009 4246.80 4191.40 4182.86 4221.76 4150.97
13-MAR-2009 6088.50 5846.80 6058.02 13-MAR-2009 4437.00 4215.80 4289.33 4473.13 4354.95
16-MAR-2009 6298.30 6096.20 6322.54 16-MAR-2009 4541.40 4292.60 4455.78 4685.67 4605.23
17-MAR-2009 6206.60 6323.90 6265.31 17-MAR-2009 4504.60 4368.60 4539.99 4666.87 4626.68
18-MAR-2009 6308.50 6133.80 6331.08 18-MAR-2009 4562.40 4385.00 4529.72 4651.10 4505.22
19-MAR-2009 6355.20 6318.10 6407.08 19-MAR-2009 4575.40 4426.40 4535.12 4691.10 4570.97
20-MAR-2009 6368.00 6313.40 6415.65 20-MAR-2009 4558.60 4434.60 4594.27 4655.78 4637.39
23-MAR-2009 6775.40 6353.10 6823.19 23-MAR-2009 4775.60 4410.30 4600.63 4856.47 4569.78
24-MAR-2009 6772.80 6848.90 6892.79 24-MAR-2009 4784.80 4580.40 4779.09 5023.05 4955.80
25-MAR-2009 7020.90 6751.70 7099.39 25-MAR-2009 4886.00 4703.90 4818.57 5019.31 4830.35
26-MAR-2009 7190.70 6982.90 7304.55 26-MAR-2009 5036.40 4749.40 4921.18 5155.17 4989.71
27-MAR-2009 7181.80 7049.10 7311.15 27-MAR-2009 5091.60 4969.10 5030.01 5278.78 5147.70
30-MAR-2009 6992.40 7087.20 7119.09 30-MAR-2009 4861.00 5065.70 5030.27 4940.90 5094.32
31-MAR-2009 7053.00 6919.90 7133.05 31-MAR-2009 4942.50 4841.20 4891.13 4808.99 4740.67
01-APR-2009 7256.00 6977.40 7344.40 01-APR-2009 5028.40 4859.90 4942.23 5094.47 4996.14
02-APR-2009 7671.80 7074.20 7795.20 02-APR-2009 5253.10 4953.00 5117.91 5363.00 5206.57
06-APR-2009 7747.60 7609.40 7960.75 06-APR-2009 5348.20 5243.00 5243.84 5543.93 5379.26
08-APR-2009 7998.80 7757.70 8191.81 08-APR-2009 5467.30 5369.50 5436.97 5535.91 5542.77
09-APR-2009 7978.10 7672.40 8175.60 09-APR-2009 5500.80 5381.90 5385.92 5564.59 5418.97
13-APR-2009 8083.20 7780.20 8296.84 13-APR-2009 5593.10 5484.30 5575.56 5670.65 5692.91
15-APR-2009 8273.40 7787.80 8499.08 15-APR-2009 5780.00 5525.50 5658.77 5850.96 5681.28
16-APR-2009 7901.40 7795.90 8172.23 16-APR-2009 5588.10 5566.30 5644.33 5754.09 5767.80

Table 15. Comparison between the Actual and

Predicted Values Generated by ANN, ARIMA,
Hybrid ANN_ARIMA and Hybrid ARIMA_ANN
Models for S&P CNX Nifty.
Predicted Close Value of S&P CNX

Actual Values of
S&P CNX Nifty
Nifty Using Following Models

Date
ARIMA_
Hybrid
Hybrid
ARIMA
ARIMA
ANN_
ANN
ANN
26-FEB-2009 2785.70 2750.90 2763.00 2785.70 2755.55
27-FEB-2009 2763.70 2763.60 2750.75 2798.50 2775.04
02-MAR-2009 2674.60 2753.90 2732.20 2684.40 2753.26
03-MAR-2009 2622.40 2715.00 2672.34 2577.57 2624.76
04-MAR-2009 2645.20 2681.50 2649.49 2585.52 2608.59
05-MAR-2009 2576.70 2677.90 2619.97 2545.20 2623.93
06-MAR-2009 2620.20 2665.70 2591.73 2550.42 2558.31
09-MAR-2009 2573.20 2662.30 2595.33 2530.72 2624.30
12-MAR-2009 2617.40 2660.30 2619.95 2550.34 2607.91
13-MAR-2009 2719.30 2666.10 2663.82 2674.98 2664.47
16-MAR-2009 2777.30 2698.20 2724.20 2785.15 2781.09
17-MAR-2009 2757.50 2745.60 2759.13 2783.97 2812.85
18-MAR-2009 2794.70 2759.20 2787.89 2790.69 2795.22
19-MAR-2009 2807.20 2775.90 2776.86 2817.22 2792.45
20-MAR-2009 2807.10 2789.70 2787.23 2817.32 2823.17
23-MAR-2009 2939.90 2791.10 2890.55 2944.78 2914.73
24-MAR-2009 2938.70 2881.10 2959.90 3013.36 3010.07
25-MAR-2009 2984.40 2958.10 2932.74 3018.16 2921.73
26-MAR-2009 3082.30 2984.10 3036.67 3101.88 3088.70
27-MAR-2009 3108.60 3087.60 3067.77 3166.06 3119.52
30-MAR-2009 2978.20 3147.50 3024.81 2999.19 3066.33
31-MAR-2009 3021.00 3061.20 3018.87 2943.59 2982.25
01-APR-2009 3060.40 3041.10 3011.76 3041.89 3020.31
02-APR-2009 3211.10 3074.30 3170.05 3221.46 3221.02
06-APR-2009 3256.60 3203.60 3249.85 3328.29 3291.46
08-APR-2009 3342.90 3276.90 3237.04 3376.84 3252.16
09-APR-2009 3342.10 3296.40 3365.83 3384.32 3472.01
13-APR-2009 3382.60 3324.20 3378.66 3416.34 3365.20
15-APR-2009 3484.20 3332.80 3406.00 3525.46 3412.40
16-APR-2009 3369.50 3348.10 3432.62 3460.00 3512.79

2010 Mohsen Dastgir, Vali Khodadadi and Maryam Ghayed 45
CASH FLOWS VALUATION USING CAPITAL CASH

FLOW METHOD COMPARING IT WITH FREE CASH
FLOW METHOD AND ADJUSTED PRESENT VALUE
METHOD IN COMPANIES LISTED ON TEHRAN
STOCK EXCHANGE
Mohsen Dastgir, Vali Khodadadi and Maryam Ghayed
Abstract
One firm valuation method is to use discounted cash flow. In this paper the valuation method of Capital
Cash Flow discounted at the Weighted Average Cost of Capital (WACC) before tax is represented and, as
a proof to its efficiency in Iran market, it is compared with two common methods, i.e. Free Cash Flow
discounted at the weighted average cost of capital after tax and the adjusted present value. For the same
purpose, 54 firms from among those companies listed on Tehran Stock Exchange were selected as sample
and their financial information for three-year financial period from 2004 to 2006 were collected and
analyzed by paired Students t-test. Research results showed that using an appropriate discount rate will
make the value calculated by Capital Cash Flow method become twice as much when using two methods
mentioned above.
Dastgir M., Khodadadi V., Ghayed M. - Cash Flows Valuation Using Capital Cash Flow Method Comparing it with Free Cash Flow Method and
Adjusted Present Value Method in Companies Listed on Tehran Stock Exchange
Introduction all of which is the using of tax shield in

calculations.
Historically, human beings face with One of the latest methods cited in this area
restrictions for using resources. Such is the Capital Cash Flow (CCF) method. This
restrictions always make them seeking the research aims at the representation of Capital
best way to use these resources with highest Cash Flow method and demonstration of its
return and least cost. The manifestation effectiveness in Iran market through making
of this in the domain of financial issues comparison with Free Cash Flow and
was discussions cited about financing Adjusted present value methods.
decisions and investment management.
Investors, analyzers and other users of Literature Review
financial information attempt to achieve
the best investment opportunity relating Useful information builds the foundation
their accessible resources. Firms valuation of decisions made by people participating
is a tool which can contribute them in this in capital market. Compilers of accounting
respect. The obvious thing is that the higher standards attempt to provide financial
the firms value is the higher return will be reporting and accounting system of capital
gained by holders of capital. market data requirements. Therefore,
Due to importance of this issue, to date analysis, investigation and, finally, using
numerous researches have been conducted suitable methods resulting in fair and correct
thereon and various methods have been valuation, can lead to optimum allocation of
introduced for the development and capital resources and selection of investment
supplementation of firms valuation which opportunities.
attempt not to include the weak points of Numerous efforts have been taken by
previous methods and rather to improve researches and practitioners in the field
their strong points. Thus, many researchers of various methods of firms valuation a
and practitioners in accounting, particularly compendium of which will be discussed
in recent decades, have involved in this below:
domain and, also, in finding more applied
and modern methods. Comparison of Various Firms
Firms valuation with using cash flow Valuation Methods
has been allocated a special position in
modern scientific accounting discussion. In a market-based valuation, Tham (2000)
This is confirmed by researches efforts discounted Free Cash Flow at the weighted
and presentation of new models in cash average cost of capital (WACC) and the
flows valuation aiming at supplementation effect of financing is taken into account by
and improvement of previous models and adjusting the WACC.
convenience in application. He specified the following conditions for
In new discussions of firms valuation, firms valuation:
taking into consideration of the value and a. Multi-period investments and rein-
position of tax shield in calculating method, vestments,
especially in those countries with efficient
debt markets, is of particular importance. b. Finite cash flows, with variable
This, in turn, has made a range of firms growth rates,
valuation methods the advert point of

c. Variable debt-equity ratios, and discounted at this rate. The obtained

d. Losses carried forward. result shows that Capital Cash Flow is equal
to Free Cash Flow plus tax shield.
With the inclusion of these conditions, the In section two, the same assumptions
analysis is more realistic. In this model, the were reviewed for risky debt and tax shield
present value of the tax shield is discounted at and in a dichotomous model for a period the
the required return with all-equity financing. formulas relating to the discount rate of cash
Using this method, the impacts of inflation flows were concluded.
are directly incorporated into the analysis. During another survey, these two
In addition, the model shows that the NPV researchers used Free Cash Flow, Capital
of the Free Cash Flow, discounted at the Cash Flow and Adjusted present value
WACC, is equal to the NPV of the Free Cash methods for the valuation of levered firms
Flow to the equity holder, discounted at the and tried to show that with using Miles and
annually adjusted return to equity. Ezells model in calculation of tax shield,
The general approach applied by Tham the three methods above will give similar
can be easily modified to take into account responses. In Free Cash Flow method, the
the varied circumstances and complexities tax shield is obtained by using discount
which are often encountered. rate of the weighted average cost of capital.
Tham and X. Wonder (2002) argue that Capital Cash Flow method might directly
traditional formulas of WACC calculation add tax shield to the Free Cash Flow and
would assume both debt and tax shield as the adjusted present value method would
risk-free. But even when the debt is risk-free, calculate it separately.
the tax shield can be risky. In addition, debt In section one the assumptions were
and tax shield can be both risky as well. In cited concerning three methods. In next
their survey, these two researches presented section, with using dichotomous model for
a non-conventional new WACC for a period a five-year period the value of non-levered
with risky debt and tax shield and concluded firms was calculated by those three methods
formulas relevant to return to equity and mentioned above. And, in the final section
debt. In contrast to preliminary formulas, in of research, the value of tax shield was
method introduced for calculating WACC, introduced in risk-free condition and, then,
discount rate of tax shield is not limited to the value of levered firms was calculated
risk-free rate and rate of return on capital. through those three methods.
They reviewed two conditions in their Research results indicate that the value
survey: of levered firms increasingly depend upon
the assumption relating to the discount
1. Risk-free tax shield and debt rate of tax shield. With using dichotomous
(conventional conditions) model for the calculation of Free Cash
Flow and using Milles & Ezells theory in
2. Risky tax shield and debt tax shield calculation, it was shown that all
three methods will give similar responses to
In section one of research a new formula levered firms valuation.
is presented for calculating WACC and Free Cooper & Nyborg (2006) propose the
Cash Flow and Capital Cash Flow in first following four methods for firms valuation:
condition; i.e. where the debt and the tax
shield which are both risk-free are calculated
1. Discounting operating Free Cash Flow tax shield which is very far from its correct
at the weighted average cost of capital value.
In a survey, Fernandez took measures
2. Discounting equity Free Cash Flow at for firms valuation through four methods
the cost of equity of: Discounting operating Free Cash Flow
at the weighted average cost of capital;
3. Adjusted present value discounting equity Free Cash Flow at the
cost of equity; discounting the Capital
4. Discounting the Capital Cash Flow at Cash Flow at the unlevered cost of capital;
the unlevered cost of capital Adjusted present value.
The research results indicated that firms
All above methods are based on a similar valuation with using above four methods
theory: levered firms value is equal to will give similar responses. This conclusion
unlevered firms value plus present value of is logic as all of these methods analyze
tax shield arising from financing due to debt. a similar fact with similar assumptions.
However, none of above methods can be These methods difference in institutions
defined without discount rate for valuation. valuations is resulted from difference in tax
These discount rates depend upon firms shield calculation. Hence, their difference
leverage policy. The firms leverage policy in term of various methods of tax shield
determines that what discount rate should calculation is presented below:
be applied for the proper estimation of tax
shield value and, consequently, firms value. Value of Tax Shield (VTS)
Therefore, all valuation methods should
start from a clear assumption about leverage Valuable researches have been conducted
policy of firm. about cash flows valuation. Discrepancy
In this survey, different hypotheses among these methods and different theories
about levered firms policy and that how on firms valuation using discounted cash
they are affected by following factors are flows is arising from difference in the
investigated: calculation of tax shield. Interest paid on
debt is a cost subtracted from profit but no
1. Applicable discount rate tax is deducted for it, while tax is received
on dividend or the accumulated profit
2. Present value of tax shield dependent upon tax share. Thus, in presence
of debt, total payments to the debt holders
3. Applicability of above methods and shareholders will be higher.
In this paper, different theories are
Then, different valuation methods and presented for calculating present value of tax
the way to select from among them are shield. Upon analysis of the results obtained
explained. Lastly, through model presented from these theories it will be proved that
by Booth (2002) and Fernandez (2004) it VTS is not the present value of the tax shield
was shown that how using wrong methods discounted at a certain rate, rather it is the
will lead to mistakes which despite of their difference between two present values:
being small, they will have major impacts present value of taxed paid in the unlevered
on calculations. Using wrong formulas can company minus the present value of the
result in an estimation of the present value of taxes paid in the levered company. The

taxes of the unlevered companies have a risk (3) Myers [1974]

lower than that of the levered companies.
In following section, some of the major Myers introduced the Adjusted Present
researches and theories taken in this respect Value (APV) method. According to Myers,
are presented: the value of the levered firms is equal to the
value of the firm with no interest (Vu) plus
(1) Gordon & Shapiro [1956] the present value of the tax shield. On the
same basis, Mayers presented the following
The model developed by Gordon & formula:
Shapiro suggests that companys profit will
grow in future by means of new investments. VTS = PV 6 Kd; TDKd @
This growth rate will be varied in predictable
periods and it will be constant and close to He believes that the risk of tax saving
economic growth rate for periods onwards. due to debt is equal to the risk of the debt.
In 1997, Luehrman also recommended that
(2) Modigliani and Miller [1963] firms may be evaluated with using APV
method and they can calculate tax shield
They studied the effect of leverage on based on Mayers method where the firms
the firms value. The results of their study value is as the same as:
indicated that in the absence of taxes, the
firms value is independent of its debts. PV = D + E = Vu + VTS
But, in the presence of taxes, required = PV6 Ku; FCF @ + 6 Kd; TDKd @
return to equity will vary in proportion to the
debt/capital ratio: (4)Benninga and Sarig [1997]
Ke = Ku + ( DE ) (1 - T) (Ku - Kd) Benninga and Sarig claimed that in
E0 += Vu + DTD0 the presence of personal taxes (PT) in
calculations, Tax advantages of debt should
Where DT is the value of tax shield for be discounted at discount rate after deduction
the case of perpetuities. of personal taxes. According to their theory:
The goal of Modigliani and Miller was to
show the impact of tax on the value of debt. VTS = PV 6 Kd ^1 - TPDh;
In 1963, they used the following formulas DK 6^1 - TPDh^1 - T h^1 - TPAh@ @
for firms valuation:
(5) Arditti and Levy [1977]
@ WACC = Ku 6 1E-+TDD
WACCBT = Ku - DT ; E
^ Ku - Kd h In their research results, Arditti and Levy
^ E + Dh
explained that firms value may be calculated
by Capital Cash Flow instead of Free Cash
In the last presented equation, Modigliani Flow. And, in this respect, it is necessary
and Miller used book value instead of market to discount the Capital Cash Flow at the
value for calculating debt/total debt ratio weighted average cost of capital before tax
and shareholders equity which is obviously (WACCBT). In 1977, they had a substantial
wrong.
problem
D
in their
E
paper: for the calculation of reduce their taxes through other tools (e.g.
and E + D they used book value of depreciation).
t-1 t-1
E +D
t-1 t-1 t-1 t-1
the debt and shareholders equity instead of

market value. (9) Miles & Ezell [1980]
(6) Miller [1977] According to Miles and Ezell, D

a firm
trying to have a constant ratio of , should
E
Miller argues that the optimum debt not use a valuation method similar to that
structure of collective companies indicates used by a firm with default debt.
the existence of such a structure for each In this respect, in firms with target debt
of them individually. Thereafter, Miller ratio the Free Cash Flow will be discounted
introduces personal income tax and at the following rate:
corporate income tax. According to Miller,
WACC = Ku - 8
D+EB
; E
D Kd T (1 + Ku)
firms value when no debt is assumed is 1 + Kd
equal to:
Et - 1 + Dt - 1 = FCFt + Kd TDt - 1
Ku - g Ku - g
Vu = FCF (1 - TPA)
Ku
(10) Miles & Ezell [1985]
Then he adds that firms attempts
made for making increase in their debts is The following formula shows the
inconsistent with market balance. Increase relationship between the levered beta (L)
in debts causes changes in rate of return with the asset beta (u) (assuming a risk-
on debt and rate of return to equity and, free debt and a debt beta of zero):
thus, firms value under such conditions is
independent of rate of debts. bL = bU + DbU ;1 - TRF E /E
1 + RF
(7) Miller & Scholes [1978] (11) Chambers, Harris & Pringle
[1982]
When the rate of income tax is higher
than rate of income on capital profit, most This group of researchers compared
investors will have to pay taxes higher than four valuation methods for discounted cash
their receivable dividend. Miller and Scholes flows: equity cash flow discounted at the
conclude that if corporate repurchases its required return to levered equity (Ke), Free
shares, it will make no preference between Cash Flow discounted at WACC, Capital
dividend and realized gain on capital. Cash Flow discounted at WACC before tax
According to these researchers, firm value is and Adjusted present value methods. They
independent of the policy of firms dividend argue that in case of target debt the first three
payment. methods will give similar results. But, when
there is no target debt, these methods will
(8) De Angelo and Masulis [1980] result in different values. Only the Adjusted
present value will give the same results as
De Angelo and Masulis extended Millers other three methods under both conditions.
study. With considering that the tax final Of course, firms are simply analyzed for one
rate varies for different firms, they predict financial period. The reason of such result
that firms, instead of using debt, try to for their researches was a mistake: they had

used book valueD

instead of market value for (16) Taggart [1991]
the ratio of E + D .
In his researches, Taggart gives a
(12) Harris and Pringle [1985] summary of all valuation methods with or
without concerning personal income tax.
These two researchers suggest that He suggests to use Miles & Ezzells when
the present value of the tax shield should the company adjusts to its target debt ratio
be discounted at the required return to once a year and Harris & Pringles approach
unlevered equity (Ku). when the company continuously adjusts to
its target debt ratio.
VTS = PV (Ku; DKd T)
(17) Damodaran [1991]
Also in their calculations they consider
that WACC before tax is equal to the required If all the business risk is borne by the
return to equity. Thus, in their opinion: corporate, then the formula relating the
levered beta (L) with the asset beta (u)
WACC = Ku - DKd T/ (D + E) will be as follows:
(13) Ruback [1995] bL = bU + ` D j bU (1 - T)

E
Ruback uses the following formula to According to Damodaran, identification

give of corporate. of the beta of the debt under situations where
it has been dropped is not the same as when
bL = bU 8 D + E B - bD 8 D B the beta of the debt is assumed zero.
E E
When the beta of the debt is zero, the
(14) Tham & Valez-Pareja [2001] required return to debt should be the risk-
free rate. The purpose of dropping of the
Following an arbitrage argument, Tham beta of the debt is to obtain a higher levered
& Valez-Pareja believe that the appropriate beta.
discount rate for tax shields is Ku, the Another formula is presented below to
required return to unlevered equity. Of show the relationship between levered beta
course, later it was shown that this approach (L) and the asset beta (u):
also comes to mistakes.
bL = bU + ` D + E j
E
(15) Lewellen and Emery [1986]
We call this method the Practitioners
In general, Lewellen and Emery believed Method. This method is used by financial
that Miles & Ezzells is the most logically consultants and investment institutions.
presented method until that time. According to this mehtod, given the same
But, in method introduced by Modigliani value for u, a higher L is obtained
& Miler the tax shield was equal to: than according to Fernandez (2004) and
according to Damodaran (1994).
VTS = PV 6 Ku; DTKu @
(18) Inselbag and Kaufold [1997] And, this value will be lower when levered
costs exist.
They believe that in a case where
monetary value of debt is fixed, the value Research Design
of tax shield (VTS) may be given by Myers
method. If the firm has a constant debt/value The objective of this research is to
ratio, the value of tax shield (VTS) will be introduce Capital Cash Flow and to compare
calculated through Miles and Ezells. it with two methods of Free Cash Flow and
According to these researchers, present Adjusted present value. Therefore, research
value of tax shield of firms planning based variables are presented as follows:
on amount of nominal debt is higher than
firms emphasizing on debt ratio. This theory Studied Variables
cannot be accepted for two reasons: firstly,
no firm has output operating cash flow for Three studied variables including Capital
having target debt ratio (instead of a target Cash Flow, Free Cash Flow and adjusted
debt outstanding). And, secondly, as we present value are introduced as follows:
know, the tax shield is the difference between
two present values of taxes in levered and Capital Cash Flow
unlevered firms.
Inselbag and Kaufold argue that the risk of Capital Cash Flow includes all cash flows
target debt ratio is higher than that of having paid or payable to investors. In this method,
target debt outstanding. If so, the present Capital Cash Flows are the cash flows
value of taxes paid by levered firms should available for all holders of the companys
be higher than that of firms with target debt securities equivalent to the equity cash flow
and, in consequence, the present value of tax after deduction of companys assets tax.
shield for the second-group firms should be Capital Cash Flow is calculated as
lower and this is opposed to the theory cited follows:
by above-mentioned researchers. Capital Cash Flow = Net Income +
Depreciation - Capital expenditures ! 9
(19) Copeland, Koller and Murrin
Working Capital + Interest
[2000]
Net profit contains tax savings due to
The studies of this group of researches interest expense of debt. Then, the impact of
confirmed the theories of Harris & Pringle cash flow adjustments including depreciation
(1985) and Myers (1974) relating to present expense, capital expenditures and capital
value of tax shield and they concluded that turnover are taken into account. At last,
we leave it to the readers judgment to upon addition of cash interest, the Capital
decide which approach best fits his or her Cash Flow is obtained which indicates the
situation. after-tax cash flow for investors. For the
same reason, it is discounted at the before-
(20) Fernandez [2001] tax weighted average cost of capital which
is calculated as follows:
This researcher believes that where the
levered costs are zero, the value of tax shield WACCAT = E Ke + D Kd
V V
is equal to tax rate multiplied by debt value.

Where D/V is the debt-to-value ratio; E/V present value. This method was introduced
is the equity-to-value ratio, and KD and KE by Mayers and separates firms value into
are the respective expected debt and equity two parts: unlevered operating cash flows
returns. and cash flows depending upon project
On the other hand, capital assets financing.
pricing model (CAPM) should be used for Where:
calculating the required returns.
APV = / tn= 0 FCF + / tn= 0 Int. t (Tc) t
1 + Kg (1 - Kd)
Ke = RF + be RP
Int. t = expense interest of debt in time t
Ke = RF + be RP
Tc = effective income tax rate
Where Rf is the risk-free rate, Rp is the Other components of this equation were
risk premium, Kd and Ke are the debt and represented before.
equity betas, respectively. In this model, firms value is initially
calculated with using firms capital expense
Free Cash Flow in the absence of debt and, then, the present
value of financing tax savings due to debt
The basic assumption in determination is added to it. The reason of this separation
of firm value with using Free Cash Flow is is to allow using different discount rates
that a trading units value is resulted from depending upon risk rate for two parts.
its power in making the operating cash flow
and other cash flows due to investment. This Methodology of Research
is calculated as follows:
Considering the structure of hypotheses
Free Cash Flow = earnings before interest
and the method used for data finding and
and taxes + estimated taxes + collection within the finite time of research,
cash adjustments statistical population was selected for three
consecutive years from 2004 to 2006 from
Since tax shield arising from interest among companies listed on Tehran Stock
expense of debt has not been taken into Exchange.
account in the calculation of Free Cash Cochrans (1977) sample size formula1
Flow, it is necessary to use the weighted was used for determining sample size
average cost of capital after tax deduction. of firms for stratified random sampling.
Where: Finally, 54 firms were selected as sample
within three years of testing course.
WACCAT = E Ke + D Kd (1 - T) In next step, all data and information
V V
needed for calculating firm value by three
T= effective tax rate represented methods were extracted from
Other components of this equation were financial information existing in stock
represented before. exchange, sites and other resources.
Adjusted Present Value Method 1

NZ + Pq
n=
Ne2 + Z2 Pq
And, finally, the third method having
been presented in this survey is the adjusted
With considering the structure of survey between two methods of Capital Cash
hypotheses and the collected data, paired Flow and Free Cash Flow. Therefore, in
Students t-test was applied for data analysis. general:
Hypothesis H0 is accepted. That is, -level
Research Findings of 5%, using Capital Cash Flow method in
the cash flows valuations would result in
Researchs First Hypothesis Testing similar results of Free Cash Flow.
Results of Students t-test by SPSS Researchs Second Hypothesis

statistics software concerning the first Testing
testable hypothesis are formulated in the
following tables: Results of Students t-test by SPSS
statistics software concerning the second
Table 1. Paired samples Test for First Hypothesis testable hypothesis are formulated in the
following tables:
Paired Differences
95% Table 3. Paired samples Test for Second Hypothesis
Confidence
Interval of
Std. Error Mean
Paired Differences
Std. Deviation
the Difference
Sig. (2-tailed)
Variables
95%
Dower
Upper
Mean
Confidence
Df
t
Interval of
Std. Error Mean

CCF & FCF 0.016 0.15 0.2 -0.39 0.42 0.82 53 0.935 Std. Deviation the Difference
Sig. (2-tailed)
Variables
Dower
Upper
Mean
Table2. Paired Samples Correlation
Df
t
CCF & APV 0.76 0.89 0.12 0.032 0.16 -0.63 53 0.530
Correlation N Variables
0.99 54 CCF & FCF Table 4. Paired Samples Correlation
Results show that there is a significant Variables N Correlation
relationship up to level sig=0.935 between CCF & APV 54 0.996

two variables of Capital Cash Flow and
Free Cash Flow which is higher than level Results show that there is a significant
of 5%. On the other hand, the correlation relationship up to level sig=0.530 between
between two variables is 0.99 indicating that two variables of Capital Cash Flow and
two methods above are fully convergent and Adjusted present value which is higher
when one of them is increased the other will than level of 5%. On the other hand, the
certainly increase and, also, with a decreased correlation between two variables is 0.996
Capital Cash Flow the Free Cash Flow will indicating that two methods above are
also decrease. Also, as the upper and lower fully convergent and when one of them is
limit at 95% confidence level are 0.042 and increased the other will certainly increase
0.039, respectively, and the logarithmic and, also, with a decreased Capital Cash
difference of research variables is almost Flow the Adjusted present value will also
close to zero it may accept that hypothesis decrease. Also, as the upper and lower limit
stating no difference at 95% confidence level are 0.016 and 0.032,
respectively, and the logarithmic difference

of research variables is almost close to zero challenging aspects to be subject of further

it may accept that hypothesis stating no researches.
difference between two methods of Capital There are numerous methods and models
Cash Flow and Adjusted present value. concerning firms valuation with using
Therefore, in general: discounted cash flows none of which neither
Hypothesis H0 is accepted. That is, -level introduced in practice in Iran market nor
of 5%, using Capital Cash Flow method in tested in terms of applicability. It is even
the cash flows valuations would result in possible to analyze the efficiency of each
similar results of adjusted present value. of these models upon the market situations
and the existing firms structure in term of
Summary and Conclusion financing method.
In addition to discounted cash flows, there
To date various methods of firms are other methods and models for firms
valuation have been presented by researches valuation the comparison of which can also
and practitioners of financial issues, among be subject of many future researches.
which one can mention using discounted In this survey, some theories about
cash flows. present value of tax shield are cited in brief.
The objective of this survey is to Determination of the value of tax shield in
introduce Capital Cash Flow method for discussions on firms valuation is a very
firms valuation and confirmation of its important issue to form the starting point for
efficiency in Iran market through comparing further researches.
it with two common methods available
in market: Free Cash Flow and Adjusted References
present value. For the same purpose, 54
firms listed on Tehran Stock Exchange were Arditti, F.D. and H. Levy (1977), The
selected as statistical sample. The required Weighted Average Cost of Capital as
information and data were collected through a Cutoff Rate : A Critical Examination
stock exchange and other existing financial of the Classical Textbook Weighted
resources and sites and they were tested by Average,Financial Management
paired Students t-test. (Fall),pp.24-34.
Test results showed that by using
appropriate discount rate and considering Arzac, E.R. and L.R. Glosten (2005), A
the value of tax shield in calculations, the Reconsideration of Tax Shield Valuation,
application of Capital Cash Flow in firms European Financial Management,
valuation would lead to the same results as forthcoming.
those of two above-mentioned methods.
Benninga, S. and O. H. Sarig (1997),
Suggestions and Prospective of Corporate Finance: A Valuation
Future Researches Approach. McGraw Hill.
Firms valuation with using discounted Booth, Laurence, 2007, Capital cash flow,
cash flows is an effective and extensive topic APV & valuation, European financial
in financial decision making and investment management, Vol.13, No.1, pp 29-48
opportunities which, in author opinion, has
Chambers, D.R., R.S. Harris and J.J. Pringle Fernandez, Pablo, 2007, Valuing
(1982), Treatment of Financing Mix in companies by cash flow discounting: ten
Analyzing Investment Opportunities, methods and nine theories, Emerald
Financial Management (Summer), pp. managerial finance, Vol.33, No.11,
24-41. pp853-876
Copeland, T.E., T. Koller and J. Murrin Fernandez, Pablo, 2007, Equivalence

(2000), Valuation: Measuring and of the different discounted cash flow
Managing the Value of Companies. Third valuation methods. Different alternatives
edition. New York: Wiley. for determining the discounted value
of tax shields and their implications for
Cooper, Jan and Nyborg, kjell, 2006, the valuation, IESE business school-
Consistent methods of valuing university of Navarro
companies by DCF: Methods and
assumptions, The paper available on the Gordon, Myron and E. Shapiro (1956),
social science research network (SSRN) Capital Equipment Analysis: The
Required Rate of Profit, Management
Damodaran, A. (1994), Damodaran on Science, 3 (Oct.), pp. 102-110.
Valuation, John Wiley and Sons, New
York. Graham (2000), How Big Are the Tax
Benefits of Debt?, Journal of Finance,
DeAngelo, L. and R. Masulis (1980), Vol. LV, pp. 1901-1941.
Optimal Capital Structure under
Corporate and Personal Taxation, Graham (2001), Taxes and Corporate
Journal of Financial Economics 8, March, Finance: A Review, Working Paper,
pp. 3-29. Duke University.
Fernndez, Pablo (2001), The correct Harris, R.S. and J.J. Pringle (1985), Risk-
value of tax shields. An analysis of 23 Adjusted Discount Rate Extensions
theories, Working Paper No. 276051, form the Average-Risk Case, Journal of
Social Science Research Network. Financial Research (Fall), pp. 237-244.
Fernandez, Pablo (2002), Valuation Inselbag, I. and H. Kaufold (1997), Two

Methods and Shareholder Value Creation. DCF Approaches for Valuing Companies
Academic Press, San Diego, CA. under Alternative Financing Strategies
(and How to Choose Between Them),
Fernandez, Pablo (2004), The Value of Tax Journal of Applied Corporate Finance
Shields is NOT Equal to the Present Value (Spring), pp. 114-122.21
of Tax Shields, Journal of Financial
Economics (July), Vol. 73/1, pp. 145-165. Kaplan, S. and R. Ruback (1995), The
Valuation of Cash Flow Forecast: An
Fernandez, Pablo, 2005, Financial Empirical Analysis, Journal of Finance,
literature about discounted cash flow Vol. 50, No. 4, September.
valuation, IESE business school-
university of Navarro

Lewellen, W.G. and D.R. Emery (1986), Miller, M. and M. Scholes (1978), Dividend
Corporate Debt Management and the and Taxes, Journal of Financial
Value of the Firm, Journal of Financial Economics (Dec.), pp. 333-364.
Quantitative Analysis (December), pp.
415-426. Modigliani, F. and M. Miller (1958), The
Cost of Capital, Corporation Finance and
Luehrman, Timothy A. (1997), Whats It the Theory of Investment, American
Worth: A General Managers Guide to Economic Review, 48, pp. 261-297.
Valuation, and Using APV: A Better
Tool for Valuing Operations, Harvard Modigliani, F. and M. Miller (1963),
Business Review, (May-June), pp. 132- Corporate Income Taxes and the Cost
154 of Capital: A Correction, American
Economic Review (June), pp. 433-443.
Massari, Mario and Roncaglio, Francesco
and Zanetti, Laura, 2007, On the Myers, S.C. (1974), Interactions of
equivalence between APV and the WACC Corporate Financing and Investment
approach in a growing leveraged firms, Decisions Implications for Capital
European financial management, Vol.14, Budgeting, Journal of Finance (March),
No.1, 2007, pp 152-162 pp. 1-25.
Miles, J.A. and J.R. Ezzell (1980), The Ruback, Richard S. (1995), A Note on
Weighted Average Cost of Capital: Capital Cash Flow Valuation, Harvard
Perfect Capital Markets and Project Life: Business School, 9-295-069.
A Clarification, Journal of Financial and
Quantitative Analysis (September), pp. Ruback, Richard S. (2002), Capital
719-730. Cash Flows: A Simple Approach to
Valuing Risky Cash Flows, Financial
Miles, J.A. and J.R. Ezzell (1985), Management, Summer, pp. 85-103.
Reformulating Tax Shield Valuation:
A Note, Journal of Finance, Vol. XL, 5 Taggart, R.A., Jr. (1991), Consistent
(December), pp. 1485-1492. Valuation and Cost of Capital. Expressions
With Corporate and Personal Taxes,
Miller, M.H. (1977), Debt and Taxes, Financial Management (Autumn), pp.
Journal of Finance (May), pp. 261-276. 8-20.
Miller, M. and F. Modigliani (1961), Tham, joseph and Wonder. X, 2002,

Dividend Policy, Growth and the Equivalence between FCF method, the
Valuation of Shares, Journal of Business, CCF method and the APV approach,
34, pp. 411-433. The paper available on the social science
research network (SSRN)
Tham, joseph and Wonder. X, 2002, Non- Tham, joseph, 2000, Practical equity
conventional WACC with risky debt valuation: A simple approach, Wp
and risky tax shields, working paper available on the social science research
available on the social science research network (SSRN)
network (SSRN)

