40
1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from U.S. Life and Property-Casualty Insurance Companies Abstract This study uses a unique data set of the insurers to examine the effects of credit default swaps (CDSon the risk profile and financial performance of the insurance companies from 2001 to 2007. The analyses are based on a simultaneous equations model to capture the endogeneity between risks and CDS utilization. We have two interesting results. First, we find that use of CDS increases the total risk, idiosyncratic risk, and market risks for both Life and Property Property-Casualty (PC) insurers. In addition, taking positions as net-buyer or net-seller also increases the three types of risks. Second, use of CDS also appears to lower the financial performance in terms of Tobin q, market value of equity to book value, and return on assets of Life and PC insurers. Furthermore, participating in CDS as net buyers has more significant effects on reducing Life insurer’s firm performance than net-seller positions. On the other hand, PC insurers participating in CDS as either net buyers or net sellers significantly lower their firm performance.

Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

  • Upload
    others

  • View
    3

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

1

Will the Use of Credit Default Swaps Affect Risks and Firm Value?

Evidence from U.S. Life and Property-Casualty Insurance Companies

Abstract

This study uses a unique data set of the insurers to examine the effects of credit default swaps (CDS)on the risk profile and financial performance of the insurance companies from 2001 to 2007. The analyses are based on a simultaneous equations model to capture the endogeneity between risks and CDS utilization. We have two interesting results. First, we find that use of CDS increases the total risk, idiosyncratic risk, and market risks for both Life and Property Property-Casualty (PC) insurers. In addition, taking positions as net-buyer or net-seller also increases the three types of risks. Second, use of CDS also appears to lower the financial performance in terms of Tobin q, market value of equity to book value, and return on assets of Life and PC insurers. Furthermore, participating in CDS as net buyers has more significant effects on reducing Life insurer’s firm performance than net-seller positions. On the other hand, PC insurers participating in CDS as either net buyers or net sellers significantly lower their firm performance.

Page 2: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

2

Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from U.S. Life and Property-Casualty Insurance Companies

1. Introduction

This study examines the effects of credit derivative swaps (CDS) on risk profiles and thus

the market value of Life and Property-Casualty (PC) insurance companies. The CDS market has

grown enormously in recent years. The notional amounts of credit derivatives reached $17.1

trillion as of year-end 2005, a 25-fold increase from the level at mid-year 2001 (International

Swaps and Derivatives Association, 2006), and further reached a notional outstanding value of

over $60 trillion by the end of the first half of 2008 (BIS 2008). According to British Bankers’

Association (2006), insurers worldwide held an 18 percent market share for selling credit default

swaps (CDS) protection in 2006 and 6 percent of the CDS market for buying credit protection.

Thus, an examination of how CDS affect the insurance companies on their risk and the market

value is of interest to risk management and policy makers.

Credit derivatives enable investors to separate the origination of credit, the funding of credit,

and to manage credit risk. Two major effects of credit derivatives innovation are (1) to enable

risk sharing by hedging, and (2) to offload risks of investments easily (Instefjord, 2005). That is,

CDS markets provide insurers a mechanism for risk management and a tool for risk-taking.

The existing literature has primarily focused on examining banks’ risk-hedging and risk-taking

behaviors using CDS. As insurers are active participants in the CDS markets, there is limited

research on how insurers utilize CDS. This study intends to bridge the gap by examining how

CDS affect the insurers.

Apparently, insurers rely extensively on derivatives in general to manage interest rate risk,

market risk, credit risk, and liquidity risk. Because the insurance industry is highly-regulated at

the state level, the application of CDS is mandated by regulators for hedging, replication, and

income generation.1 With such features embedded in CDS utilization, CDS theoretically can

allow insurers to increase or decrease risk of their companies, depending on their underlying

motives. Receiving insurance premiums as their future liabilities payments and investing the

1 Hedging transactions are to reduce price, quantity, currency and other risks associated with their assets and liabilities. Income generation involves writing derivatives. Replication transactions replicate the performance of one or more assets that the insurers allowed to acquire.

