Big picture: Development of Secondary Corporate Bond and CDS Markets

Madhu Kalimipalli PhD

Associate Professor-Finance & Director, Financial Services Research Centre School of Business and Economics | Wilfrid Laurier University | Waterloo| Ontario| Canada

mkalimipalli@wlu.ca | www.wlu.ca/sbe/mkalimipalli IGIDR, Mumbai Dec 16, 2011

Current literature: Overview

Overview of the Indian Corporate debt market

Subir Gokarn:

http://www.careratings.com/Content/CDMS/Dr%20%20Subir%20Gokarn.pdf

http://www.un.org/esa/ffd/events/2007debtworkshop/shyamala%20 gopinath.pdf

Importance of corporate bond markets in the Indian context

India's Bond Market Needs to Bulk Up: Indias ambitious $1 trillion infrastructure program won't succeed without a more robust corporate bond market

http://www.businessweek.com/magazine/content/11_08/b4216011325565.htm

Broad basing and deepening the bond market in India

http://fic.wharton.upenn.edu/fic/papers/01/0132.pdf

Size of Debt Market Relative to GDP

Source: Chakrabarti, 2007

Corporate bond issues as a % of new capital issues: India

Source: Chakrabarti, 2007

4

Corporate bond markets in India:

(Charabarti, Rajesh, 2007, SSRN WP)

High taxes and stamp duties (lack of uniformity in the latter across states) Issue size is small, and hence preponderance of private placements ( over 90%).

Low issues costs and disclosure requirements in private placement

Lack of centralized information on bond trading, prices and defaults Lack of uniformity in market practices like lot size and conventions for coupon calculations Lack of multilateral trading and central counterparty mechanism Lack of market makers, prepared to provide two-way quotes in the corporate bond market
5

The market for corporate bonds is rather limited: Financial institutions like provident and pension funds, are not allowed to hold anything below top-rated bonds; and this has led to an overwhelming proportion of high rated corporate bonds being issued. FIIs typically use corporate debt only for short-term parking of funds and also have quantity restrictions in holding corporate bonds. Also restrictions on corporate bonds being used as collateral for repo transactions.

Current literature: Recent developments

SEBI sets up 16 member committee on corporate bond market

http://articles.economictimes.indiatimes.com/2011-03-31/news/29366219_1_corporatebond-market-securitised-debt-securitised-instruments http://www.indianexpress.com/news/sebi-eyes-setting-up-corporate-bond-mkt/769961/

RBI has released final guidelines for CDS in corporate bonds

http://www.emecklai.com/Market_Resources/Regulatory_Update/CDS_Guidelines.pdf

Developments in the Corporate Bonds and Securitization Markets

http://203.199.12.51/debt/expertreport.pdf http://www.sebi.gov.in/cms/sebi_data/statistics/corporate_bonds/corpstatus.html http://www.iflr1000.com/pdfs/Directories/3/India.pdf http://www.adb.org/Documents/Books/Mortgage_Backed_Securities_Markets/2_india.pdf

Did CDS Trading Improve the Market for Corporate Bonds?

S A N J I V DA S Santa Clara University M A D H U K A L I M I PA L L I Wilfrid Laurier University S U B H A N K A R N AYA K Wilfrid Laurier University IGIDR, Mumbai, Dec 15

Agenda
1. Introduction & Motivation 2. Background & Related Literature

7. Conclusions

6.Tests of Liquidity

3. Data and summary stats

5. Tests of Market quality

4. Tests of Bond Efficiency

Financial Crisis of 2007-2009

Four key developments:

Asset securitization Loan syndication Proprietary Investing Increased use of credit derivatives to transfer risk

Collateralized Debt Obligations CDOs

XYZ BANK
Rating: BBB Assets: Installment sales receivables of $75m

Create a Special Purpose Vehicle (SPV) or a Special Purpose Corp. and sell the $75 mi receivables

Get a credit enhancement Internal: Reserve funds Over-collateralization Senior/subordinated structures

Finally, SPV Issues TRANCHED securities at a better rating External: 3rd party guarantee: Corporate Guarantee Line of credit Bond insurance

Sold as CDOs to other risky investors

10

Structured Investment Vehicles SIVs

SIV
Long term Loan #1 Long-term Loan #2

Pooled monthly cash flows: Net interest Scheduled principal payment Principal prepayments
(asset backed commercial paper)

ABCP

Long-term Loan #3

CMO (tranches)

Long-term Loan #n

Short-term instruments

Monthly cash flows Sold to risky investors

11

Loan Syndication

A syndicated loan is one that is provided by a group of lenders and

is structured, arranged, and administered by one or several commercial or investment banks known as arrangers.