2010 Eduardo Mrquez Pea 59
EFFECTS OF VALUE ADDED TAX ON MEXICAN

ECONOMY A MULTISECTORIAL ANALYSIS
BY DEVELOPING AN APPLIED GENERAL
EQUILIBRIUM MODEL. MEGA
Eduardo Mrquez Pea
Abstract
For this work, we developed a General Equilibrium Model for the Mexican economy where the role of
the indirect tax rate on value added is important. The model is used to analyze the effect of the tax reform
proposed by the President in 2001. Without any doubts, the results obtained are important for discussion
of actual themes in economic politics in Mexico. Particularly, evidence suggests that applying a tax rate
of 15% on food, medicines, educational services and press and editorial products will represent 93,000
million pesos of additional resources for the government, of which 11,079 million pesos could be used to
compensate the first three deciles of the population in order to maintain their original consumption level.
Mrquez-Pea E. - Effects of Value Added Tax on Mexican Economy A multisectorial analysis by developing an Applied General Equilibrium Model. MEGA
1. Introduction Trying to find the best way to model

the Mexican economy implied introducing
The approval of fiscal policies that new formulas and data which were not
foster economic growth is the main focus generated or were not available from current
of discussion and controversy among the literature. For instance, a new idea of how to
different politic actors in Mexico. For represent the investment sector was used;
instance, the Fiscal Reform proposed in it considered that not only it is equal to the
2001 by the President for applying a 15% tax addition of private savings, government
on food, medicines, education services and saving and foreign saving, but also that it is
good from press and editorials was rejected made up by of gross fixed investment and
by the Congress arguing that the policy will inventory variations. Through Tobin system,
affect the poor. But, will it really affect the we obtained that gross fixed investment is
families with lower income? And if so, how proportional to the difference between
much? the return on capital in the economy and
The purpose of this work is to analyze the interest rate of foreign assets. Another
this tax policy and its effects on production, contribution of the model was the generation
private consumption, government, of a private consumption matrix divided by
investment, foreign sector and intermediate population deciles of total income per capita,
demand. The experiment consists on to know the quantities demanded by each
simulating two types of fiscal policies. population decile of the goods provided by
The first one considers taxing at 15% food, the 23 productive sectors modeled.
medicines, education services and good Analysis is based on data for transactions
from press and editorials. The second one, in the Mexican economy for year 2000, so
in order to cover the concern of protecting a Social Accounting Matrix was constructed
individuals with lower income, consists on from the information of the Input Production
taxing the same way these sectors but with Matrix (MIP) for year 2000 calculated with
an increase on the subsidy given to the first the RAS method by Hernandez Laos (2004).
three populations deciles through wages. At the same time, the information of the MIP
The structuring of the work is through was not enough to model the foreign sector
the construction of a General Equilibrium and private consumption. It was necessary
Model (MEGA) whose applications let us to statistically manipulate the data from
know the adjustments made to relative prices the National Survey of Home Income and
and quantities as a result of any change Spending 2000 (ENIGH 2000) to obtain
on the exogenous variables. The general consumption for each decile in producer
equilibrium is computed from the system of prices. Other missing data included exports
equations developed by the model. Due to price elasticities, domestic and foreign
the complexity for solving such system of goods substitution elasticities, so we used
non-linear equations the General Algebraic the Commercial Information System of
Modeling System (GAMS) software was Mexico, from the World Trade Atlas, to
used. This software is specifically design obtain the quantities and prices for each
for solving complex linear, non-linear and customs tariff number and connect them to
multiple integrals optimization problems their respective productive branch (catalog
allowing for multiple solutions for one provided by Banco de Mexico). With this
element and generating summary reports for information it was possible to perform
each case. econometric estimations of elasticities for

the 63 productive branches in the Mexican For the production sector, both
economy, whose goods are tradable. The simulations show a decrease in education
estimation of the elasticities constitutes one services. However, for the fiscal policy
of the most important contributions for this without subsidy this effect is less and also
work. promotes a greater production of goods in
The simulation results of both fiscal the construction and installation sectors.
policies can be summarized as follows: Besides, the fiscal policy without subsidy
For private consumption, an increase in the generates more value added and the
return on capital, an increase in the subsidy reallocation of production factors is less
to the first population deciles through wages strong as a consequence of a greater increase
and the increase in the price of compounded in the return on capital, of 1.1%, than in the
goods, affect the income and consumption fiscal policy with subsidy.
decisions of population deciles. In both Total investment increases more in the
simulations, income raises due to the increase fiscal policy with no subsidies for the first
in return on capital and to the reallocation of three population deciles because government
production factors. But changes in income saving is greater than in the fiscal policy with
are greater for the simulation of the second subsidies. Gross fixed investment increases
fiscal policy where the first population more for the fiscal policy with subsidy for
deciles are benefited by the subsidy in the first three population deciles because the
their income through wages. Changes in return on capital corresponding to the rent
income and in price of compounded goods price of capital pieces is greater due to the
led to a decrease in total consumption for increase on wages.
all population deciles. In both simulations
the most affected population deciles are the 2. Model
first three in terms of their consumption,
and the least affected is the tenth population 2.1. Considerations about the model
decile; however, it is important to point out
that the second fiscal policy, where deciles The following model represents a small
with lower income are protected, generates open economy that is price acceptant, with
a larger decrease in total consumption than five economic sectors: companies, families,
the policy without the subsidy. For the first government, foreign sector and investment
fiscal policy, the impact in total consumption sector.
for the first three deciles is 11,079 million Equations within the model include
pesos. endogenous and exogenous variables, as
In the government sector, revenue from well as parameters. Endogenous variables
the collection of income tax, tax on capital will be identified by capital letters;
and value added tax is 13.19% greater for exogenous variables will be in capital letters
the policy without subsidy, which represents and hats, and parameters in lower-case or
an additional 93,300 million pesos to the Greek letters
original equilibrium, whereas for the second
fiscal policy (with subsidy) 20,000 million 2.2. Price System
pesos more are collected respect to the
original equilibrium. Four prices are computed in the price
system: price for domestic goods (Pi ) ,
price for compounded goods (Pi ) , price of
m
imports (Pi ) and the international price of for final consumption goods which are
we
exports (Pi ) . The first one is obtained from made of intermediate inputs from other
the zero-gain equation for companies1. sectors and of value added. The decision
of how many domestic goods are produced
(1)
(Xis) is calculated by aggregating, through a
vai (1 + tiiva)` wK j ` wl j Leontief function, the intermediate inputs
1-a a i i
n
Pi = d 1 - ai ai
+ / a j,i Pj and demand for labor and capital.
ci j=1 (5)
The second one is obtained by equaling II II II II

Xis = min c VAi , 1,i , 2,i , 3,i , ..., j,i m
vai a1,i a2,i a3,i a j,i
the production value of the compounded
good to the sum of values of imported
and exported goods, considering that the Being IIj,i the quantity of intermediate
function for the compounded goods has input j used in the fabrication of final good i
constant substitution elasticity (CES). which is denoted by the following equation:
(2) (6)
d vi
^1 - nih P i II j,i = Xis a j,i
P im c m m + P id
ni Pi
Pi = 1 v
bi ^1 - nih P id v v
i
where aj, i represents the technical

i-1
i-1
c ni + ^1 - nihc m m
ni Pim coefficient of intermediate input j, that is, it
corresponds to the minimum input for good
Price for imports of good i (Pim) will be j used for the fabrication of unit i.
determined in an exogenous way by the In the generation of value added in the
foreign price of good i% (Pi*) , the exchange different sectors (VAi), productive services
rate (TC) and the import tariffs for good of capital and labor are used through a
i (tmi) . This will satisfy the assumption of a Cobb-Douglas function.
(7)
small open economy. (3) ai 1 - ai
VAi = ci L K
S
i i
Pim = Pi* ^1 + tmih TC

To obtain demands for labor (Li) and
International price for exports of good capital (Ki), companies make a minimization
i (Piwe) is determined by the domestic price of capital and labor costs subject to a function
of good i (Pi ) , by the exchange rate (TC) and on value added.
d
(8)
by the subsidy on exports for good i (si) .
wL ^1 - aih a i
(4) VAi c m
Ki = wk ai
we P id ci
Pi = (9)
^1 + sih TC
2.3. Supply VAi c wK ai 1-a i
m
wL ^1 - aih
Li =
ci
In the supply side we consider n
different productive sectors in the economy Where ci corresponds to the efficiency
1
This consideration derives from the assumption that in a small
factor of sector i.
open economy companies are price acceptant, so that revenue Total demand for capital (Kd) and labor
for each additional unit produced is equal to its cost the sum of demands in each sector i.

(10) (15)
n S
n
K d = / Ki S g = Zg - / Yig Pi
i=1
i=1
(11)
n
Ld = / Li 2.5. Demand
i=1
Private consumption demand is

In the same manner, usage of productive represented by H family groups with
services of capital and labor by productive different income levels so each group h
industries of different goods, must match will have a determined level of factors
the total quantity available of those factors (capital and labor), that is, its budget
in the market: constraint (Zph).
(12) (16)
S S S
K d = K0s Zph = wL Lsh ^1 - tiISRh + wK K hs ^1 - ticaph
(13) where:
S WL = Labor wages.
Ld = Ls0
W K = Capital retribution.
Lsh = Labor supply for group h.
Where % K0s represents the availability of K hs = Capital supply for group h.
capital and %Ls0 the availability of labor.
It is considered that each groups h
2.4. Government demand (Yihc) for final goods in the economy
h responds to a maximization process of its
For the purpose of this work, government Cobb-Douglas type utility function subject
consumption is assumed to be determined to its respective budget constraint
% (17)
exogenously (Yig) , so that the Federal h
Government assigns a fixed percentage of b gh Zph
i
Y ic =
h
its income to spending in the different goods Pi

produced in the economy.
Income not used for consumption is used
S for saving (S h )
p
Y i = Y ig
g
(18)
p
Government income (Zg) comes from S = ^1 - ghh Zph
h
value added tax, capital tax, income tax and

from import tariffs. Total private consumption for good i
(14)
(Yic) will be represented by the sum of
S S S
demands for the H groups of consumers.
H H n
Zg = / t hISR `wL Lshj + / t hcap `wK K hsj + / tmi Pi* TC Mi +
h=1
n
h=1
n
i=1
In the same token, total private saving (Sp)
+ / t ^wL Li + wk Kih - / si Pi TC Ei
i
IVA we
will be the sum of savings of the H groups
of consumers.
i=1 i=1
(19)
Additionally, government saving (Sg) H
corresponds to the income not used for Y ic = / Y ic h
consumption. h=1
(20) (24)
H
P imq
S p = / S hp
ic
Pk = c Pic m c imq m
h=1 ic imq
where:
2.6. Investment
ic = Preference parameter for construction good.
imq = Preference parameter for machinery and equipament
For the investment side a new system
was used, which was not found in other Consequently, price for new capital
general equilibrium problems in the revised pieces (Pk) in the economy will be a weighted
literature2. Total investment is considered average of the prices of its two inputs:
not only to be the sum of public, private construction price and that of machinery
and foreign savings, but also equal to the and equipment. This price and the rent price
fixed gross investment (IBF) plus inventory of capital services allows to determine the
variation (YiVE) . return on capital (t) and the value of fixed
(21)
gross investment (IBF).
S The vector of demand for goods that fixed
I = S p + S g + F TC
gross investment generates is the following:
The value of fixed gross investment made
(25)
in the economy (IBF) responds to a Tobins BF
q process, by which the amount of fixed YiIBF = ii I
Pi
gross investment is proportional to the
difference between return on capital in the Where qi represents the preference
domestic economy (t) and the real interest parameter of sector i for fixed gross
rate of foreign financial assets %
(r*) . investment goods.
(22)
The vector of demand for total investment
S will correspond to a percentage qi of total
I BF = d ` t - r*j
investment amount (I/Pi) for each investment
Return on capital (t) refers to the good i.
(26)
rent price of capital pieces or the price of
productive services of capital earned by its YiI = zi I
Pi
owners (wK), relative to its price (Pk).
(23)
Therefore, demand for inventory
t = wK variations will correspond to the difference
PK
between the demand of total investment
Due to prices of construction goods (Pic) and the demand for fixed gross investment
and to prices of machinery and equipment goods, allowing demand for inventory
(Pimq), and considering that the marginal variations to adjust to changes in demand
cost of producing new capital pieces in of total investment and in demand for fixed
the economy must match capital price, and gross investment goods.
(27)
applying Shepards theory, we obtain:
YiVE = YiI - YiIBF
2
See: Serra (1979), Cordova (1991), Balzarotti and Cicowiez
(2004) and Cicowiez (2004)

2.7. Foreign Sector 2.9. Equilibrium in Market of goods
Demand for exports was obtained To calculate total demand of domestic

through an exponential function where the produced goods (Xi ) , it is necessary to
d
quantity demanded of exports of good i (Ei) calculate the demand of domestic goods
is determined by the foreign price of exports (Di) and the demand for export goods (Ei).
%
of good i(Pi*) , by the international price of To compute the first one, we need to find
exports of good i (Piwe) and by the price the ratio of domestic use (UDi) through
elasticity of exports (hi). a minimization problem to determine
(28)
the quantity of domestic and imported
S Si*
hi
goods needed by companies to produce
Ei = Ei f Pwe p
compounded goods.
Pi (32)
Demand for imports responds to a UDi = 1

v i
minimization process that aggregates ^1 - nih P id v - 1 ^v - 1h

i
i
bi c ni + ^1 - nihc m m
domestic and imported goods in a CES-type ni Pim
function. Quantity demanded of imports
of good i depends on the domestic price of With this, the internal demand for
good i (Pid) , on the quantity demanded of domestic goods i (Di) will correspond
domestic good i (Di) and on the substitution to a fraction UDi of the total investment
elasticity between domestic and imported demand, private consumption, government
goods of sector i (si). consumption and of intermediate demand.
(29)
^1 - nih P i d vi (33)
S
Mi = c m Di
ni Pim
Di = UDi `YiI + Yic + Yig + Vij
Balance of payments is made up by
the current account and foreign savings Total demand of good i produced
considering that the exchange rate is flexible domestically (Xi ) is equal to the sum of
d
and it adjusts to maintain equilibrium in the internal demands for domestic produced
balance of payments. goods (Di) and of the demand for exports of
(30)
good i.
n S n S (34)
BP = / Pi* Mi - / Piwe Ei - F = 0 Xid = Di + Ei
i=1 i=1
2.8. Intermediate Demand Equilibrium in the market of goods will

be reached when no excess of supply or of
Intermediate demand for each productive demand for good i exists.
(35)
sector (Vi) is equal to the sum of sales in
each of the n sectors of the economy for Xid - Xis = 0
the production of all compounded goods.
(31)
n
Vi = / ai,j Xis
j=1
2.10. Rules for closure the number of equations. The following

table describes the endogenous variables of
In order to maintain consistency in the the model and the corresponding equation.
model, the number of variables must match
Table 1. Rules for Closure
3. Calibration An important assumption to calibrate the

model is that the base year was in equilibrium
In order to obtain the value of the with prices equal to 1. This does not mean
equation parameters in the model described that all prices equal one in absolute terms,
previously, firstly, it is necessary to select but rather that they will serve as indexes
the base year for data; secondly, it is and at the base year their value is one. By
necessary to formulate the assumption that this, we can obtain the percentage change in
equilibrium prices are set to 1, and lastly, process and quantities, but not in absolute
a social accounting matrix must be built to terms.
reflect the transactions in the economy given The social accounting matrix helps to
the previous assumption. reflect the transactions in the economy during
Base year selection is important because a period of time. For its construction it was
all information used must correspond to necessary to use the Input Product Matrix
that particular year. For the purposes of this for 2000 calculated by the RAS method by
model, we consider 2000 as the base year Hernandez Laos (2004)3, that includes sales
due to the particular interest of analyzing of goods and services for each sector to other
the application of the fiscal policies during sectors which supply to final demand as well
the subsequent years, and due to the fact that as other purchases to other sectors and the
most information is available for that year in rent of productive services of factors and the
order to calculate the general equilibrium of payments associated to taxes on production.
the economy. Additionally, for the analysis of demand
3
According to the handbook of the MIP 2000, information was we constructed a matrix of demand by each
updated from the Mexican domestic transaction matrix, base
year 1980, in its version of 93 sector made by INEGI. From
population decile of per capita income, using
the matrix, three concepts of the gross domestic product were data from the National Survey of Home
divided in four parts according to the methodology developed
by Adriaan Ten Kate and Sergio Escamilla Snchez. Actual-
Income and Spending 2000 (ENIGH 2000).
izacin de Matrices de Insumo Producto con el Mtodo RAS. Matrix construction procedure is detailed in
Revista Estadstica, Vol. III, No. 5, Mxico, 1989. Exhibit II of Mrquez (2006).

The construction of the social accounting Calibration of parameters is made based

matrix and the calibration of parameters on equilibrium assumptions and from data
is done for 23 productive sectors of the of the social accounting matrix. There are
Mexican economy in which are included ten parameters, whose calculation is based
those whose production and consumption from the equations for supply, private
will be directly affected by imposing a 15% consumption, investment, foreign sector
tax rate on their goods consumption: food and market equilibrium, all previously
sector, medicine sector, education services, described.
press and editorials. All 23 sectors are shown
in Table 2.
Table 2. Productive Sectors
In the supply side four parameters are good j used to produce one unit of good i. Its
calculated for each of the 23 sectors. The value is obtained from equation (6):
first one corresponds to the amount of
(37)
aggregated value used in the production of
one unit of good i (vai). For this, we need to II j,i
a j,i =
work out its value from equation (5): Xis
The efficient factor parameter (ci) is
(36)
obtained from the equation (7), working out
vai = wK Ki +s wL Li its value we obtain:
Xi (38)
The second parameter corresponds to ci = wk Kai +1 -waL Li

Li Ki i i
the technical coefficient of the intermediate

good j (aj;i), that is, the minimum input of The last parameter, computed from the
supply equations, is the ratio of wages paid
(43)
in sector i respect to the total production IBF VE
costs in that sector (ai). Its value is obtained zi = Pi Yi + Pi Yi
I
from the relative price expression relative to
the equilibrium of factors. From the equations for the foreign sector,
we obtain just one parameter mi from the
wK = 1 - ai Li expression of relative equilibrium prices of
wL ai Ki
domestic and imported goods.
where ai is:
(39) 1
P d = ni M iv i
ai = wL Li P m 1 - ni D v1 i
wL Li + wK Ki i
In the private consumption sector two Where the value of mi is the following:
parameters are calculated: the preference
(44)
of each group h for good i (bih) and the 1
percentage each group h allocates of their Pid D iv i
ni =
income to consumption (gh). 1
Pim M + Pid D iv
vi
1
(40)
i
i
c
bih = nPi Yi h
From the equations of market equilibrium
/ Pi Yic h of goods we get parameter bi for each
i=1
productive sector from the minimization
The second parameter is obtained from problem used to obtain demand for domestic
dividing total consumption of group h by goods, where:
(45)
total income of group h:
(41) vi - 1 vi - 1
vi
Pid Di + Pim Mi = bi ` ni Di j vi - 1
n
vi
+ ^1 - nih Mi vi
/PY i
c
ih
gh = i=1
And working out the value of bi we get:
Z hp
For the investment sector we also calculate bi = Pid Di + Pim Mi
vi
two parameters. One is the preference of
vi - 1 vi - 1
` ni Di vi
+ ^1 - nih Mi vi j vi - 1
sector i for fixed gross investment goods (qi)

and the other is the preference of sector i for The values of each parameter for the 23
total investment (fi). productive sectors are shown in Exhibit I.
The first one is calculated from equation
(25) from dividing the value of demand of 4. Calculation of the General
fixed gross investment in sector i by the total Equilibrium
value of fixed gross investment.
(42)
IBF
After explaining the model and the
ii = Pi YBFi calibration of parameters, we can apply
I
the model to the Mexican economy for the
transactions of year 2000. To get the original
he second one is calculated from equation equilibrium, that is the prices and quantities
(26) from dividing the sum of demand of that satisfy the general equilibrium, it is
fixed gross investment in sector i, and the necessary to solve the system of equations
value of inventory variations divided by worked out for the model. Due to the fact
total investment in the economy. that equations are non-linear, manual

solving is complex, so we need software demand for capital and labor, as well as the
that helps us solve this type of equations. quantity produced for year 2000 for each of
It could be argued that the calculation of the 23 productive sectors shown in Table 2.
the general equilibrium is not necessary The level of detail of the private demand
because all prices will be equal to one allowed obtaining data for consumption by
since the parameters of the utility function each population decile of total income per
and production were calibrated under this capita for the 23 final goods. At the same
assumption; however, obtaining these time, GAMS showed data about income
results will show that the model is indeed and savings for each population decile.
well calibrated. If the quantities obtained for In the investment sector we got quantities
the original equilibrium do not match those demanded for exports and imports for the
used in the social accounting matrix and/or 23 sectors due to the estimation of price
the prices were not equal to one, an error has elasticities for exports and substitution
occurred in the calibration process to obtain elasticities between domestic and imported
the equilibrium. products, for the 63 tradable goods in the
The software used is General Algebraic Mexican economy.
System (GAMS), which is specifically At the same time, GAMS produced
designed for linear, non-linear and multiple results for intermediate demand of the
integration optimization4. 23 productive sectors corresponding to
The calculation of the general equilibrium intermediate inputs used in the fabrication of
in the Mexican economy for year 2000, the final goods. Finally, data on prices allowed
following assumptions were satisfied: to check that prices from the calculation of
the general equilibrium by GAMS were the
a. Supply is equal to demand in all markets. same as those found in the social accounting
matrix.
b. No sector has positive benefits. All results for the general equilibrium are
shown in Exhibit II.
c. All actors fulfill their budgetary
constraint. 5. Simulations of Fiscal Policies
d. Foreign sector is in equilibrium. After the calibrating parameters and

calculating the general equilibrium of
e. Original equilibrium has prices equal to Mexican economy for the year 2000,
one. the model is ready for use to analyze the
economic policies in pairs with equilibrium
In the supply side we obtained as base reference.
information on the aggregated value of the The experiment consists on simulating
two types of fiscal policies. The first one
4
AThe system is commonly used in large and complex problems, considers taxing at 15% the consumption of
allowing for model programming to solve for different values of
a specific element while generating a report of the outcome for food, medicines, education services, press
each solution. The model program must be entered in algebraic and editorials. The second one, besides
form and we must select a specific solver for the problem. For
this work we used the CONOPT solver. Results are generated
the 15% tax, considers an increase in the
after the solver performs several iterations to find the solution income subsidy through wages for the first
that satisfies the equilibrium in all markets. Please refer to the
webpage: http://www.gams.com/docs/intro.htm
three population deciles, being 3% to 9% for
the first decile and 2% to 7% for the second the subsidy through wages in the first three
and third deciles. population deciles. Decrease in demand
The first simulation has as main goal for capital among sectors is explained by
to expose the effects of the 15% tax on the different utilization intensities of that
production, consumption, investment, trade, factor. Table 3 shows that the Non-alcoholic
tax colleting and on gross domestic product. beverage sector absorbed all the capital
This will be helpful since the policy was quantity freed by other sectors. Demand
proposed by the President and was sent for capital was 61.3% more in the second
for review to the Congress in 2001, but simulation than in the first one.
has neither been approved or rejected due In the case of the labor market, for
to the unknown economic effects of such both simulations we see an increase in the
policy. The second simulation, besides of demand for labor in all sectors except for
having the same taxing as the first one, is the education service sector, due to the
important because it can prove if protecting substitution among production factors.
lower income population deciles through Nevertheless, changes in quantity demanded
subsidies, as mentioned by different political of labor are smaller in the first simulation,
actors, will help to implement the fiscal but in the same direction as in the second
policy proposed by the President. simulation. Table 4 shows that for the
For both simulations, the description construction and installation sector, the
the effects is detailed for all 23 sectors in increase in quantity demanded of labor was
the economy. The following tables show larger for the first simulation in 3.9% than
the percentage variations obtained respect the second simulation.
to the initial equilibrium and also include
the variations between the first and second Table 3. Capital Demand
simulations.
Simulation Simulation
Sector % Change
1 2
5.1. Effects on Production Sector
sec1 -0.1% -0.3% 0.2%
sec2 -0.1% -0.2% 0.2%
In both simulations we get distortions in
sec3 -0.1% -0.2% 0.1%
the price of compounded goods and salaries.
For both cases, wages have no change since sec4 0% -0.4% 0.4%
they are fixed in order to calculate the new sec5 17.6% 78.9% -61.3%
equilibriums; nevertheless, the economy sec6 0% 0% 0%
experienced and increase in the capital sec7 -0.3% -0.6% 0.3%
retribution that affected the decisions on the sec8 0% -0.4% 0.4%
quantity demanded for production factors sec9 -0.5% -0.5% 0%
for the fabrication of compounded goods. sec10 -0.6% -1.1% 0.6%
In the case of capital markets, the increase sec11 -0.3% -0.6% 0.3%
in the return for that production factor was sec12 0% -0.3% 0.3%
0.8% and 1.9% for the first and second sec13 0% -0.3% 0.3%
simulations respectively. This means that the sec14 -0.2% -0.2% 0%
decrease in quantity demanded for capital sec15 -0.2% -0.5% 0.3%
was smaller for the first simulation than sec16 -0.4% -0.8% 0.4%
for the second simulation, which besides sec17 3.8% -1.3% 5%
the new tax of 15% included an increase in

Simulation Simulation Simulation Simulation

Sector % Change Sector % Change
1 2 1 2
sec18 -0.5% -1.4% 1% sec4 0% 0% 0%
sec19 -0.2% -0.5% 0.3% sec5 17.7% 79.6% -61.9%
sec20 -0.2% -0.5% 0.3% sec6 0% 0% 0%
sec21 -0.1% -0.2% 0.1% sec7 0% 0% 0%
sec22 -0.3% -0.7% 0.4% sec8 0% 0% 0%
sec23 -6.6% -11.1% 4.6% sec9 0% 0% 0%
sec10 0% 0% 0%
Table 4. Labor Demand sec11 0% 0% 0%
sec12 0% 0% 0%
Simulation Simulation sec13 0% 0% 0%
Sector % Change
1 2
sec14 0% 0% 0%
sec1 0.6% 1.4% -0.8%
sec15 0% 0% 0%
sec2 1.3% 1.9% -0.6%
sec16 0% 0% 0%
sec3 0.6% 1.4% -0.9%
sec17 4.2% -0.2% 4.4%
sec4 2.2% 2.2% 0%
sec18 0% 0% 0%
sec5 18.1% 81.9% -63.9%
sec19 0% 0% 0%
sec6 0% 0% 0%
sec20 0% 0% 0%
sec7 0.6% 1.5% -0.9%
sec21 0% 0% 0%
sec8 1.4% 1.4% 0%
sec22 0% 0% 0%
sec9 0% 1.6% -1.6%
sec23 -6% -9.8% 3.8%
sec10 0% 1.1% -1.1%
sec11 0.5% 1.4% -0.9%
sec12 1% 2.1% -1% Regarding value added, we can see in
sec13 0% 1.6% -1.6% Table 5 that the sector of Alcoholic beverages
sec14 1.4% 2.8% -1.4% the increase is larger by 61.9% for the
sec15 0.6% 1.4% -0.8% fiscal policy with subsidy than that for the
sec16 0% 0.9% -0.9%
one without it, whereas in the construction
sec17 4.5% 0.7% 3.9%
and installation sector the first simulation
generates 4.4% more value added than the
sec18 0.4% 1.2% -0.8%
second one. For the education sector we can
sec19 0.6% 1.4% -0.9%
see a decrease in value added generation
sec20 0.5% 1.4% -0.9%
which is 3.8% less in the first simulation
sec21 0.6% 1.8% -1.1%
than in the second simulation.
sec22 0.4% 1.2% -0.7%
Table 6 presents the changes in production
sec23 -5.8% -9.5% 3.7% of compounded goods for both simulations.
Results indicate that the first simulation
Table 5. Value Added Generation promotes a greater quantity of compounded
goods produced in the construction and
Sector
% Change installation sector (4.4% more) and that the
1 2
decrease in education services is smaller for
sec1 0% 0% 0%
the first simulation.
sec2 0% 0% 0%
Wrapping up, the effects in the production
sec3 0% 0% 0%
sector for the first policy (only 15% tax on
consumption of food, medicines, education 5.2. Effects on Foreign Sector

services and goods from press and editorials)
are better since this policy promotes a greater Changes produced in the quantity
production of goods in the construction and demanded of domestic, imported and export
installation sector and reduces the effect goods, as consequence of the fiscal policies
of the decrease in production goods in the simulated, are similar in both cases. Table
education sector. Besides, the first policy 7 shows that changes in quantity demanded
generates a larger quantity of value added of imported goods are the same for both
and reallocation of production factors among simulations, except for the sector of Other
sectors is not very intense as consequence of manufacture industries where the quantity
the increase in the return on capital, which is demanded of imported goods is 0.2% larger
1.1% smaller for the first simulation than in in the first simulation than in the second
the second one. simulation.
In the case of demand for domestic goods,
Table 6. Production of Compounded Goods Table 8 reveals that the first policy generates
a greater demand for domestic goods in the
Sector
% Change construction and installation sector which is
1 2
4.4% bigger than in the second fiscal policy.
sec1 0% 0% 0%
Besides, the decrease in quantity demanded
sec2 0% 0% 0%
for domestic goods in the education sector is
sec3 0% 0% 0%
3.8% smaller in the first simulation than in
sec4 0% 0% 0% the second one.
sec5 0% 0% 0% In the commercial, restaurants and hotels
sec6 0% 0% 0% sector we can see that the application of the
sec7 0% 0% 0% first policy leads to a decrease of 0.01% in
sec8 0% 0% 0% the quantity demanded of domestic goods in
sec9 0% 0% 0% this sector, situation that is not present for
sec10 0% 0% 0% the fiscal policy with subsidy.
sec11 0% 0% 0%
sec12 0% 0% 0% Table 7. Demand of Imported Goods
sec13 0% 0% 0%
sec14 0% 0% 0% Sector % Change
1 2
sec15 0% 0% 0% sec1 0% 0% 0%
sec16 0% 0% 0% sec2 0% 0% 0%
sec17 4.2% -0.2% 4.4% sec3 0.8% 0.8% 0%
sec18 0% 0% 0% sec4 0% 0% 0%
sec19 0% 0% 0% sec5 0% 0% 0%
sec20 0% 0% 0% sec6 0% 0% 0%
sec21 0% 0% 0% sec7 0% 0% 0%
sec22 0% 0% 0% sec8 0% 0% 0%
sec23 -5.9% -9.7% 3.8% sec9 0% 0% 0%
sec10 16.7% 16.7% 0%
sec11 0% 0% 0%

Simulation Simulation Table 9. Demand for Exportable Goods

Sector % Change
1 2
sec12 12.9% 12.9% 0% Simulation Simulation
Sector % Change
sec13 0% 0% 0% 1 2
sec14 0% 0% 0% sec1 0% 0% 0%
sec15 0% 0% 0% sec2 0% 0% 0%
sec16 0.2% 0% 0.2% sec3 -0.7% -0.7% 0%
sec17 0% 0% 0% sec4 0% -0.8% 0.8%
sec18 0% 0% 0% sec5 0% 0% 0%
sec19 0% 0% 0% sec6 0% 0% 0%
sec20 0% 0% 0% sec7 0% 0% 0%
sec21 -3.2% -3.2% 0% sec8 0% 0% 0%
sec22 0% 0% 0% sec9 0% 0% 0%
sec23 0% 0% 0% sec10 -5.9% -5.9% 0%
sec11 0% 0% 0%
Table 8. Demand for Domestic Goods sec12 -7.7% -7.7% 0%
sec13 0% 0% 0%
Sector % Change sec14 0% 0% 0%
1 2
sec1 0% 0% 0% sec15 0% 0% 0%
sec2 0% 0% 0% sec16 0% 0% 0%
sec3 0.03% 0.03% 0% sec17 0% 0% 0%
sec4 0.2% 0.2% 0% sec18 0% 0% 0%
sec5 0% 0% 0% sec19 0.05% 0.05% 0%
sec6 0% 0% 0% sec20 0% 0% 0%
sec7 0% 0% 0% sec21 0% 0% 0%
sec8 0% 0% 0% sec22 0% 0% 0%
sec9 0% 0% 0% sec23 0% 0% 0%
sec10 0.6% 0.6% 0%

sec11 0% 0% 0% Changes in international prices of exports
sec12 0.9% 0.9% 0% for both fiscal policies made distortions
sec13 0% 0% 0% in demand for export goods equal in both
sec14 0% 0% 0% cases. Just for the Alcoholic beverage sector,
sec15 0% 0% 0% the second fiscal policy generates a greater
sec16 0% 0% 0%
decrease of 0.08% in the amount demanded
sec17 4.2% -0.2% 4.4%
of export goods than in the first simulation.
Furthermore, the exchange rate
sec18 0% 0% 0%
depreciates 7% in the first fiscal policy and
sec19 -0.01% 0% -0.01%
1.6% in the second one, both in relation to the
sec20 0% 0% 0%
original equilibrium. This is explained by an
sec21 0% 0% 0%
increase in total investment, which is greater
sec22 0% 0% 0%
in the first simulation, causing the deficit in
sec23 -5.9% -9.7% 3.8% the balance of payments to be greater and
thus, the depreciation of the exchange rate
to be greater to allow the equilibrium in the deciles. For the fourth decile the difference
balance of payments5. between both policies is 0.7%, for the fifth
and sixth deciles is 0.6%, for the seventh,
5.3. Effects in Private Consumption eighth and ninth deciles is 0.8%, and finally
1.1% for the tenth decile.
The fiscal policy that, besides applying Table 11 shows the change in consumption
the tax of 15% to the consumption of food, by decile for each fiscal policy. Data reveal
medicines, education services and goods that the fiscal policy with subsidy affects
from press and editorials, also increases the more total consumption of deciles, except
subsidy through wages, has a larger impact, for the tenth decile which does not change
than the policy without the subsidy, in the for both simulations. Figures for the second
change of income of population deciles fiscal policy indicate that the decrease in
as consequence of an increase in return consumption for the first decile is 0.6%
on capital and a greater reallocation of greater than in the first fiscal policy, 0.7%
production factors among the sectors. greater for the second and third deciles, 0.6%
greater for the fourth and fifth deciles, 0.4%
Table 10. Income of Population Deciles greater for the sixth decile, 0.7% greater for
the seventh decile, and 0.8% greater for the
Decil
Simulation
1
Simulation
2
% Change eighth and ninth deciles.
Regarding consumption of goods
dec1 0.5% 14.7% -14.2%
produced in the food and education sectors,
dec2 0.4% 17.2% -16.7%
we can observe that population deciles
dec3 0.4% 20% -19.6%
reduce their consumption for these goods
dec4 0.4% 1.1% -0.7%
with the second fiscal policy, even though
dec5 0.4% 1% -0.6% the increase in subsidy through wages.
dec6 0.4% 1.1% -0.6% Comparing to the first fiscal policy, the first
dec7 0.5% 1.2% -0.7% and second deciles reduce their consumption
dec8 0.5% 1.2% -0.7% 0.3% more with the second fiscal policy,
dec9 0.5% 1.3% -0.8% 0.5% more in the third decile, 0.2% more in
dec10 0.8% 1.9% -1.1% the fourth and fifth deciles and 0.4% more in
the fifth decile.
Table 10 clearly shows that changes in For education services consumption
income of population deciles are larger in in the first simulation (without subsidy) is
the second simulation benefiting the first reduced for the first decile in 15.4% more
three deciles in 14.2%, 16.7% and 19.76%, than in the second simulation, 11.11%more
respectively, more than in the first policy. for the second decile, 13.3% more for the
The change in income of the rest of the third decile, 12.2% more for the fourth
deciles is still greater for the second fiscal decile, 11.4% more for the fifth and seventh
policy, but the change compared to the first deciles, 10.2% more for the sixth decile,
fiscal policy is not as big as in the first three 11.9% more for the eight decile, 11.7% more
for the ninth decile and 11.4% more for the
5
According to the Mudell-Flemming model, in an open economy
tenth decile.
with a flexible exchange rate and imperfect capital flow, an in-
crease in the IS leads to a deficit in the balance of payments, so
that the exchange rate must depreciate to reach equilibrium. For
additional reference see Appleyard/Field (1997).