Page 3: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

3

funds as their asset holdings, insurers intend to manage effectively their assets-liabilities

durations. The application of CDS plays an essential role in such duration management and can

possibly alter insurers’ risk characteristics. Intuitively, insurers can purchase CDS to hedge the

credit risk of their bond holdings. On the other hand, insurers can sell CDS to alter the credit

risks from bond holdings.. As the CDS market has more liquidity, replicating bond portfolios for

flexibility and greater diversification with CDS (Goldfried, 2003) can possibly shift risks to a

lesser degree. Therefore, either hedging or replication can possibly reduce insurers’ risks (Blanco

et al. 2005; Cummins et al., 2001). In fact, Guay (1999) has demonstrated that using derivatives

can reduce the risk of companies. For example, Metlife in its 2009 10-Q report confirms that it

uses certain credit default swaps intending to hedge against credit-related changes in the value of

its investments and to diversify its credit risk exposure in certain portfolios. However, as Metlife

realistically utilizes credit default swaps in non-qualifying hedging relationships, which can even

further increase firm risks.

Apparently, CDS can increase the risks of insurers. Selling CDS protection for income

generation purpose is a direct and intuitive link to increased risk because protection sales entail

intrinsic credit risk of default by the counterparty. An insure writes credit default swaps for which

it receives a premium to insure credit risk for the CDS buyer. CDS market can promptly reflect

credit risk of the overall market movement (Fung, et al., 2008); therefore, when insurers use

CDS extensively through income generation motive, they would expect to display greater risks

as illustrated by AIG’s large sale of CDS.2 It is true particularly for the beta risk of the insurers

writing CDS. Also, insurers writing protection CDS increase their credit risk if the CDS are

underpriced or their risks are not well diversified away. Given the complexity of the CDS

products, it is not straight forward to price them properly and thus CDS require expertise for the

correct pricing. Insurers are generally not experts in CDS trading as compared to investment

banks or commercial banks, leading to the sale of likely underpriced CDS. There is also a

potential spillover from counterparty risks among issuers because one failure may spillover to

other issuers too (Gaiyan –your citation?).

Income generation activity increases risk of the insurers from trading CDS portfolios to

generate profits on short-term differences in CDS price (see for example, Metlife 2009 10-Q

report). Such trading strategy may include a more subtle strategy of having a long position in

2 “How AIG fell apart,” by Adam Davidson, September 18, 2008, Reuters.

Page 4: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

4

CDS initially. That is, insurers purchase CDS initially in anticipation of widening CDS spreads

and then make a subsequent sale of CDS if credit spreads are expected to narrow. Profit, if any,

from these speculative credit-risk transactions may indeed increase risk of the insurers in light of

a wrong bet.

Consequently, it is an empirical question to examine how the use of CDS through the

application of its features of hedging, replication, and income generation affects insurers’ risk

profile. The objectives of this study are twofold. First, we use both Life and PC insures as net

buyers and sellers to examine the effects of CDS on three types of risk: total risk, idiosyncratic

risk and systematic risk. Life insurers’ liabilities have a longer maturity and thus a greater

duration mismatch risk. With such different underwriting skills and investment portfolios, Life

and PC insurers may utilize CDS to manage their balance sheet differently. We therefore

distinguish Life insurers from PC insurers in our analysis.3

Second, we evaluate the effectiveness of CDS using an outcome evaluation approach.

That is, we investigate the effect of credit derivatives on three performance measures of an

insurer: Tobin’s q (TQ), the market value of equity to book, and the return of asset. These

measures capture the different goals of using the CDS.

The relationship between firm risk and the use of CDS may be endogenous, which has been

raised in the earlier literature (e.g., Graham and Rodgers, 2002). To this end, we tend to mitigate

model biases arising from the issue of endogeneity using a simultaneous-equation model that is

estimated by a generalized method of moment (GMM) method. Our results show that for both

samples of Life and PC insurers, CDS participations consistently increase their total risk,

idiosyncratic risk, and market risk. In addition, life insurers’ participation positions as net sellers

also increase each dimension of risk and the results are mostly similar to PC net sellers except no

significant effects on market risk. To our surprise, insurers acting as CDS net buyers, each type

of firm risks (total risk, beta risk or idiosyncratic risk) is increased, especially for the sample of

PC insurers. Net-buyer CDS positions taken by Life insurers have marginally significant effect

on increasing market risk, but not on total and idiosyncratic risk. As CDS protection purchase

3 The existing insurance literature commonly discusses these two types of insurance companies separately for their underwriting, investing, or risk managing activities. For example, Cummins et al. (1997) discuss the different use of derivatives by life and PC insurers and find that the percentages of life and PC insurers engaging in derivatives transactions are 7% and 12%, respectively. In addition, according to the NAIC SVO report (2007), the investment portfolio of life insurers consists of 76.1 percent of total invested assets in 2006, while property/casualty insurers allocate 62.1 percent of their assets to bond holding.