Starting with the large leveraged buyout (LBO) loans of the mid-1980s,

the syndicated loan market has become the dominant way for issuers to tap banks and other institutional capital providers for loans. LBO: issue low grade debt to buy out the equity

The reason is simple:

Syndicated loans are less expensive and more efficient to administer than traditional bilateral, or individual, credit lines.

12

Bloomberg: Example IBM bond

13

IBM bond

14

What happens to the Syndicated loans?

After loan syndication is closed two things can happen:

Syndicate members can sell their loan syndication shares in the secondary market for syndicated bank loans

However the lead syndicate bank/arranger still retains a large stake in the loan to provide monitoring and market making services to the syndicate members

Syndicated loans are often pooled together and securitized in the form of CLOs

15

Proprietary Investing

As traditional on-balance sheet investing in loans became less attractive, banks continued to seek out profit opportunities ( break down of Glass Steagall) This has taken the form of increased level of trading of securities within the bank s portfolio Banks established specialized off-balance sheet vehicles and subsidiaries

A.

B.

to engage in investments and investment strategies that might be viewed as being too risky if conducted on their balance sheets For e.g. banks established through lending or equity participations hedge funds, private equity funds, or venture funds Such off-balance sheet vehicles (often operating as off-shore entities and with limited number of investors) can often act outside the controls of regulators such as SEC

16

CDS market

Credit Derivatives in general are contracts whose payoffs depend on the credit quality of the underlying reference company or sovereign entity The CDS market started to grow fast in the late 1990s

By 2006 notional principal totaled over $35 trillion At the end of 2007, the CDS notional amount outstanding was about $62 trillion, equivalent to roughly five times the US annual GDP.

BIS data indicates that more than 95 percent of credit default swap transactions are between financial institutions.
17

Notional Amounts Outstanding (in bi USD)

Source: ISDA

2.5 times US annual GDP

18

OTC derivative market

19

Bond market overview: US debt market

US Outstanding Public and Private Debt: 2007 in %

Federal Agencies 10% Money market 14% Asset backed 8% Municipal 9%

Source:

Corporate 19%

US Treasury 16%

mortgage related/ MBS 24%

The Securities Industry and Financial Markets Association (SIFMA)

20

Trading volume in US secondary markets (2007) Source- SIFMA

Trading volume in secondary markets in %

Federal Agencies 8% Asset backed 3% Municipal 2%

Corporate 2%

MBS 30%

US Treasury 55%

21

Single name CDS

Source: BBA
22

Market shares of main credit derivative instruments (British Bankers Association 2002)
TRS 7% Credit Linked Notes 8% spread options 5%

Asset swaps 7%

CDS 45%

basket swaps 6%
Portfolio CDS/Synthetic CDOs 22%

23

Reference entities by credit ratings : Fitch ratings 2002

<BBB 8% BBB 28%

AAA 21%

AA 15% A 28%

24

Buyers/Sellers of CDS protection (table 18.2)

25

CDS contract

The credit default swap market is generally divided into three sectors:

corporates, bank credits and emerging market sovereigns.

CDS can reference a single credit or multiple credits. Multi-credit CDS can reference a custom portfolio of credits agreed upon by the buyer and seller, or a CDS index. The credits referenced in a CDS are known as "reference entities." CDS range in maturity from one to 10 years although the five-year CDS is the most frequently traded

26

CDS contract

27

Most common credit events

Credit event could also include a credit rating downgrade

28

Credit Default Swap (CDS) Structure

Premium is known as the credit default spread. It is paid for life of contract or until default

Payments are usually made quarterly or semiannually in arrears

90 bps per year Default Protection Buyer, A

Payoff if there is a default by reference entity=100(1-R)

Default Protection Seller, B

Settlement can be specified as delivery of the bonds or a cash equivalent amount. Recovery rate, R, is the ratio of the value of the reference bond29 immediately after default to the face value of the bond

CDS are unfunded

CDS are unfunded credit market instruments In the cash market buying bonds/extending loans involves upfront funds or capital (and hence need to be funded on the investors balance sheet)

whereas at the inception of CDS no funds change hands

The investors may find it hard to invest in cash market if the funding costs are very high

30

Main uses of credit derivatives

1. 2. 3. 4. 5.