Table 11. Consumption of Population Deciles Table 12. Total Private Consumption
Sector % Change
Sector
I II III IV V VI VII VIII IX X 1 2

sec1 -0.7% -0.9% 0.2%
1.06
Sec 1
1.2% 0% 1.1% 0% 0% 0% 0% 0% 0%
% sec2 -0.2% -0.5% 0.3%
Sec 2
0.3% 0.5% 0.5% 0.3% 0.4% 0.4% 0.3% 0.2% 0.2% -0.1% sec3 -7.2% -7.4% 0.2%
sec4 -0.5% -0.8% 0.3%
Sec 3
0.3% 0.3% 0.5% 0.2% 0.4% 0.2% 0.2% 0% 0% -0.2%
sec5 -1.2% -1.3% 0.1%

Sec 4
0% 0% 0% 0% 0% 0% 0% 0% 0% -1.9%
sec6 0% 0% 0%
Sec 5
0% 0% 0% 0% 0% 0% 1.2% 0% 0% -0.3%
sec7 -0.3% -0.3% 0%
sec8 0% 0% 0%
Sec 6
0% 0% 0% 0% 0% 0% 0% 0% 0% 0%
sec9 -0.9% -0.9% 0%

Sec 7
0% 0% 0% 0% 0% 0% 0% 0% 0% -0.4%
sec10 -6.8% -6.8% 0%
Sec 8
0% 0% 0% 0% 0% 0% 0% 0% 0% 0% sec11 -0.2% -0.3% 0.1%

sec12 -7.5% -7.7% 0.2%
Sec 9
0.2% 0.3% 0.4% 0% 14.3% 0% 0% 0% 0% 0%
sec13 0% 0% 0%
Sec 10
0% 0% 0% 0% 0% 0% 0% 0% 0% 0%
sec14 0% 0% 0%
sec15 0% 0.1% -0.1%
Sec 11
0% 0% 0% 0% 0% 0% 1.2% 0.9% 0% -0.3%
sec16 0% 0% 0%
Sec 12
0% 0% 0% 0% 0% 0% 0% 0% 0% 0%
sec17 0% 0% 0%
sec18 0% 0% 0%
Sec 13
0% 0% 0% 0% 0% 0% 0% 0% 0% 0%
sec19 -0.04% 0% -0.03%

Sec 14
0.3% 0.4% 0.5% 0.3% 0.3% 0.3% 0.2% 0.2% 0.2% -0.2%
sec20 -0.1% -0.1% 0%
sec21 -0.1% -0.2% 0.1%
Sec 15
0% 0% 0% 0% 0% 0% 0% 0% 0% -0.1%
sec22 0% 0.3% -0.3%

Sec 16
0.2% 0.3% 0% 0% 0% 0% 0% 0% 0% 0%
sec23 -17.6% -29.1% 11.4%
Sec 17
0% 0% 0% 0% 0% 0% 0% 0% 0% 0%
Besides the changes described previously,

Sec 18
0% 0% 0% 0% 0% 0% 0% 0% 0% 0%
when the second fiscal policy is applied,
we can observe larger decreases in the
Sec 19
0% 1.1% 0% 0% 0.3% 0.2% 0.2% 0% 0.1% -0.2%
consumption of goods from the basic-metal

sector, paper and cardboard sector, mining
Sec 20
0% 0% 0.5% 0.3% 0.3& 0.2% 0.2% 0.1% 0.1% -0.3%
sector and agriculture, forestry and fishing

Sec 21
0% 0% 0% 0.7% 0.6% 0.4% 0.3% 0.2% 0.2% -0.1%

sector.
Specifically, in the first simulation
Sec 22
0% 0% 2% 0% 0% 0% 1.3% 0% 0% -0.3%
15.4 11.1 13.3 12.2 11.4 10.2 11.4 11. 11.7 11.4
the first three deciles decrease their total
Sec 23
% % % % % % % 9 % % consumption in 11,079 million of pesos,

whereas, for the second simulation, the
decrease in total consumption for the same
Change %
0.6% 0.7% 0.7% 0.6% 0.6% 0.4% 0.7% 0.5% 0.8% 0%
deciles was 13,657 million of pesos.

Table 12 summarizes the changes in total
private demand of compounded goods in
the economy. Data clearly show that both
policies generate a reduction in consumption 5.4.Effects on Government Sector

of most compounded goods fabricated in the
different productive sectors. Nevertheless The first fiscal policy considers only
the reduction in total private consumption is a tax of 15% on the consumption of food,
smaller with the first fiscal policy than with medicines, education services and goods
the second one. Only in the commercial from press and editorials; the second fiscal
sectors of restaurants and hotels and in the policy, besides considering this tax, increases
metal, machinery and equipment sectors the de subsidy on income through wages for the
quantity demanded is larger when applying first three population deciles. This difference
the second fiscal policy, which includes de makes revenue for government (Zg) to be
subsidy. 13.19% higher for the first fiscal policy
Table 13. Saving for Population Deciles than for the second one. The first fiscal
Simulation Simulation policy provides an additional 93,300 million
Decil % Change
1 2 pesos of revenue to the government than in
dec1 0.6% 40.2% -39.6% the initial equilibrium, whereas the second
dec2 0.5% 39.5% -39% policy just an additional 20,000 million
dec3 0.4% 46.6% -46.2% pesos. Due to the fact that government
dec4 0.4% 1.1% -0.7% spending is fixed for both simulations,
dec5 0.4% 1% -0.5% government savings (Sg) 78.4% greater in
dec6 0.4% 1.1% -0.7% the first simulation than in the second one.
dec7 0.4% 1.2% -0.7%
dec8 0.5% 1.2% -0.7%
Table 14. Government Revenues and Saving
dec9 0.5% 1.2% -0.8%
% Change
dec10 0.7% 1.8% -1.1% 1 2
Zg 16.79% 3.6% 13.19%
Since the proportion of income that
Sg 28.6% -49.8% 78.4%
population deciles allocate to consumption
is smaller with the second fiscal policy,
savings of those population deciles will be 5.5. Effects on Investment Sector
larger in this simulation than in the first
simulation. We observed that for the first simulation,
From Table 13 we can see that due to total investment rose 4.7% compared to the
the increase of subsidies in income through initial equilibrium and 1% for the second
wages for the first three deciles, savings simulation. This is explained by higher
of these three deciles are greater in 39.6%, government savings in the first simulation
39% and 46.2% respectively in the second than in the second one, thus generating more
simulation than in the first one. The fourth, investment.
sixth, seventh and eight deciles show a
larger increase of 0.7% than in the first Table 15. Total Investment
simulation, whereas the fifth shows a 0.5%
larger increase, the ninth 0.8% and the tenth Simulation Simulation
% Change
1 2
1.1%. Total private savings (SPT) are 7.1%
Total
greater in the second simulation than in the Investment
4.7% 1% 3.7%
first simulation.

Table 16. Fixed Gross Investment Figure 1. Tobins q
Decil % Change
1 2
r* 0% 0% 0%
P 0.2% 0.9% -0.7%
Pk 0.4% 1.1% -0.6%
IBF 0.4% 1% -0.7%
d 0% 0% 0%
F 0% 0% 0%
Concerning fixed gross investment (IBF),

it is higher for the second simulation in 0.7%
than in the first simulation, due to the fact
that return on capital corresponding to the Table 17. Intermediate Demand
price of rent of capital pieces (r) was 0.7%
higher for the second simulation. Sector
% Change
1 2
sec1 0% 0% 0%
5.6. Effects on Intermediate Demand
sec2 0% 0% 0%
sec3 21.4% 22.4% -1.1%
Effects on intermediate demand are
different between both simulations. Table 17 sec4 4.2% 12.5% -8.3%
indicates that the simulation of the second sec5 450% 500% -50%
fiscal policy promotes a greater quantity sec6 0% 0% 0%
of intermediate demand in the sectors of sec7 -0.3% 0.2% -0.5%
textiles, clothing and leather industry; wood sec8 -1% 0% -1%
industry; chemical substances, petroleum sec9 0% 0.1% -0.1%
by-products, rubber and plastic; non-metallic sec10 6% 6% 0%
mineral products; basic metal industries; sec11 -0.03% 0.1% -0.1%
metal products, machinery and equipment; sec12 19.3% 19.7% -0.3%
other manufacture industries, transport and sec13 -0.1% 0.1% -0.2%
storage and communications; whereas the sec14 -0.1% 0% -0.1%
second simulation resulted in a reduction sec15 -0.8% 0.1% -0.9%
of the quantity demanded of intermediate sec16 -0.4% 0.1% -0.6%
goods in those same sectors. sec17 0% 0% 0%
For all other productive sectors that
sec18 0% 0% 0%
present changes between both simulations,
sec19 0% 0% 0%
results are in the same direction. However,
sec20 -0.4% 0.3% -0.7%
for the simulation of the second fiscal policy
sec21 0.3% 0.5% -0.2%
increases in the quantity demanded of
sec22 -0.02% -0.3% 0.3%
intermediate goods are greater than in the
simulation of the first fiscal policy. sec23 0% 0% 0%
5.7. Effects on Gross Domestic reduction in consumption by all population

Product deciles and a reduction in total investment. If
we want to compensate the first three deciles
To determine GDP, the expenditure and, at the same time, increase government
approach with constant prices was used for revenue, a policy of taxing with 15% the
both simulations. Table 18 shows that GPD consumption on food, medicines, education
is 0.8% higher in the second simulation than services and goods from press and editorials
in the first one due to the fact that prices for may be done, thus generating an additional
compounded goods turn out to be higher for 93,300 million pesos from which 11,079
the second simulation. million pesos could be applied to maintain
the consumption of the first three deciles
Table 18. Gross Domestic Product (GDP) as a compensation, leaving 82,221 million
pesos as extra revenue. Extensions of this
Sector
% Change work could focus on determining the best
1 2
compensation ways in a direct link between
GDP 2.21% 3.01% -0.80%
government and individuals that generate
certainty and credibility when exercising
6. Conclusions those compensatory measures.
The main motivation for this work was to 7. References

represent the Mexican economy through a
model of general equilibrium to analyze the lvarez y Gonzlez 2001, Estimacin
effects on production, private consumption, de Elasticidades de Armington: Una
investment, foreign sector and intermediate aplicacin al caso de Mxico, Mxico,
demand, caused by the implementation Secretara de Economa.
of a fiscal policy that taxes at 15% the
consumption on food, medicines, education Appleyard/Field 1997, Economa
services and goods from press and editorials. Internacional, Espaa, McGraw-Hill.
Results show that applying such a
policy will generate an increase in the Balzarotti, Nora/Cicowiez, Martin 2004,
return on capital; an increase in the price Zonas Francas: su efecto sobre el bienestar
of domestic, import and export goods and econmico, Argentina, Universidad de la
on compounded goods; a reallocation of Plata.
production factors utilization; an increase in
the exchange rate; a higher total investment Banco de Mxico, Subgerencia de Anlisis
and a higher intermediate demand, as well del Sector Externo 2005, Catlogo que
as a reduction in consumption of population conecta fracciones arancelarias con
deciles where the first three deciles are the ramas productivas, Mxico, Banco de
most affected. Nonetheless, applying a Mxico.
policy with an increase in the subsidy on
income for the first three deciles through Cmara de Diputados, Centro de Estudios
wages does not improve their condition, de Finanzas Pblicas 2005, Recaudacin
nor maintains their original consumption total de ingresos tributarios, Mxico.
level, but in the contrary, it causes bigger
distortions as a reduction in tax collection,

Cicowiez, Martn/Di Gresia, Luciano Mankiw, Gregory 1997, Macroeconoma,

2004, Equilibrio General Computado: Espaa, Antoni Bosch.
Descripcin de la Metodologa,
Argentina, Universidad de la Plata. Mrquez Pea, Eduardo 2006, Efectos del
IVA en la Economa Mexicana. Un anlisis
Crdova Gutirrez Alberto 1991, Efectos de multisectorial a travs de un Modelo de
la Apertura Comercial sobre la Economa Equilibrio General Aplicado. MEGA,
Mexicana: un Enfoque de Equilibrio Mxico, Tesis de Licenciatura. ITAM.
General, Mxico, Tesis de Licenciatura.
ITAM. Parkin, Michael 1995, Macroeconoma,
Mxico, Addison-Wesley, 2a. Edicin.
Francois, Shiells 1994, Modeling Trade
Policy: Applied General Equilibrium Parkin, Michael 1995, Microeconoma,
Assessment of North American Free Mxico, Addison-Wesley, 2a. Edicin.
Trade, USA, Cambridge University
Press. Serra Puche, Jaime 1979, A General
Equilibrium Model for the Mexican
Hernndez Laos, Enrique 2004, Matriz de Economy, USA, Tesis Doctoral. Yale
Insumo-Producto, Mxico, Consultora University.
Internacional Especializada.
Sistema de Informacin Comercial de
Hicks, John Richard 1979, Value and Capital: Mxico 2005. World Trade Atlas,
an Inquiry into some Fundamental Mexico, Bancomext.
Principles of Economic Theory, USA,
Clarendon Press. 2a. Edicin. Varian 1998, Microeconoma Intermedia,
Espaa, Antoni Bosch, 4a. Edicin.
INEGI 2000, Encuesta Nacional de Ingresos
y Gastos de los Hogares 2000, Mexico,
INEGI.
A. Appendix I
Table 19. Parameters of Supply Equations Table 20. Parameters of Investment Equations
Sector a i
vai ci Sector q i
f i
sec1 0.17561 0.63128 1.59149 sec1 0.00042 0.00037

sec2 0.11444 0.77617 1.42716 sec2 0.00012 0.00026
sec3 0.13295 0.42451 1.47988 sec3 0.00027 0.01519
sec4 0.16480 0.50405 1.56447 sec4 0.00001 0.00102
sec5 0.22899 0.45909 1.71267 sec5 0.00002 0.00181
sec6 0.20440 0.58636 1.65939 sec6 0.00001 0.00018
sec7 0.35239 0.36942 1.91346 sec7 0.00067 0.01785
sec8 0.22739 0.40344 1.70934 sec8 0.00926 0.01070
sec9 0.25078 0.35198 1.75627 sec9 0.00004 0.00150
sec10 0.33532 0.45691 1.89248 sec10 0.00002 0.00112
sec11 0.30917 0.37475 1.85602 sec11 0.00012 0.00789
sec12 0.23615 0.55920 1.72737 sec12 0.00001 0.00171
sec13 0.16418 0.54481 1.56289 sec13 0.00013 0.00364
sec14 0.12368 0.38328 1.45381 sec14 0.00012 0.00303
sec15 0.27426 0.30853 1.79939 sec15 0.17141 0.19482
sec16 0.30558 0.37464 1.85062 sec16 0.00547 0.01164
sec17 0.56895 0.45477 1.98101 sec17 0.63421 0.56809
sec18 0.55096 0.35955 1.98962 sec18 - -
sec19 0.25227 0.72381 1.75911 sec19 0.14124 0.12651
sec20 0.27815 0.65722 1.80615 sec20 0.03495 0.03131
sec21 0.11137 0.73392 1.41818 sec21 - -
sec22 0.37352 0.69601 1.93635 sec22 0.00151 0.00135
sec23 0.84334 0.89316 1.54356 sec23 - -

Table 21. Parameters of Private Consumption Equations
Sector I II III IV V VI VII VIII IX X

Sec1 0.09799 0.08956 0.07335 0.06197 0.05404 0.04300 0.03766 0.03383 0.02644 0.01210
Sec2 0.00006 0.00009 0.00004 0.00004 0.00002 0.00002 0.00001 0.00003 0.00002 0.000005
Sec3 0.50182 0.37903 0.34771 0.27098 0.24123 0.19280 0.16301 0.13130 0.09464 0.03761
Sec4 0.03132 0.02610 0.02712 0.02343 0.02251 0.01826 0.01647 0.01356 0.00949 0.00345
Sec5 0.04372 0.04384 0.01811 0.02856 0.03821 0.01910 0.02870 0.01855 0.01516 0.01878
Sec6 0.00528 0.00215 0.00324 0.00310 0.00292 0.00157 0.00249 0.00232 0.00153 0.00105
Sec7 0.02171 0.02609 0.02429 0.02551 0.02429 0.02530 0.02401 0.02536 0.02574 0.01635
Sec8 0.00196 0.00218 0.00340 0.00333 0.00354 0.00521 0.00456 0.00468 0.00587 0.00329
Sec9 0.00015 0.00016 0.00034 0.00123 0.00347 0.00246 0.00281 0.00390 0.00496 0.00350
Sec10 0.01385 0.01057 0.01019 0.00802 0.00776 0.00685 0.00684 0.00627 0.00588 0.00390
Sec11 0.01887 0.02828 0.03215 0.02879 0.03025 0.02906 0.02913 0.02893 0.02852 0.01995
Sec12 0.03951 0.02635 0.02813 0.02299 0.01773 0.01511 0.01592 0.01313 0.01224 0.00841
Sec13 0.00341 0.00806 0.00375 0.00674 0.00574 0.00587 0.01221 0.00794 0.00573 0.00576
Sec14 0.00090 0.00064 0.00052 0.00039 0.00031 0.00025 0.00021 0.00016 0.00012 0.00004
Sec15 0.00702 0.01082 0.01082 0.00923 0.01537 0.02975 0.01762 0.01947 0.03236 0.04396
Sec16 0.00116 0.00107 0.00111 0.00117 0.00140 0.00133 0.00135 0.00175 0.00175 0.00143
Sec17 - - - - - - - - -
Sec18 0.00990 0.01477 0.01945 0.01991 0.01749 0.01643 0.01600 0.01444 0.01441 0.00855
Sec19 0.04470 0.09451 0.09667 0.12393 0.15585 0.19207 0.21978 0.24946 0.27691 0.23591
Sec20 0.09811 0.15723 0.17270 0.19480 0.19174 0.20790 0.20171 0.19970 0.18216 0.10218
Sec21 0.02803 0.03574 0.05615 0.09913 0.09154 0.10583 0.13487 0.15334 0.16494 0.20364
Sec22 0.01031 0.01276 0.04403 0.03943 0.05111 0.06126 0.02729 0.03549 0.03044 0.24764
Sec23 0.02020 0.02970 0.02672 0.02730 0.02347 0.02057 0.03733 0.03639 0.06070 0.02248
Parameter I II III IV V VI VII VIII IX X
fi 0.65035 0.57965 0.58087 0.63930 0.67122 0.72102 0.67650 0.70500 0.77608 0.82623
Table 22. Parameters of Foreign Sector Equations Table 23. Parameters of Equilibrium Market of
Goods Equations
Sector m i
Sector bi
sec1 1.00000 sec1 1.11512

sec2 0.99997 sec2 1.06075
sec3 1.00000 sec3 1.04768
sec4 1.00000 sec4 1.01857
sec5 1.00000 sec5 1.00000
sec6 1.00000 sec6 1.00000
sec7 0.99147 sec7 1.28283
sec8 0.99920 sec8 1.12771
sec9 0.98527 sec9 1.47036
sec10 0.92398 sec10 1.17911
sec11 0.99829 sec11 1.54715
sec12 0.89820 sec12 1.48363
sec13 0.99907 sec13 1.12076
sec14 0.88410 sec14 1.66428
sec15 0.99999 sec15 1.36485
sec16 0.63145 sec16 1.94048
sec17 1.00000 sec17 1.00000
sec18 1.00000 sec18 1.03592
sec19 1.00000 sec19 1.00077
sec20 1.00000 sec20 1.03320
sec21 1.00000 sec21 1.00477
sec22 1.00000 sec22 1.00081
sec23 1.00000 sec23 1.00000
Table 24. Exogenous Variables
% % % % d
K0s Ls0 F r*
3,156,911 1,526,705 -357,433 0.07 18.596

B. Appendix II
Table 25. Exogenous Variables Table 26. Production Sector. (Hundred Billion of
Pesos)
Sector Pi* Ei Ygi Sector Li Ki VAi Xsi
sec1 1 28,825 2,118 sec1 0.356 1.669 2.025 3.207

sec2 1 24,649 39 sec2 0.155 1.203 1.359 1.75
sec3 1 29,492 582 sec3 0.345 2.247 2.592 6.106
sec4 1 12,385 1 sec4 0.045 0.23 0.275 0.546
sec5 1 682 15 sec5 0.083 0.279 0.362 0.789
sec6 1 1,486 - sec6 0.01 0.038 0.048 0.082
sec7 1 57,579 2,507 sec7 0.343 0.631 0.975 2.638
sec8 1 8,200 36 sec8 0.07 0.238 0.309 0.765
sec9 1 3,630 1,811 sec9 0.061 0.182 0.243 0.689
sec10 1 5,123 907 sec10 0.088 0.174 0.261 0.571
sec11 1 62,287 2,140 sec11 0.431 0.963 1.393 3.718
sec12 1 7,810 34 sec12 0.097 0.315 0.413 0.738
sec13 1 16,258 2,265 sec13 0.125 0.634 0.759 1.393
sec14 1 23,756 17 sec14 0.072 0.513 0.585 1.527
sec15 1 488,702 3,423 sec15 1.227 3.246 4.473 14.498
sec16 1 30,675 6,717 sec16 0.116 0.263 0.378 1.009
sec17 1 - - sec17 1.476 1.118 2.594 5.705
sec18 1 488 6,280 sec18 0.253 0.207 0.46 1.28
sec19 1 219,276 5,844 sec19 2.634 7.808 10.443 14.427
sec20 1 51,980 9,497 sec20 1.518 3.939 5.457 8.303
sec21 1 - 11,024 sec21 0.616 4.918 5.534 7.54
sec22 1 18,117 129,486 sec22 2.681 4.496 7.177 10.311
sec23 1 - 217,799 sec23 2.465 0.458 2.923 3.272
Table 27. Government Expenditures. (Hundred Table 29. Saving and Income by Population Deciles.
Billion of Pesos) (Hundred Billion of Pesos)
Sector Ygi Decil Zpi Spi

sec1 0.021 dec1 1.005 0.351
sec2 0.000 dec2 1.584 0.666
sec3 0.006 dec3 1.946 0.816
sec4 0.000 dec4 2.356 0.85
sec5 0.000 dec5 2.808 0.923
sec6 0.025 dec6 3.31 0.923
sec7 0.000 dec7 4.152 1.343
sec8 0.018 dec8 5.241 1.546
sec9 0.009 dec9 6.515 1.459
sec10 0.021 dec10 18.45 3.208
sec11 0.000
sec12 0.023
sec13 0.000
sec14 0.034
sec15 0.067
sec16 0.063
sec17 0.058
sec18 0.095
sec19 0.110
sec20 1.295
sec21 2.178
sec22 0.000
sec23 0.000
Table 28. Government Revenues and Saving.

(Hundred Billion of Pesos)
Variables Valuesi
Zg 5.557
Sg 1.531
Revenues as % of GDP 10.5%

Table 30. Consumption of Population Deciles (Hundred Billion of Pesos)
Sector I II III IV V VI VII VIII IX X

Sec1 0.064 0.082 0.083 0.093 0.102 0.103 0.106 0.125 0.134 0.185
Sec2 4.00E-05 8.00E-05 5.00E-05 6.00E-05 4.00E-05 4.00E-05 4.00E-05 1.00E-04 9.00E-05 7.00E-05
Sec3 0.328 0.348 0.393 0.408 0.455 0.46 0.458 0.485 0.479 0.574
Sec4 0.02 0.024 0.031 0.035 0.042 0.044 0.046 0.05 0.048 0.053
Sec5 0.029 0.04 0.02 0.043 0.072 0.046 0.081 0.069 0.077 0.286
Sec6 0.003 0.002 0.004 0.005 0.005 0.004 0.007 0.009 0.008 0.016
Sec7 0.014 0.024 0.027 0.038 0.046 0.06 0.067 0.094 0.13 0.249
Sec8 0.001 0.002 0.004 0.005 0.007 0.012 0.013 0.017 0.03 0.05
Sec9 1.00E-04 1.50E-04 3.90E-04 0.002 0.007 0.006 0.008 0.014 0.025 0.053
Sec10 0.009 0.01 0.012 0.012 0.015 0.016 0.019 0.023 0.05 0.059
Sec11 0.012 0.026 0.036 0.043 0.057 0.069 0.082 0.107 0.144 0.304
Sec12 0.026 0.024 0.032 0.035 0.033 0.036 0.045 0.045 0.062 0.128
Sec13 0.002 0.007 0.004 0.001 0.011 0.014 0.034 0.029 0.029 0.088
Sec14 5.90E-01 5.90E-04 5.90E-04 5.90E-04 5.90E-04 6.00E-04 6.00E-04 6.00E-04 6.00E-04 6.00E-04
Sec15 0.005 0.01 0.012 0.014 0.029 0.071 0.049 0.072 0.164 0.67
Sec16 7.60E-04 9.80E-04 0.001 0.002 0.003 0.003 0.004 0.006 0.009 0.022
Sec17 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00
Sec18 0.006 0.014 0.022 0.03 0.033 0.039 0.45 0.058 0.073 0.13
Sec19 0.029 0.087 0.109 0.187 0.294 0.458 0.617 0.922 1.4 3.598
Sec20 0.064 0.144 0.195 0.298 0.361 0.496 0.567 0.738 0.921 1.558
Sec21 0.018 0.033 0.063 0.149 0.173 0.258 0.379 0.567 0.834 3.106
Sec22 0.007 0.012 0.05 0.059 0.096 0.146 0.077 0.131 0.154 3.777
Sec23 0.013 0.027 0.03 0.041 0.044 0.049 0.105 0.134 0.307 0.343
Table 31. Investment. (Hundred Billion of Pesos) Table 33. Foreign Sector. (Hundred Billion of Pesos)
Sector YiBF YiE Yil Sector Mi Ei
sec1 0.004 0 0.004 sec1 0.311 0.288

sec2 0.001 0.002 0.003 sec2 0.063 0.246
sec3 0.002 0.15 0.153 sec3 0.247 0.295
sec4 7.00E-05 0.01 0.01 sec4 0.005 0.124
sec5 1.40E-04 0.018 0.018 sec5 0 0.007
sec6 5.00E-05 0.002 0.002 sec6 0 0.015
sec7 0.006 0.173 0.179 sec7 0.36 0.576
sec8 0.083 0.024 0.107 sec8 0.056 0.082
sec9 3.20E-04 0.015 0.015 sec9 0.219 0.036
sec10 2.00E-04 0.011 0.011 sec10 0.006 0.051
sec11 0.001 0.078 0.079 sec11 1.453 0.623
sec12 8.00E-05 0.017 0.017 sec12 0.132 0.078
sec13 0.001 0.035 0.037 sec13 0.092 0.163
sec14 0.001 0.029 0.03 sec14 0.509 0.238
sec15 1.542 0.415 1.956 sec15 3.126 4.887
sec16 0.049 0.068 0.117 sec16 0.442 0.307
sec17 5.705 0 5.705 sec17 0 0
sec18 0 0 0 sec18 0.041 0.005
sec19 1.27 0 1.27 sec19 0.007 2.193
sec20 0.314 0 0.314 sec20 0.234 0.52
sec21 0 0 0 sec21 0.031 0
sec22 0.014 0 0.014 sec22 0.007 0.181
sec23 0 0 0 sec23 0 0
Table 32. Tobin System. (Hundred Billion of Pesos)
Variable Original Equilibrium

r* 0.07
0.554
Pk 1.806
I BF 8.995
18.596
F
-3.574

Table 34. Intermediate Demand. (Hundred Billion of Table 35. Investment. (Hundred Billion of Pesos)
Pesos)
Sector Vi Sector Di Xid XiS

sec1 2.13 sec1 2.919 3.207 3.207
sec2 1.563 sec2 1.504 1.75 1.75
sec3 1.512 sec3 5.811 6.106 6.106
sec4 0.024 sec4 0.422 0.546 0.546
sec5 0.002 sec5 0.782 0.789 0.789
sec6 0.003 sec6 0.067 0.082 0.082
sec7 1.467 sec7 2.062 2.638 2.638
sec8 0.49 sec8 0.683 0.765 0.765
sec9 0.723 sec9 0.653 0.689 0.689
sec10 0.301 sec10 0.52 0.571 0.571
sec11 3.566 sec11 3.096 3.718 3.718
sec12 0.305 sec12 0.66 0.738 0.738
sec13 1.034 sec13 1.231 1.393 1.393
sec14 1.762 sec14 1.289 1.527 1.527
sec15 9.65 sec15 9.611 14.498 14.498
sec16 0.909 sec16 0.703 1.009 1.009
sec17 0 sec17 5.705 5.705 5.705
sec18 0.808 sec18 1.275 1.28 1.28
sec19 3.211 sec19 12.235 14.427 14.427
sec20 2.27 sec20 7.783 8.303 8.303
sec21 1.887 sec21 7.54 7.54 7.54
sec22 4.32 sec22 10.13 10.311 10.311
sec23 0 sec23 3.272 3.272 3.272

2010 Vali Khodadadi, Abolfazl (Parviz) Zandinia and Marzieh Nouri 89
APPLICATION OF ANTS COLONY SYSTEM FOR

BANKRUPTCY PREDICTION OF COMPANIES
LISTED IN TEHRAN STOCK EXCHANGE
Vali Khodadadi, Abolfazl (Parviz) Zandinia and Marzieh Nouri
Abstract
One of the most recent researches in financial field is using an Ant Colony System (ACS) in the
corporate bankruptcy prediction.
In this paper, the Ant Colony System is used to predict the bankruptcy of listed companies in Tehran
Stock Exchange. The research findings shows that models prediction power is about 59% with taking into
account the rations proposed by Altman (1968). After examining the causes of significant errors in model,
a proposed model is introduced. This model which is based on four financial ratios fits the economical
infrastructure of Iran. The results obtained from testing of the suggested model indicate that this model is
able to predict the corporate bankruptcy one year prior to bankruptcy with an accuracy of 93% for large-
sized firms and about 90% for small-sized ones, respectively.
Khodadadi V., Zandinia A., Npuri M. - Application of Ants Colony System for Bankruptcy Prediction of Companies listed in Tehran Stock Exchange
Introduction Ants behavior
Human foraging behavior always In real world ants run more or less at
seeks to approach an integrated set of random around their colony to search for
information and to discover relationships food. Ants deposit a chemical substance
between phenomena for the explanation called pheromone along the travelled paths
of these phenomena and the prediction of (Dorigo and Caro, 1999). After rainfall,
their behaviors. These attempts extended to these paths get white and become visible.
financial field and have induced researchers Other ants upon finding these paths would
to discover relationships between financial follow the trail. Then, if they discover
information in order to provide the financial food source they will return to the nest and
information users with more powerful deposit another path along with the previous
control tools. one, and will be strengthened the preceding
A highly valuable effort made by path.
financial researches has been the application When it turns to select a path, ants
of financial ratios in bankruptcy prediction. will prefer one containing high density of
The research conducted in this field pheromone, i.e. paths on which more ants
resulted in proposing numerous models have travelled. Pheromone is gradually
for the prediction of corporate bankruptcy. evaporated. This evaporation is useful for
The continuity of these researches led to three reasons:
introduction of using artificial intelligence It makes the next trail less attractive for
(AI) and computer science techniques in the following ants. Ants mostly take shorter
field of finance. paths; hence, the shorter path between nest
One of the techniques introduced is and food source is highly strengthened and
the application of Ant Colony System in any farther path is less strengthened.
bankruptcy prediction (Milea, 2005). This Unless pheromone did not evaporate at
model was initially proposed by Dorigo all, the paths having been passed many times
(1992) to solve the Traveling Salesman would become so attractive that the random
Problem (TSP) (Dorigo and Stutzle, 2004). searching for food would be highly limited.
Thereafter, researches tended to benchmark When food finished at the end of path,
this model to solve similar optimization evaporation of the remaining paths would
problems in their field of study. The not mislead ants in their searching for that
researchers of financial field were also to track leading to food source.
test this model in order to find out a solution Therefore, when an ant finds a shorter
for financial problems. (optimal) path from nest to food source, other
In this research, the applicability of ants will more likely follow the same path
this model in the prediction of corporate and, over time, the continuous strengthening
bankruptcy will be explained and the of the path and the evaporation of other trails
coefficient of success of this model will be will make all ants become unidirectional
tested. For this reason, the financial ratios (Dorigo and Colorni, 1996).
of corporate have filed in Tehran Stock This behavior in ants has a kind of swarm
Exchange in the period 2001-2007 were intelligence having been recently considered
used. by scientists. But it should be reassured that
there is a major difference between swarm
intelligence and social intelligence. In social

intelligence the agents have a degree of Probabilistic Transition Rule

intelligence. But, in swarm intelligence the
agents have random behavior and there is no A probabilistic transition rule pij (t) , the
k
direct relationship between them and they probability that ant k will go from i to j at
are in indirect contact with each other by iteration t, is used by the ants in building
signals. This behavior was first studied by their tours. This probability depends on
Grass, a French scientist. It was referred 2 parameters: a heuristic measure of the
to as Stigmergy which is a particular form desirability of adding edge (i,j) to the current
of indirect communication used by social tour, ij, and the amount of pheromone
insects to coordinate their activities via currently on edge (i,j), ij. Probabilistic
changes made in the local environment transition rule is given as (2) below:
(Dorigo et.al, 2000).
(2)
Ant Behavior System
Ant Colony System was initially

proposed by Dorigo and Caro to solve the
Traveling Salesman Problem (TSP) and Where and are adjustable parameters
with benchmarking of ants behavior he and Jk(i) is the set of cities that remain to be
designed a powerful algorithm for solving visited by ant k.
optimization problems. In TSP the aim is to Because in the beginning of the simulation
find a closed tour of optimal path connecting the paths generated by the ants are mostly
n cities (Dorigo and Stutzle, 2004). random, pheromone evaporation should
Therefore, the path in TSP is a directional take place, thus avoiding convergence to a
graph, also referred to as Hamilton cycle. local optimum.
(3)
An optimal tour in weighted directional
graph means a path with minimal length.
Thus, the TSP is, in fact, finding an optimal
tour for a weighted directional graph. Also, Where r is the coefficient of evaporation
the length of optimal tour does not depend (0 # t # 1) and Oxijk is given by:
(4)
on the selection of Start vertex.
Pheromone in Ant Colony