Page 5: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

5

can embed both hedging and replication features, our results suggest that the increase in risk

through replication dominates the risk reduction from hedging, thereby leading to an overall risk

increase for CDS net buyer position.

In terms of the effects of CDS utilization on insurers’ firm performance, the extant literature

has shown mix results on the usage of financial derivatives. Our results suggest that CDS

participation and participation positions reduce firm value for both Life and PC insurance

samples. In particular, Life insurers, as net buyers of their CDS participation and their

participation positions reduce their market-based firm performance (Tobin’s Q and ratio of

market to book value of equity) and also accounting-based firm performance— return on assets.

Net sellers do not show significant lower firm performance. PC insurers, on the other hand,

show lower Tobin’s Q and market to book value of equity for both net CDS buyers or sellers.

This study demonstrates that CDS utilization adversely increases Life and PC insurers’ risk

profile and reduces firm performance. Our findings support the ongoing effort of the National

Association of Insurance Commissioners (the “NAIC”) to reexamine the role of insurance

regulators who must monitor closely insurance companies that engage in derivative transactions.

While insurance companies are primarily regulated at the state level, our results confirm the need

of an amended regulation for derivatives from NAIC as a national standard to be implemented by

all states.

The rest of the paper is organized as follows. The literature and related hypotheses are

summarized in Section 2. Data description and the construction of variables are illustrated in

Section 3. Section 4 presents methodology application and empirical findings. The last section is

the conclusion.

2. Literature and Model Development

2.1 CDS and Risks

Many theoretical studies (e.g., Stulz, 1984; Smith and Stulz,1985) have shown that, in the

presence of market imperfections such as taxes, financial distress costs and agency conflicts,

companies are motivated to adopt corporate hedging policy. However, the existing literature

show mixed empirical results for the effects of derivative applications on the firm’s risk.

Hentschel and Kothari (2001) use a panel of 425 large U.S. firms as a research sample and do

not detect an economically or statistically significant relation between firms' risk characteristics

Page 6: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

6

and the extent of their participation in derivatives markets, concluding that derivative use does

not measurably increase or decrease firm’s return volatility. Adam and Fernando (2006)

examine a unique database that contains gold derivatives positions for a sample of 92 North

American gold mining firms from 1989 to 1999. They conclude that derivatives usage has no

measurable impact on exposure or volatility.

A different line of research provides evidence that the derivative use can increase or decrease

risks. DeMarzo and Duffie (1992), and Froot, Scharfstein, and Stein (1993), among others,

construct models of corporate hedging, predict firms to reduce the risks if they have poorly

diversified and risk-averse owners, suffer large costs from potential bankruptcy, or have funding

needs for future investment projects in the face of strongly asymmetric information. Such risk

reduction can be achieved through derivatives. Tufano (1996a) conducts a detailed study of risk

management in the gold mining industry and find evidence supporting the hypothesis that firms

in the gold mining industry use derivatives to reduce risks. The primary motivation for this

hedging seems to be managerial and owner risk aversion.

Firm owners might use derivatives to take on additional risks because of agency problem.

Jensen and Meckling (1976) and Myers (1977) point out that the owners of leveraged firms can

have incentives to increase the firms' riskiness to transfer wealth from bond holders to stock

holders. Froot and Stein (1998) find that active risk management can allow banks to hold less

capital and to invest more aggressively in risky and illiquid loans. Cebenoyan and Strahan (2004)

find that actively managing credit risk through loan sale allows banks to have lower risk-adjusted

capital than banks that don’t actively manage their credit risk. These studies suggest that active

users of credit derivatives may have incentives to reduce capital and increase their holdings of

risky loans. Instefjord (2005) argues that the risk sharing benefits from credit derivatives may

encourage banks to take more risk, thus creating a potential for greater bank instability. Morrison

(2005) finds that credit derivatives could reduce banks’ incentives to monitor their loan portfolios.

In addition, Shao and Yeager (2007) conclude that the use of CD increases overall risk for a

sample of U.S. bank holding companies between 1997 and 2005 by showing that credit

derivatives usage motivates BHCs to shift from safer loans to riskier ones.