Credit management by banks Managing bank regulatory capital Yield enhancement and portfolio diversification Shorting corporate bonds Other uses:
Hedging vendor financed deals Hedging by convertible bond investors Selling protection as an alternative to loan origination Credit derivatives as market indicators

1.Credit management by banks

Suppose Citibank would like to reduce credit exposure wrt its important client XYZ without hurting its business/client relationships it can use CDS market to lay off its risk Cheaper alternative ( lower legal and set up costs, and tax, accounting implications) to secondary market loan sales and asset securitizations.

2. Managing bank regulatory capital

Basel (1998) stipulated same capital requirement for all corporate entities.
Type of exposure OECD govts OECD banks Corporate and non OECD banks and Govts Risk weight 0 20 100 Capital charge % 0 1.6 8

Regulatory capital change= r capital charge notional exposure

Managing bank regulatory capital contd.

However banks could consider some entities to be more or less risky than others. Notion of regulatory capital was misaligned with that of economic capital, or the capital that prudent bank would want to hold in reserve given its overall credit risk exposure So a bank would buys CDS from OECD bank for highly rated obligors; thereby lowering the capital requirement from 8% to1.6%.

34

3. Yield enhancement and portfolio diversification

CDS are unfunded credit market instruments In the cash market buying bonds/extending loans involves upfront funds or capital (and hence need to be funded on the investors balance sheet)

whereas at the inception of CDS no funds change hands

The investors may find it hard to invest in cash market if the funding costs are very high

Synthesize long positions in corporate debt:

If Citibank wants to take exposure to XYZ corps debt, which however is all locked up in loans on books of different banks. Then Citi could simply short a CDS on such bonds
35

4.Shorting corporate bonds

Shorting corporate bonds ( in event of expected negative news) is not easy. Simply buy a CDS now. If the credit event does happen, the CDS price does goes up and we can short the CDS at a high price.

36

5. Other uses-I

Hedging vendor financed deals

A manufacturing firms provides capital/credit financing for its clients And buys credit protection on its clients

Hedging by convertible bond investors

Investors often find that call options embedded in the convertibles is often underpriced. If an investor wants to lay off the underlying credit risks, he could by a CDS on the issuer: the net position would be a long call position on the stock

37

Other uses-II

Selling protection as an alternative to loan origination

Frictions and barriers in cash and loan markets could lead to the bank selling credit protection

Credit derivatives as market indicators

Forward looking credit risk information ( such as option IVs) in modeling bond or loan spreads Facilitates a greater integration between debt and loan markets because of their increased use by banks

38

CLN Structure: reference entity XYZ

Sell CLNs

CLN investors

Pay the principal of $100 mi

Coupon $C per year XYS bond issuer Invested in XYZ bonds.

BMO Trust CLN Seller (issuer)

Collect the $100 mil principal and invest in high quality collateral

39

CLN Structure: no default of XYZ

Coupon-Servicing fee

CLN investors

Pay the principal of $100 mi

Coupon $C per year XYS bond issuer Invested in XYZ bonds.

BMO Trust CLN Seller (issuer)

Collect the $100 mil principal from collateral investment

40

CLN investors

CLN Structure: default of XYZ

Coupon-Servicing fee Payoff of 100R if there is a default by reference entity

Coupon $C per year XYS bond issuer

Value recovered is100R or lost for the Trust =100(1-R) (physical delivery)

BMO Trust CLN Seller (issuer), CDS seller

Sell and collect the principal from collateral investment

Recovery rate, R, is the ratio of the value of the bond issued by reference entity immediately after default to the face value of the bond
41

CDS: Evaluation
Pros

CDS strengthen the financial system (ISDA)

Cons

CDS market may have contributed to

CDS enable banks to transfer risk to other risk takers, so banks can make more loans. CDS help distribute risk widely throughout the system and thus prevent large concentrations of risk that otherwise would occur. CDS provide important information about credit conditions, helping bankers and policymakers to supervise traditional banking activities. CDS serve a valuable signaling function CDS prices produce better and more timely information.

the naked shorting and speculation (empty creditors) lowering capital requirements for FIs the lower underwriting standards in ABS market the lower monitoring incentives for banks

The recent OTC Derivatives Reform spearheaded by CFTC and SEC

Dodd-Frank Act involves

collateral requirements for dealers promote transparency : exchanges Vs dealer markets setting up centralized clearing houses position limits

42

Further: CDS offer an additional channel of information on corporate bonds.