System Where m being the number of ants (Dorigo
and Stutzle, 2004). In Ant Colony System,
In Ant Colony System, the quantity of the parameters should take the values: =1,
pheromone Oxijk (t) deposited by ant k on =5 and =0.5 (Aziz et.al, 1998).
each edge (i,j) of the tour Tk(t) is as follows:
(1) Ant Colony System Algorithm
The algorithm for Ant Colony System is

as follows:
Where Q is an adjustable parameter
and Lk is the length of the tour (Dorigo and
Stutzle, 2004).
1. The initialization step: Current Ratio are more effective than others
(Balcaen and Ooghe, 2006).
In this step ants are placed in the initial Followed by these studies, a great number
position and an amount of 0 is taken for of researchers involved in examining
initial pheromone. the prediction of corporate bankruptcy
based on financial ratios which resulted
2. Tour calculation: in the presentation of four univariate
analysis methods, risk index model,
Distance between parameters is multiple discriminant analysis models
calculated. and conditional probability models for
bankruptcy prediction. In 1967, Beaver was
3. Model estimation: pioneer to examine the corporate bankruptcy
using univariate analysis method (Balcaen
If the system output is not satisfactory, and Ooghe, 2006).
then ants will move to the next location Tamari in 1966 presented risk index
based on the probabilistic transition function model to predict bankruptcy and thereafter
and pheromone is converged. Moses and Liao (1987), also applied this
model to predict corporate bankruptcy.
4. Repeating Criterion: In reference to studies made by Beaver,
Altman proposed a more accurate model to
Procedure continues until the system predict firms risk of bankruptcy with using
output reaches a satisfactory position Multiple Discriminant Analysis (MDA)
(Dorigo and Stutzle, 2004). (Altman, 1968). After his work, some
extensive researches were also conducted
Literature Review about bankruptcy prediction based on
financial ratios and applying multiple
1. Classic Methods of Bankruptcy discriminant analysis (MDA). Altman
Prediction (1968), Deakin (1972), Edmister (1972),
Blum (1974), Altman et al. (1977), Deakin
A primary study was made on bankruptcy (1977), Taffler and Tisshaw (1977), Van
prediction by Whitaker and Smith in 1935. Frederikslust (1978), Bilderbeek (1979),
They involved investigation about the Dambolena and Khoury (1980), Taffler
sufficiency of financial ratios as predictors of (1982), Ooghe and Verbaere (1985), Taffler
financial failures. The two aforementioned (1983), Micha (1984), Betts and Belhoul
researches had been analyzing the average (1987), Gombola et al. (1987), Gloubos and
procedure of 21 rations for 10 years since Grammatikos (1988), Declerc et al. (1991),
then and concluded that Working Capital/ Laitinen (1992), Lussier and Corman
Assets has more stability and accuracy and (1994), Altman et al. (1995) are among those
it is significant as an index to identify failure. researchers who one after the other proposed
Merwin (1942) also examined several bankruptcy prediction models using multiple
ratios within first 6 years of the lifespan discriminant analysis technique.
of corporate with continuous activity and Beginning in the 1980s, the application
those whose activity had been stopped. of Multiple Discriminant Analysis (MDA)
He concluded that three ratios of Working method was declined but it still remained
Capital/Total Assets; Net Equity/Debt and

as a standard accepted method and was Discriminant Analysis Models (ADM)

continuously applied for comparative with artificial neural networks in terms of
studies. Then, MDA technique was bankruptcy prediction (Cybinski, 2001).
substituted by other statistical techniques One of the most recent studies in the field
such as Logit and Probit analysis and of bankruptcy prediction is the application
linear probability modeling. These methods of Ant Colony System introduced by Dorigo
resulted in the presentation of conditional to solve the travelling salesman problem.
probability models including a combination With application of this model to predict
of variables which highly discriminated the financial status of corporate for the first
bankrupted firms from non-bankrupted time, Wang, Zhao and Kang, 2004, took a
ones (Balcaen and Ooghe, 2006). Ohelson new step to introduce artificial intelligence
(1980); Zavgren (1983); Zmijewski (1984); to the field of financial researches. Then,
Gentry et al (1985); Zavgren (1985); Keasey in 2005, Viorel Milea made some small
and Vatson (1987); Peel and Peel (1987); changes in Ant Colony System model to
.Swanson and Tybout (1988); Aziz et al. predict firms bankruptcy and, then, he
(1988);Keasey and McGuinness (1990); compared this model with Altmans classic
Platt and Platt (1990); Ooghe et al (1993); model. The choice for the financial ratios
Sheppard (1994); Lussier (1995); Mossman used in this work is based on the study by
et al. (1998); Chritou and Trigeorgis (2000); Altman. Results obtained in Milea research
Becchetti and Sierra (2002) as well as Chritou showed that Ant Colony System is able to
et al. (2000) are among those researchers predict firms bankruptcy with a coefficient
involved in bankruptcy prediction based on of success of 80% and it, as compared with
conditional probability models. Altman model, provides a better prediction
in some of the studies. Thus, users are able
2. New Methods to predict Corporate to predict firms financial status with having
Bankruptcy very limited statistical data at hand and in
the shortest time possible (Milea, 2005).
As science and technology advances, the
financial information users need to utilize Research Design
new and stronger models and methods for
decision making is raised. As described In this research we have decided to
above, an issue on which a proper and predict the bankruptcy of listed companies
accurate decision should be made is the in Tehran Stock Exchange. This is done
corporate bankruptcy prediction for the by first running test data for measuring the
purpose of right logical investment. accuracy of the algorithm in separating
Artificial neural networks have been the known bankrupt firms from the non-
considered as a proposed model for corporate bankrupt firms. To reach our goal we
bankruptcy prediction and its application fund that the current algorithm does not
in bankruptcy prediction proved that this fit our situation and should be modified to
method is able to predict corporate financial have reasonable answer. In our modified
status with a very high accuracy. Odom and algorithm there is one point corresponding
Sharda (1990); Koster, Sandak and Bourbia to each firm. Therefore, in this research the
(1990); Cadden (1991); Coatsand Fant number of ants is as the same as the number
(1993); Lee, Han and Kwon (1996) were of the firm under the study.
among those researchers who compared
Studied Variables shall be obliged to call the shareholders

extraordinary general meeting to decide for
Independent variables used in this model the survival or winding up of the company.
and included in financial ratios applied by If the said meeting does not decide
Altman (1968) in his ZETA model are as for winding up of the company, it will be
follows: required to decrease the companys capital
X1 = Working Capital/Total Assets, by the current amount at the same meeting
X2 = Retained Earnings/Total Assets, in virtue of the regulations as stipulated in
X3 = EBIT/Total Assets, article 6 of the Commercial Law of Iran.
X4 = Market Value of Equity/Book Value Otherwise firm considered as Bankrupt.
of Total Debt,
X5 = Sales/Total Assets Studied Sample Firms
Classification Rules In this research three sample firms

consisted of 36 large-size firms, 36 small-
In Ant Colony System applied in present size firms and a control sample composed of
research, ants, after running at random, 42 firms (21 bankrupt and 21 non-bankrupt
would find the shortest route among the firms) are tested. The goal has been not only
paths (financial parameters) to identify to test the prediction power of the algorithm,
bankruptcy. The resulting path will be but to test its effectiveness on both small and
presented as a criterion to classify firms into large size firms.
bankrupt and non-bankrupt as follows:
(5)
Research Findings
R = {CX1, CX2, CX3, CX4, CX5} Application of Ant Colony System model
in the studied sample of large and small size
(6)
firms based on the five financial ratios of
Xki # CXi, i ! "1, 2, 3, 4, 5, sample firms led to the following results as
shown in Tables 1 and 2.
Where CXn represents the cut-point value
of financial ratio Xn for which the fitness Table 1. Models Results in Large- sized firms
function is maximized. So, a firm with values
smaller or equal to R for each financial ratio
2002
2003
2004
2005
2006
2007
Description (No.)
is predicted to go bankrupt otherwise not;
and k is the given firm (Milea, 2005). Actual Bankrupt 2 2 2 1 5 5
Situation
Non-Bankrupt 34 34 34 35 31 31
Explanation of Bankruptcy Model Bankrupt 0 0 0 0 2 1
Concept in Iran Prediction
Non-Bankrupt 36 36 36 36 34 35
According to article 141 of commercial

law of Iran corporate are classified in two
groups of bankrupt and non-bankrupt firms
as follows:
Where, due to incurred losses, the half of
corporate capital is lost the board of directors

Table 2. Models Results in Small- sized firms Proposed model
The failure of the original model in
2002
2003
2004
2005
2006
2007
Description (No.)
predicting the bankruptcy status of the
Actual Bankrupt 5 6 8 9 8 5 Tehran Stock Exchange, especially in the
Situation
Non-Bankrupt 31 30 28 27 28 31 case of small-size firms, forced us to closely
Model Bankrupt 0 0 1 1 1 0 examine the model. We found out that in the
Prediction
Non-Bankrupt 36 36 35 35 35 36 original model:
The tables show that this model nearly is 1. Non-fitness of bankruptcy identification
not able to predict bankrupt firms, but it can criterion in Iran with the ratio of Market
only predict non-bankrupt firms. Value of Equity/ Book Value of Total
To be confident of the accuracy of the Debt:
obtained results and to examine whether
these results are affected by unequal number As described in previous sections,
of bankrupt and non-bankrupt firms or not, the bankruptcy of companies in Iran is
we take a sample consisting of 42 bankrupt identified in article 141 of the Commercial
and non-bankrupt firms (21 bankrupt and 21 Law. Hence, the classification criterion to
non-bankrupt firms) as control sample and identify firms into two groups of bankrupt
the results are as shown in Table 3. and non-bankrupt is the amount of firms
capital. With glancing over the concerned
Table 3. Models Results in Control sample firms independent variables in order to test Ant
Colony System for bankruptcy prediction we
Description (No.) Bankrupt
Non- will find out that this criterion has not been
Bankrupt
considered in any of independent variables
Actual Situation 21 21
and, instead, Market Value of Equity ratio
Model Prediction 4 38
has been expressed in terms of fluctuation
Correct prediction 4 21
which, in spite of firms belonging to
Percentage of correct
prediction
59.52% samples with equal sizes, would be high in
studied samples.
2. Inflation Impact on Sales/Book Value of

The results of the tested sample show Total Assets:
that for even for equal number of bankrupt
and non-bankrupt samples the coefficient of In the calculation of Sales/Book Value of
success of the model has reduced to 59.52%. total assets, sales at current value as well
This result shows that the prediction power as assets value are historical costs. Under
concerning bankrupt firms is as low as 19%. the Iran economic conditions with very
Therefore, it is concluded that the original high inflation rate as well as instability of
Ant Colony System model, based on the five economic conditions, sales are very inflated
original financial rations cannot discriminate rather than book value of total assets. In
bankrupt firms registered in Tehran stock order to make modifications to the influence
exchange above 59.52%. of inflation on Sales/Book Value of Total
Assets, revaluation of firms assets enabling
this ratio to express firms status in a more
reasonable manner is necessary. But, due of firms, we, firstly, test the model for the
to lacking any legal obligation concerning control sample composed of 42 corporate
listed companies in Tehran Stock Exchange (21 bankrupt and 21 non-bankrupt firms).
to revaluate their assets, this ratio will The results obtained from control testing are
remain as a problematic parameter whenever as shown in Table 4.
financial problems are to be solved. Also,
as a result of monopoly for some products Table 4: Results of proposed model in Control
such as steel and automobile, the above ratio sample firms
cannot be an appropriate scale to compare
firms. Description (No.) Bankrupt
Non-
Bankrupt
The following measures were taken to
Actual Situation 21 21
resolve the above-mentioned issues:
Model Prediction 21 21
Correct prediction 19 19
1. Due to using amount of Capital for
classifying corporate into two groups prediction
90.47%
of bankrupt and non-bankrupt firms, we
have used Book Value of Equity/Book The results displayed in Table 4 indicate
Value of Total Debts instead of Market that this model is able to predict up to 90%
Value of Equity/ Book Value of Total of corporate financial status.
Debts. After examining of proposed model in
2. As a result of very highly inflated Sales/ control sample and being confident of the
Book Value of Total Assets as well as model accuracy within the same sample,
lacking any legal obligation for Listed we tested the new suggested model for two
Companies in Tehran Stock Exchange to main samples, i.e. sample of large- and
revaluate their assets and the monopoly small-sized firms. The results obtained from
for some products, we eliminate this testing of proposed model are as shown in
ratio from the model. Tables 5 and 6.
Thus, we examine Ant Colony System Table 5: Proposed models Results in Large- sized
model with new conditions for the samples firms
having been previously tested. The new
suggested model to predict the bankruptcy of
2002
2003
2004
2005
2006
2007
Description (No.)
listed companies in Tehran Stock Exchange
is composed of four independent variables Actual Bankrupt 2 2 2 1 5 5
Situation
as follows: Non-Bankrupt 34 34 34 35 31 31
X1= Working Capital/Total Assets Model

Prediction
Bankrupt 5 1 3 1 2 4
X2= Retained Earnings/Total Assets Non-Bankrupt 31 35 33 35 34 32
X3= EBIT/Total Assets Correct prediction (No.) 33 35 33 34 33 33

X4= Book Value of Equity/Book Value of
91.67
97.22
91.67
94.44
91.67
91.67
prediction
Total Debts Average of correct
Due to unequal number of bankrupt and predictions per.
93.06%
non-bankrupt corporate in both studied

samples (large- and small-sized firms) and The results of the test made on proposed
in order to be confident that the proposed model indicate a coefficient of success of
model is not affected by unequal number

93% for the model to predict the bankruptcy After these variables were tested in the
of large-sized corporate. model we found out that this model is not
After testing of proposed model in the likely able to predict non-bankrupt firms
sample of large-sized firm, the model was and its real prediction power is about 59%.
also tested for small-sized firms and the The investigations showed that the system
following results were obtained: error was caused by the two ratios of Market
Value of Equity/Book Value of Total Debts
Table 6: Proposed models Results in Small- sized as well as Sales/Total Assets. In virtue
firms of article 141 of the Commercial Law of
Iran and choosing capital as a criterion for
2002
2003
2004
2005
2006
2007
Description (No.) corporate bankruptcy, using Market Value
Actual Bankrupt 5 6 8 9 8 5
of Equity is not logical. Moreover, this will
Situation
Non-Bankrupt 31 30 28 27 28 31
still remain as a problematic ratio because
Model Bankrupt 3 5 5 4 7 1 of lacking any legal obligation for listed
Prediction
Non-Bankrupt 33 31 31 32 29 35 companies in Tehran Stock Exchange for
Correct prediction (No.) 34 35 33 29 33 30 assets revaluation. Also, as a result of very
Percentage of correct high inflation rate in Iran, instable economic
94.44
97.22
91.67
80.56
91.67
83.33
prediction
situations as well as monopoly for some
Average of correct
89.82% products such as steel and automobile, the
predictions per.
ratio of Sales/Total Assets is very inflated
The results of the test made on proposed and may not be regarded as an appropriate
model indicate a coefficient of success scale to compare firms.
of 89.82% for the model to predict the In order to resolve above-mentioned
bankruptcy of small-sized firms. matters, we consider Book Value of Equity
Upon comparison of the results of the test instead of Market Value of Equity and, also,
made on proposed model we found out that Sales/Total Assets was omitted from among
this system is slightly affected by corporate variables. At last, a proposed model consisted
size. So that the models coefficient of of four variables as described below was
success in large-sized corporate group is proposed to predict the bankruptcy of listed
about 93% while in small-sized corporate companies in Tehran Stock Exchange based
group is about 90%, respectively. on Ant Colony System:
Consequently, as the results of Working Capital/Total Assets; Retained
investigations show, the proposed model is Earnings/Total Assets; EBIT /Total Assets;
able to predict the bankruptcy of corporate Book Value of Equity/Book Value of Total
one year prior to bankruptcy. Debts.
Conclusion and Summarization The results produced by the testing of
This study aims at the application of Ant new suggested model indicated that the
Colony System for bankruptcy prediction prediction power of this model within the
of corporate accepted in Tehran Stock sample of large-sized firms is about 93%
Exchange. and in the sample of small-sized firms is
The primary variables used in this model about 90%, respectively. Therefore, this
include Working Capital/Total Assets; model is able to predict the bankruptcy of
Retained Earnings/Total Assets; EBIT/Total listed companies in Tehran Stock Exchange
Assets; Market Value of Equity/Book Value with a very high coefficient of success and it
of Total Debts; Sales/Total Assets. is slightly affected by corporate size.
Suggestions and Prospective of and their related problem. The British

Future Researches Accounting Review. 38: 63-93.
As the Ant Colony System is a dynamic Becchetti L, Sierra J (2003). Bankruptcy

model having a very high coefficient of risk and productive efficiency in
accuracy to search the answer of optimization manufacturing firms. Journal of Banking
problems, this model is recommended for and Finance. 27:20992120.
these types of problems. Also, in order to
be confident of the percentage of influence Betts J. Belhoul D (1987). The effectiveness
made by the elimination of Sales/Total of incorporating stability measures in
Assets it is suggested to reconsider this company failure models. Journal of
ratio within the model with considering the Business Finance and Accounting. 14
assets revaluation problem. Also, it has been (3):323334.
proposed to investigate about the corporate
bankruptcy with using Ant Colony System Bilderbeek J (1979). An empirical study of
and applying phase rules in Iran. the predictive ability of financial ratios
in the Netherlands. Zeitschrift Fur
References: Betriebswirtschaft: 388407.
Altman E I (1968). Financial ratios, Blum M (1974). Failing company

discriminant analysis and the prediction discriminant analysis. Journal of
of corporate bankruptcy. The Journal of Accounting Research. 12 (1):125.
finance.23: 589-609.
Charitou A, Trigeorgis L (2000). Option-
Altman E I, Haldeman R G, Narayanan P based bankruptcy prediction. Paper
(1977). ZETA analysis: a new model to presented at 6th Annual Real Options
identify bankruptcy risk of corporations. Confernce, Paphos.Cyprus.46 July:
Journal of Banking and Finance. 1 (1): 125.
2951.
Cybinski P (2001). Discription, Explanation,
Altman E I, Eom Y H, Kim D W (1995). Prediction the Evolution of Bankruptcy
Failure prediction: evidence from Studies. Faculty on International
Korea. Journal of International Financial Business and Politics. Giftin University.
Management and Accounting. 6 (3):230 Brisbane 27(4): 29-44.
249.
Dambolena I, Khoury S (1980). Ratio
Aziz A, Emanuel D C, Lawson G H (1988). stability and corporate failure. Journal of
Bankruptcy prediction: an investigation Finance. 33(4):10171026.
of cash flow based models. Journal of
Management Studies. 25 (5):419437. Deakin E (1972). A discriminant analysis of
predictors of business failure. Journal of
Balcaen S, H Ooghe (2006). 35 years of Accounting Research 10(1):167179.
studies on business failure an overview
of the classic statistical methodologies Deakin E (1977). Business failure
prediction: an empirical analysis. In:

Altman, B Sametz, A. (Eds.). Financial Gombola M, Haskins M, Ketz, J, Williams

crisis: institutions and markets in a fragile D (1987). Cash flow in bankruptcy
environment. Wiley, New York: 7298. prediction. Financial Management: 55
65.
Declerc M, Heins B, Van Wymeersch Ch
(1991). The use of value added ratios in Keasey K, Watson R (1987). Non-financial
statistical failure prediction models: some symptoms and the prediction of small
evidence on Belgian annual accounts. company failure: a test of Argentis
Paper presented at the 1991 Annual hypotheses. Journal of Business Finance
Congress of the European Accounting & Accounting. 14 (3): 335354.
Association. April 10-12. Maastricht
(The Netherlands): 124. Keasey K, McGuinness P (1990). The
failure of UK industrial firms for the
Dorigo M, E Bonabeau, G Theraulaz (2000). period 19761984. Logistic analysis and
Ant algorithms and Stigmergy. Future entropy measures. Journal of Business
Generation Computer System. 16: 851- Finance & Accounting. 17 (1): 119135.
871.
Laitinen E K (1992). Prediction of failure
Dorigo M, G di Caro (1999). The ant colony of a newly founded firm. Journal of
optimization meta- heuristic. McGraw- Business Venturing. 7:323340.
Hill: 11-32.
Lussier R N, Corman J (1994). A success
Dorigo. M, A Colorni (1996).The ant system vs. failure prediction model of the
optimization by a colony of cooperating manufacturing industry. Paper presented
agents. IEEE Transactions on system, at the Conference of the Small Business
Man, and Cybernetics- Part B. 26: 1-13. Institute Directors Association, San
Antonio, Texas. 48: 15.
Dorigo M, T Stutzle (2004). Ant Colony
Optimization. Lussier R N (1995). A nonfinancial business
success versus failure prediction model
Edmister R (1972). An empirical test of for young firms. Journal of Small
financial ratio analysis for small business Business Management. 33 (1): 820.
failure prediction. Journal of Financial
and Quantitative Analysis: 14771493 Milea V (2005). An ant system for bankruptcy
prediction. Bachelors thesis. Erasmus
Gentry J A, Newbold P, Whitford D T University Rotterdam, Informatics.
(1985b). Predicting bankruptcy: if cash
flows not the bottom line, what is? Micha B (1984). Analysis of business
Financial Analysts Journal. 41 (5):4756. failures in France. Journal of Banking
and Finance. 8: 281291.
Gloubos G. Grammatikos T (1988). The
success of bankruptcy prediction models Moses D, Liao S (1987). On developing
in Greece. Studies in Banking and models for failure prediction. Journal of
Finance. 7: 3746. Commercial Bank Lending.69:2738.
Mossman Ch E, Bell G, Swartz L M, Turtle Taffler R J (1982). Forecasting company

H (1998). An empirical comparison of failure in the UK using discriminant
bankruptcy models. Financial Review. analysis and financial ratio data. Journal
33 (2):3554. of the Royal Statistical Society. 145:
342358.
Ohlson J (1980). Financial ratios and the
probabilistic prediction of bankruptcy. Taffler R J (1983). The assessment of
Journal of Accounting Research.18: 31- company solvency and performance
109. using a statistical model. Accounting and
Business Research. 15 (52):295307.
Ooghe H, Verbaere E (1985). Predicting
business failure on the basis of Tamari M (1966). Financial ratios as a means
accounting data: The Belgian experience. of forecasting bankruptcy. Management
The International Journal of Accounting. International Review.4:1521.
9 (2):1944
Van Frederikslust R A I (1978). Predictability
Ooghe H, Joos P, De Vos D (1993). Risico- of corporate failure: models for prediction
indicator voor een onderneming aan of corporate failure and for evaluation of
de hand van falings predictiemodellen. corporate debt capacity. PhD thesis in
Accountancy en Bedrijfskunde Economic Sciences, Martinus Nijhoff
Kwartaalschrift. 18 (3):326. Social Science Division, Leiden/Boston,
Erasmus University, Rotterdam (The
Peel M J, Peel D A (1987). Some further Netherlands).
empirical evidence on predicting private
company failure. Accounting and Whitaker R (1999). The Early Stage of
Business Research. 18: 5766. Financial Distress. Journal of Economics
and Finance. 23(2):123-133.
Platt H D, Platt M B (1990). Development of
a class of stable predictive variables: the Zavgren C (1983). The prediction of
case of bankruptcy prediction. Journal corporate failure: the state of the art.
of Business Finance & Accounting. 17 Journal of Accounting Literature. 2.
(1):3151.
Zavgren C V (1985). Assessing the
Sheppard J P (1994). Strategy and bankruptcy: vulnerability to failure of American
an exploration into organizational death. industrial firms: a logistic analysis.
Journal of Management.20 (4):795833. Journal of Business Finance and
Accounting. 12 (1):1945.
Swanson E, Tybout J (1988). Industrial
bankruptcy determinants in Argentina. Zmijewski M E (1984). Methodological
Studies in Banking and Finance.7:125. issues related to the estimation of financial
distress prediction models. Journal of
Taffler R J, Tisshaw H (1977). Going, Going, Accounting Research. 22 : 5986.
GoneFour Factors Which Predict.
Accountancy. 88: 5054.

2010 Zahra Nikbakht and Abbas Aflatooni, 101
THE VALUE RELEVANCE OF EARNINGS

COMPONENTS IN TWO DIFFERENT GAAPs
Zahra Nikbakht and Abbas Aflatooni,
Abstract
The purpose of this study is to explore attribute differences between U.S. GAAP and Iranian GAAP
(IRA-GAAP) earnings components. We compare the relative and incremental value relevance of IRA-
GAAP and U.S. GAAP based operating cash flows and accruals. The issue is investigated by regressing
listed Iranian firms stock returns on the levels of operating cash flows and accruals in IRA-GAAP and
same items adjusted to U.S. GAAP.
The results indicate that there are no significant differences between IRA-GAAP and U.S. GAAP based
earnings components in explaining stock returns. Thus, operating cash flows and accruals components
based on IRA-GAAP are not more value relevant than operating cash flows and accruals components
adjusted to U.S. GAAP.
The results of this study have important implications for accounting standards setters in Iran and other
countries about how cash flow statement classifications affect the value relevance of their items.
Nikbakht Z, Aflatooni A. - The Value Relevance of Earnings Components in Two Different GAAPs
1. Introduction regression models are compared to assess

the relative and incremental value relevance
Recently, Audit Organization in Iran of operating cash flows and accruals.
has provided Iranian Accounting Standards To assess the significance of differences
(IRA-GAAP). In order to harmonize between the value relevance of items, the
accounting standards in Iran with global Vuong (1989) Z-statistics are used. We
standards, Audit Organization has selected estimate discretionary accruals using the
the path to standard setting toward IFRS modified Jones model. Our analysis is based
but yet, there are differences between IRA- on a sample of 1579 firm-years for listed
GAAP based and IFRS based accounting Iranian firms during 2002-2007.
standards in several aspects. One of these The rest of paper proceeds as follows.
differences is on operating cash flows. Section 2 presents the institutional
The second Iranian accounting standard is background. Section 3 describes the
on cash flows statement. In this standard, cash Literature review and section 4 presents the
flows are classified into these five categories: hypotheses and research design. Section 5
(1) Operating activities, (2) Returns on reports the empirical findings and finally,
investments and servicing of finance, (3) conclusion remarks are presented in last
Taxation, (4) Financing activities and (5) section.
Investing activities. The statement of cash
flows that are prepared in accordance with 2. Institutional background
International Accounting Standards (and
also in accordance with US GAAP) requires 2.1. The emergence of Tehran Stock
classification into these three categories: (1) Exchange
Operating activities, (2) Investing activities
and (3) Financing activities. Operating The idea of having a well-organized
cash flows that are provided based on stock market and accelerating the process of
IRA-GAAP includes all items of operating industrialization dates back to the 1930s in
cash flows that are provided based on U.S. Iran when the Melli Bank undertook a study
GAAP except for Taxation, Interest payable, on the subject. A report completed in 1936
Interest receivable and dividends. Thus, worked out the details for the formation
operating cash flows based on IRA-GAAP of a stock market and laid the foundations
include recurring items but operating cash for the plan. The outbreak of World War
flows based on U.S. GAAP includes some II and subsequent economic and political
non-recurring items, too. From this point events delayed the establishment of the
of view, the first category of cash flows stock exchange until 1967, when the Stock
statement that are provided based on IRA- Exchange Act was ratified. The TSE opened
GAAP, is more similar to U.K. GAAP than in April 1968. Initially only government
U.S. GAAP/IFRS operating cash flows. bonds and certain state-backed certificates
This study compares the relative and were traded on the market. During the 1970s,
incremental value relevance of IRA-GAAP the demand for capital boosted the demand
and U.S. GAAP based operating cash flows for stocks. At the same time, institutional
and accruals. The issue is investigated by changes, like the transfer of stocks of
regressing stock returns on the levels of public companies and large monopolies
operating cash flows and accruals. The to employees and the private sector led to
coefficient and explanatory power of the the expansion of stock market activities.

The restructuring of the economy which have become mandatory since 2001 and
followed after the Iran Revolution, expanded 2002.
public-sector control over the economy and After finalization, the committee assigned
reduced the need for private capital. At the them to the Technical Committee. These
same time, the interest-bearing bonds were standards were studied by the Technical
abolished. Because of these events, the TSE Committee and sent to the board. On the
experienced a period of standstill. basis of the Audit Organizations public
This stagnation ended in 1989 with the convention approval, these new standards
revival of the private sector through the became mandatory in 2002 (Mashayekhi
privatization of state-owned enterprises and and Mashayekh, 2008). On the date of
the promotion of private-sector economic writing this paper, the number of Iranian
activities based on the First Five-year Accounting Standards was 29.
Development Plan of the country. Since
then the TSE has expanded continuously 3. Literature review
(Mashayekhi and Mashayekh, 2008). Thus,
TSE is an emerging market in the world Prior empirical research on quality
economy. differences between accounting-standard
regimes is scarce and provides mixed results.
2.2. A brief History of accounting For example, Barth, Landsman, Lang and
standards setting in Iran Williams (2006) compare measures of
accounting quality for firms applying IAS
Prior to 1979, financial reporting in Iran with US firms to investigate whether IAS are
was influenced heavily by Anglo-American associated with less earnings management,
practices (Mirshekari and Saudagaran, more timely loss recognition, and higher
2005). Following the Iranian revolution, in value relevance of accounting amounts
1980, an amendment to the Direct Tax Law than US GAAP. They find that IAS firms
disbanded the Official Accountants Institute. exhibit lower accounting quality relative to
During 19801982, the government US firms in terms of earnings smoothing,
established audit enterprises and the correlation between accruals and cash flows,
Budget, Planning and National Industries timely loss recognition, and the association
Organization. These new organizations between accounting amounts and share
recruited a large number of accountants prices. Also, they find that IAS accounting
who were previously employed with private amounts are of similar quality to reconciled
auditing firms. In 1987, following the merger US GAAP amounts. Their results suggest
of the public-sector audit entities, the Audit that although IAS accounting amounts
Organization was established as the sole may not be of higher quality than those of
audit organization with public ownership US GAAP applied comprehensively, they
and as the only regulatory body for setting are of comparable quality to reconciled US
national accounting and auditing standards. GAAP amounts reported by cross-listed
The Audit Organization is responsible for firms. Harris and Muller (1999) suggest
compiling and determining principles and that investors have realized that US-
rules of auditing and accounting in Iran. GAAP provides more investment-related
Twenty five accounting standards were information than a standard IAS.
issued initially after approval by the Audit Alford, Jones, Leftwich and Zmijewski
Organizations public convention. They (1993) investigate the value relevance
and timelines of earnings reported under asymmetry. He exploits a unique setting

different accounting standards in their own where the two sets of standards are put on a
country and then compares levels of value level playing field. Results show that at least
relevance and timelines. They conclude for New Market firms, the choice between
that earnings based on Danish, German, IAS and U.S. GAAP appears to be of little
Italian, Singaporean and Swedish GAAP consequence for information asymmetry
contain less information and are less timely and market liquidity. These findings do not
than U.S. GAAP earnings, while earnings support widespread claims that U.S. GAAP
based on local GAAP of Australia, France, produce financial statements of higher
The Netherlands and the U.K. are relatively informational quality than IAS.
more informative and timely. Bartov, Goldberg and Kim (2005)
Sawabe (2005) examines the co- investigate comparative value relevance,
evolutionary relationship between measured as the slop coefficient of the return/
accounting rules and creative accounting earnings regression. They find that value
instruments under the rules-based approach. relevance of U.S. GAAP based earnings
The case of financial instruments evolved is higher than that of IAS based earnings,
from convertible bonds in the U.K. is used which in turn is more value relevant than
as an illustrative example to show that the earnings produced under German GAAP.
co-evolutionary structure is the source of Barth and Clinch (1996) synthesize
proliferation of complexity and diversity and extend research exploring differences
of accounting rules and creative accounting between U.S. and other countries GAAP by
instruments. The proliferation of accounting investigating whether differences between
rules and creative accounting instruments domestic and U.S. GAAP for U.S. listed
triggered initiatives to alter the accounting U.K., Australian, and Canadian firms are
standard setting approach towards being associated with firms returns and prices.
more principles-based in the U.K. and with Their findings suggest that the SEC-required
a decade lag in the U.S. GAAP reconciliation reflects information
Chan and Seow (1996) examine the useful to investors for U.K., Australian, and,
association between stock returns and to a more limited extent, Canadian firms.
foreign GAAP earnings versus earnings Pope and Rees (2007) investigate the
adjusted to U.S. GAAP. Using a sample information content of two alternative
of foreign firms with common stock or accounting earnings measures constructed
American Depositary Receipt (ADR) traded under U.K. and U.S. GAAP. The research
in U.S. exchanges, they compare the returns- design involves testing the association
earnings relations between U.S. and foreign between U.K. stock returns and alternative
GAAP-based earnings. Results indicate that accounting numbers. The evidence suggests
earnings based on foreign GAAP are more that, for the sample examined, U.K.
closely associated with contemporaneous GAAP earnings changes have incremental
stock returns than earnings reconciled to information content after controlling for U.S.
U.S. GAAP. They find evidence that their GAAP earnings changes, but that earnings
results may be driven by institutional factors levels measured under U.S. GAAP have
which are specific to foreign markets. some independent incremental information
Leuz (2003) investigates whether firms content after controlling for U.K. GAAP
using U.S. GAAP vis a vis IAS exhibit earnings. Their results are consistent with
differences in several proxies for information GAAP adjustments having a significant

transitory component. The empirical results 4. Research design

display explanatory power which is broadly
consistent with previous work and the 4.1. Hypotheses
GAAP earnings adjustments add marginally
to the ability of earnings to explain returns. Previous studies (e.g., Bernard and
Meulen, Gaeremynck and Willekens Stober, 1989; Subramanyam, 1996; Wilson,
(2007) explore attribute differences between 1986, 1987, Haw, Qi and Wu, 2001) examine
U.S. GAAP and IFRS earnings. They test the value relevance of cash flows and
two market-based earnings attributes, i.e., accruals in a regression where the dependent
value relevance and timeliness, as well as variable is stock returns. Based on the US
two accounting-based earnings attributes, data, Subramanyam (1996) and Wilson
i.e., predictability and accrual quality. These (1986, 1987), among others, report results
attributes are tested for German New Market consistent with both components having
firms as they are allowed to choose between incremental information content, while
IFRS and U.S. GAAP for financial reporting Bernard and Stober (1989) find little of such
purposes. Overall, they find that U.S. evidence. Given emerging capital market
GAAP and IFRS only differ with regard to and relatively incomplete financial reporting
predictive ability. The fact that U.S. GAAP systems and low quality of auditing in Iran,
accounting information outperforms IFRS however, the value relevance of information
also holds after controlling for differences contained in accruals and operating cash
in firm characteristics, such as size, leverage flows remains to be an empirical issue.
and the audit firm. However, their results This study investigates the relative
also seem to suggest that these differences and incremental information content of
are not fully valued by investors, as they operating cash flows in IRA-GAAP and U.S.
do not observe significant and consistent GAAP, and also examines whether accrual
differences for the value-relevance attribute. earnings (in IRA-GAAP and U.S. GAAP)
Amir, Harris, and Venuti (1993) find that provides incremental information beyond
the 20-F reconciliations made by Non-U.S. that contained in operating cash flows in
filers are reflected in stock prices and thus Iranian capital market. There is evidence
are valued by the market. Harris, Lang, that discretionary accruals have incremental
and Moller (1994) is the only study that information content in the mature market
provides an analysis of value relevance of the United States (Subramanyam, 1996).
across exchanges. Similar to the U.S. stock We also investigate whether discretionary
exchange studies, Harris et al. (1994) assess accruals provide incremental information
quality by looking at the association between content beyond that provided by the
prices and earnings (or shareholders nondiscretionary component of earnings in
equity). These associations are however Iranian emerging capital market.
not calculated for the entire sample, but Finally, as operating cash flows in IRA-
for the German and U.S. stock market GAAP include recurring items but operating
separately. They find that the explanatory cash flows in U.S. GAAP include some non-
power of German earnings is comparable recurring items (Tax, Returns on investments
to U.S. earnings, but the explanatory power and servicing of finance), it is hypothesized
of shareholders equity in Germany is that operating cash flows and accruals in
significantly lower than in the United States. IRA-GAAP, be more value relevant than
same items in U.S. GAAP.
4.2. Research models ACCR, has a higher association (R2)

with stock returns is interpreted as more
In order to test the hypotheses, annual effectively summarizing firm performance
market adjusted returns are regressed on the or relatively more value-relevant.
levels of earnings components (operating Alternatively, the coefficients of earnings
cash flows and accruals), which is consistent components are compared to assess the
with Dechow (1994) and Subramanyan incremental information content of the
(1996). We first develop the following linear variables. We estimate the models in cross
regression models (Eqs. (1) and (2)) to test sectional and pooled approach.
relative and incremental information content
of operating cash flows (CFO) and total 4.3. Measurement of discretionary
accruals (ACCR) that are provided based on and non-discretionary accruals
IRA-GAAP and adjusted to U.S. GAAP:
While nondiscretionary accruals are
(1)
accounting adjustments to operating cash
Retit =+ b1 CFOitUS GAAP + b2 ACCRitUS GAAP + fit flows as mandated by accounting standard-
(2)
IRA GAAP IRA GAAP
setting bodies, discretionary accruals
Retit =+ b1 CFO it + b2 ACCR it + fit are adjustments subject to management
discretion. Managers choose discretionary
We then decompose total accruals accruals from an opportunity set of
to test incremental information content generally accepted procedures defined by
of discretionary (DACCR), and accounting standard-setting bodies and,
nondiscretionary accruals (NDACCR), thus, discretionary accruals are often
separately as below (Eqs. (4) and (5)): used as a measure of managers earnings
manipulation (Haw et al., 2001).
(3)
We compute total accruals (ACCR) as the
Retit =+ b1 CFOitUS GAAP + b2 DACCRitUS GAAP + difference between earnings and cash flow
b3 NDACCRitUS GAAP + fit from operations. Operating cash flows for
firm i are computed based on IRA-GAAP and
(4)
then adjusted to U.S GAAP. Discretionary
Retit =+ b1 CFOitIRA GAAP + b2 DACCRitIRA GAAP + accruals (DACCR) are determined using the
b3 NDACCRitIRA GAAP + fit modified Jones model:
(5)
All of the explanatory variables are ACCRit = TACCRit /Ait - 1 = a 61/Ait - 1@ +

scaled by lagged total assets for firm i, b6^3 Revit - 3 Revith /Ait - 1@ +
consistent with prior studies. The dependent c6 PPEit /Ait - 1@ + fit
variable, RET, is annual market-adjusted
stock returns measured over a 12-month Where TACCRit is total accruals in year
period ending 4 months after the fiscal year- t for firm i, Revit is revenues in year t less
end. In Iran, listed firms have to release their revenues in year t-1 for firm i, Recit is
annual financial statements within 4 months receivables in year t less revenues in year t-1
from the fiscal year-end. for firm i, PPEit is gross property, plant and
Whichever measures, U.S. GAAP based equipment in year t for firm i, Ait-1 is total
CFO vs. IRA-GAAP based CFO and U.S. assets in year t-1 for firm i, and it is error
GAAP based ACCR and IRA-GAAP based term in year t for firm i.