In light of the insurance literature, the study on the use of CDS is limited. This study intends

to bridge this gap by examining the effects of CDS on risk characteristics of both Life and PC

insurers in three dimensions: total risk, systematic (market) risk, and idiosyncratic risk. Using

Page 7: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

7

total risk allows us to capture the overall risk profile of an insurer after identifying individual risk.

The examination on market risk is to capture the market responses to insurers’ participation in

CDS. Instefjord (2005) argues that when the credit markets are highly price-inelastic, the

systemic risk is also likely to decrease with an access to a richer set of credit derivatives,

particularly when the benefits to consumers of credit are low. Consequently, it is of interest to

examine the effects of CDS on insurers’ systematic risks. Idiosyncratic risk reflects the

firm-specific risk profile. Low idiosyncratic risk implies large information asymmetry and

opaqueness (Bali et al., 2005; Durnev, 2004; Morck, 2000). It is important to examine how CDS

utilization affects idiosyncratic risk because the literature has raised its concern on the

information flow and opacity problems embedded in CDS (Acharya and Johnson, 2007, Nicolo

and Pelizzon, 2008). Nicolo and Pelizzon (2008) indicate that the recent sub-prime crisis has

highlighted the growth in volume and diversity of credit derivative products, which do not

appear to mitigate the problem of the transparency in the insurance markets. Acharya and

Johnson (2007) reports significant incremental information revelation in the credit default swap

market; however, the early detection of credit risk in the CDS market is often attributed to insider

trading by large institutional investors with information advantage. It is possible that insurance

companies may have sophisticated skills and resources to obtain and process information.

However, insurance industry may have an information disadvantage on the transactions

compared to bank holding companies, which are likely to have insider information on the

reference entity. Thus, through the idiosyncratic risk analysis, this study intends to explicitly

investigate the information asymmetry issue in insurance industry.

Buying CDS protection intuitively is to reduce credit risks. It is an empirical question on

whether the reduction in credit risks can be transferred into the reduction of firm’s overall risk,

market risk or firm-specific risk. As developed in Fairley (1979), insurers may set up a target

firm risk-level, which is a result of the tradeoff between the risks inherent inside the insurer.

Therefore, we conjecture that it is likely that given a maintained target risk level, the reduction in

credit risks through CDS protection purchase may simultaneously increase insurer’s capacity to

take more other risks (other than credit risks).

Specifically, we examine (1) whether purchasing CDS protection to reduce credit risk can

simultaneously reduce insurers’ overall risks or otherwise, (2) whether the market can recognize

the reduction of credit risks attributed to CDS protection purchase and thus reflect in the

Page 8: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

8

reduction of systematic risks or the market prefers insures not to hedge credit risks so that the

reduction of credit risks is transferred to higher systematic risks, and (3) whether purchasing

CDS protection signals the investors the opaqueness on the risks embedded inside the firm

thereby reflecting on higher level of information asymmetry, namely, lower idiosyncratic risk or

otherwise.

Selling CDS protection is expected to increase insurers’ exposure to credit risks. Hopman

(2007) points out that insurance companies have been carrying significant credit risks because of

the investment portfolio and thus selling CDS protection is simply an extension of credit risks to

insurance companies. As a result, insurance companies that engage more on selling CDS

protection can be viewed as an alternative way of taking extra credit risks. The regulators

consider insurers’ use of replication as being speculative and thus require higher regulatory

capital level for the added risk. This line of research suggests that selling CDS would increase

insurers’ risks.

While it is intuitively to link CDS sell positions to the increase of risk, engaging in CDS sell

position can possibly reduce firm risks from the asset replication feature embedded in CDS as

argued in Goldfried (2003) that credit derivatives allow insurers for precise portfolio

management, risk mitigation, and for optimal capital management. Credit derivatives enables

insurers to create a security structure similar to the corporate bonds by taking a short position in

CDS provided the fact that the existing bonds may not provide adequate flexibility to the insurers

for duration consideration.4 As a result, they may benefit from the CDS asset class.

Therefore, it is a matter of empirical question on how insurer’s risk profile is affected

through the utilization in CDS sell position. We examine specifically whether taking CDS sell

position increases or decreases the three measures of risk: total risk, systematic risk, and

idiosyncratic risk.

2.2. CDS and Firm Value

CDS is a special type of general derivatives for risk management. The existing literature,

however, has also provided mixed evidences of the effects of derivatives on firm value.