CFTC Chairman Gary Gensler, OTC Derivatives Reform Conference (Mar 9, 2010):

We need broad regulatory reform of over-the-counter derivatives to best lower risk and promote transparency in the marketplace. While similar to other derivatives, credit default swaps have unique features that require additional consideration. Only with comprehensive reform can we be sure to fully protect the American public..
43

Key stakeholders in the US reform debate

Chris Dodd: D Conn

44

Global OTC derivative market reform

45

In this paper

We examine whether the introduction of CDS improved the quality of the bond market in terms of underlying efficiency, quality and liquidity. Taking a time-series perspective we ask:

Did an issuers bonds become more efficient and liquid after CDS trading was instituted on the underlying bonds?

From a cross-sectional perspective we ask:

Are bonds of issuers with traded CDS more efficient and liquid than bonds of issuers with no traded CDS?

46

Fig 1: Corporate bond trading before and after the commencement of CDS trading

47

Fig 2: Bond turnover before and after the commencement of CDS trading

48

Specifically we ask 3 questions

Did CDS trading improve the efficiency of bond prices?

we consider multivariate lead-lag relationships of corporate bond returns

to returns on various other securities that would also be incorporating issuer-specific information.

Did CDS trading improve the accuracy of bond prices?

Following Hotchkiss and Ronen (2002) we implement the market quality measure (q) of Hasbrouck (1993).

Did bond market liquidity respond favorably to the inception of CDS?

We computed several proxies for bond liquidity before and after the introduction of CDS trading.

We employ an extensive data set with over 1.3 mi time-series observations spanning 20022008, a period that witnessed the explosion of the CDS market
49

Agenda
1. Introduction & Motivation 2. Background & Related Literature

7. Conclusions

6.Tests of Liquidity

3. Data and summary stats

5. Tests of Market quality

4. Tests of Bond Efficiency

50

Extant literature: overview

I.

Relative efficiency of stock vs. bond markets

Stocks are more efficient
Kwon(1996); Downing, Underwood and Xing (2009); Gebhardt, Hvidkjaer and Swaminathan (2005);

Bonds are more efficient

Hotchkiss and Ronen (2002); Ronen and Zhou (2009);

II.

Price leadership role of CDS vs. stock and bond markets

the CDS market leads the bond market in determining the price of credit risk Stocks lead CDS

III.

Impact of CDS markets on debt and loan markets

Impact of CDS on cost of borrowing and market quality of equity

Complements earlier work on Impact of options on the underlying equity markets

The effects are mainly positive (lower volatility, higher trading volume and relaxing short-sale constraints)

51

II.Price leadership role of CDS vs. stock and bond markets

Blanco, Brennan and Marsh (2005):

the CDS market leads the bond market in determining the price of credit risk.

For 27 of the firms they examined, the CDS market contributes on average around 80% of price discovery

Hull, Predescu and White (2005):

CDS market effectively anticipates credit rating down grades or negative credit rating changes in the market

52

II. Contd..

Norden and Wagner (2008):

They find that CDS spreads explain loan rates much better than spreads of similar-rated bonds.

Forte and Pena (2009):

find that stocks lead CDS and bonds more frequently than the reverse, and CDS market leads the bond market.

Norden and Weber (2009):

They find that stock returns lead CDS and bond spread changes, and the CDS market contributes more to price discovery than the bond market

(stronger for US than for European firms)

Baba and Inada (2009): subordinated bond and CDS spreads for Japanese banks are largely cointegrated, and the CDS spread plays a bigger role in price discovery than the bond spread
53

III. Impact of CDS markets on debt and loan markets

Ashcraft and Santos (2009):

CDS introduction has not lowered the cost of debt financing or loan funding for the average borrower

Boehmer, Chava and Tookes (2010):

Examine the implications of derivatives and corporate debt markets on equity market quality. They find that listed options have more liquid equity and more efficient stock prices. By contrast, firms with traded CDS contracts have less liquid equity and less efficient stock prices. Overall, they find that the impact of CDS markets is generally most negative, followed by corporate bond markets, and then options.