As you see, all variables are scaled by population of Iranian firms listed in Tehran
lagged total assets to reduce heteroscedsticity. Stock Exchange for 2002-2007, but because
Using the estimated coefficients ( of data problems, only 1579 firm-years
) from Eq. (5), Non-discretionary accruals are used in research. For data collection
(NDACCR) are defined as: purpose, we use the electronic archival
(6)
data provided by TSE. In some cases that,
the required data is incomplete, the manual
archives in the TSEs library are used. Also,
a part of data is acquired from Tadbirpardaz
And then discretionary accruals and Sahra (two Iranian) databases.
(DACCR) are defined as the residuals of Eq.
(5): 5.1. Descriptive statistics
(7)
The means, medians and standard

deviations for the regression variables are
reported in Table 1. The means and medians
As in prior studies, the level of gross of operating cash flows (CFO), total
property, plant, and equipment and the accruals (ACCR), and nondiscretionary
change in revenues are included to control for accruals (NDACCR) in IRA-GAAP and
changes in nondiscretionary accruals caused U.S. GAAP are positive, and discretionary
by the change of economic conditions. accruals (DACCR) in both GAAPs are, in
general, negative for each of the 6 years and
5. Empirical results the pooled (full) sample.
The sample selection starts with the entire

Table 1: Descriptive statistics
U.S. GAAP IRA-GAAP P-Value for Difference in
NDACCR
NDACCR
NDACCR
DACCR
DACCR
DACCR
ACCR
ACCR
ACCR
CFO
CFO
CFO
RET
N
Meana
2002 272 0.15 0.09 0.07 -0.28 0.36 0.15 0.01 -0.32 0.33 (0.00)*** (0.00)*** (0.38) (0.42)
2003 245 0.26 0.07 0.13 -0.18 0.32 0.12 0.08 -0.21 0.29 (0.02)** (0.10) (0.60) (0.49)
2004 255 -0.02 0.06 0.26 0.05 0.22 0.11 0.21 0.01 0.20 (0.75) (0.76) (0.86) (0.49)
2005 325 -0.41 0.06 0.08 -0.13 0.21 0.10 0.03 -0.15 0.18 (0.08)* (0.13) (0.57) (0.27)
2006 365 0.10 0.07 0.02 -0.20 0.22 0.11 -0.02 -0.22 0.19 (0.00)*** (0.03)** (0.70) (0.54)
2007 297 0.18 0.09 -0.06 -0.28 0.22 0.17 -0.14 -0.34 0.20 (0.75) (0.77) (0.85) (0.71)
Full sample 1759 0.03 0.07 0.08 -0.17 0.26 0.12 0.02 -0.20 0.23 (0.28) (0.31) (0.59) (0.13)
Medianb
2002 272 0.00 0.07 0.07 -0.19 0.23 0.14 0.01 -0.22 0.20 (0.00)*** (0.00)*** (0.08)* (0.15)
2003 245 0.00 0.06 0.08 -0.12 0.21 0.12 0.03 -0.15 0.18 (0.00)*** (0.00)*** (0.15) (0.14)
2004 255 -0.01 0.04 0.08 -0.09 0.15 0.09 0.04 -0.11 0.13 (0.00)*** (0.00)*** (0.29) (0.15)
2005 325 -0.14 0.07 0.05 -0.11 0.14 0.10 0.01 -0.12 0.12 (0.01)** (0.00)*** (0.27) (0.09)*
2006 365 0.00 0.06 0.01 -0.12 0.13 0.10 -0.02 -0.13 0.11 (0.00)*** (0.00)*** (0.21) (0.05)*
2007 297 0.00 0.06 0.04 -0.07 0.11 0.10 -0.01 -0.09 0.09 (0.00)*** (0.00)*** (0.11) (0.06)*
U.S. GAAP IRA-GAAP P-Value for Difference in
NDACCR
NDACCR
NDACCR
DACCR
DACCR
DACCR
ACCR
ACCR
ACCR
CFO
CFO
CFO
RET
N
Full sample 1759 0.00 0.06 0.05 -0.11 0.15 0.11 0.01 -0.14 0.14 (0.00)*** (0.00)*** (0.00)*** (0.00)***
Standard deviationc
2002 272 1.19 0.23 0.23 0.48 0.47 0.23 0.22 0.47 0.44 (0.78) (0.40) (0.64) (0.33)
2003 245 1.92 0.27 0.39 0.55 0.50 0.27 0.39 0.53 0.47 (0.83) (0.97) (0.48) (0.43)
2004 255 2.42 0.32 2.45 2.51 0.43 0.27 2.29 2.34 0.42 (0.28) (0.24) (0.23) (0.62)
2005 325 1.16 0.34 0.40 0.51 0.29 0.33 0.39 0.49 0.27 (0.62) (0.53) (0.56) (0.39)
2006 365 0.98 0.19 0.25 0.62 0.51 0.19 0.26 0.59 0.48 (0.77) (0.87) (0.37) (0.23)
2007 297 3.81 0.34 3.23 3.29 0.82 0.42 3.63 3.73 0.83 (0.01)** (0.03)** (0.03)** (0.85)
Full sample 1759 2.13 0.29 1.67 1.69 0.52 0.30 1.77 1.80 0.51 (0.00)*** (0.01)** (0.01)** (0.26)
*, **and *** Significant at the 0.1, 0.05, 0.01 level, respectively. RET is annual market-adjusted stock returns measured over
a 12-month period ending 4 months after the fiscal year-end, CFO is operating cash flows that are measured based on U.S.
GAAP and IRA-GAAP, ACCR is total accruals that are measured by subtracting CFO from earnings. DACCR and NDACCR
are discretionary accruals and nondiscretionary accrual that are measured by using of modified Jones model.All variables
are scaled by lagged total assets.a, b and c: P-values for differences in means, medians and standard deviations are based on
t test, Wilcoxon test and F test.
Due to the economical-political situation (1995), Subramanyam (1996), and Haw et

during 2004-2005 periods, the Iranian al. (2001). This negative correlation could
financial market indices declined sharply. arise due to accrual accounting or income
Therefore, the mean and median of the smoothing.
market-adjusted returns in 2005 and 2006
are negative. Table 2: Pearson correlation coefficients
Comparison between mean, median and
standard deviation of items in IRA-GAAP U.S. GAAP CFO ACCR DACCR
and U.S. GAAP show that the differences ACCR -0.93 (0.00)***
between median of CFO and ACCR are DACCR -0.93 (0.00)*** 0.95 (0.00)***
significant in all years and full sample but, NDACCR -0.01 (0.60) 0.17 (0.00)*** -0.13 (0.00)***
the differences between mean, median IRA- GAAP CFO ACCR DACCR
and standard deviation of variables are, in ACCR -0.94 (0.00)***

general, insignificant. In full sample, the DACCR -0.94 (0.00)*** 0.95 (0.00)***
standard deviation of U.S. GAAP based NDACCR -0.00 (0.67) 0.16 (0.00)*** -0.11 (0.00)***
CFO is lower than IRA-GAAP based CFO
*** Significant at the 0.01 level, respectively.
and difference between them is significant
at the 0.01 level. Reported are Pearson correlation coefficients (p-value)
among the variables. Reported probabilities are from a
Table 2 reports contemporaneous Pearson two-tailed test.
correlation coefficients between various
components of earnings that are provided In U.S. GAAP and IRA-GAAP, the
based on U.S. GAAP and IRA-GAAP. In correlation between operating cash flows
both GAAPs operating cash flows (CFO) is and discretionary accruals is -0.93 and
negatively (and significantly) correlated with -0.94, respectively. Also, there are no
total accruals (ACCR) and discretionary significant correlation between operating
accruals (DACCR), consistent with prior cash flows and nondiscretionary accruals
evidence in Dechow (1994), Dechow et al. in U.S. GAAP and IRA-GAAP. Thus,

discretionary accounting choices explain 5.2. Regression results of market-

a larger portion of the negative correlation adjusted returns on operating cash
between operating cash flows and total flows and total accruals
accruals than nondiscretionary accruals.
While in U.S. GAAP and IRA-GAAP, Table 3 examines incremental value
discretionary accruals (DACCR) are relevance of earnings components in multiple
negatively (and significantly) correlated with regressions. In Panel A and B, earnings are
nondiscretionary accruals (NDACCR) with decomposed into operating cash flows and
a mean coefficient of -0.13 and -0.11, the total accruals that are provided based on
correlation is significant at the conventional U.S. GAAP and IRA-GAAP.
level. Thus, there is strong evidence that
listed Iranian firms smoothen income.
Table 3. Incremental information content of operating cash flows and total accruals
Full Sample 2002 2003 2004 2005 2006 2007
N 1759 272 245 255 325 365 297
Regression results of market-adjusted returns on operating cash flow s and total accruals
Panel A: U.S. GAAP
-0.08 -0.05 -0.19 -0.75 -0.35 -.00 0.23
Intercept
(-2.39)** (-0.65) (-1.62) (-5.57)*** (-5.16)*** (-0.01) (1.01)
Operating 0.62 1.85 3.95 4.28 -0.76 1.24 -0.23
cash flows (6.61)*** (6.27)*** (8.88)*** (11.99)*** (-2.93)*** (4.79)*** (-1.34)
Total 0.63 0.59 1.64 4.35 -0.13 0.57 -0.23
Accruals (6.97)*** (1.90)* (4.81)*** (12.24)*** (0.58) (3.00)*** (-1.42)
Adjusted R2 0.027 0.127 0.251 0.370 0.037 0.079 0.006
Panel B: IRA-GAAP
-0.08 -0.13 -0.30 -0.74 -0.33 -0.02 0.23
Intercept
(-2.42)** (-1.62) (-2.40)** (-5.52)*** (-4.70)*** (-0.42) (1.01)
Operating 0.64 1.85 3.75 4.32 -0.73 1.22 -0.23
cash flows (6.83)*** (6.24)*** (8.12)*** (12.10)*** (-2.80)*** (4.62)*** (-1.36)
Total 0.62 0.54 1.76 4.36 -0.11 0.59 -0.23
Accruals (6.89)*** (1.69)* (5.09)*** (12.21)*** (-0.52) (3.14)*** (-1.42)
Adjusted R2 0.027 0.126 0.227 0.367 0.033 0.078 0.006
Vuong Z 0.026 0.087 2.090** 1.037 0.627 0.258 -0.033
statisticsa
*, **, *** Significant at the 0.1, 0.05, 0.01 level, respectively. Reported numbers are regression coefficients (t value). a Vuongs
(1989) Z-statistic compares the non-nested models.
Results indicate that operating cash 5.3. Regression results of market-

flows and accruals based on IRA-GAAP adjusted returns on operating cash
are not more closely associated with flows, nondiscretionary accruals and
contemporaneous stock returns than discretionary accruals
operating cash flows and accruals reconciled
to U.S. GAAP. Thus operating cash flows Table 4 decomposes earnings into three
and accruals based on IRA-GAAP are not parts: operating cash flows, discretionary
more value relevant than operating cash accruals, and nondiscretionary accruals
flows and accruals reconciled to U.S. GAAP. based on U.S. GAAP and IRA-GAAP.
Table 4. Incremental information content of operating cash flows, nondiscretionary accruals and discretionary
accruals.
Full Sample 2002 2003 2004 2005 2006 2007
N 1759 272 245 255 325 365 297
Regression results of market-adjusted returns on operating cash flows, nondiscretionary accruals and discretionary
accruals
Panel A: U.S. GAAP
-0.06 -0.07 -0.27 -0.48 -0.43 -0.06 0.44
Intercept
(-1.63) (-0.84) (-1.85)* (-2.92)*** (-5.20)*** (-1.03) (1.86)*
Operating cash 0.63 1.89 4.02 3.84 -0.75 1.33 0.91
flows (6.66)*** (6.24)*** (8.94)*** (9.48)*** (-2.87)*** (5.06)*** (2.16)**
Nondiscretionary 0.56 0.64 1.94 3.12 0.24 0.93 -1.29
accruals (4.83)*** (1.97)** (4.31)*** (5.09)*** (0.77) (3.81)*** (-3.30)***
Discretionary 0.64 0.56 1.67 3.91 -0.10 0.69 0.92
accruals (7.08)*** (1.80)* (4.85)*** (9.75)*** (-0.48) (3.51)*** (2.19)**
Adjusted R2 0.029 0.128 0.255 0.391 0.045 0.093 0.036
Panel B: IRA-GAAP
-0.06 -0.14 -0.38 -0.43 -0.39 -0.08 0.40
Intercept
(-1.54) (-1.46) (-2.54)** (-2.61)*** (-4.77)*** (-1.33) (1.72)*
Operating cash 0.66 1.86 3.82 3.71 -0.72 1.32 1.06
flows (6.92)*** (6.20)*** (8.19)*** (9.05)*** (-2.76)*** (4.92)*** (2.48)**
Nondiscretionary 0.52 0.57 2.10 2.74 0.23 0.97 -1.39
accruals (4.41)*** (1.67)* (4.51)*** (4.31)*** (0.69) (3.96)*** (-3.59)***
Discretionary 0.64 0.53 1.79 3.76 -0.09 0.72 1.07
accruals (7.05)*** (1.65)* (5.14)*** (9.23)*** (-0.42) (3.65)*** (2.50)**
Adjusted R2 0.029 0.127 0.232 0.396 0.039 0.093 0.042
Vuong Z -0.948 0.149 2.07** -0.634 0.750 0.116 -0.594
statisticsa
*, **, *** Significant at the 0.1, 0.05, 0.01 level, respectively. Reported numbers are regression coefficients (t value). a Vuongs
(1989) Z-statistic compares the non-nested models.
In panel A (U.S. GAAP), for the full of earnings, and discretionary accruals
sample, the coefficients on operating provide incremental information over
cash flows, discretionary accruals, and that contained in the nondiscretionary
nondiscretionary accruals are 0.63, 0.56 component of earnings. In panel A and B,
and 0.64, respectively and, in panel B (IRA- the adjusted R2 for full sample is 2.9%
GAAP) for the full sample, the coefficients and 2.9% and Based on Vuong Z-statistics
on operating cash flows, discretionary (-0.984), the difference between them is
accruals, and nondiscretionary accruals not significant. Except for 2003, the annual
are 0.66, 0.52 and 0.64, respectively. All regression results present a similar pattern.
coefficients are significant at the 0.01 level, Results indicate that operating cash flows
but they are not statistically different from and discretionary and nondiscretionary
each other (except for 2003). accruals based on IRA-GAAP are not more
These results suggest that in addition closely associated with contemporaneous
to operating cash flows, both discretionary stock returns than operating cash flows and
and nondiscretionary components of accruals components adjusted to U.S. GAAP.
accruals contribute to the value relevance Therefore operating cash flows and accruals

components based on IRA-GAAP are not of U.S. versus non-U.S. GAAP

more value relevant than operating cash accounting measures using Form 20-F
flows and accruals components adjusted to reconciliations, Journal of Accounting
U.S. GAAP. Research, 31, 230264 (Supplement).
6. Conclusions Barth, M., Landsman, W., Lang, M., and

Williams, C. (2006), Accounting quality:
In this study we report empirical evidence International accounting standards and
about attribute differences between U.S. U.S. GAAP, Working paper series.
GAAP and Iranian GAAP operating cash
flows and accruals. In fact, we compare the Barth, M., and Clinch, G. (1996),
relative and incremental value relevance International accounting differences and
of IRA-GAAP and U.S. GAAP based their relation to share prices: Evidence
operating cash flows and accruals. The issue U.K., Australian, and Canadian Firms,
is investigated by regressing stock returns Contemporary Accounting Research,
on the levels of operating cash flows and 13(1), 135179.
accruals. As operating cash flows in IRA-
GAAP include recurring items but operating Bartov, E., Goldberg, S., and Kim, M.
cash flows in U.S. GAAP include some (2005), Comparative value relevance
non-recurring items, it is hypothesized that among German, U.S., and international
operating cash flows and accruals in IRA- accounting standards: A German
GAAP, be more value relevant than same stock market perspective, Journal of
items in U.S. GAAP. Accounting, Auditing and Finance,
The results show that, in explaining 20(2), 95119.
the stock return, there are no significant
differences between IRA-GAAP based Bernard, L.V., and Stober, T.L. (1989), The
earnings components and earnings nature and amount of information in
components that are adjusted to U.S. GAAP. cash flows and accruals, The Accounting
This means that operating cash flows and Review, 4, 625651.
accruals components based on IRA-GAAP
are not more value relevant than operating Chan, K., and Seow, G. (1996), The
cash flows and accruals components adjusted association between stock returns and
to U.S. GAAP. foreign GAAP earnings versus earnings
adjusted to U.S. GAAP, Journal of
References Accounting and Economics, 21(1),
139159.
Alford, A., Jones, J., Leftwich, R., and
Zmijewski, M. (1993), The relative Dechow, P. (1994), Accounting earnings
informativeness of accounting and cash flows as measures of firm
disclosures in different countries, Journal performance: The role of accounting
of Accounting Research, 31, 183201 accruals, Journal of Accounting and
(Supplement). Economics, 18(1), 342.
Amir, E., Harris, T., and Venuti, E. (1993), Dechow, P., Sloan, R., and Sweeney, A.
A comparison of the value-relevance (1995), Detecting earnings management,
The Accounting Review, 70, 193226. financial statement users in developing

countries: Evidence from Iran, Journal of
Harris, M., and Muller, K. (1999), The International Accounting, Auditing and
market valuation of IAS versus U.S.- Taxation, 14, 33-54.
GAAP accounting measures using
Form 20-F reconciliations, Journal of Pope, P., and Rees, W. (1992), International
Accounting and Economics, 26(13), differences in GAAP and the pricing
285312. of earnings, Journal of International
Financial Management and Accounting,
Harris, T., Lang, M., and Moller, H. (1994), 4(3), 190219.
The value relevance of German accounting
measures: An empirical analysis, Journal Sawabe, N. (2005), Co-evolution of
of Accounting Research, 32(2), 187210. accounting rules and creative accounting
instruments-The case of a rules-based
Haw, I.M., Qi, D., and Wu, W. (2001), The approach to accounting standard setting,
nature of information in accruals and cash Evolutionary and Institutional Economics
flows in an emerging capital market: The Review, 1(2), 177195.
case of China, The International Journal
of Accounting, 36, 391406 Subramanyam, K.R. (1996), The pricing
of discretionary accruals, Journal of
Leuz, C. (2003), IAS versus U.S. GAAP: Accounting and Economics, 22, 249
Information-asymmetry based evidence 281.
from Germanys new market, Journal of
Accounting Research, 41(3), 445472. Vuong, Q.H. (1989), Likelihood ratio tests
for model selection and non-nested
Mashayekhi B. and Mashayekh S. (2008), hypotheses, Econometrica, 57, 307333.
Development of Accounting in Iran, The
International Journal of Accounting, 43, Wilson, G.P. (1986), The relative
66-86. information content of accruals and cash
flows: combined evidence at the earnings
Meulen, S.V., Gaeremynck, A., and announcement and annual report release
Willekens M. (2007), Attribute date, Journal of Accounting Research,
differences between U.S. GAAP and 24, 165203.
IFRS earnings: An exploratory study,
The International Journal of Accounting, Wilson, G.P. (1987), The incremental
42, 123142 information content of the accrual and
funds components of earnings after
Mirshekari S. and Saudagaran S. (2005), controlling earnings, The Accounting
Perceptions and characteristics of Review, 62, 293322.

2010 Jos Carlos Arias 113
A WORD ON GAME THEORY APPLICATION TO

CORPORATE FINANCE
Jos Carlos Arias (PhD, DBA)
Arias J.C. - A Word on Game Theory Application to Corporate Finance

Finance in general is concerned with how an incentive to take on debt. Modigliani

the savings of investors are allocated through and Miller corrected their original model
financial markets and intermediaries to to include corporate income taxes showing
firms, who use them to fund their activities. that a firm could increase its equity, or
Finance can broadly be broken down shareholder value, by taking on debt and
into two fields. The first is asset pricing, taking advantage of tax shields. Their model
which is concerned with the decisions of then showed all firms stood to gain the most
investors. The second is corporate finance, if they were 100% debt financed; however
which is concerned with the decisions of this is not observed in reality. In fact, some
firms. This paper will focus on the latter companies and industries thrive with no debt
field and how game theory can be used to at all. Different game theory models have
explain certain behaviors that are regularly been used to explain the actions of managers
witnessed. Traditional financial thinking in determining their companys capital
relies on assumptions of certainty, complete structure, the most influential deals with
knowledge and market efficiency and in the signaling effects attributed to debt vs.
this context, financial decisions should be equity financing. In 1984 Myers and Majluf
relatively straightforward. In the real world developed a model based on asymmetric
though, many times what is observed deviates information that insists managers are better
greatly from what would be expected using informed of the prospects of the firm than
traditional financial thinking. This paper the capital markets. If management feels
will show how different game theory models that the market is currently undervaluing
can be used to more accurately explain its firms equity then it will be unwilling to
observed financial decisions dealing with raise money through an equity issue because
capital structure, corporate acquisitions and it will be selling the stock at a discount.
initial public offerings (IPOs). On the other hand, management might be
Game theory has made great strides eager to issue equity if it feels its stock is
in explaining many of the observed overvalued, because it will be selling its
phenomena falling under corporate finance. stock at a premium. Investors are not dull
One example is the capital structure decided and will predict that managers are more
upon by a firms management. Capital likely to issue stock when they think it is
structure deals with the firms decision to overvalued while optimistic managers may
raise funds through debt versus equity and cancel or defer issues. Therefore, when an
what ratio of debt to equity should the firm equity issue is announced, investors will
maintain. Modigliani and Miller in 1958 mark down the price of the stock accordingly.
showed that in perfect capital markets (i.e. Thus equity issues are considered a bad
no frictions and symmetric information) and signal, even companies with overvalued
no taxes a firm could not change its total stock would prefer another option to raise
value by altering its debt/equity ratio; thus money to avoid the mark down in stock
capital structure is irrelevant. However in price. Firms prefer to use less information
the real world, capital structure is carefully sensitive sources of funds. This leads to
thought about by every company, and it is in the pecking order of corporate financing:
fact not irrelevant because taxes do exist and Retained earnings are the most preferred,
capital markets are not perfect. In the United followed by debt, then hybrid securities such
States, interest paid by a company is a tax- as convertible bond and lastly equity. Some
deductible expense. This tax shield creates industries by their nature support companies

that finance most of their growth through much when the company is near default
retained earnings. Airlines however are an because most of the payoff will be paid out
example of an industry that is characterized to the debt holders. A financially sound
by its high debt level. In general, capital company would not have this agency
structure is similar within industries with problem because equity holders stand to
differences resulting from weighing the lose more from risky projects when the
benefits of a higher tax shield versus the company is not in risk of going bankrupt,
benefits of the less information sensitive and thus want to avoid them along with
financial of retained earnings. bondholders. The second conflict arises
A second application of game theory to when equity holders cannot fully control
capital structure is concerned with agency the actions of managers. This occurs when
costs. In 1976 Jensen and Meckling managers have an incentive to pursue their
described two kinds of agency problems in own interests rather than those of the equity
corporations: One between equity holders holders. Executive compensation in the form
and bondholders and the other between of option contracts can create incentives
managers and equity holders. The first arises for managers to make risky decisions in
because the owners of a levered firm have an attempt to gain the highest payoff from
an incentive to take risks at the expense of the call options. Higher risk increases the
debt holders. Stockholders of levered firms value of an option, but risk can also cause a
gain when business risk increases because stock price to take a nosedive. A manager
they receive the surplus when returns are with options is not hurt nearly as much as a
high but the bondholders bear the cost when worker with his/her retirement savings in a
default occurs. Bondholders value does company whose stock plummets because of
not increase with the value of the firm, thus risky bets. Option contracts were meant to
they would like the firm to take safe bets better align the interests of managers with
to minimize the risk of default. Equity stockholders, but it is obvious that this is not
holders on the other hand, receive whatever so easily achieved.
is leftover after paying back debt holders. Game theory can also be used to explain
They would like to see the upside potential what is observed in the course of many
of the company maximized and this occurs corporate acquisitions. If markets are
through taking on risky projects (higher efficient then one would expect a company
returns are generated though greater risk to pay fair value when acquiring another
taking.) It is obvious that there is a conflict company; however in many instances
of interest between equity holders desire for the acquirer pays a large premium to buy
business risk and bondholders aversion to the other company. In 1986 Shleifer and
business risk. Financial managers who act Vishny provide one explanation of this
strictly in the interests of shareholders will phenomenon, the free rider problem. One of
favor risky projects over safe ones. It is the concepts behind efficient markets is the
important to note that this agency cost does market for corporate control. The market
not occur in financially sound companies. It for corporate control says that in order for
mainly occurs when the odds of default or resources to be used efficiently, companies
high and equity holders feel they can make need to be run by the most able and
one last gamble to avoid bankruptcy and get competent managers. One way to achieve
a big payoff at the same time. An average this is through corporate acquisitions. A
payoff would not benefit the stockholders company is not likely to purchase another

company unless it feels it can run it better between informed buyers and uninformed
than the current management. This can be buyers. The informed buyers know the true
through new management or synergies that value of the stock and will only purchase
occur because of the operational efficiency shares at or below its true value. The
of the combined firms. Shareholders of implication of this is that the uninformed
the company to be acquired realize that the buyers will receive a high allocation of
new management will run in their company overpriced shares since they will be the
more efficiently and thus their shares will only people in the market when the offering
be worth more than before the purchase. price is above the true value. Knowing this,
Since shareholders will be able to benefit uninformed buyers would be unwilling to
from the improvements implemented by the purchase the stock; forcing the informed
new management they have an incentive to buyers to hold onto the stock because there
hold out and say no to the bid. They will is no one they can sell it to. Therefore, to
only be willing to tender their shares if the induce the uninformed to participate they
offer price fully reflects the value under the must be compensated for the overpriced
new management. Therefore, a bidding firm stock they end up buying. One way to do
cannot make a profit from tendering for the this is to under-price the stock on average.
target firm, and in truth, after the costs of This means that on average the uniformed
acquiring information in preparation for the will buy a stock that started out undervalued
bid, the bidding firm will make a loss. and thus they are still able to buy the stock
A second puzzle associated with corporate at or below its true value. Since all investors
takeovers that has been documented is the know that an IPO will likely be under priced
fact that bidding in takeover contests occur they all try to buy the stock as quick as
through several large jumps rather than many possible creating a demand for the stock that
small ones. The logic behind such behavior results in substantial price gain in the initial
as reported by Fishman in 1988 is relatively days of trading.
straightforward. Observing a bid alerts the Another interesting implication of IPOs
market to the potential desirability of the pointed out by Ritter in 1991 is the fact that
target. If the initial bid is low, a second bidder while they experience high returns in the
will find it worthwhile to spend the cost to short run they typically under-perform the
investigate the target. This second firm may market in the long run. One argument for
then bid for the target and push out the first this behavior is that the market for IPOs is
bidder or force a higher acquisition price. subject to fads and that investment banks
By starting with a sufficiently high bid, the under-price IPOs to create the appearance of
initial bidder can reduce the likelihood of excess demand. This leads to a high price
this competitor. initially but subsequently underperformance;
Initial Public Offerings (IPOs) have long therefore companies with the highest initial
been known to provide a significant positive returns should have the lowest subsequent
return in the initial days of trading. This returns. There exists evidence of this in the
occurrence directly conflicts with the theory long run.
of market efficiency because the companies Game theory has been extremely useful
should be fairly valued at their IPO and in explaining certain financial decisions.
any return in the initial days should be This paper has only highlighted a few
minimal. In 1986 Rock explained that this of the aspects where behavioral analysis
phenomenon was due to adverse selection has helped explained observed behavior.

Specifically, game theory has helped explain Modigliani, F. and M.H. Miller: The Cost
the reasons companies might choose various of Capital, Corporation Finance and
capital structures and the agency costs the Theory of Investment, American
between managers, equity holders, and Economic Review, 48:261-297
debt holders. In addition, the existence of (June1958).
free rider problems and bidding wars in
corporate acquisitions has been made clear Myers, S.C. and N.S. Majluf: Corporate
through game theory applications. Lastly, Financing and Investment Decisions
IPOs exhibit behavior contrary to the When Firms Have Information Investors
efficient market theory, and game theory can Do Not Have, Journal of Financial
be utilized to help show why this behavior Economics, 13:187-222 (June 1984).
occurs.
Ritter, J. R.: The long-run performance
References of initial public offerings, Journal of
Finance 46:3-27 (1991).
Fishman, M.: A Theory of Pre-Emptive
Takeover Bidding, Rand Journal of Rock, K.: Why new issues are underpriced,
Economics, 19:88-101 (1988). Journal of Financial Economics, 15:187-
212 (1986).
Jensen M.C. and W.H. Meckling: Theory
of the Firm: Managerial Behavior, Shleifer, A. and R. Vishny: Large
Agency Costs and Ownership Structure, Shareholders and Corporate Control,
Journal of Financial Economics, 3:305- Journal of Political Economy, 94:461-488
360 (October 1976). (1986).


2010 Niki Geiersbach 119
THE IMPACT OF INTERNATIONAL BUSINESS ON

THE GLOBAL ECONOMY
Niki Geiersbach
Abstract
With the dawn of globalization, international business is becoming increasingly popular. Multinational
organizations are among the most profitable in the world. A company needs to be aware of the language
and culture of the country where it plans to embark with its investment. Politics and laws of the nation
can either make international business easy or hard. With the success of international business, its future
is gleaming, on a global scale.
Geiersbach N. - The Impact of International Business on the Global Economy

The International Business A 30-second advertising spot proved to

Imperative be a costly mistake for Doubletree Hotels
Corporation. The advertisement was the
Revolutionary changes in technologies announcement of a $31 million marketing
have provided the mechanisms that propel campaign to illustrate the warm, friendly
the growth of international business. The atmosphere of the hotels. Deemed offensive
intensification of competition at both domestic to the Muslim community, the spot portrayed
and international levels has driven firms to employees of the hotel dressed in Arab-style
look beyond their domestic markets for new clothing and bowing to the guests. This was
opportunities. The progressive removal interpreted as the employees worshipping or
of barriers to trade and capital movements praying to hotel visitors. As the Muslims
has stimulated greater flows of exports, worship the one true God, this advertisement
imports and foreign direct investment (FDI). was seen as ridicule of reality.
Multinational enterprises have emerged as Moreover, translation errors are the cause
the key agents of international economic co- of the majority of blunders in global trade.
ordination. They provide the capability to The Coors slogan Turn it Loose, turned into
generate innovations and deliver new goods Drink Coors and Get Diarrhea. Even the
and services to the market; they also provide largely popular Got Milk? campaign lost
the capability to exploit these technological its edge when it was introduced in Mexico as
advances at a global level; and they are a Are you Lactating? As is obvious, these
depiction of the capacity of international mistakes and others like them can result in
managerial co-ordination to operate a devastating loss of revenue for companies
efficiently across international boundaries. in todays global marketplace. Hence, it is
Furthermore, the growing economic strength very important to know its culture before
of the newly-industrializing countries (e.g. conducting international business with a
Taiwan, Hong Kong, Singapore, Korea) and country (International Business Customs).
the opening up of China and Eastern Europe
have provided an additional stimulus to National Trade and Investment
international business activities (Wei). Policies
Culture and International The 1996 National Trade Estimate of

Business the European Union explains the notions of
national trade and investment policies. The
Chevrolet was unsuccessful at marketing European Union and the United States share
the Chevy Nova in Mexico, and the result the largest two-way trade and investment
was lost revenue because inn Spanish, No relationship in the world. In 1995, the
va means it doesnt go. The American US had a trade deficit with the EU of $8.3
Motor Corporation named one of its products billion, $3.4 billion less than that recorded
the Matador, to create images of strength in 1994. In 1995, US merchandise exports
and virility. But Matador means killer to the EU were $123.6 billion, an increase
in Puerto Rico, where the traffic fatality of $15.9 billion from those in 1994. US
rate is exceedingly high. It is therefore imports from the EU totaled $131.9 billion
imperative to conduct an evaluation of in 1995, or 10.4 percent greater than those
names prior to introducing a product on the in 1994. The stock of US foreign direct
market (International Business Customs). investment in the EU was $256.1 billion in

1994, an increase of 6.9 percent over that in Italys highly fragmented and sometimes
1993 (1996 National Trade). non-transparent government procurement
As an example of import policies, when practices created obstacles to US firms
Austria, Finland, and Sweden joined the EU participation in Italian government contracts.
on January 1, 1995, these countries adjusted Procurement in certain areas was heavily
their tariffs to the EUs common external directed toward Italian suppliers. In 1994, the
tariff, resulting in increased tariffs on $3 Italian parliament enacted legislation aimed
billion of US industrial and agricultural at providing more transparent procurement
exports. The European Commission was procedures, including establishment of a
required to negotiate with the US and central body to monitor implementation.
other affected trading partners a package Due to its complexity, the bill was not fully
of compensating tariff cuts. During 1995, implemented.
the EU adopted interim compensation for The US cling peach industry complained
the US under which the EU continued to in 1994 that the EU had failed to observe
apply pre-accession tariff levels on imports and enforce a commitment made in the 1985
into the three countries on most of the US-EU Canned Fruit Agreement (CFA) to
affected industrial products, but provided no not subsidize EU processing operations for
compensation in agriculture. In December peaches in syrup. The US industry claimed
1995, the EU and US concluded negotiations implementation of the EUs minimum
on the permanent compensation owed to grower price and fruit withdrawal programs
the US. Some of the concessions were in was undermining the no-processing
the form of acceleration of tariff reductions subsidies commitment made by the EU in the
agreed in the Uruguay Round, while others CFA, and that the sale of subsidized Greek
involved reductions of tariffs beyond levels canned peaches in the US and a number of
agreed in the Round or the establishment of foreign markets, including Japan, Mexico,
tariff-rate quotas. and Canada, was harming the US industry.
EU member states had widely differing The level of software piracy continued
standards, testing and certification to be a source of concern in Germany, as
procedures in place for some products. in other large developed markets. The
These differences served as barriers to the effects of Germanys 1993 implementation
free movement of these products within the of the EUs software directive, as well as
EU and could cause lengthy delays in sales an educational campaign by the software
due to the need to have products tested and industry, might have helped reduce piracy
certified to account for differing national from previous levels.
requirements. Nonetheless, the political US express package services like UPS and
will and the advent of the New approach, Federal Express remained concerned that
which streamlined technical harmonization the prevalence of postal monopolies in many
and the development of standards for EU countries restricted their market access
certain product groups, based on minimum and subjects them to unequal competitive
health and safety requirements, generally conditions. Proposals to liberalize many
pointed toward the harmonization of laws, postal services and to otherwise constrain
regulations, standards, testing, quality and the advantages enjoyed by the monopolies
certification procedures in the EU. The might not be sufficient to fully redress these
European standardization process had been problems.
closed to US firms direct participation.