4 Insurance company may use credit derivatives to diversify their credit risks by selling CDS contracts in the two-, three-, five- or 10-year maturities or in names that do not trade in the bond or cash market. Thus, engaging in CDS transactions enables insurers to have access to duration management that might not be available in the cash market in achieving a more effective asset/liability management.

Page 9: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

9

Apparently, risk management policies could be relevant to firm valuation under an imperfect

market.

One line of research supports value creation. For example, the study by Mackay and Moeller

(2007) finds that risk management can increase firm value if the cost and revenues functions are

not linearly related to input and output prices. Focusing on hedging specific risks and on specific

financial instruments, Smithson (2005) finds that the management of interest rate and foreign

exchange risks adds value to the firm and Alyanis and Weston (2001) conclude that users of

foreign currency derivatives have higher market value than those non-users. In addition, the

research on specific industry’s risk management provides similar conclusions. For example,

Dionne and Triki (2006) study gold mining industry and find that hedging increases returns on

asset. In addition, Nelson et al. (2005) examine non-financial firms hedging behaviors and

conclude that hedgers with derivatives outperform non-hedgers.

The other line of research provides no evidence of firm value creation. In particular, Adam

and Fernando (2006) find no evidence that the use of derivatives has increased the systematic

risk for the firms in their sample, thereby implying no benefit to shareholder value from

derivatives transactions. Jin and Jorion (2006) use oil/gas producers as a research sample and

show that hedging has no significant effects on increasing firm value. Moreover, Shao and

Yeager (2007) use bank holding companies as research sample and find that their use of

derivatives increase risks and lower return.

These above studies use specific industry as research sample to examine the effects of

specific instruments of risk management on examining firm value creation or firm value

reduction hypothesis. Our study specifically uses insurance industry to examine the effects of

CDS on firm value. Using single homogenous insurance industry as compared to a sample of

U.S. multinationals lessens the possibility of spurious results from confounding factors. In

recognizing the differentia of underwriting operations between Life and PC insurers, we analyze

their CDS utilization separately. In sum, we investigate the effects of CDS utilization on

financial performance based on the measures of Tobin’s Q, a market-based firm value measure,

and on the measure of return on assets (ROA), an internally accounting-based profitability

measure.

3. Data and Variable Construction

Page 10: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

10

3.1 Data and Sample Selection

As a highly-regulated industry, insurers are required to report their investment activities

including derivatives use to the National Association of Insurance Commissioners (NAIC). The

detailed nature of the reported data on CDS utilization provides a unique opportunity for

analyzing insurers’ practices of credit derivatives and evaluating the effects on their risks and

firm values. We compiled the data from the regulatory annual statements filed by insurers with

the National Association of Insurance Commissioners (NAIC) for the period from 2001 to 2007.

Insurers mainly practice credit derivatives for risk swap and asset replication. Detailed

information on the use of credit derivatives can be found in Schedule DB. From which, we

manually identify the transaction information on CDS positions for purchasing CDS protection

or selling protection. Moreover, the following trading information can be handily collected : the

notional amount, date of opening position, date of termination date, date of maturity,

consideration received or paid, gain (loss) on termination, and individual within-year and

year-end transactions volume. We conduct the analysis based on individual firm-level data.5 To

our best knowledge, no prior study has done such a detailed examination on how insurers apply

credit default swaps and investigating the effects of CDS utilization and participation positions

on insurers’ risk profile and firm value.

To analyze the effects of CDS use on risks and firm value by controlling firm-specific

characteristics, we compile the data from CRSP and CompuStat databases and merge them with

CDS data for the period from 2001 to 2007. The merged dataset results in 44 distinct insurers

identified as CDS users including 11 PC insurers and 33 Life insurers. In addition, we also

include those publicly-listed insurers with no CDS activities in our sample. The final sample size

is with 85 and 127 distinct Life and PC insurers, respectively and their respective firm-year

observations are 427 and 666. When each CDS transaction is viewed as an individual

observation, the total number of firm-year-transaction observations of the life insurer sample is

4,889 and 1,639 for PC insurers.

3.2 Variable Construction

3.2.1 CDS Variables

5 Many insurers are members of groups that operate under common ownership. Cummins et al. (1997, 2001) found that the group level analysis group level analysis provided virtually no information concerning the derivatives participation decision and thus may loss importance information.