Ismailescu and Phillips (2011) :

Most recently, in sovereign bond markets, provide evidence that the

introduction of sovereign CDS swaps improved efficiency in the underlying bonds.

54

Our goals in this paper

To assess what role the CDS markets played vis-a-vis the bond markets To determine whether CDS trading was beneficial or detrimental to the bond markets on criteria such as

efficiency quality liquidity.

55

Agenda
1. Introduction & Motivation 2. Background & Related Literature

7. Conclusions

6.Tests of Liquidity

3. Data and summary stats

5. Tests of Market quality

4. Tests of Bond Efficiency

56

Data sources

CDS:

Assumption :

Bloomberg Single-name 5yr CDS TRACE and FISD CRSP

Bond:

CDS starting date in Bloomberg is the date of CDS introduction

Stocks:

CDS introduction:

Swap and VIX:

Datastream

CDS market is OTC and hence decentralized CDS introduction is initiated by the dealer banks depending on factors such as

Time frame: 2002-2008

57

size of outstanding debt on an issuer, underlying credit risk of the issuer, and demand for credit protection.

BOND DATA

TRACE
34,900 bonds issued by 4,869 firms (5.76 mi time series obs) between 08/01/2001 - 09/30/2009

FISD
11,950 bond issues (19942007) *

2,806 bond issues by 967 issuing firms 843,442 time-series observations (i.e., # of days with bond trades)

Publicly traded: 8,291 bond issues with matching CRSP perm numbers

58

*Exclude convertibles, sinking fund provision, 144A,TIPS, Treasuries, Munis, STRPS, Agency, MTNs, foreign bonds, Canadian,Yankees. Only consider US-domestic, dollar-denominated, straight bonds, and bonds with call, and put features

Bond and CDS DATA

TRACE-FISD
2,806 bond issues by 967 issuing firms

CDS
620 CDS issued (598,221 obs of CDS spreads)

TRACE-FISD-CRSP 2,155 of 2,806 bond issues have matching stock returns

TRACE-FISD-CRSPCDS Intersect with CDS Issues (+ additional filters) 59

1,545 bond issues have corresponding CDS issues 350 issuing firms have CDS issues

Variables used

60

We employ two data sampling approaches

Approach 1

Approach 2

we retain those days on which we have three consecutive observations of all traded securities in our sample. Hence, we will be focusing on periods of active trading,

Stock data is available every day. The gaps in the data occur because of the absence of consecutive days when both CDS and bonds trade, precluding return calculations.

which are more likely when information is being released these are exactly the periods when we want to test for market efficiency.

In this approach we use all dates on which both bonds and CDS were traded (and had observations) these dates need not be consecutive

61

Table 1: Data & Summary statistics

62

Table 1: contd.

1545 total bond issues: 158 not rated

1545 total bond issues: 776 short-term (maturity < 7 years)

152 below investment grade (below BBB/Baa) 1,235 investment grade (BBB/Baa and above)

277 medium-term (maturity 7-15 years) 492 long-term (maturity > 15 years)

63

CDS and Bond data

CDS introductions
120 300,000

All bond issues

100

250,000

80

200,000

60

150,000

40

100,000

20

50,000

2002

2003

2004

2005

2006

2007

2008

2002

2003

2004

2005

2006

2007

2008

64

Table 2: Contd.

65

Table 2: Contd. Post-CDS correlations

Bond returns are negatively correlated to lagged bond returns, suggesting that there may be frequent return reversals in the bond markets. Bond returns are positively correlated to stock and Treasury returns, and negatively correlated to volatility and CDS spreads, both contemporaneously and lagged.

66

Agenda
1. Introduction & Motivation 2. Background & Related Literature

7. Conclusions

6.Tests of Liquidity

3. Data and summary stats

5. Tests of Market quality

4. Tests of Bond Efficiency

67

Empirical Analysis of Bond Efficiency

68

Bond Efficiency tests

1.

2.