In 1992, the EU established a calendar divided nationalist from internationalist

for liberalizing the cabotage practice. While business groupings. Following World War
cabotage within peninsular Spain had been II, internationalist firms were attracted
liberalized, the EU had allowed Spain to to Central America as a potential site
restrict merchant navigation to and within for low-wage manufacturing and these
the Balearic Islands, the Canary Islands interests sought three policy changes to
and Ceuta and Mililla to Spanish flag aid the movement of foreign investment to
merchant vessels until 1999. The benefits the region: (1) expanded aid programs to
of EU law in the aviation and maritime improve Central Americas poor economic
areas were reserved to firms majority- infrastructure, (2) changes in US tax law
owned and controlled by EU nationals. to encourage American direct foreign
In addition, the EU Commission had investment abroad, and (3) reductions in
proposed that companies wishing to benefit US tariff rates so that goods produced by
from the mutual recognition of licenses US firms in Central America and elsewhere
for the provision of satellite network or could be profitably exported back to the US.
communications services be 75 percent Nationalist firms, fearing increased import
owned, and effectively controlled by, EU competition, opposed all three initiatives.
nationals. But they succeeded in blocking only the
French regulations prohibited the import latter. US tariff levels remained high until
of poultry products, except offal, from the implementation of the Kennedy Round of
US. A French decree of 1962 banned imports tariff cuts in the Sixties. Discouraged from
of poultry products from countries using using Central America as an export platform
arsenicals in poultry feed, as is the case with by US tariff barriers, internationalist firms
American poultry. The US had renewed its began to advocate Import Substitution
objection to this barrier, which was imposed Industrialization in Central America and
only by France. While harmonization of other Third World countries (Cox).
policies within the EU may end this ban, World business has consistently
the US may continue to monitor this issue recommended to governments that they
closely (1996 National Trade). could best support further growth of
electronic commerce by focusing their
Politics and Laws energies on providing a basic legal and
institutional framework that ensures
Political risk in international business effective competition as well as general
entails discontinuities occurring in the trust through more predictable and media-
business environment that are: difficult to neutral rules. The European Commission
anticipate; and that result from political has stated that its primary aim in issuing
action(s) or changes(s) that possess the the draft -- Directive on Certain Legal
potential to significantly affect profit or Aspects of Electronic Commerce -- is to
business goals. Political risk may be related respond to this call from the private sector
to political instability but it need not be and to eliminate certain legal obstacles that
(Kluyver). remain to the online provision of services,
Coxs account of business conflict particularly for small and medium sized
over US policy toward Central America enterprises (Policy Statement). Politics
progressed through three stages. During and laws are therefore, very important in
the 1950s and 60s, the principal cleavage international business. They determine how

and where international business may be The International Economic

conducted by an organization or country. In Activity of the Nation: The
terms of international business, brotherhood, Balance of Payments
friendship and universality of laws are the
best ideas. Economists keep score by looking at
income statements and balance sheets and in
The Theory of International the area of international economics, the key
Trade and Investment accounts are a nations balance of payments.
A countrys balance of international
International trade is not limited to payments is a systematic statement of all
commodities that some countries produce economic transactions between that country
and others do not. Countries sometimes and the rest of the world, the statement of
import goods that they themselves could international economic activity of a nation.
produce more cheaply than the countries The major components of the balance of
from which they get them. It has been payments account are the current account
claimed, for example, that Britain could raise and the capital account (Samuelson and
dairy produce more cheaply than Denmark. Nordhaus, 682).
But Britain nevertheless imports part of
its supplies from that country and devotes International Financial Markets
its main energies to producing machinery,
electrical equipment, motor vehicles and Todays financial market is truly
other manufactures, because its advantages international in scope but the international
over Denmark in producing these things legal system continues to be based on the
are greater than its advantages in producing principle of national sovereignty. The result
dairy produce. This concentration on is an absence of international institutions
manufacturers involves what is known capable of regulating this global financial
as the principle or the law of comparative market and criminal organizations taking
costs, or simply comparative advantages advantage of this situation in pursuit of
in this theory of international trade. As their interests. The activities of criminal
applied to international trade this means that organizations on the global financial market
a country tends to concentrate on producing should be of great concern because they
those things that will give it the best return constitute formidable obstacles to law
for any given investment of its productive enforcement and represent a risk to the
resources. The law of comparative costs stability of these markets that are especially
is an extension of the principle of division sensitive to exchange rates and taxes. The
of labor to the international field (Gartside, international financial market is an extremely
195). The theory of international investment important mechanism in the global economy,
explains international capital movements as it enables the international allocation of
in the context of international production capital, and as such should be protected by
and trade. International investment creates appropriate institutions.
international production and is integrated via Criminal organizations are capable of
international trade. Knowledge, know-how using the international financial market
and technology are generally transferred to take advantage of the limitations of an
between countries along with financial international legal system based on the
capital (Wei). sovereignty of states; and a wide variety of

economic crimes can be facilitated by using of transportation and communication have

the international financial market to move reduced the costs of transporting
capital throughout the globe quickly and goods, services, and factors of production
often anonymously. The primary motive and of communicating economically useful
of criminal participants in the international knowledge and technology. Secondly, the
financial market, to avoid law enforcement tastes of individuals and societies have
efforts, can lead to rapid shifts of capital that generally, but not universally, favored
have the potential of disrupting the stability taking advantage of the opportunities
of this market. Institutions designed to provided by declining costs of transportation
regulate the international financial market and communication through increasing
must be based on voluntary cooperation, economic integration. And thirdly, public
and the international community should policies have significantly influenced the
recognize its collective interest in regulating character and pace of economic integration,
the global financial market and establish although not always in the direction of
institutions for that purpose (Sussman). increasing economic integration (Mussa).
Economic Integration Market Transitions and

Development
International economic integration is not
a new phenomenon. Some communication In Seattle, there was a manifestation of
and trade took place between distant popular opposition to globalization under
civilizations even in ancient times and since the rules of free trade. What has been
the travels of Marco Polo seven centuries happening in Puerto Rico is a result of that
ago, global economic integration same process in a specific national context.
through trade, factor movements, and This is the country where the free trade model
communication of economically useful of development was implemented decades
knowledge and technologyhas been on before it became the dominant global
a generally rising trend. This process of paradigm. For over 100 years, the history of
globalization in the economic domain has the Puerto Rican people has been dominated
not always proceeded smoothly; nor has it by the economic imperatives of the US,
always benefited all whom it has affected. living under an evolving colonialism that has
However, despite occasional interruptions, always manifested itself in significant socio-
such as following the collapse of the Roman political movements against it. The general
Empire or during the interwar period in this strike in 1998 was a representative of such
century, the degree of economic integration movements. This strike made transitions in
among different societies around the world relation to international market integration
has generally been rising. Indeed, during difficult. Thus, new international businesses
the past half century, the pace of economic were not allowed to integrate in the market
globalization has been particularly rapid. healthily (Diaz).
With the exception of human migration, The Brattle Groups extensive experience
global economic integration today is greater with all sectors of the natural gas industry,
than it ever has been (Crafts). from wellhead to burner-tip, allowed them to
Three fundamental factors have affected offer a wide array of services to firms looking
the process of economic globalization. to thrive in this complex environment. In
Firstly, improvements in the technology addition to their North American practice,

The Brattle Group as their way into Multinational Corporations

international market development, began
assisting clients in the United Kingdom, A multinational corporation is a company
Europe, Australia, and other countries or enterprise operating in several countries,
in policy debates over privatization and usually defined as one that has 25% or
restructuring of their natural gas industries. more of its output capacity located outside
They also provided testimony in their its country of origin. The worlds four
regulatory and legal proceedings regarding largest multinationals in 1994 were General
such matters as pricing and terms of access, Motors, Ford, Exxon, and Shell. Their
and conflicts over contract performance total sales exceeded the gross national
(The Brattle Group). product of all of Africa, and the top 100
multinational corporations controlled
International Business Research $3.4 trillion in financial assets. In 1993,
multinational corporations accounted for
International business research is about one-third of the worlds industrial output,
enterprise development in another country. with sales of $4,800 billion. They are
Enterprise development, especially as it seen in some quarters as posing a threat
relates to micro, small and medium enterprise to individual national sovereignty and as
development is a complex endeavor with exerting undue influence to secure favorable
many facets such as looking into the policy operating conditions. Unsuccessful efforts
environment, entrepreneurship, innovation, were made 1992, under UN auspices, to
competitiveness, subcontracting, etc. negotiate a voluntary code of conduct
These facets of international enterprise for multinationals, but governments and
development require investigation and corporations alike were hostile to this idea.
research (Enterweb). In 1993, 11 of the 100 largest multinational
corporations were British (Hutchinson
International Business Entry Family Encyclopedia).
Many countries make it attractive to Strategic Planning in

incorporate in their area, even when activities International Business
are to be conducted elsewhere. In fact, there
are so many tax efficient jurisdictions that The strategic plan must be developed and
an initial problem for most organizations owned by the management team that has the
wanting to form an international business job of implementing it. Strategic planning
company, is how to select from the available in international business must be objective.
options. Belize is such a country that After international business research, the
entered the offshore industry after carefully international business must define a first year
analyzing and adopting the best features of operating budget, build an infrastructure
some of the best offshore jurisdictions in flexible enough to meet expansion needs,
existence. Its long history of democracy and and prepare the company for expansion. It
stability, enhanced by its legal system which must understand different growth options
is based on English common law, have available for developing transatlantic
made it the premier source for easy market operations, and know the risks and rewards
transition (Belize Offshore Consultants). of each. The international business must also
build a long-range plan, and then adhere to

that plan. It must find the right mix of direct factors needed for lucrative international
and indirect operations, and the right rollout marketing. These factors include knowledge
sequence to keep risk low while maximizing of culture, political risk evaluation, etc.
longer-term market share and revenue
potential. To craft a good business strategy International Services
for international success, the company
must have broadened awareness, coupled International businesses like Shell provide
with business experience. Specialized international services. Shell Services
organizations such as Atlas Venture help International provides, among many other
international businesses formulate the specialized services, electricity in the US.
strategic plan (Atlas Venture). Shell Energy was formed in 1997 to pursue
new growth opportunities in the retail
International Marketing electricity and gas markets that have recently
been opening to competition. Today, Shell
International Marketing Services, Inc, Energy serves more than 300,000 gas and
(IMS) is a uniquely positioned international electricity customers in Ohio and Georgia
marketing firm. Since 1986, it has assisted (Shell Services International).
over five hundred US, European and Russian
companies export products, develop joint International Logistics and
ventures, locate foreign investment, and Supply-Chain Management
form strategic relationships. Its extensive
experience with business practices and International logistics are about
culture in North America, East-Central international freight forwarding, moving
Europe and the C.I.S. is the foundation of and storage, warehousing and storage,
its success. IMS cross-cultural perspective project shipping, office building and letters
allows it to rapidly overcome significant of credit. Currently, there are organizations
obstacles to export sales and joint ventures that specialize in making the task of
that most firms are unequipped to deal with international logistics easier for international
alone. The company also brings broad business (International Logistics
analytical skills to bear on its global projects. Management). A supply chain is a network
This includes substantial international trade of facilities and distribution options that
and joint venture negotiations expertise, performs the functions of procurement of
market and competitive assessments, materials, transformation of these materials
macroeconomic, financial and statistical into intermediate and finished products, and
analyses, and political risk evaluations. In the distribution of these finished products to
addition to broad skills, IMS employees hail consumers. The geographic placement of
from diverse backgrounds and industries. production facilities, stocking points, and
The company offers a full spectrum of sourcing points is the first step in creating
services ranging from one-time primary a supply chain. The strategic decisions
market research assignments to marketing include what products to produce, and which
and strategy consultation where it walks plants to produce them in, allocation of
its clients through each difficult stage of suppliers to plants, etc. Inventory decisions
expanding their business presence abroad refer to means by which inventories are
(International Marketing Services). The managed. And the mode choice aspect of
successful operation of IMS indicates the

transportation decisions is the more strategic tax paid by the Maltese company on income
ones (Ganeshan and Harrison). arising from these foreign holdings. Such
refund is triggered upon a distribution of
Multinational Financial this income to the non-resident shareholders
Management of the Maltese firm (Accounting and
Taxation).
International trade, financing and
investments have grown at an extremely International Human Resource
rapid pace in recent years, and the operations Management
of corporations are increasingly becoming
multinationalized. Corporate executives Effective management of expatriate
buying and selling goods and services, and managers is one of the most important
making financing and investment decisions areas in human resource management, and
across national boundaries, have formulated there is abundant research on expatriates
policies and procedures for managing cash of American, European, and Japanese
flows denominated in foreign currencies. multinationals with large expatriate
These policies and procedures, and the populations. In the age of globalization,
related managerial actions of executives, strategic management of human resources
change as new relevant information is becoming critical for organizational
becomes available and this field is that of survival. Global business environments
multinational financial management (Reid). demand flexibility and rapid response and
there is a growing realization that the human
Countertrade dimension provides the key to flexibility and
adaptability in organizations. Multinational
Countertrade simply refers to listening and transnational corporations have been
to the companys international customers increasingly aware of the growing necessity
and meeting their needs. This could be to have not only international business
in the areas of hard currency generation, strategies, but also international human
technology transfer, or marketing assistance resource strategies. Just as international
(The American Countertrade). business strategy is likely to have unique
features and needs to be understood well,
International Accounting and international management of human
Taxation resources has its own peculiarities that have
to be managed well to be able to flourish in
In Malta, taxation of an international overseas operations (Naresh).
trading company (ITC) is based on the
Tax Refund Mechanism. An ITC is taxed Organization, Implementation,
at the normal company rate of tax, which and Control of International
is currently about 35%. Non-resident Operations, and their Future
shareholders of an ITC are taxed at a flat
rate of 27.5% on all distributions received Effective management of international
from the firm. Non resident shareholders business operations includes efficient
of an international holding company which management of finance, personnel,
has a participating holding in a non-resident product development, marketing, and
company qualify for a full refund of the Malta communication. This is so that the

organization, implementation and control Gartside, L. 1987. Commerce A Guide

of the operations go well. The future to the Business World. London: Pitman
of international business is bright as Publishing.
globalization and the need for universality
continue. International Business Customs. 21 June
2001. <www.eglobalgreetings.com>.
References
International Logistics Management.
1996 National Trade. 1996. 1996 National International Logistics Management,
Trade Estimate European Union. 21 Ltd. 21 June 2001. <www.ilm-vietnam.
June 2001. <www.ustr.gov/>. com>.
Accounting and Taxation. 21 June 2001. International Marketing Services. 2000.

Accounting and Taxation of International International Marketing Services, Inc. 21
Companies. 21 June 2001. <www.inter- June 2001. <www.consultims.com>.
shore.com>.
Kluyver, de. Global Trade. The Political
Atlas Venture. Strategic Planning. 21 June Imperative in International Business. 21
2001. <www.atlasventure.com>. June 2001. <www.gmu.edu/departments/
t-icp/>.
Belize Offshore Consultants. 1999. Belize
Offshore Consultants, Inc. 21 June 2001. Mussa, Michael. 25 August 2000. Factors
<www.offshorebiz.com>. Driving Global Economic Integration.
IMF Working Paper. Washington, DC:
Cox, Ronald W. 1994. Power and Profits: International Monetary Fund.
U.S. Policy in Central America.
Lexington: University Press of Kentucky. Naresh, Khatri. 1998. Singaporean
expatriates in China: success factors.
Crafts, Nicholas. March 2000. Does Asian management matter?
Globalization and Growth in the Inaugural Conference of the Asian
Twentieth Century. IMF Working Academy of Management. Hong Kong.
Paper, WP/00/44. Washington, D.C.: 147-150.
International Monetary Fund.
Policy Statement. 27 July 1999.
Diaz, Jean. Latin American Perspectives. 21 International business comments on
June 2001. <csf.colorado.edu/lap/>. the proposed EU directive on certain
legal aspects of electronic commerce.
Enterweb. 2000. International Business Commission on Telecommunication and
Research. 21 June 2001. <www.enterweb. Information Technologies. 21 June 2001.
org>. <www.iccwbo.org/home/>.
Ganeshan, Ram, and Harrison, Terry P. 22 Reid, Cynthia. 19 March 2000. Journal of
May 1995. An introduction to supply- Multinational Financial Management. 21
chain management. 21 June 2001. <http:// June 2001. <http://gort.ucsd.edu/>.
silmaril.smeal.psu.edu/>.

Samuelson, Paul A., and Nordhaus, William The American Countertrade. 4 June 2001.
D. 1998. Economics. 16th ed. United The American Countertrade Association.
States: Irwin/McGraw-Hill. 21 June 2001. <www.countertrade.org>.
Shell Services International. February The Brattle Group. 2000. 21 June 2001.
2001. Shell Energy: Ready to Serve <www.brattle.com>.
Texas Electric Customers. 21 June 2001.
<www.shellservices.com>. Wei, Annie. International Business and
Globalization. 21 June 2001. <www.
Sussman, Joshua. 1997. The International lancs.ac.uk/staff/>.
Financial Market, Organized Crime, &
International Law. U.S. Department of
Justice. 21 June 2001. <www.usdoj.gov>.


2010 Isola Oluwabusuyi 131
REASONING DISPARITIES BETWEEN HK AND US

MANAGERS
Isola Oluwabusuyi
Abstract
The purpose of this study was to examine and compare the factors that influence intuition as a decision-
making tool for leaders/managers in Hong Kong and in the United States. This study examined the
relationships among gender, management level, extent of management experience, country of operation,
and the reported use of intuition in decision making. Existing empirical research in this field is sparse. In
this research, attempt was made to contribute to empirical research on the viability and reported use of
intuition as a decision-making skill of leaders.
Agors Intuitive Measurement Survey (AIM) survey was adapted (with permission from copyright
owner) from Weston Agors study to measure the relationship between a managers reported use of
intuition in decision making and the managers management level, his level of management experience,
the managers gender, and the managers country of operation. Each participant was electronically sent
a link that led to a web page containing the survey questions. Once the respondent clicks submit, the
questionnaire was mailed directly to the researcher.
The research shows significant relationship between research variables. Administrative managers in
Hong Kongs reported use of intuition in decision making was significantly lower than US managers
reported use of intuition in decision making. The paper concludes by examining the implications of these
significant findings to global business management and management education.
Oluwabusuyi I. - Reasoning Disparities Between Hk And Us Managers

Introduction more and more studies are showing that

effective executives are less rational in
A new study conducted at Argosy their managerial decision making especially
Universitys school of business showed when faced with dynamic environments.
that Business owners with 10 years or less While most business schools are focused
management experience in Hong Kongs on teaching rational decision making
reported use of intuition in decision making techniques, more and more studies are
is lower than US supervisors with 10 years or showing that the impact of intuitive or non-
less management experiences reported use rational decision making techniques can no
of intuition in decision making. The study longer be ignored without costs.
had four different management categories Douglas Dean and John Mihalasky
including supervisors, managers, executives (1974) discovered through their ten year
and business owners. research that executives whose companies
The researcher (Isola, 2009) surveyed increased the most profits, scored higher
100 participants from each country involved on tests of precognition than average.
in the comparative study. In commenting They tested approximately 165 CEOs and
on the findings of the study, Dr Isola said presidents of companies in the United
this finding could represent a significant States and found that 80% of the leaders
departure from conventional thinking. who doubled or nearly doubled their profits
While the few studies in this specialized in a five-year period had above-average
area of International Business Management scores on intuitive ability (Church, 2005).
has remained inconclusive, one would at The findings of their research showed that
least expect the Chinese to be more intuitive intuitive managers were more successful
in their decision making given their very than their more analytical counterparts.
high score on Hofstedes LTO (Long term Studies by Agor (1989) also showed that top
orientation) index. executives rated significantly higher than
Managerial decision making has always middle or low level managers in intuitive
being a subject of intense academic abilities.
discuss. What is the most effective way of Intuition, like many other popular
choosing between alternatives in managing concepts is quite elusive to define. It has
companies? Should the companys new plant been defined by many scholars in various
be located in New Jersey or Guangzhou? ways. Jung (1924) defined intuition as that
In answering these questions management psychological function which transmits
scholars focused on the rational approach perception in an unconscious way. In
for several years. The idea of utilizing commenting on the diversity of definitions,
intuitive skills to make such decisions is a Isaack (1978) wrote an examination of
recent phenomenon and it remains largely the literature on intuition disclosed that
unpopular. authorities on the subject had different
Peter Drucker (1967) identified six major concepts about it Behling & Eckel (1991)
steps that should be followed by executives found eighty-seven descriptions of intuition
making decisions for their companies. during a search of two electronic data bases
In 2004, he presented another paper that in twenty-four books and articles published
further narrowed the whole process to three between 1976 and 1987. In trying to make
steps. Both approaches emphasized rational sense out of the descriptions, they recruited
decision making processes. However,

six MBA students to group the descriptions. (Myers-Briggs Type Indicator). The
Six clusters eventually emerged. instrument, therefore, uses the reliability
In one cluster, authors described and validity of MBTI (Agor, 1984). Studies
intuition as a paranormal phenomenon. Top have found strong support for construct
managers who relied on psychics to make validity, internal consistency, and test
decisions were cited as examples in this related reliability of MBTI instrument
cluster. H. L. Hunt relied on a psychic to (Thompson & Borello, 1986). Further, the
help pick oil properties (Rowen, 1986) instrument was designed to best measure all
A second cluster consists of descriptions the variables in this study. The questionnaire
of Intuition as a personality trait. Scholars was modified and simplified so it contained
in this category saw intuitive decision clear instructions, questions, and possible
making as genetic or acquired very early in answers.
life (Behling & Eckel, 1991). Three of the
remaining four clusters described intuition Research Questions
as an unconscious process, a set of actions,
and distilled experience. The fourth cluster The studys four research questions
was named the residual category. Authors in explored the relationship between the studys
this category defined intuition as decisions independent and dependent variables.
not made using the rational decision making
process. Research Question 1: What is the
relationship between a leaders
Methodology management level and the leaders
reported use of intuition in decision
The quantitative research employed making?
Agors Intuitive Measurement Survey (with
permission by copyright owner). The AIM Research Question 2: What is the
survey was administered to 100 participants relationship between a managers
from the US and 100 participants from work experience as a manager and the
Hong Kong. Questions on the survey were managers reported use of intuition in
developed to measure all independent and decision making?
dependent variables.
Questions on the survey measured the Research Question 3: What is the
following variables: Gender, Management relationship between a managers sex
Level, Management Experience, Country of and a managers reported use of intuition
Operation, and Reported use of intuition in in decision making?
decision making. Reported use of intuition
in decision making was the only dependent Research Question 4: What is the
variable in the study, the remaining four relationship between a managers
variables were independent variables. country of operation and a managers
reported use of intuition in decision
Validity and Reliability of the making?
Instrument
AIM Survey (Agors Intuitive

Measurement Survey) is a modified MBTI

Assumptions and Limitations than US supervisors with 10 years or less

management experiences reported use of
The following assumptions were intuition in decision making. Mean scores
formulated that were central to the design of of Hong Kong Business owners with 10
this research. years or less management experience
Respondents understand the questions and US supervisors with 10 years or less
and are able to answer all of them in the management experience are 6.0 and 6.35
questionnaire. respectively. This result is significant at the
Answers to the questionnaire are given .022 level.
with the respondents knowledge and Additional Finding 3: Supervisors with
honesty. 10 years or less management experience in
Data collection process was authentic. Hong Kongs reported use of intuition in
Analysis tools were accurate decision making is lower than US supervisors
Below are some limitations that may with over 10 years management experiences
influence the results. These limitations reported use of intuition in decision making.
include: Mean scores of Supervisors with 10 years
The samples were a combination or less management experience in Hong
of criterion, maximum variation and Kong and US supervisors with over 10 years
convenience based selection of companies management experience are 6.6363 and 8.00
in Hong Kong and the United States. respectively. This result is significant at the
The United States samples were selected 0.01 level.
from the East and West coast of the United Additional Finding 4: US supervisors with
States. 10 years or less management experiences
reported use of intuition in decision making
Findings is lower than US supervisors with over 10
years management experiences reported use
The studys seven null hypotheses were of intuition in decision making. Mean scores
not rejected but a further analysis of gathered of US supervisors with 10 years or less
data uncovered forty-one significant findings management experience and US supervisors
discussed below. with over 10 years management experience
Additional Finding 1: Administrative are 6.35 and 8.00 respectively. This result is
managers in Hong Kongs reported use of significant at the 0.01 level.
intuition in decision making is lower than Additional Finding 5: Managers with 10
US administrative managers reported use years or less management experience in Hong
of intuition in decision making. Mann- Kongs reported use of intuition in decision
Whitney U results showed the mean rank of making is lower than US supervisors with
Hong Kong and US administrative managers over 10 years management experiences
are 13.17 and 20.94, respectively. Mean reported use of intuition in decision making.
scores of Hong Kong and US administrative Mean scores of Managers with 10 years or
managers are 6.333 and 8.11 respectively. less management experience in Hong Kong
This result is significant at the .023 level. and US supervisors with over 10 years
Additional Finding 2: Business management experience are 6.8636 and
owners with 10 years or less management 8.00 respectively. This result is significant at
experience in Hong Kongs reported use the 0.01 level.
of intuition in decision making is lower

Additional Finding 6: US managers with respectively. This result is significant at the

10 years or less management experiences 0.01 level.
reported use of intuition in decision making Additional Finding 10: US executives
is lower than US supervisors with over 10 with 10 years or less management
years management experiences reported experiences reported use of intuition in
use of intuition in decision making. Mean decision making is lower than US supervisors
scores of US managers with 10 years or less with over 10 years management experiences
management experience and US supervisors reported use of intuition in decision making.
with over 10 years management experience Mean scores of US executives with 10 years
are 7.1304 and 8.00 respectively. This result or less management experience and US
is significant at the 0.01 level. supervisors with over 10 years management
Additional Finding 7: Managers with experience are 6.666 and 8.00 respectively.
more than 10 years management experience This result is significant at the 0.01 level.
in Hong Kongs reported use of intuition in Additional Finding 11: Executives with
decision making is lower than US supervisors more than 10 years management experience
with over 10 years management experiences in Hong Kongs reported use of intuition in
reported use of intuition in decision making. decision making is lower than US supervisors
Mean scores of Managers with more than with over 10 years management experiences
10 years management experience in Hong reported use of intuition in decision making.
Kong and US supervisors with over 10 years Mean scores of Executives with more than
management experience are 7.6666 and 8.00 10 years management experience in Hong
respectively. This result is significant at the Kong and US supervisors with over 10 years
0.01 level. management experience are 7.25 and 8.00
Additional Finding 8: US managers with respectively. This result is significant at the
more than 10 years management experiences 0.01 level.
reported use of intuition in decision making Additional Finding 12: US executives
is higher than US supervisors with over 10 with over 10 years management experiences
years management experiences reported use reported use of intuition in decision making
of intuition in decision making. Mean scores is lower than US supervisors with over 10
of US managers with more than 10 years years management experiences reported
management experience and US supervisors use of intuition in decision making. Mean
with over 10 years management experience scores of US executives with over 10 years
are 8.5714 and 8.00 respectively. This result management experience and US supervisors
is significant at the 0.01 level. with over 10 years management experience
Additional Finding 9: Executives with are 7.6666 and 8.00 respectively. This result
10 years or less management experience in is significant at the 0.01 level.
Hong Kongs reported use of intuition in Additional Finding 13: Business
decision making is lower than US supervisors owners with 10 years or less management
with over 10 years management experiences experience in Hong Kongs reported use
reported use of intuition in decision making. of intuition in decision making is lower
Mean scores of Executives with 10 years or than US supervisors with over 10 years
less management experience in Hong Kong management experiences reported use of
and US supervisors with over 10 years intuition in decision making. Mean scores
management experience are 7.125 and 8.00 of Business owners with 10 years or less
management experience in Hong Kong

and US supervisors with over 10 years experiences reported use of intuition in

management experience are 6.00 and 8.00 decision making. Mean scores of US business
respectively. This result is significant at the owners with 10 years or less management
0.01 level. experiences and Hong Kong supervisors
Additional Finding 14: Business owners with over 10 years management experience
with over 10 years management experience are 8.00 and 6.0 respectively. This result is
in Hong Kongs reported use of intuition in significant at the 0.01 level.
decision making is lower than US supervisors Additional Finding 18: US male
with over 10 years management experiences managers with more than 20 years
reported use of intuition in decision making. management experiences reported use of
Mean scores of Business owners with over intuition in decision making is lower than
10 years management experience in Hong US female managers with between 6 and
Kong and US supervisors with over 10 years 10 years management experiences reported
management experience are 7.7142 and 8.00 use of intuition in decision making. Mean
respectively. This result is significant at the scores of US male managers with more
0.01 level. than 20 years management experience
Additional Finding 15: US business and US female managers with between 6
owners with over 10 years management and 10 years management experience are
experiences reported use of intuition in 7.1875 and 7.238 respectively. This result is
decision making is lower than US supervisors significant at the .013 level.
with over 10 years management experiences Additional Finding 19: Female
reported use of intuition in decision making. managers with between 6 and 10 years
Mean scores of US business owners with management experience in Hong Kongs
over 10 years management experience reported use of intuition in decision making
and US supervisors with over 10 years is higher than Hong Kong female managers
management experience are 6.18181 and with between 11 and 20 years management
8.00 respectively. This result is significant at experiences reported use of intuition in
the 0.01 level. decision making. Mean scores of Female
Additional Finding 16: Business owners managers with between 6 and 10 years
with over 10 years management experience management experience in Hong Kong and
in Hong Kongs reported use of intuition Hong Kong female managers with between
in decision making is higher than US 11 and 20 years management experience are
managers with 10 years or less management 6.3809 and 6.333 respectively. This result is
experiences reported use of intuition in significant at the .032 level.
decision making. Mean scores of Business Additional Finding 20: US female
owners with over 10 years management managers with more than 20 years
experience in Hong Kong and US managers management experiences reported use of
with 10 years or less management experience intuition in decision making is higher than
are 7.7142 and 7.1304 respectively. This Hong Kong female managers with 5 years or
result is significant at the 0.01 level. less management experiences reported use
Additional Finding 17: US business of intuition in decision making. Mean scores
owners with 10 years or less management of US female managers with more than 20
experiences reported use of intuition in years management experience and Hong
decision making is higher than Hong Kong Kong female managers with 5 years or less
supervisors with over 10 years management management experience are 7.666 and 7.1

respectively. This result is significant at the intuition in decision making. Mean scores
.020 level. of US female managers with between 6
Additional Finding 21: Female and 10 years management experience and
managers in Hong Kong with 5 years or US male managers with 5 years or less
less management experiences reported use management experience are 7.238 and 9.5
of intuition in decision making is lower respectively. This result is significant at the
than US male managers with 5 years or less .010 level.
management experiences reported use of Additional Finding 25: Female managers
intuition in decision making. Mean scores of in Hong Kong with between 11 and 20
Female managers in Hong Kong with 5 years years management experiences reported
or less management experience and US male use of intuition in decision making is lower
managers with 5 years or less management than US male managers with 5 years or
experience are 7.1 and 9.5 respectively. This less management experiences reported
result is significant at the .010 level. use of intuition in decision making. Mean
Additional Finding 22: US female scores of Female managers in Hong Kong
managers with 5 years or less management with between 11 and 20 years management
experiences reported use of intuition in experience and US male managers with 5
decision making is lower than US male years or less management experience are
managers with 5 years or less management 6.333 and 9.5 respectively. This result is
experiences reported use of intuition in significant at the .010 level.
decision making. Mean scores of US female Additional Finding 26: US female
managers with 5 years or less management managers with between 11 and 20 years
experience and US male managers with 5 management experiences reported use of
years or less management experience are intuition in decision making is lower than
6.35 and 9.5 respectively. This result is US male managers with 5 years or less
significant at the .010 level. management experiences reported use of
Additional Finding 23: Female intuition in decision making. Mean scores of
managers in Hong Kong with between 6 and US female managers with between 11 and 20
10 years management experiences reported years management experience and US male
use of intuition in decision making is lower managers with 5 years or less management
than US male managers with 5 years or experience are 7 and 9.5 respectively. This
less management experiences reported result is significant at the .010 level.
use of intuition in decision making. Mean Additional Finding 27: Female
scores of Female managers in Hong Kong managers in Hong Kong with more than 20
with between 6 and 10 years management years management experiences reported
experience and US male managers with 5 use of intuition in decision making is lower
years or less management experience are than US male managers with 5 years or less
6.380 and 9.5 respectively. This result is management experiences reported use of
significant at the .010 level. intuition in decision making. Mean scores
Additional Finding 24: US female of Female managers in Hong Kong with
managers with between 6 and 10 years more than 20 years management experience
management experiences reported use of and US male managers with 5 years or less
intuition in decision making is lower than management experience are 6.875 and 9.5
US male managers with 5 years or less respectively. This result is significant at the
management experiences reported use of .010 level.