Page 11: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

11

We use a dummy variable, CDS_Dummy, which equals to 1 if the CDS activity is identified

and 0 otherwise to proxy for CDS participation. We measure the extent of CDS participation by

the notional amount of CDS positions held at year end on insurers' annual statement to measure

the transaction volume engaged. We further create two separate dummy variables, Net_Buyer

and Net_Seller based on the annual aggregate notional amount in Buy and Sell position. An

insurer is indentified as a net buyer (i.e., Net_Buyer = 1) if the aggregate notional amount of

CDS buy positions is greater than that of CDS sell positions and zero elsewhere. A net seller (i.e.,

Net_Seller = 1) is denoted as its aggregate notional amount of sell positions is greater than the

buy positions. To control for the effects from CDS position changes, we use a CDS_Change as a

dummy variable with its value equals to 1 if insurers change their CDS position from (1) CDS

users to non-users, (2) from net-seller positions to net-buyer positions, or (3) from net-buyer

positions to net sellers. Spread_Vol measures the volatility of daily CDS spread.

3.2.2 Risk Variables

To obtain risk measures, we first estimate the capital asset pricing model (CAPM) for each

of the firms in our sample. The market risk is the beta coefficient estimate from the following

regression model:

rj,t – rf,t = αj,t + βj,t×(rm,t – rf,t) + εj,t (1)

where rj,t is firm j's monthly return at time t; rm,t is a value-weighted average market return; and

rf,t is the risk-free rate. Our estimation is based on a rolling regression model over a five-year

period. The regression coefficient, β, from the regression model represents the market risk.6 The

total risk is the total standard deviation of monthly returns while the idiosyncratic risk is the

standard deviation of residuals from the regression model.

3.2.3 Performance Variables

We apply Tobin’s q (TQ) as a market-based measure of firm value and also use an

accounting-based internal profitability measure of return on asset (ROA) to examine the

6 Some previous studies on insurer risk-taking employ the NAIC database or AM Best reports to compute asset risk, product risk, and portfolio risk (Cummins and Sommer 1996; Baranoff and Sager 2002). However, our study extracts market-based variables such as stock returns from publicly-traded insurance companies, hence the utilization of market-based risk is more appropriate for this research. Moreover, the use of market-based risk measurements such as total risk and residual risk captures the overall risk profile as well as the firm-specific risk profile of the insurers.

Page 12: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

12

relationship between the financial performance and CDS transactions. In addition, we use the

ratio of market value of equity to book value of equity (MV(Eqty)_BV(Eqty)) to measure firm

performance. That is, we test whether the market rewards the insurers on their use of CDS.

Tobin’s q (TQ) is a market-based proxy for firm value and is defined as the market value of

equity plus the book value of liabilities divided by the book value of assets, i.e.,

)(

)()()(

assetstotalBV

equitycommonMVequitycommonBVassetstotalBVTQ

(2)

where MV (common equity) is the product of the stock price and the number of shares

outstanding. This measure has been widely used in the existing literature to measure the

value-effects of factors (Yermack, 1996; Morck, Schleifer, and Vishny, 1988; Servaes, 1996;

Smithson and Simkins, 2005;Cummins, Lewis, and Wei, 2006; Jin and Jorin, 2006). Tobin’s Q is

used in our study to measure the financial performance because it can reflect future expectations

of investors. Cummins et al. (2006) contend that this version of Tobin’s Q is appropriate for

insurance companies because the book value of their assets is a much closer approximation of

replacement costs than would be the case for non-financial firms.

4. Methodology and Empirical Results

4.1 Descriptive Statistics

Table 1 shows a summary of CDS transaction activities of Life insurers in Panel A and PC

insurers in Panel B. For the Life insurers, the frequency of CDS participation is about 21.78%,

among which, about 30% are net buyers and 70% are net sellers.7 On the other hand, for the

sample of PC of insurers, CDS participation is about 5.86% in which 64% and 36% are

net-buyers and net-sellers respectively. This provides preliminary evidence supporting that (1)

life insurers are more active in CDS utilization and (2) life insurers tend to participate in CDS

transactions as net sellers that PC insurers, which tend to be net buyers. In addition, among all

the 4889 CDS transactions over the sample period, 30% and 70% are for buy and sell

transactions, respectively for Life. Among all the 1639 CDS transactions for PC, the respective

percentages for buy and sell transactions are 68% and 32%. In terms of the notional amount of

7 Among 427 (666) observations for life (PC) insurers, 93 (39) are the observations for CDS users over the sample period and of which 28 (25) are for net-buyer and 65 (14) are for net-sellers.