Contemporaneous & lagged data

Hou and Muskowitz (RFS,2005)

69

Table 3: Individual bond regressions

70

Panel regressions

Diagnostic tests
1.

heteroskedasticity:
exists, highly significant

2.

auto-correlation in residuals:
minor/marginal issue, weak significance

3.

multicollinearity:
doesn't exist, no significance

4.

clustering effect:
clusters based on either year or issuing firm with multiple bonds doesn't exist, no significance

Only (1) and (2) are relevant, that's why we carried out regressions with Newey West HAC correction

71

Table 4: Partitioned panel regressions

72

Table 5:Joint panel regressions

73

Table 6:Joint panel regressions with CDS dummy interaction

74

Table 9: Robustness tests

75

Summary

F-statistics and D1 metrics both suggest that

bond market efficiency did not increase with the advent of CDS trading. Table 6 suggests even drop in market efficiency

Results robust to alternative data sampling approaches Results robust to

Issue size; crisis-period; ratings and maturity based samples; fixed effects; difference-difference approach

76

Agenda
1. Introduction & Motivation 2. Background & Related Literature

7. Conclusions

6.Tests of Liquidity

3. Data and summary stats

5. Tests of Market quality

4. Tests of Bond Efficiency

77

Market quality

78

Market quality contd.

79

Table 7:Market Quality Before and After Introduction of CDS

pre-cds median q = 0.892 post-cds median q = 0.878

Z-statistic of difference = 0.378 (p-value = 0.7053)

80

Market quality for the equity and CDS

The sample comprised 107 stocks with at least 30 trading days of data pre- and post-CDS. Pre-CDS,

on average across stocks,

q = 0:98 and post-CDS, q = 0:99 (the difference is not statistically significant).

The quality of equity markets is thus much higher than that of bonds.

For the post-CDS period, we also examined the quality of the CDS market.

We used 325 individual CDS with at least 30 trading days data to compute q.

The average is q = 0:92 (s:d: = 0:07).

Hence, CDS markets are of higher quality than bond markets, though not as high quality as equities.

81

Agenda
1. Introduction & Motivation 2. Background & Related Literature

7. Conclusions

6.Tests of Liquidity

3. Data and summary stats

5. Tests of Market quality

4. Tests of Bond Efficiency

82

83

Agenda
1. Introduction & Motivation 2. Background & Related Literature

7. Conclusions

6.Tests of Liquidity

3. Data and summary stats

5. Tests of Market quality

4. Tests of Bond Efficiency

84

Summary

We find no evidence that corporate bonds became more efficient after the introduction of CDS trading. Hasbrouk market quality measure does not improve after CDS trading begins, suggesting that there is no evidence that CDS markets enhanced bond market quality. Whereas the mean number of daily trades increased with the growth of bond markets over time, many other measures such as :

the mean size of the trades daily turnover, LOT, covariance illiquidity, and Amihuds metric No evidence of liquidity improvement after the CDS were introduced

Taken together, the results suggest that CDS introduction did not necessarily improve the corp. bond market efficiency, liquidity or quality.
85

Implications
1.

The CDS markets are very active and mostly ( about 95%) dominated by institutional and hence a venue for all the informed trading At the same time corporate bond markets

(a) witnessed proliferation of CDO -securitization market , whereby bonds were sitting inside the pools and not actively traded , and (b) captured most of the buy-hold investors

For these reasons, as the institutional investors migrated to the CDS markets, the bond markets became less illiquid and active...

86

( though TRACE mandate improved bond market liquidity somewhat: Harris & Piwowar 2006; Bessembinder et al., 2006)

Implications
2.

Our findings provide insights into how the bond markets may be impacted following the CDS introduction and have bearings on the recent reforms in the OTC derivatives market. Our findings also have bearings on

3.

where informed trading and hence price discovery might take place, and thereby indicating that excessive regulations in CDS markets may be costly. Blanco, Brennan and Marsh (2005): Price discovery occurs in the CDS market

because of (micro) structural factors that make it the most convenient location for the trading of credit risk, and because there are different participants in the cash and derivative markets who trade for different reasons.

Easley, O.Hara and Srinivas (1998)

show that price discovery role of options should be more pronounced when the liquidity of the option market is higher compared to that of the stock market when options provide higher leverage and when the probability of informed trading is high.

87