Additional Finding 28: US female Hong Kong and US administrative managers

managers with more than 20 years are 6.8571 and 9.5 respectively. This result
management experiences reported use is significant at the .010 level.
of intuition in decision making is lower Additional Finding 32: US male
than US male managers with 5 years or managers with between 11 and 20 years
less management experiences reported management experiences reported use of
use of intuition in decision making. Mean intuition in decision making is lower than
scores of US female managers with more US male managers with 5 years or less
than 20 years management experience and management experiences reported use of
US male managers with 5 years or less intuition in decision making. Mean scores of
management experience are 7.666 and 9.5 US male managers with between 11 and 20
respectively. This result is significant at the years management experience and US male
.010 level. managers with 5 years or less management
Additional Finding 29: Male managers experience are 7.2857 and 9.5 respectively.
in Hong Kong with between 6 and 10 years This result is significant at the .010 level.
management experiences reported use of Additional Finding 33: Male managers
intuition in decision making is lower than in Hong Kong with more than 20 years
US male managers with 5 years or less management experiences reported use of
management experiences reported use of intuition in decision making is lower than
intuition in decision making. Mean scores of US male managers with 5 years or less
Male managers in Hong Kong with between management experiences reported use of
6 and 10 years management experience intuition in decision making. Mean scores
and US male managers with 5 years or less of Male managers in Hong Kong with more
management experience are 7.0625 and 9.5 than 20 years management experience and
respectively. This result is significant at the US male managers with 5 years or less
.010 level. management experience are 8.125 and 9.5
Additional Finding 30: US male respectively. This result is significant at the
managers with between 6 and 10 years .010 level.
management experiences reported use of Additional Finding 34: US male
intuition in decision making is lower than managers with over 20 years management
US male managers with 5 years or less experiences reported use of intuition in
management experiences reported use of decision making is lower than US male
intuition in decision making. Mean scores of managers with 5 years or less management
US male managers with between 6 and 10 experiences reported use of intuition in
years management experience and US male decision making. Mean scores of US male
managers with 5 years or less management managers with over 20 years management
experience are 7.00 and 9.5 respectively. This experience and US male managers with 5
result is significant at the .010 level. years or less management experience are
Additional Finding 31: Male managers 7.187 and 9.5 respectively. This result is
in Hong Kong with between 11 and 20 significant at the .010 level.
years management experiences reported Additional Finding 35: US male
use of intuition in decision making is lower managers with over 20 years management
than US male managers with 5 years or less experiences reported use of intuition
management experiences reported use of in decision making is higher than male
intuition in decision making. Mean scores of managers in Hong Kong with between 11

and 20 years management experiences in decision making is higher than US

reported use of intuition in decision making. managers with 5 years or less management
Mean scores of US male managers with experiences reported use of intuition
over 20 years management experience and in decision making. Mean scores of US
male managers in Hong Kong with between managers with between 11 and 20 years
11 and 20 years management experience are management experience and US managers
7.1875 and 6.8571 respectively. This result with 5 years or less management experience
is significant at the .025 level. are 7.4 and 6.93478 respectively. This result
Additional Finding 36: US male is significant at the .009 level.
executives reported use of intuition in Additional Finding 41: US managers
decision making is lower than US female with more than 20 years management
managers reported use of intuition in experiences reported use of intuition
decision making. Mean scores of US male in decision making is higher than US
executives and US female managers are managers with between 6 and 10 years
7.388 and 7.461 respectively. This result is management experiences reported use of
significant at the .005 level. intuition in decision making. Mean scores
Additional Finding 37: US female of US managers with more than 20 years
business owners reported use of intuition management experience and US managers
in decision making is higher than Hong with between 6 and 10 years management
Kong female executives reported use of experience are 7.44 and 7.0 respectively.
intuition in decision making. Mean scores This result is significant at the .046 level.
of US female business owners and Hong
Kong female executives are 7.25 and 6.307 Conclusion
respectively. This result is significant at the
.038 level. The study discovered that United States
Additional Finding 38: Hong Kong managers/ leaders were less rational in their
male executives reported use of intuition managerial decision making than Hong
in decision making is higher than Hong Kong leaders/managers in twenty of twenty-
Kong male business owners reported use of two categories. These findings can be quite
intuition in decision making. Mean scores significant to global business managers
of Hong Kong male executives and Hong intending to shift more of their activities
Kong male business owners are 7.933 and to Asia in the near future. Multinational
6.583 respectively. This result is significant corporations would have to provide more
at the .022 level. data for their Chinese managers operating
Additional Finding 39: US business out of Hong Kong in order to make them
owners reported use of intuition in decision more comfortable with their decision
making is lower than Hong Kong executives making tasks as they tend to rely more on
reported use of intuition in decision making. data based decision-making techniques.
Mean scores of US business owners and Business schools in the West may also
Hong Kong executives are 7.0476 and need to redesign their curriculum as more
7.1785 respectively. This result is significant business leaders feel more comfortable
at the .002 level. with intuitive decision making techniques.
Additional Finding 40: US managers This will make business schools more
with between 11 and 20 years management relevant to what obtains outside school
experiences reported use of intuition walls and produce graduates that will be

more amenable to top executive position industrysolutions//newsArchive/

appointments. news2006/august/femaleIntui.htm
The author would also like to recommend
more comparative studies in managerial Drucker, P. F. (1967). The effective decision
decision-making so that a clearer picture of [Electronic version] Management
intuitive managerial decision making can Review, 56, 52-55
emerge.
Drucker, P. F. (2004). What makes an
References effective executive? [Electronic version]
Harvard Business Review, 82, 58-63
Agor, W.H. (1984). Intuitive management:
Integrating left and right brain Drucker, P. F. (2007). Management: Tasks,
management skills. Englewoods Cliffs, responsibilities, practices New York, NY
NJ: Prentice Hall. Harper Collins Publishers
Agor, W. H. (1986). The logic of Intuitive Fields A. F. (2003) A Study of Intuition in

decision-making: A research- based Decision-Making using Organizational
approach to top management. New York: Engineering Methodology. Dissertation
Quorum Books. Abstracts International, 66 (07), 2474.
(UMI No. 3182016)
Agor, W. H. (1989). Intuition in
organizations: Leading and managing Hodgetts R.M., Luthans F., & Doh J.P.
productively. Newbury Park, CA: Sage. (2006) International management:
Culture, strategy and behavior New York,
Agor, W.H. (1991). The logic of Intuition: NY McGraw Hill/Irwin
How executives make important
decisions: Intuition in organizations. Hofstede, G. (2003) Geert Hofstede cultural
Newbury Park, CA: Sage Publications. dimensions Itim International Retrieved
August 10, 2006 from http://www.geert-
Atsunyo, M. (1992). A comparative study hofstede.com/hofstede_china.shtml
of executive decision making in the
United States and Ghana University of Isenberg, D. (1984). How senior managers
Massachusetts, Amherst, MA. think. Harvard Business Review, 62, 81-
90.
Church, M. J. (2005) Intuition, leadership,
and decision making: A phenomenon Robbins, S. P. (2004) Organizational
Dissertation Abstracts International, behavior (11th ed.). Upper Saddle River,
66(05), 1847. (UMI No. 3177390) N.J: Prentice Hall.
Downey, L. (2006) Female intuition Robbins, S. P., & Judge T.A. (2007)
and emotional intelligence linked to Organizational behavior (12th ed.).
management success Swinburne industry Upper Saddle River, New Delhi: Prentice
solutions Retrieved December, 2006 from Hall of India.
http://www.swinburne.edu.au/corporate/

2010 Arpan Kumar Kar, Ashis Kumar Pani and Supriya Kumar De 141
A STUDY ON USING BUSINESS INTELLIGENCE FOR

IMPROVING MARKETING EFFORTS
Arpan Kumar Kar, Ashis Kumar Pani and Supriya Kumar De
Abstract
A major success criteria for the marketing department of the company is the understanding of the
consumers and the success of the campaigns. Based on the outputs of certain data mining tasks, a marketer
would be able to chalk out his strategy. But the major barrier against proper utilization of business
intelligence is that often the marketer is unable to see the actual benefits the system can provide him
in meeting his objective. Firstly this article provides a basic knowledge of the various data-mining tasks
which can create value to the marketer. A major concern becomes, even if the marketer is convinced of
the benefits to be obtained, he is unaware of the process involved to get it done. The data requirements
are huge for a business intelligence system. This article also provides some insights on what kind of data
requirements the data-mining system would require to perform such business intelligence tasks. Thirdly,
this article maps how business intelligence can be collaborated with marketing strategy to create value in
each stage of the product development life cycle following Kotlers product life cycle framework. The
entire paper provides these conceptualizations based on theoretical understanding of the areas under
discussion.
Kumar Kar A., Kumar Pani A., Kumar De S. - A Study on Using Business Intelligence for Improving Marketing Efforts
1. Introduction 2. Theoretical Discussion
A major expense associated with Traditional views of marketing, like that

the marketing department is that from of Kotler and Kelly (2006) have mainly
advertisements. Total global advertising focused on the physical and human aspects
expenditure reached US$290 billion in 1999, of the organization. The information view
an increase of over US$15 billion on 1998. of marketing started getting conceptualized
In real terms, global ad-spend increased by with contributions from Haeckel and Nolan
more than 3% from 1998 onwards. It has been (1993), Naude and Holland (1995), Rayport
established empirically and mathematically and Sviokla (1995). More recently Holland
that investment on advertising improves and Naude (2004) argued that marketing
consumer knowledge about the product should increasingly be viewed as information
and hence affects product sales positively. handling problem rather than the classical
Kaldor and Silverman (1948) established transaction driven or the relationship driven
that advertising as a percentage of sales is approach.
remarkably similar by product in the U.S. The management approach to the study
and U.K. Telser (1961) established that of marketing can be traced to such concepts
advertising affects the profitability of firms as the marketing concept, marketing mix,
and a firm maximizing profit spends an product life cycle and market segmentation.
amount on advertising and chooses a price According to Moller (1994), this
such that the price elasticity of the demand for management approach can be characterized
its product equals the value of the marginal as trying to solve the problem of how to
sales effect of advertising. Multiple studies develop an optimal marketing mix consisting
starting as early as late 1960s (Comanor and of Product, Place, Price and Promotion
Wilson, 1967) established the relationship solutions for the competing preferences
between profit performance, advertising of a chosen target segment of consumers,
intensity, and market structure. households or organizational buyers. The
Advertisement has a huge spiraling concept of the marketing mix focuses on
cost as has already been indicated. Due to the need for marketing managers to view
the huge nature of the possible number of the marketing task as the process of mixing
consumers, it makes sense to understand or integrating several different functions
which advertisement would impact which simultaneously, as was postulated by Sheth,
consumer the most. Knowing this would Gardner and Garett (1988). To improve
not only help in serious cost cutting of the fit between the 4 Ps, understanding the
advertisement expenditure, but also improve customers become crucial, for which not
the take up rates of advertising campaigns. only data is needed on the customers, but
This is where information technology plays what can be done with the information which
a crucial role by processing information and can deliver key insights on marketing.
thus providing key business intelligence This focus on the mentioned problem
for the marketer by providing key insights. domain has opened up a new area of
This paper talks about some of the business study, known as database marketing. This
intelligence techniques which help to do so offers benefits from increased revenue to
and provides insight on which technique the marketer from two areas, increase in
may be most optimally used under which revenue by better targeting and increase
conditions. in revenue from better understanding of

customer needs and hence designing more The data-mining tasks which are used
suitable promotions. for the purpose are clustering, classification,
It has been established by Webster (1988) pattern association, summarization,
and Gronroos (1990) that marketing can no predictive modeling, link analysis and social
longer be an area of the marketing specialists network analysis. These tools have been
but that everyone in the organization classified under query tools, descriptive
must be charged with responsibility for statistics, visualization tools, regression type
customers and contributing to developing models, association rules, decision trees,
and delivering value for them. According case based reasoning, genetic algorithms
to Webster (1992), this customer focus may and graph theory.
require increasingly large investments in In this paper, a brief introduction has
information management and information been provided on how the following data
technology. The next frontier of automation mining tasks can help to reduce marketing
will be marketing and sales functions as expenditure and increase the take up rate of
was argued by Moriarty and Swartz (1989). campaigns. While there are multiple studies
Brooks (1989) argued that these functions in the area of each of these tasks in data
are likely to receive the largest investment mining, there is no study which dictates how
of technological resources in the future. It these tasks can be used effectively by the
is also evident that many companies have marketer at different stages of the product
started, or are planning to build, marketing life cycle, based on which, promotion
related IS. It is essential now to study and strategies are actually taken. This paper
classify these systems in more detail so as strives to bridge this gap.
to be able to design better and more cost- In the following part, a brief description
effective IS in marketing for the future. of the major tasks of data mining is provided
There are various data mining or business with the description of their possible
intelligence techniques which are used in application for a marketer.
marketing as a tool to reduce marketing
expenditure and increase the take up rate 2.1 Cluster analysis / Clustering
of campaigns. These techniques are used
individually or are clubbed to do certain Clustering or cluster analysis is the
tasks. The challenge for every customer- process of grouping the data into classes or
oriented organization consists of identifying clusters, so that objects within a cluster have
potential customers and satisfying high similarity in comparison to one another
and retaining existing customers. This but are very dissimilar to objects in other
necessitates a detailed understanding of the clusters (Han and Kamber, 2006). Clustering
peoples needs and expectations. Adequately is a method of unsupervised learning, and
addressing these needs and at the right time a common technique for statistical data
is crucial to grow and maintain a long-lasting analysis. Data clustering algorithms may be
and mutually profitable relationship. In hierarchical which find successive clusters
applying data mining methods to marketing using previously established clusters. These
problems, there are several critical issues algorithms can be either agglomerative (also
within the knowledge discovery process, called bottom-up) or divisive (also called
from Business Understanding over Data top-down). Agglomerative algorithms begin
Preparation and Modeling to Deployment of with each element as a separate cluster and
the model in a marketing environment. merge them into successively larger clusters.
Divisive algorithms begin with the whole to specific classes of customers which the
set and proceed to divide it into successively latter would be able to relate to better, and
smaller clusters. Partition algorithms thus in the process, increase the take up
typically determine all clusters at once, but rate and the success of such advertisement
can also be used as divisive algorithms in campaigns.
the hierarchical clustering. Density-based
clustering algorithms are devised to discover 2.3 Pattern association
arbitrary-shaped clusters. In this approach, a
cluster is regarded as a region in which the From the early 1970s, pattern association
density of data objects exceeds a threshold. was primarily studied in the context of
Clustering may typically be used in market research model development. Pattern
marketing for advertisements when the association is used to predict patterns based
customers are yet to be segmented. After on past patterns or data sequences on which
running a cluster analysis, the clusters the tool has been trained upon. Perhaps the
may be examined for characteristics based most common precursors to the exploration
on which advertisement campaigns may of the associative relationship between two
be directed at the customer base. After variables in marketing involve the use of
segmentation, based on the characteristics bivariate cross-tabulations or multivariate
of the clusters, product positioning, product analysis (DeSarbo and Hildebrand 1980;
repositioning and product development may Green 1978; Perreault and Barksdale 1980).
be done, to improve its fit with the targeted Several alternative measures are available
customers. Cluster analysis may also be for assessing the extent of association
done to selecting test markets. in a contingency table. Current pattern
association studies in data mining started
2.2 Classification developing from 1993. Agrawal, Imielinski
and Swami (1993); Holsheimer, Kersten and
A pattern classification problem is Mannila (1995); Houtsma and Swami (1995)
essentially mapping an input pattern, studied association in mining approaches
represented as an input vector, to a particular and are highly cited works.
class or category. Thus given a database Pattern association may be extensively
D={t1,t2,,tn} and a set of classes used to predict customer preferences
C={C1,,Cm}, the classification problem when very little data about the customer
is to define a mapping f:DC where each is available to the marketer. Tools for
ti is assigned to one class (Dunham, 2006). pattern association would help a marketer
Traditionally classification (Duda, Hart to predict which product or advertisement
and Stork, 2001) has been studied using the customer may be interested in solely by
Bayesian decision theory and parameter the current buying behavior of the customer
estimation, non-parametric techniques, and matching it with the buying behavior
linear discriminant functions, multi-layer of similar customers (who bought similar
neural networks, stochastic methods and products) even when no information is
non-metric methods. available for the customer.
Pattern classification may be used
efficiently in marketing and advertisements
by first segmenting the entire customer base
and then delivering selective advertisements

2.4 Summarization for each customer, as a function of a set of

explanatory variables (predictors). If the
Summarization refers to methods response level exceeds a certain cutoff point,
that collapse large amounts of data into the customer is selected for the promotion;
the summary measures that provide otherwise the customer is rejected. In
general description of variables and most practical applications, the response is
their relationships (Peacock, 1998). measured by a discrete, often a binary yes/
Summarization maps data into subsets with no variable, such as buy/do not buy, pay/do
associated simple descriptions (Dunham, not pay, loyal/non-loyal, and the likes. But in
2003). Simple cross tabs, counts, averages, many cases, the response is continuous, with
graphs, pie charts and descriptive statistics a degree of belongingness to both responses.
provide summarization. Summarization This task enables careful selection of
maps data into subsets with associated customers for targeted advertising, and thus
simple descriptions like characterization brings down overall ad-spend considerably.
and generalization.
Summarization can help marketers 2.6 Social network analysis / Link
identify the profitability of segments and analysis
then focus resources accordingly, such that
segments with higher profitability receive From the point of view of data mining,
for advertisement efforts. By this task, those a social network is a heterogeneous and
segments which are non-profitable may be multi-relational data set represented by a
neglected thus saving a lot of advertisement graph. The graph is typically very large,
expenses. Also, using this task, those with nodes corresponding to objects and
segments may be identified for who the take edges corresponding to links representing
up rate of campaigns are faster, and also relationships or interactions between objects
push new products to early adopters. and both nodes and links have attributes.
Social Network Analysis or Link Analysis
2.5 Predictive modeling is a methodology for mapping and measuring
the information flows through interactions
Predictive modeling is the process by among people in groups, represented as
which a model is created or chosen to try to nodes on the graph. Social network analysis
best predict the probability of an outcome. is a set of scientific techniques for modeling
In many cases the model is chosen on the and assessing social relationships which
basis of detection theory to try to guess the depicts nodes and the types and strengths of
probability of a signal given a set amount of links between them. Marketers use it to gain
input data. Predictive modeling is assuming a detailed understanding of how people in
an increasing role in the database marketing groups interact. Software tools are emerging
(DBM) industry to analyze customers that simplify the process of creating the
response and drive the decision process. The link analysis diagrams. The task uses graph
range of predictive modeling covers a variety theory to analyze social networks.
of models, including statistical models and Using this task, the marketer is able to
artificial intelligence-based models (such as isolate early adopters and influencers from
neural networks). a group, to push new campaigns. Also, the
Predictive modeling is typically used to marketer is able to identify thought leaders
predict some type of a response measure and influencers in a social group, who if
takes up a campaign, will influence others 3. Proposition

in their social network to do the same. This
task may also be used effectively for low As has been argued and established in
cost but highly effective viral marketing. previous studies, information and how it
Also, the same may be used to identify is mined plays a crucial role in marketing,
potential churners and thus impact the especially in the success of campaigns,
cost of customer acquisition. Thus Social promotions and advertisements. Based
network analysis can help a company save on the insights developed from the study
a lot on their marketing expenditure from made on the various tasks used to meet
advertisement campaigns. various marketing utilities, a matrix is being
proposed for mapping which task would
2.7 Task discussion summary create value for the marketer for which type
of product or at which stage of the product
Each of the tasks described earlier life cycle the same is. The proposition is
has a very specific data requirement. The entirely conceptual in nature and aims to
following matrix summarizes few possible help marketer lower the huge expenses
utilities of each task and when they can be incurred in advertising.
used, by the marketer, based on the preceding The product life cycle framework being
discussions. Also the data requirements of used is to further develop the information
each task are also provided in the same. needs based on the task is as depicted by
This matrix has been drawn from conceptual Kotler and Kelly (2006). They presented a
understanding of the theory in the area. comprehensive strategy based on product
lifecycle, for each product, but the same
Table 1. Mapping of Task, Data requirement and did not focus on information needs being
Utility to the marketer extremely crucial, as it is in the current
information age.
Task
Type of customer Utility to Information can be used by a marketer in
information needed marketer
Cluster analysis Psychographics, Segment
many ways using the previously mentioned
demographics, product identification, tasks. Identification of early adopters can be
preferences need identification
done by link analysis, predictive modeling
Classification Psychographics, Segmentation of
demographics, product customers, and and summarization. This can help a firm
preferences, other related
data
future purchase
prediction
kick-start the campaign of a new product.
Pattern Purchase behavior records, Predicting future
Similarly identifying influencers can help
association basket data information purchases a company focus their marketing efforts
Summarization Segment purchase details, Profitable and non- on those few people who if accepts the
customer details profitable segment
identification product will make the campaign successful
Predictive Psychographics, Predict campaign by influencing others to take it up as well.
modeling demographics, past trends take-up, loyalty,
in behavior, acceptance of and hence Similarly, identifying needs and thus the
campaign/new products profitability basis for segmentation is one of the crucial
Social network
analysis
Social network with whom
the customer interacts
Identifying
early adopters,
tasks of a marketer, which may be done by a
influencers for cluster analysis, and then customers may be
product launches,
viral marketing mapped to their segment by a classification
tool, and then more focused campaigns
may be possible. Similarly, clustering may
help to identify the gap between expected

benefits and perceived benefits of a product, covering all aspects of the product life cycle.
and thus help the marketer to reposition his As is evident in the following matrix,
product. Again, if nothing is known about various tasks can help to provide various
the customer, just on the basis of mining benefits for the company based on the stage
his immediate shopping patterns, it is of the product in the product development
possible to predict future purchases, through life cycle. The tasks can be used to identify
association rule mining, and then create an the segments for a focused targeting and
immediate campaign based on generated also predict which customers belong to
rules, thus increasing sales by cross-selling which segment, thus automating the entire
of multiple products. process. Mining data can help a company
The list of tasks and outputs are not position and reposition their products based
comprehensible, but the matrix given in on the needs of their customer. Also the
the following part gives a good indication tasks may help to improve the products
of the data-mining tasks and their business and even extend the product line by better
impact. The matrix identifies which task can understanding of customers needs. The
have what type of impact at which stage of tasks may also be used to identify which
the product life cycle. The proposed outputs customers should actually be targeted by
have been developed based on the suggested campaigns to derive higher returns on the
classical strategies for the different stages huge marketing expenditure made by the
of the product life cycle which can be firm. Also, it may help to identify potential
impacted by information processing, as churners and help prevent loss of revenue
has been mentioned in Kotler and Kelly for the client and also help to reduce cost
(2006). The same has been chosen due to the of customer acquisition by better targeted
comprehensive nature of the framework in campaigning.
Table 2. Mapping of Tasks, Product life cycle and utility to the marketer.
Link analysis Identify innovators, Identify early Identify influencers & Identify churners & Identify churners
early adopters adopters & laggards laggards
influencers
Predictive modeling Identify innovators, Identify early dentify profitable & loyal Identify loyal customers,
early adopters adopters & customers Calculation of CLTV
influencers
Summarization Identify early adopters Identify new Identify profitable Identify profitable
customer customer segment, customer segment,
segments Calculation of CLTV Calculation of CLTV
Pattern association Identify potential Identify potential Identify potential Identify possible
associated associated purchases associated purchases product line
purchases diversification
Classification Push products to target Push products to Push products to target Push products to target Push products to
segments target segments segments segments target segments
Clustering Identify segments and Identify product Identify product Reposition based on Reposition based
size, market needs extensions extensions segment needs if needed on segment needs
if needed
Tasks - Introduction Growth Maturity Saturation Decline

Stage $
Product life cycle ( increasing maturity )
The mapping of the task can have 5. References

important implications of how a practitioner
may use the information available to him, Agrawal R, Imielinski T, Swami A (1993).
at each stage of the product life cycle, to Mining association rules between sets
fulfill the business needs. Knowing which of items in large databases. ACM,
task can fulfill which business need and SIGMOD Intl. Conf. Management of
what sort of data would be needed to do so, Data, Washington.
would enable a practitioner chalk out a plan
of action to implement a business strategy. Brooks NAL (1989). Marketing technology:
This mapping has been conceptualized by new marketing information systems
studying the theory in the related discipline. enhance service and profitability. Bank
Administration. 65(5): 52-54.
4. Conclusion
Comanor WS, Wilson TA (1967).
Information plays a major role in Advertising, Market Structure, and
the current times in improving business Performance. Review of economic
processes. To become productively statistics. 49 (4): 423-440.
engaged in knowledge discovery activities,
marketing managers should possess a good DeSarbo WS, Hildebrand DK (1980). A
understanding of what these activities Marketers Guide to Log-Linear Models
consist of. For doing the same, there has for Qualitative Data Analysis. Journal of
to be a good understanding of the various Marketing. 44 (summer): 40-51.
tasks, their requirements and their outputs,
to be able to appreciate their relevance in the Duda RO, Hart PE, Stork DG (2001). Pattern
business context. Classification. Wiley publications.
This paper provides a basic understanding
of the various tasks and indicates how each Dunham MH (2003). Data mining:
task can provide value to the marketer Introductory and advanced topics.
based on theoretical understanding of both Pearson Education.
areas under discussion. It also maps the
data-mining tasks with the product life- Gartner Research (2008). Findings: Social
cycle stages and also provides insights of a Network Analysis Is coming into the
few possible business benefits that may be Limelight. ID Number: G00157280.
extracted by engaging in the tasks. The focus
has been on the product life cycle as various Green PE (1978). An AID/Logic
strategies actually depend on the stage of Procedure for Analyzing Large Multi-
the product in the PLC curve, with which way Contingency Tables. Journal of
the data-mining tasks have been mapped. Marketing Research. 42(4): 92-100.
This paper also provides a keen insight on
how information technology can be used Gronroos C (1990). Service Management
to enable business processes and strategies and Marketing: Managing the Moments
better by using the keen insights developed of Truth in Service Competition.
from processing information. Lexington Books, Lexington, MA.

Haeckel SH, Nolan RH (1993). Managing Moriarty RT, Swartz GS (1989). Automation
by wire. Harvard Business Review. pp. to boost sales and marketing. Harvard
122 132. Business Review. 67(1): 100-108.
Han J, Kamber M (2001). Data Mining: Naude P, Holland CP (1996). Business

Concepts and Techniques. 2nd edition, to business marketing. Relationship
Morgan Kaufmann Publishers. marketing, theory and practice. Paul
Chapman Series, 40-54.
Holland PC, Naude P (2004). The
metamorphosis of marketing into Peacock PR (1998a). Data mining
an information-handling problem. in marketing: Part 1. Marketing
The Journal of Business & Industrial Management. 6(4): 8-19.
Marketing. 19(3): 167-178.
Peacock PR (1998b). Data mining
Holsheimer M, Kersten M, Mannila H, in marketing: Part 2. Marketing
Toivonen H. (1995). A perspective on Management. 7(1): 14-26.
databases and data mining. In 1st Int.
Conf. Knowledge Discovery and Data Perreault WD, Barkswale HC (1980). A
Mining. Model-Free Approach for Analysis of
Complex Contingency Data in Survey
Houtsma M & Swami A (1995). Set-oriented Research. Journal of Marketing Research.
mining of association rules in relational 17: 503-515.
databases. Proceedings, 11th Int. Conf.
Data Engineering. Rayport JR, Sviokla JJ (1995). Exploiting
the virtual value chain. Harvard Business
Kaldor N, Silverman R (1948). A Statistical Review. Nov-Dec: 75-85.
Analysis of Advertising Expenditures and
of the Revenue of the Press. Cambridge Sheth JN, Gardner DM, Garrett DE (1988).
University Press. Marketing Theory: Evolution and
Evaluation. John Wiley & Sons, New
Kotler P, Keller KL (2006). Marketing York.
Management. 12th Edition, Prentice hall,
New York. Telser GL (1961). How Much Does it Pay
Whom to Advertise. American Economic
Levin N, Zahavi J (1999). Continuous Review. 51: 194-205.
predictive modeling - A comparative
analysis. Journal of Direct Marketing. The World Advertising Research Center
12(2): 5-22. (2001). World advertising expenditure.
International Journal of Advertising. 20:
Moller KE (1994). Inter-organizational 266 268.
marketing exchange: Meta-theoretical
analysis of current research approaches. Webster FE (1988). The rediscovery of the
Research Traditions in Marketing. marketing concept. Business Horizons.
Kluwer Academic Publishers, Boston. 31: 29-39.
29. Webster FE Jr. (1992). The changing

role of marketing in the corporation.
Journal of Marketing. 56(4): 1-17.

2010 Vali Khodadadi, Soheila Khazami and Abbas Aflatooni 151
THE EFFECT OF CORPORATE GOVERNANCE

STRUCTURE ON THE EXTENT OF VOLUNTARY
DISCLOSURE IN IRAN
Vali Khodadadi, Soheila Khazami and Abbas Aflatooni
Abstract
This study examines the relationship between corporate governance and the extent of voluntary
disclosure in listed firms of Tehran Stock Exchange.The corporate governance attributes are the percentage
of independent directors on the board, the existence of dominant personalities (CEO/Chairman duality)
and the percentage of institutional investors. The research sample was selected among listed companies in
Tehran Stock Exchange during 2001-2005 and totally is consist of 106 observations. In this research, the
size and the type of audit institute (Audit organization or other institutes) are used as control variables.
The research results show that there is no significant relationship between percentage of independent
directors on the board, CEO/Chairman duality and the extent of voluntary disclosure. Also the research
results provide significant evidence on relationship between the percentage of institutional investors and
the extent of voluntary disclosure.
Khodadadi V., Khazami S., Aflatooni A. - The Effect of Corporate Governance Structure on the Extent of Voluntary Disclosure in Iran
1. Introduction information, are affected by many factors

(Archambault and Archambault, 2003).
The occurrence of Industrial Revolution So, the empirical assessment of variables
and its merging in 19th Europe required the effective on disclosure may create space
implementation of great projects needing and constraints contributing to improved
enormous monetary capitals. Since it was disclosure quality.
far beyond financial facilities of one or Corporate governance is a factor
more investors and even governments of the which brings higher quality for corporate
time, the first corporates were established. performance and, especially, the information
This modern format (i.e. corporate) was an submitted by management. Therefore, this
appropriate solution to enormous capitals survey endeavors to analyze the influence
supply and trading risks distribution. of some structure features of corporate
When this happened, the topic dealt with governance, e.g. percentage of independent
separation of ownership from management directors on the board, the existence of
and, as a consequence, the issue of financial dominant personalities (CEO/Chairman
reporting and providing information to duality) and percentage of institutional
take decisions and to assess directors investors ownership.
performance was considered. General
objective of financial reporting is to provide 2. Review on Research
information resuling in making useful Background
trading and economic decisions. Today,
economic development is based on access to There has been no research conducted
financial resources for a stable investment. on this area in Iran and no background
Investors dicision on investing their is available. Forker (1992) analyzed the
resources in companies, in turn, depends on relationship between corporate governance
appropriate financial reporing existing in and disclosure quality. The results of this
company. Financial reportings disclose the study showed that the sign of relationship
allocation of capital resources in a trading between corporate size and disclosure
company and its profitability. This economic quality is negative, but it is not significant.
information contributes to the recognition Results indicated that there was negative
of financial status and health of a trading relationship between corporate audit by
unit. Financial information disclosure is an one of the six large audit committees and
important factor to decrease information disclosure quality. The relationship between
asymmetry. Weak financial disclosure often percentage of share held by managers and
results in misleading of shareholders and disclosure quality was also direct, but this
has adverse effect on their wealth. was not statistically significant as well.
In view of above description it is revealed There was a direct relationship between
that the extent of information disclosure for existence of an auditor and disclosure
contributing optimal allocation of restricted quality. But this relationship was also
economic resources may have a critical weak. Relationship between non-executive
role to help investors in their taking useful (independent) managers and disclosure
decisions. Beside, decisions made on the quality was unexpected (direct) but was not
extent of financial information disclosure, statistically significant.
firms reporting, composition of information Chen and Jaggi (2000) analyzed the
resources, quantity and quality of disclosed relationship between independent non-

executive directors and firms family factors of culture and corporate governance
control with financial disclosure in Hong and extent of disclosure. This study
Kong. Results showed that proportion of included several hypotheses and variables
the independent directors is directly related among which they tested only six variables
to financial disclosure. This relationship as corporate governance variables. These
is weaker in firms which are under family variables include proportion of independent
control. Thus, researchers stated that the directors on the boars, proportion of family
presence of independent non-executive members on the board, joint position
directors in board composition would of Chairman and CEO, non-executive
enhance the probability of compliance with chairperson, proportion of board of directors
disclosure requirements by firm and may with cross-directorships and chairperson
result in information transparency. with cross directorships. In their research,
Ho and Wong (2001) analyzed the they used firm-specific characteristics (size,
relationship between corporate governance leverage, profitability, industry type, auditor
structures and the extent of voluntary type, listing status) as control variables.
disclosure in Hong Kong Stock Exchange. Results from this survey only refer to
They represented four major corporate the existence of a significant relationship
governance attributes. These corporate between two variables of non-executive
governance attributes are the proportion of directors on the board and proportion of
independent non-executive directors on the family members on the board with firms
board, the existence of an audit committee, extent of voluntary disclosure both of which,
the existence of dominant personalities of course, have negative relationship with
(CEO/Chairman duality), and the proportion dependent variable.
of family members on the board. The results Lakhal (2003) analyzed the relationship
indicate that the existence of an auditor between voluntary disclosure by directors of
is significantly and positively related to French firms and the corporate governance
the extent of voluntary disclosure, while characteristics. Results indicated that the
the proportion of family members on the ownership structure is rather dispersed and
board is negatively related to the extent of in case where non-executive directors have
voluntary disclosure. more share in board, the level of disclosure
But, statistical tests showed that two other increases voluntarily. In firms where the
independent variables, i.e. the proportion of CEO holds the position of the chairperson
independent directors on the board, which as well, voluntary disclosure would be less
was expected to be directly associated with likely. There is an insignificant and weak
voluntary disclosure and the joint position relationship between non-executive board
of members and decisions about voluntary
Chairman of the Board (Chairman) and disclosure and, also board size and disclosure
Chief Executive Officer (CEO) held by a decisions. A hypothesis of this research
single individual which was expected to be indicating the negative relationship between
negatively associated with disclosure, have board size and voluntary disclosure was not
no significant relationship with disclosure. confirmed.
In a survey titled Culture, Corporate Eng and Mack (2003) analyzed the
governance and Disclosure in Malaysian relationship between corporate governance
Corporates, Hanifa and Cooke (2002) and voluntary disclosure in Singapore. In this
investigated the relationship between two investigation, corporate governance testing
is examined in two aspects of ownership pay and board size as have been considered
structure and the board composition as independent vaiables of research. Among
where ownership structure by means of control variables of this research one can
the proportion of managerial ownership, mention firm size, situation in stock market,
blockholder ownership and government industry type and shares dispersion. Results
ownership and board composition are showed that proportion of independent
measured with the criteria of the proportion directors on the board and existence of
of independent directors. Control variables audit committtee in firm and, also, board
used in this research include firm size, participation in the capital of the company
leverage, industry type, reputation of as well as the approval of stock option
audit committee. Results of statistical plans as directors pay are directly related to
tests showed that as much as managerial disclosure.
ownership becomes lower and government Chen and Courtenay (2006) tested the
ownership higher, the voluntary disclosure relationship between board supervision
of information is increased. It means that and the extent of voluntary disclosure. In
government ownership enhances the moral this study, the relationship among three
hazard and agency problems and disclosure independent variables, i.e. proportion of
is utilized to reduce these problems. Results independent directors on the board, board
of this research showed that blockholder size and, also, joint position of Chairman and
ownership has a significant relationship CEO are analyzed on the dependent variable
with the extent of disclosure and as much (voluntary disclosure). Reuslts showed
as the proportion of independent directors that proportion of independent directors
increases the voluntary disclosure is is directly associated with the extent of
decreased. voluntary disclosure and the board size and
Gul and Leung (2004) examined and the joint position of Chairman and CEO
tested the relationship between board also have no relationship with voluntary
structure and voluntary disclosure. For disclosure. They found that existence of
the same purpose, two factors of the joint a corporate governance mechanism and
position of CEO and the chairperson and environmental rules would increase the
the proportion of independent directors power of the relationship between proportion
on the board (as determinant of board of non-executive directors on the board and
structure) were studied with using voluntary the extent of voluntary disclosure.
disclosure. Results showed that the joint Arcay and Vazquez (2005) investigated
position of CEO and the chairperson is the relationship of firm characteristics, rules
negatively and weakly associated with of corporate governance and the extent of
inormation voluntary disclosure. voluntary disclosure in Spain. Proportion
Arcay and Vazquez (2005) investigated of independent directors on the board,
the relationship of firm characteristics, rules existence of audit committee, splitting the
of corporate governance and the extent of responsibilities of the Chairman and CEO,
voluntary disclosure in Spain. Proportion board participation in the capital of the
of independent directors on the board, company, stock option plans as directors
existence of audit committee, splitting the pay and board size as have been considered
responsibilities of the Chairman and CEO, as independent vaiables of research. Among
board participation in the capital of the control variables of this research one can
company, stock option plans as directors mention firm size, situation in stock market,

industry type and shares dispersion. Results 3.1. Independent (Non-Executive)