Page 13: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

13

CDS transactions, each transaction is about $15 million dollars for Life insurers and $23 M for

PC. The aggregate notional CDS amount to total asset is 0.40% for Life and 0.60% for PC.

Table 2 presents the descriptive statistics of the three risk variables and other control

variables used in the analysis. Panels A and B are for the samples of Life and PC insurers that

include both CDS users and non-users. Panel A shows that, for the Life insurers on average,

the idiosyncratic risk is 0.099, the systematic risk is 0.713, and total risk is 0.104. For the PC

insurers, the mean values for idiosyncratic risk, systematic risk, and total risk are 0.095, 0.679,

and 0.100, respectively. Tobin’s q, on average, is 1.15 units for Life insurers and 1.10 units for

PC insurers. In terms of profitability, the mean value of ROA is 3.7% for Life insurance and

3.4% for PC. On average, Life insurers have a higher total risk, systematic risk, and idiosyncratic

risk than that of PC insurers. In addition, both market-based firm value (Tobin’s Q) and

accounting-based internal performance (ROA) are larger in the sample of Life insurers than PC.

[Insert Table 2 here]

Table 3 compares medians and means of the risks, the firm value, and other firm characteristics

variables between insurers who are CDS users, CDS net buyers, and CDS net sellers and those

insurers who do not participate in CDS at all. Panels A and B report the comparison results for

the samples of Life and PC, respectively.

Panel A shows that Life insurers with CDS transactions, on average, have a larger

systematic risk, lower idiosyncratic risk, and lower total risk than those of non-CDS users. On

average, the values of systematic risks are greater for CDS users (0.861) than non-users (0.671),

while the average total risk (0.11) and idiosyncratic risk (0.105) for non-CDS users are larger

than the corresponding figures of CDS users (0.085 and 0.077). In addition, non-CDS users

have larger Tobin’s Q and return on asset than CDS users. Particularly, the average value of

Tobin’s Q is 1.183 for non-CDS users and 1.033 for CDS-users.

The Tobin’s Q for net-buyers is at an average value of 1.018 and for net-sellers is, on

average, 1.042, both of which are smaller than that of non-CDS users with an average value of

1.183. As expected, CDS users, on average, are large Life insurers with a firm size (natural

logarithm of total assets) of 11.18 compared to 9.03 for non-CDS users. In addition, CDS life

Page 14: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

14

insurers are with a higher financial leverage, a lower investment opportunity and a larger

proportion of short-term debts.

[Insert Table 3 here]

Panel B of Table 3 shows the comparison results between users and non-users for PC

insurers. The total risk of the CDS users is similar to the non-CDS users. CDS-users in PC

sample have higher systematic risk and lower idiosyncratic risk than those of non-CDS users. In

particular, the average values of systematic risks are 1.37 and 0.64 for CDS users and non-users,

respectively; in addition, the respective mean values of idiosyncratic risks are 0.086 and 0.096.

In addition, consistent with Life sample, net buyers of PC sample have significantly lower

idiosyncratic risk (0.080) and insignificantly smaller total risk (0.098) than those of non-CDS

users. On the other hand, net sellers of PC insurers show significantly higher idiosyncratic risk

(0.108) and significantly larger total risk (0.118) than those of non-CDS users with the respective

values of 0.096 and 0.099.

The firm value comparison between CDS users and non-users of PC sample also show

consistent conclusions as discussed in Life sample. CDS users of PC sample have lower

Tobin’s Q and lower ROA than those of non-users. The respective mean values of Tobin’s Q and

ROA are 1.01 and 2% for CDS users compared to the values of 1.10 and 4% for non-users.

Moreover, CDS net buyers and net sellers also have lower market-based firm value and

accounting-based firm performance than those of non-users.

Univariate analyses for both Life and PC insurers reveal several findings. First, there is a

positive relationship between insurers’ market risks and CDS participation and CDS net-buyer

positions, i.e., CDS users or net buyers have higher market risk than non-CDS users. Second,

there appears to have a negative relationship between CDS utilization and total risk and

idiosyncratic risk. That is, CDS users or net buyers have lower total and idiosyncratic risks

than non-users. Finally, a lower firm value appears to be associated with more CDS utilization.

A non-tabulated univariate correlation analysis shows a positive correlation between CDS

utilization and market risk, but a negative correlation of CDS use with total risk, idiosyncratic

risk, and firm market value. The following sections are to examine the relationships on a

multivariate basis.