showed that proportion of independent Directors Role in Corporate
directors on the board and existence of Governance
audit committtee in firm and, also, board
participation in the capital of the company In view of Agency theory, the presence
as well as the approval of stock option of non-executive directors in the board of
plans as directors pay are directly related to the firms and their supervisory performance
disclosure. as independent individuals, remarkably
Chen and Courtenay (2006) tested the contribute the declined conflicts of interests
relationship between board supervision existing between shareholders and directors
and the extent of voluntary disclosure. In of the firm. Of course, it should be noted that
this study, the relationship among three the executive directors of firm play a major
independent variables, i.e. proportion of role in making an appropriate composition
independent directors on the board, board of executive and non-executive directors,
size and, also, joint position of Chairman and among board members. Such a composition
CEO are analyzed on the dependent variable is regarded as a main element of efficient
(voluntary disclosure). Reuslts showed and effective board; since though executive
that proportion of independent directors directors offer valuable information about
is directly associated with the extent of firm activities, non-executive directors
voluntary disclosure and the board size and in the firms board and their supervisory
the joint position of Chairman and CEO performance as independent individuals
also have no relationship with voluntary markedly helps to declined conflicts of
disclosure. They found that existence of interest between shareholders and directors
a corporate governance mechanism and of firm (Hassas Yeganeh and Baghoomian,
environmental rules would increase the 2006).
power of the relationship between proportion Of course, it should be taken into
of non-executive directors on the board and account that executive directors of firm
the extent of voluntary disclosure. play major role in making an appropriate
composition of executive and non-executive
3. Literature and Theoretical directors among board members. Such
Framework of Research a composition makes a board having
specialty, independence and required legal
Board of Directors of the firm, as a power, be cosidered as powerful potential
leading position, is of great importance. It mechanism of corporate governance (Bryd
seems that firms success key depends on its and Hickman, 1992).
desired guidance. According to information
theory and agency theory, those hypotheses 3.2. Splitting the responsibilities of
of this research improving the control on the the Chairman and CEO
board, would result in increased voluntary
disclosure. In Cadbury report it has been
recommended that there should be balance
of power among board members not to allow
anyone having unconstrained control over
decision process in the firm and it is required
to clearly identify responsibilities splitting in
senior levels of firm to be sufficiently assured agency problem and their existence in the
of the balance of power and the liabilities of position of sharehloder result in separation
board members. The importance of splitting of ownership and control.
the responsibilities of the Chairman and There are two thinking schools about
CEO of England companies has been also institutional investors in firm. The first
emphasized in report by Higz. Higz declared says that institutional investors have short-
that with considering the recommendations term viewpoint. This short-term horizon
of Cadbury report and until the time of his inhibits institutional investors from taking
report, about 90% of firms listed in Stock supervisory expenses because the benefits
Exchange had splitted the responsibilities of such supervision will be unlikely won
of the Chairman and CEO. This proves the by them in a short period of time. The other
permanent impact of Cadbury report on thinking school states that when firm shares
England companies. ownership is concentrated among a few
number of investors (especially institutional
3.3. Institutional Investors Role in investors); the problems concerning the
Corporate Governance separation of ownership and control will
be decreased. When the proportion of
Numerous groups have influence on institutions ownership increases, firm
corporate governance. Among these, leaving will take more expense because
shareholders, especially institutional major shares sales usually require major
investors, play important role. In theoretical discounts (Black and Koffee, 1994).
view, the position of institutional investors Institutional investors tend to have
in corporate governance is very complicated. communication with senior managers of
From this point of view, institutional companies existing in their portfolio and
investors reveal another mechanism of to participate in behind closed doors
powerful corporate governance which could supervisory activities (Ramsay et al.,
have supervision over firm management. 2000; Stapledon, 1996). As a result of
Because they can either considerably affect their supervisory activities, these investors
firm management or coordinate interests have a better understanding of conditions
of shareholders groups. Of course, in influencing on corporates performance
literatures about corporate governance, the and it is less likely to make the managers
ownership focus is referred to as an important of companies existing in their portfolio be
mechanism to control agencys problems subjected to fine for their low profit not
and to improve investors interest supporting caused by their weak management.
(Shleifer and Vishny, 1997). However, such
focus may have negative impacts as well, 3.4. The extent of Disclosure
e.g. access to confidential information may
result in information asymmetry between Financial disclosure is an abstract
them and smaller shareholders. concept not measurable directly which
As a result of their benefiting from have no certain feature to be used for
economic advantages and variability, easily measurment of its intensity or
institutional investors can resolve agency quality (Wallace and Naser, 1995) But,
problems. Therefore, it seems that disclosure is simply defined as transfering
institutional investors, as firms shareholders, and presentation of economic information
reveal both cause and resolution/solution of in terms of financial and non-financial,

quantitative or other forms of information range from very low to very high. The
related with firms financial status and next step is performed like that of first aspect
operation. This disclosure, when becomes and here the only difference is that obtained
necessary by a law enacting and regulating scores are multiplied by the importance
source, is called involuntary disclosure and coefficients extracted from different
when information disclosure is not affected individuals comments and the weighed
by certain rules it is considered as optional criteria is applied. Singhvi and Desai (1971)
disclosure (Owusu-Ansah, 1998). and Barrett (1977) have employed this
Also, disclosure/extent of disclosure aspect in their surveys. First aspect has been
has very important and determining role in used for two reasons as follows:
theoretical and empirical areas. This concept
is so broad and extensive that consists of a. Ashton (1974) believes that individuals
all issues and topics on financial reporting. have little knowledge and information
A defined and accurate comprehension of about their decision making and
disclosure quality requires the assessment of judgement. Therefor, the need to make
restrictions governing thereon and achieving decision about the importance of each
such comprehension would undoubtedly itme is removed and different individuals
help accountant in drawing a framework to mistake in weighing is less likely.
improve disclosure problems of accounting
in the future (Buzby 1974), decrease the b. Using this method allows researcher to
heterogeneous disclosure among informed do measurements independently from
and uninformed investors (Bushman and different individuals viewpoints. When
Smith, 2001) and non-disclosure will result using weighed aspect, it is possible to
in market inefficiency and wrong pricing of take different people personal opinions
firms shares (Diamond and Verrecchina, into consideration and, in consequence,
1991; Kim and Verrecchina, 1994). In the pertinent criteria will not be
general, two aspects have been suggested independent from the viewpoint of
in measuring voluntary disclosure level in certain individuals and/or groups.
researches:
3.4.1. Using non-weighed index In this respect, Owusu-Ansah (1998)
In this aspect, a list of disclosed cases is says that using non-weighed aspect is
prepared and, then, after financial statements preferred over weighed aspect. Also, Cooke
and financial reports have been studied, the expresses that there is no difference in the
values of one, zero and not applied results of weighed and non-weighed criteria.
will be given for disclosure, non-disclosure Therefore, with studying and analysing of
or not qualified for disclosure, respectively. the considered voluntary disclosure cases in
Researchers like Wallace (1987); Cooke similar researches, a list of disclosures cases
(1989, 1991, and 1992), Ahmed and Nicols including 31 items are extracted as follows:
(1994) and Owusu-Ansah (1998) utilized 3. Brief history of company
this aspect in their researches. 4. Names of boards members
3.4.2. Using weighed index 5. Names of staff managers
In second aspect, firstly the list of 6. Names of block holder
disclosure cases is extracted and the 7. Firms strategies and its possible effects
importance of each item is identified with 8. Firms major markets
using different individuals comments and a 9. Forecasted sales
10. Forecasted cost of goods sold voluntary disclosure in financial

11. Forecasted operating profit reporting
12. Forecasted net income
13. The acted works to access the objects 2. There is significant relationship between
14. Information about various productions CEO and the chairman position duality
15. The trend of sales in last years and extent of information voluntary
16. The trend of cost of goods sold in last disclosure in financial reporting.
years
17. The trend of operating profit in last years 3. There is significant relationship between
18. The trend of financial costs in last years the percentage of institutional investors in
19. The trend of capital expenditure in last firm and extent of information voluntary
years disclosure in financial reporting.
20. Profitability ratios
21. The ratios of financial structure 5. Methodology of Research
22. The liquidity ratios
23. The number of unit sold of major product 5.1. Statistical Sample and Population
24. The per unit price of major product
25. The number of employee The Statistical population of research is
26. Description about changes in revenue composed of firms listed in Tehran Stock
27. Description about changes in cost of Exchange. Statistical sample of the research
goods sold has been gained through applying following
28. Description about changes in gross conditions:
income
29. Description about changes in 1. 1. Due to their having information
administrative expenses reporting and a nature of operation
30. Description about changes in financial different from other corporates,
costs investment and financial corporates have
31. Description about changes in interest been omitted from sample of research.
32. Information about executive and non-
executive managers 2. 2. In order to ignore the impact of
33. Information about future projects information disclosure time and to omit
the impact of financial period of each
4. Research Hypotheses and corporate on the extent of voluntary
Questions disclosure in financial reporting, the
firms with fiscal year not ending to 20th
Substantial question of this research is March (year-end in Iran), were excluded
that whether there is significant relationship from sample of research.
between corporate governance and extent
of voluntary disclosure or not? This is 3. Loss firms are excluded from our sample.
asked within the framework of following
hypotheses: 4. Required information such as financial
statements and notes to financial
1. There is significant relationship between statements, summary of decisions taken
the percentage of non-executive board by regular general meeting having
members and extent of information been published by Stock Exchange

Organization as well as boards annual we will have an ordered dependent variable

report submitted to the shareholders in sequential scale. We use the Ordered
regular general meeting are available. Logit to approach to model estimation1.
Outdir: percentage of non-executive
With applying above-mentioned directors on the board. This percentage is
conditions, sample of this research consists obtained through dividing the number of
of 106 corporates. non-executive board members by the total
number of board members at the end of
5.2. Research Model period.
Duality: CEO/chairperson duality
Strategy processing information bank means that CEO and the chairperson/
and internet sites containing information vice chairperson of a firm is not a single
on firms listed in stock exchange were individual (this position not held by one
employed for data collection. In first step, person). In case were CEO and chairperson/
after data collection, descriptive statistics vice chairperson position is held by a single
concerning research variables are presented. individual, 1 and otherwise 0 is given to this
To test our hypotheses the following model variable.
is used: Inst: Percentage of institutional investors
(1)
ownership. This amount is a percent of firm
shares kept by special investing institutions2.
Size: Firm size: various criteria like rate
Where: of assets, sales, and number of staff and
Disc: in this survey, for the purpose of are used to measure firms size. In this
measuring the extent of voluntary disclosure; research, firm size equals to logarithm of
initially, according to list presented in assets book value. This variable is used as
section 3.4.2 and with using non-weighing control variable.
method, the number of voluntary disclosures Audit: type of the committee that has
for which the firm has presented information audited the firm. In case where the considered
is defined. Then, as the variable of the committee is the audit organization, this
extent of voluntary disclosure is qualitative, variable is 1 and otherwise it will be zero.
the number of the said cases are ranked in This variable is used as control variable.
five voluntary disclosures as very low (0-5
number of cases), low (6-11 number of 6. Research Results
cases), medium (12-18 number of cases),
high (19-26 number of cases) and very high 6.1. Descriptive Statistics
(26-31 number of cases) and code 1 to 5 is
given to the said ranks, respectively. Thus, In Table (1) the number of firms existing
in each disclosure rank over the studied
1
For further study on how to estimate these kinds of models you
period is depicted.
can refer to Green (2003).
2
Institutional investor is a character or institute involved in pur-
chasing and selling of a great volume of securities, i.e. state and
private banks, pension funds, insurance companies and social
security organization (SSO), institutions and foundations and
investment companies.
Table 1. The number of firms in disclosure rank firm-years; also, the audit organization was
the auditor. Meanwhile, the number of firms
in which the position of chairperson and
Very high
disclosure
voluntary
Very low
Extent of
Medium
CEO is separated has been increased over
Total
High
Low
time (from 8 to 20 firms). As time passed, the
2001 23 23 17 19 24 106
number of firms whose auditor was the audit
2002 21 27 17 21 20 106
organization has been declined (from 66 to
2003 21 26 19 20 20 106
34 firms) and other institutions play further
2004 21 22 27 15 21 106
role in auditing of sample firms. In Table
2005 23 21 18 24 20 106
3 descriptive statistics pertaining to other
Total 109 119 98 99 105 530
research variables have been illustrated.
As seen, over the period of time, 109 firm- Table 3: Descriptive statistics
years with very low voluntary disclosure, 119
firm-years with low voluntary disclosure, 98 % of non-
% of
institutional
firm-years with medium level of voluntary executive
investors
Firm Size
directors
disclosure, 99 firm-years with high voluntary ownership
disclosure, 105 firm-years with very high Mean 0.65 0.47 5.41
voluntary disclosure were observed. You can Median 0.60 0.51 5.38
study the number of firms in disclosure ranks Max. 0.83 0.95 7.16
from 2001 to 2005 in Table (1). The number Min. 0 0 4.04
of firms with considering CEO/chairperson Std. Dev 0.19 0.33 0.54
duality and the type of auditor which audited
the firms are as shown in Table 2. The average percentage of non-executive
directors, percentage of institutional
Table 2: The number of firms with considering CEO/ investors ownership and firm size equal to
chairperson duality and the type of audit committee 65%, 47% and 5.41, respectively. Maximum
percentage of non-executive directors
and institutional investors ownership and
Variable
CEO/chairperson
Auditor
duality maximum firm size equal to 83%, 95%
and 7.16, respectively, and the minimum
percentage of non-executive directors
organization
Year Yes No Total Total and institutional investors ownership and

Others
Audit
minimum firm size equal to zero, zero and

4.04, respectively.
2001 98 8 106 66 40 106
2002 93 13 106 56 50 106
6.2. The independent 2 Test and
2003 89 17 106 44 62 106 Spearman Correlation Coefficients
2004 88 18 106 35 71 106
2005 86 20 106 34 72 106 In Table 4 it has been dealt with the
Total 454 76 530 235 295 530 significance of relationship between
two nominal variables of 1) CEO and
As observed in Table 2, over the studied chairperson duality and 2) Auditor type
period of time in 76 firm-years the CEO has and extent of voluntary disclosure. Results
held the position of chairperson and in 235 of independent 2 test ( = 2.48) show that

there is no significant relationship between investors ownership (0.15) is significant

CEO and chairperson duality and extent of at 1% level. Therefore, existence of a
voluntary disclosure. However, significant significant linear relationship between
relationship between auditor type and extent two above-mentioned variables is not
of voluntary disclosure is not rejected (2 rejected. Correlation coefficient between
= 34.18). Of course, control variables of firm size and extent of voluntary disclosure
auditor type and firm size have not been (0.06), and correlation coefficient between
taken into account in this test, as in case of percentage of independent directors and
considering mentioned variables different extent of voluntary disclosure (-0.02) is very
results will be obtained3. small and not significant. Consequently,
existence of a linear relationship between
Table 4: Results of independent 2test said variables is rejected. In other cases,
correlation coefficient is also very small and
not significant.
2 Statistics
disclosure
voluntary
Very high
Extent of
Very low
Medium
High
Table 5: Spearmans ranked correlation coefficient

Low
CEO/chairperson
% of
Yes 96 99 84 82 93 Extent of % of non-
institutional
voluntary executive
investors'
2.48 disclosure directors
ownership
duality
No 13 20 14 17 12 % of non-
-0.02
executive 1
(0.69)
directors
organization
% of
31 48 51 65 40 institutional 0.15 0.01
1
Others Audit
investors (0.00)* (0.78)

34.18*
Auditor type
ownership
Firm Size 0.06 -0.05 0.02
78 71 47 34 65 (0.13) (0.25) (0.71)
* significance at 1% level
* significance at 1% level
6.3. Estimation Results
Spearman ranked correlation coefficients
among variables of the extent of voluntary We estimate model (1) by ordered logistic
disclosure, percentage of independent method for testing the research hypotheses
directors, percentage of institutional and for taking into account the control
investors ownership and firm size are variables of auditor type and firm size.
presented in Table 54. Estimation results of model (1) are provided
Results shown in Table 5 indicate that in Table 6.
only the Spearmans ranked correlation Estimation results of model (1) show
coefficient between extent of voluntary that the coefficient of the variable of the
disclosure and percentage of institutional percentage of independent directors (-0.07)
and the coefficient of the variable of CEO
3
Regression analysis is the best method to taking control variables and chairperson duality (-0.17) are not
into consideration and this has been discussed in section (6.3).
significant, this means that the first and
4
Since the variable of voluntary disclosure level is of rank type,
Spearman Correlation Coefficients have been applied instead of the second hypotheses are rejected. The
Pearson Correlation Coefficients. coefficient of the variable of the percentage
of institutional investors ownership equals theory and this supervisory tool has little
to 0.78 which is significant at 1% level. role in improved financial disclosure.
Therefore, our third hypothesis of research Beside, in most literature published
is not rejected as a result of significant concerning corporate governance, it
relationship existing between the percentage has been emphasized on splitting the
of institutional investors and the extent of responsibilities of the chairperson and CEO.
voluntary disclosure. In addition, results In most Iranian firms the positoin of CEO
show that as the percentage of institutional and the chairperson has been separated and
investors ownership increases the extent of usually two (and not one) individuals are in
disclosure increases, too. the charge of the said positions. However,
though duties and liabilities of CEO and
Table 6: Estimation results of model (1) chairperson have not been well separated, it
was already anticipated that the said variable
makes little effect on voluntary disclosure.
Significance
Coefficient
Z- statistic
Institutioal ownership is also an issue

applicable to most securities dealers of
our country. In most countries members to
% of non-executive
directors
-0.07 -0.16 0.86 Securities & Exchange Corporations, the
CEO/chairperson duality -0.17 -0.79 0.43
institutional investors are considered as
% of institutional investors'
blockholer owners and, therefore, with their
ownership
0.78 3.20 0.00*** influence on board they considerably affect
Firm Size 0.30 2.01 0.04** disclosure level (including disclosure of
Auditor type 0.29 1.84 0.06* voluntary and additional information). This
Pseudo R-squared=1.27%
is confirmed by result of present study. It is
Likelihood Ratio statistic=21.68 (0.00)***
usually expected that larger firms arrange
for voluntary disclosure more often than
*, ** and *** significance at %10, %5 and 1%, respectively.
smaller ones. This seems reasonable because
The coefficient of the control variable of of financial facilities and benefiting from
firm size (0.30) is significant at 5% level. stronger information systems. As a result of
The coefficient of the control variable of the quality of auditing performance, auditor
auditor type (0.29) is also significant at type will affect the extent of disclosure. Two
10% level. Results show that as firm size recent cases have also been confirmed in
increases, the extent of voluntary disclosure this survey.
is also increased. Thus, it seems that larger
firms will more tend to voluntary disclosure. References
The Pseudo R-squared is 1.27% and the
likelihood ratio is 21.68 which is significance Ahmed, K. and Nicholls, D. (1994)
at 1% level and indicates the significance of The effect of non-financial company
the whole model. characteristics on mandatory disclosure
compliance in development countries:
7. Conclusion The case of Bangladesh. The
International Journal of Accounting, Vol.
Results of this survey show that non- 29, 62-77.
executive directors role in Iran does not
comply with the fundamentals of agency

Arcay, M. and Vazquez, M (2005) Corporate Chen, C. and Jaggi, B (2000) Association
Characteristics Governance Rules and Between Independent Non-Executive
The Extent of Voluntary Disclosure in Directors, Family control and Financial
Spain. Advances in Accounting, vol.21, Disclosures in Hong Kong, Journal of
299-331. Accounting and Public Policy, Vol. 19,
285-310.
Archambault, J.J, and Archambault
E.M (2003) A multinational test of Chen, C. and Courtenay, M (2006) Board
determinants of corporate disclosure, Composition, Regulatory Regime and
The International Journal of Accounting, Voluntary Disclosure. The International
Vol.38, 173-194. Journal of Accounting, Vol. 41, 262-289.
Ashton, R.H (1974) The Predictive-Ability Cooke, T. E. (1989) Voluntary Corporate

Criterion and User Prediction Models, Disclosure by Swedish Companies.
The Accounting Review. Vol. 49, 719- Journal of International Financial
732. Management and Accounting. Vol. 1,
171195.
Barrett, M.E. (1977) The Extent of
Disclosure in Annual Reports of Large Cooke, T. E. (1991) An Assessment of
Companies in Seven Countries. The Voluntary Disclosure in The Annual
International Journal of Accounting Reports of Japanese Corporations. The
Education and Research, Vol. 13, 1-25. International Journal of Accounting, 26,
174189.
Black, B. and Coffee, J. (1994) Hail
Britannia?: Institotional Investor Cooke, T.E. (1992) The impact of size,
Behavior Under Limited Regulation. stock market listing, and industry type
Michigan Law Review, Vol.92 , 1997- on disclosure in the annual reports of
2087. Japanese listed corporations, Accounting
& Business Research, Vol. 22, 229-237.
Bryd, J. and Hickman ,K. (1992) Do outside
directors monitors managers?: Evidence Diamond, D. & Verrecchia, R. (1991)
from Tender Offer Bids. Journal of Disclosure, Liquidity, and The Cost of
Financial Economics. Vol. 32: 195-221. Capital. The Journal of Finance, Vol. 66,
1325-1355.
Bushman, R. and Smith, A. (2001) Financial
accounting information and corporate Eng, L. and Mak, Y (2003) Corporate
governance. Journal of Accounting and Governance and Voluntary Disclosure,
Economics, Vol. 32, 237-334. Journal of Accounting and Public Policy,
Vol. 22, 325-345.
Buzby, S.L. (1974) Selected items of
information and their disclosure in Forker, J. (1992) Corporate Governance
annual reports, The Accounting Review, and Disclosure Quality. Accounting and
Vol. 43, 423-35. Business Research, Vol. 22, 111-124.
Gul, A. and Leung, S (2004) Board mandatory disclosure and reporting by

Leadership, Outside Directors Expertise listed companies in Zimbabwe. The
and Voluntary Corporate Disclosures. International Journal of Accounting, Vol.
Journal of Accounting and Public Policy, 33, 605631.
Vol. 23, 351-379.
Ramsay, L. and Lang, H. (2000) Corporate
Green, W.H, (2003) Econometric Analysis, Governance and Anti-Takeover
Prentice Hall, 5 ed. Devices: Evidence From Australia . An
International Review, Vol.8.
Hanifa, R. M. and Cooke T.E (2002)
Culture, Corporate Governance and Singhvi, S. and Desai H.B (1971), An
Disclosure in Malaysian Corporations. Empirical Analysis of the Quality of
ABACUS, Vol. 38, 317-349. Corporate Financial Disclosure, The
Accounting Review, Vol. 46, 621-632.
Hassas yegane, Y and Baghoomian, R
(1385) The Role Board in Corporate Shleifer, I. and Vishny, R. W. (1997) A
Governance. The Iranian Accountant Survey of Corporate Governance. The
Journal, Vol. 173, 26-29, 62-68 and 74- Journal of Finance, Vol. 52 , 737-783.
75.
Stepledon, G. (1996) Disincentives to
Ho, S and Wong, K (2001) A study Ictivism by Institotional Investors in
the relationship between corporate Listed Australian Companies. Sydney
governance structures and the extent Law Review, vol.18 No.2.
of voluntary disclosure. Journal of
International Accounting Aauditing & Wallace, R.S (1987) Disclosure in
Taxation, Vol. 10,139-156. accounting information in developing
countries: a Nigerian case study.
Kim, O. and Verrecchia, R. (1994) Market Doctoral Dissertation, University of
Liquidity and Volume Around Earnings Exeter, Devon.
Announcements. Journal of Accounting
and Economics, Vol. 17, 41-68. Wallace, R.S and Naser, K (1995)
Firm specific determinants of the
Lakhal, F (2003) Earning Voluntary comprehensiveness of mandatory
Disclosures and Corporate Governance: disclosure in the corporate annual reports
Evidence from France. www.ssrn.com. of firms listed on the stock exchange of
Hong Kong. Journal of Accounting &
Owusu-Ansah, S. (1998) The impact of Public Policy, Vol. 14, 31-68.
corporate attributes on the extent of

SRRNet
Social Responsibility
Research Network
www.socialresponsibility.biz
Social Responsibility Research Network
Who are we?
We are an international network of scholars who share similar interests in aspects of social
responsibility. Currently we have about 500 members and membership is free.
Network officers:
Chair of the Network: Professor Dr. David Crowther, De Montfort University, Leicester Business
School, The Gateway, Leicester LE1 9BH, UK davideacrowther@aol.com
Vice Chair: Professor Dr. Gler Aras, Yildiz Technical University, Institute of Social Science,
Yildiz Besiktas 34349, Istanbul, TURKEY guleraras@aol.com
What do we do?
Conferences
2008 7th conference CSR and SMEs
Durham, UK
2009 8th conference CSR and NGOs
Pretoria, South Africa
2010 9th conference CSR and Global Governance
Zagreb, Croatia
2011 10th conference CSR and the New Economy
New Orleans, USA
Publications
Social Responsibility Journal
The official refereed journal of the Network; published 4 times per year by Emerald.
Discussion Papers in Social Responsibility
An opportunity for early publication of articles. Published when necessary by SRRNet.
The Newsletter
Published 3 times per year and containing news and opinion pieces. Sent to all members.
Research Book Series: Issues in Corporate Behaviour and Sustainability
Books published in association with the conferences and given to all conference delegates.
Full details of all of our activities can be found from our website www.socialresponsibility.biz
If you share our aims then please join us. We look forward to hearing from you.
Isles Internationale Universit
School of Doctoral Studies
(European Union)
Approved by Charter of The Ministry of Education of the British Isles to act as a chartered
University outside of the United Kingdom; with Full Accreditation granted from the Acadmie
Europenne dInformatisation; established in Brussels, Belgium by Order of the King of
Belgium Albert II, with full recognition from The Ministry of Justice and Research of the
Belgian Crown:
The Isles Internationale Universit (European Union) has been commended to host the School of Doctoral Studies (EU)
in Brussels, Belgium, aiming to accomplish three fundamental missions:
Development and enhancement of elite doctoral studies programmes, with cutting edge standards on academic and
scientific research;
Enforcement of EU analogue quality standards on academic programmes developed by tuition institutions outside the
EU area: (a) By evaluating academic methodology applied on learning programmes; (b) By providing full coaching
and tutoring support to tuition institutions worldwide to upgrade, assure and maintain academic methodology quality
levels on their way towards excellence; and (3) By awarding EUASC Seal (EU Analogue Standards Certification) at
corresponding quality levels achieved at each evaluation point; and
Ensuring academic EU Analogue Standard by performing: (a) Academic Validation on degrees earned by students at
recognised and/or certified tuition institutions located outside the EU area; (b) Double Degree awarding on degrees
earned by students at recognised and/or certified tuition institutions located elsewhere the EU area; and (c) Degrees
awarding on studies programmes developed by tuition institutions, once certification at higher quality levels on applied
academic methodology has been achieved and after collaboration agreement has been executed for these purposes .
In order to achieve its missions, the Isles Internationale Universit has gathered some of the best minds in Europe, who
have developed sate-of-the-art doctoral academic programmes, elite research methodology and cutting edge technological
tools, and who act as permanent Faculty Members to strictly enforce this toolkits proper application on daily basis.
Doctoral Studies
The School of Doctoral Studies of the EUs academic structure includes four Departments (Business Management and
Economics, Engineering and Technology, Science and Social Science) which host 37 Disciplines, offering PhD studies on
practically every main field of human knowledge.
Over 355 PhD students are involved in more than 116 cutting edge research projects, most of them being currently
developed in collaboration with 12 other universities in the EU area and elsewhere; 94% of these students are engaged on
programmes designed to undertake pure research assignment and 6% are required to undertake a research assignment and
pursue theoretical studies in the form of seminars or courses.
Studies towards a doctoral degree are worth 240 higher education credits (ECTS credits) and require an average of four
years of full-time study. The research is intended to lead to a scholarly thesis; writing it will take up most of a students
time and all theses are publicly defended. Every doctoral student receives a studies grant for partial or total coverage on
full programmes term costs, as well as individual tutoring. Currently, slightly over 72% of all programmes full term
costs are covered by studies grants.
Science students spend a great deal of time in the laboratory. Some departments may require that the thesis be part of an
ongoing project within the department. In the fields of technology and natural science, researchers often work as part of a
team. If research findings are reproduced in academic journals the thesis may be a compilation of the published articles.
Forward inquiries to: admin@iiuedu.eu

Who are we?
The Universidad del Valle de Mxico (UVM) is one of the largest and most prestigious universities in
Mexico. Founded in 1960 and accredited by the Federacion de Instituciones Mexicanas Particulares
de Educacion Superior, UVM enrolls students at 32 campuses throughout Mexico.
Universidad del Valle de Mxico was founded by a Group of entrepreneurs and

academics, led by Mr. Jose Ortega in response to the professional requirements
of the Mexican labor market, aiming to provide education with quality:
High School
Undergraduate
Undergraduate for working adults
Graduate
Continuing Education
UVMs mission is to be an institution that fully educates with an equilibrium

between sciences and technology, ethics and culture, according to the social needs,
in search of truth and welfare. The institutional philosophy and educational model
seek to educate in human values, updated knowledge and the acquisition of skills
that teach the student to competitively adapt to the labor market.
UVM alumni are distinguished by their abilities, knowledge, attitudes, and social skills
that are shaped by the identity subjects.
UVM is the only global university in Mexico

Because of its prestige and quality the UVM is part of Laureate International Universities, the
most important network of universities in the world which is integrated by 24 well-known private
universities in:
Spain, Switzerland, France, Costa Rica, Panama, Honduras, Ecuador, Chile, Peru, Brazil, England,
Chipre, China, Canada, US, Germany and Mexico.
UVM has 35 campuses throughout Mexico offering 38 undergraduate degree programs in Arts
and Humanities, Social Sciences, Economic and Management Sciences, and Engineering; 11
undergraduate degree programs for working adults and 28 graduate programs. UVM has more than
100,000 students and more than 9,000 employees (teachers and staff).
UVM opens its door to the world

UVM students have access to international qualified academic opportunities through programs
offered by Laureate International Universities
Summer courses
Semester academic exchanges
Double degree
Graduate studies
Recognitions:
Through its history, UVM has received recognitions that prove its excellence:
Academic Excellence - Secretara de Educacin Pblica (SEP).
Affiliation - Asociacin Nacional de Universidades e Instituciones

de Educacin Superior (ANUIES).
Academic Quality Certification - Federacin de Instituciones

Mexicanas Particulares de Educacin Superior (FIMPES).
Second private university in the country holding the highest figure

of academic programs accredited by Consejo para la Acreditacin
de la Educacin Superior, A.C. (COPAES).
National Registree of Scientific and Technologic Institutions and

Enterprises (RENIECYT) - CONACYT.
According to the Readers Digest Intelligent Decision Markers (IDM) Guia Universitaria 2008,
UVM is one of the best upper education institutions in the country, from a range of over 100
public and private universities.
UVMs faculty training programs

UVMs commitment with its students and with the country is to enhance its academic quality and
prepare successful professionals; therefore, it invests in constant training and specialization of its
teachers through the Centre for Academic Excellence (CAE).
2008 Year of academic strengthening

UVM trained more than 3.000 teachers and staff in different areas of knowledge between January and
September of this year. It also started a program of specialization with postgraduate and doctorate
courses, as well as an English program. UVM has identified important public and private upper
education institutions in Mexico and abroad, to perform together development and training programs.
Teacher training and professional development

Nearly 2.000 UVM teachers all over the country took the Institutional Teaching and Pedagogy
Program with a focus on the correct implementation of the educational model, which underlies the
students principles: learn to learn, learn to be, learn to do, and learn to undertake.
Also, more than 400 teachers of the 35 campuses took 12 seminars in areas such as: Industrial
Engineering, Mechatronic, Animation, Marketing, Management, Communication, Law, Psychology,
Health Sciences and Hospitality.
140 academic leaders of the 6 regions in which UVM has classified its campuses in 15 states of
Mexico and Mexico City, are enrolled in the Diplomat in Leadership for Academic Management.
Postgraduate Studies in prestigious institutions

A priority for UVM is to have teachers with postgraduate studies; thus, Walden University, a
prestigious on-line upper education institution accredited in the US, has offered scholarships for
Master degrees in Education, Management, Information Technologies, Psychology and Systems
Engineering.
Sports at UVM go beyond boundaries

UVM considers Sports an important part of the students education. This has allowed some of its
students to excel at national and international events. Maria del Rosario Espinoza and Guillermo
Perez were both Tae Kwon Do gold medal winners in the Olympic Games of Beijing, China.
UVM supports young leaders of projects with social impact

From creating eco-tourism opportunities to developing a national
hotline to combat domestic abuse, young people in Mexico are
using their energy and creativity to improve their communities
and country. To support their efforts, Universidad del Valle
de Mexico (UVM), joined the Sylvan/Laureate Foundation
and the International Youth Foundation in 2006 in creating
Premio UVM por el Desarrollo Social (UVM Prize for
Social Development). Its goal: to celebrate and support
outstanding young Mexican social entrepreneurs.
Premio UVM has adapted the YouthActionNet Global Fellowship model to provide a tailor-
made, culturally-relevant, and Spanish-language centered leadership development experience for 15
young Mexican leaders, ages 18-29, annually. The Premio UVM fellowship strengthens the project
management and communications skills of young Mexican leaders, while connecting them to their
peers and experts to create a national network of youth leaders affecting positive change.
Student Development
Responsible of the education that students receive, UVM sets special emphasis on the students
integral education and the continuous improvement of its faculty. This way, it responds to the
expectations and trust of Mexican families and prepares good successful professionals with a global
vision, who will acquire the skills and knowledge that the labor market requires.
The student development area at UVM has important national and international projects as:
Institutional Fair of Entrepreneurs UVM

Congress Simulation.
United Nations Simulation Model.
Congresses.
Student Councils.

A Study On Using Business Intelligence F

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

A Study On Using Business Intelligence F

Enviado por

Direitos autorais:

Formatos disponíveis

Business Intelligence Journal

July, 2010 Vol.3 No.2

Business Intelligence Journal

Profile of authors included in this number 2

Information for Contributors 4

Luis G. Acosta-Calzado, Carlos Acosta-Calzado, Humberto Murrieta Romo

Nitin Merh, Vinod P. Saxena, Kamal Raj Pardasani

Zahra Nikbakht, Abbas Aflatooni

Jos Carlos Arias

Arpan Kumar Kar, Ashis Kumar Pani, Supriya Kumar De

Vali Khodadadi, Soheila Khazami, Abbas Aflatooni

Business Intelligence Journal - July, 2010 Vol.3 No.2

EDITORIAL NOTE Business Intelligence Journal

Business Intelligence Journal - July, 2010 Vol.3 No.2

Business Intelligence Journal

Business Intelligence Journal - July, 2010 Vol.3 No.2

Business Intelligence Journal - July, 2010 Vol.3 No.2

INFORMATION FOR CONTRIBUTORS

Business Intelligence Journal - July, 2010 Vol.3 No.2

Business Intelligence Journal - July, 2010 Vol.3 No.2

Examples: Short Communications

Gurau C (2006). Multi-channel banking Submission of a manuscript implies: that

Business Intelligence Journal - July, 2010 Vol.3 No.2

Costs for Authors Manuscripts submitted will be judged

Business Intelligence Journal - July, 2010 Vol.3 No.2

Business Intelligence Journal - July, 2010 Vol.3 No.2

REVAAM MODEL TO DETERMINE A COMPANYS

Luis G. Acosta-Calzado, Carlos Acosta-Calzado and

Introduction priced equally. Despite these arguments,

Business Intelligence Journal - July, 2010 Vol.3 No.2

We extracted information for public table shows the number of companies,

Table 1. Number of companies by sector

Business Intelligence Journal - July, 2010 Vol.3 No.2

Graph 1. P/E vs ROE Graph 2. P/E vs Net Margin

P/E multiple P/E multiple

Graph 3. EV/EBITDA vs ROC Graph 4. EV/EBITDA vs Pre-tax Operating Margin

Business Intelligence Journal - July, 2010 Vol.3 No.2

Graph 5. Ln(P/E) vs Ln(ROE) Graph 6. Ln(P/E) vs Ln(Net Margin)

Graph 7. Ln(EV/EBITDA) vs Ln(ROC) Graph 8. Ln(EV/EBITDA) vs Ln(Pre-tax Operating

We calculated equation (3) for years 2000

Table 2. Regression results for equation (3) - P/E

Table 3. Regression results for equation (4) EV/EBITDA

Table 4. Average multiples vs REVAAM multiples

Net P/E - P/E - Pre-tax In(EV/ EV/EVITDA- EV/EVITDA-

Business Intelligence Journal - July, 2010 Vol.3 No.2

Net P/E - P/E - Pre-tax In(EV/ EV/EVITDA- EV/EVITDA-

2009 P/E - average 2009 EV/EBITDA - average

from the year to year comparison of our 25.00 EV/EBITDA

multiples. As shown on Graph 11, P/E 20.00

multiple tends to be more volatile over 15.00

5.02, whereas for the latter the average has -

Further Applications discounting future cash flows, then both

Graph 12. EV/EBITDA calculation by sector for 2009

Metals & Mining (Div.)

Auto & Truck

Business Intelligence Journal - July, 2010 Vol.3 No.2

those below are considered undervalued. Metals, Homebuilding, Healthcare

Graph 13. P/E calculation by sector for 2009

Financial Svcs. (Div.)

Metals & Mining (Div.)

Table 5. Perrigos and Drug sector multiples (REVAAM and averages)

Business Intelligence Journal - July, 2010 Vol.3 No.2

Determining the Value of Any Asset Myers, Brealey. 2002. Principles of