Page 15: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

15

4.2 Multivariate Regression Models

We first examine the risk-CDS relation under a multivariate framework by controlling the

firm-specific characteristics. To mitigate the bias arising from the endogeneity between risk and

CDS utilization, we use a simultaneous equations model to investigate the relationships between

risks and CDS participation and the participation positions as either net sellers or net buyers.

4.2.1 Simultaneous Equations Model on Risk Analysis

Under a simultaneous framework, we examine the relationship between the risk level and

CDS participation (and participation positions) of insurers after controlling for the firm

characteristics. The use of simultaneous equations, as compared to independent OLS regressions,

is more appealing for the following reasons. First, the literature suggests potential endogeneity

between risk and CDS utilization and therefore a joint determination of these variables should be

simultaneously determined (Note: Literature references needed here, Gaiyan). Second, from the

econometrics theory, simultaneous equations are appropriate when the endogenous variables are

jointly determined, and the interactions of key variables have significant implications for the

parameters estimation. CDS utilization may constrain insurers’ risk-taking level due to regulatory

consideration and may also potentially lead to a higher level of risk-taking for income

enhancement purpose. Regulatory capital requirement may limit insurers to use CDS and

motivate the use of CDS for hedging and asset replication purposes.

The simultaneous equations model is described in Equation (3) and Equation (4)

representing the risk equation and the CDS equation, respectively. We estimate the system of

equations with the generalized method of moments (GMM) using the exogenous variables as

instruments in the moment conditions as follows.8

8 Note that other instrumental variables techniques, such as two-stage least squares (2SLS), are special cases of GMM. For example, in comparison with 2SLS, Greene (2002) and Kennedy (2003) observe that are more efficient than 2SLS estimates when regression errors are heteroskedastic and/or autocorrelated, and that GMM estimates coincide with 2SLS estimates otherwise. Thus, GMM ensures that the standard errors of the estimates are heteroskedasticity and autocorrelation consistent. Finally, note that we do not report the R2s for our estimated equations, since as Goldberger (1991) observes, there is no guarantee that the R2s reported in system estimation techniques lie between zero and one. Unfortunately, there is no widely accepted goodness of fit measure for nonlinear system estimation.

Page 16: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

16

Riski,t= α1 + β1,1 ×CDSi,t + tii Z ,,1 + β1,2 ×Div_yieldi + β1,3 ×CDS_Changei,t +

β1,4 ×Spread_Volti,t + ε1,i,t (3)

CDSi,t= α2 + β2,1 ×Riski,t + tii Z ,,2 + β2,2 ×NY_Dummyi,t + β2,3 ×CDS_Changei,t +

β2,4 ×Spread_Volti,t +ε2,i,t (4)

where Risk represents insurers’ risk profile characterized by total risk, systematic risk, and

idiosyncratic risk. The estimation procedure of each risk measure has been stated in Section 3.

CDS represents CDS-related dummy variables that depict insurers’ participation in CDS

transactions (CDS_Dummy), their participation positions as net buyers (Net_Buyer), or the

positions as net sellers (Net_Seller). The definition of these CDS variables is described as

follows:

CDS_Dummyi,t = 1 if insurer i participates in CDS transactions in year t and zero elsewhere;

Net_Buyeri,t = 1 if the aggregate notional amount of CDS buy position is greater than that of

the sell position for insurer i in year t and zero elsewhere;

Net_Selleri,t = 1 if the aggregate notional amount of CDS sell position is greater than that of

buy position for the insurer i in year t and zero elsewhere;

For model identification purposes, Div_yield defined as dividend yield is included in the

risk equation, but not in CDS equation; State_Dummy is included in CDS equation, but not in

risk equation and it is defined as a dummy variable equals to 1 if the insurer i operates in the

New York State. The model has also controlled for time effects. Because the regressions involve

multiple years, we also include annual dummies. For space consideration, the estimates of these

time-effect variables are not reported.

In the risk equation (equation (3)), the coefficients of CDS participation and the net position

of CDS are of interest. The extant literature (e.g. Instefjord, 2005; Morrison, 2005; Shao and

Yeager, 2007) suggests risk-taking behaviors to be observed from the use of CDS; that is a

positive coefficient of CDS_Dummy in the risk equation. On the other hand, risk-hedging

hypothesis suggests a negative coefficient. Extending this line of research, our study takes one

step further to distinguish the effects of taking CDS protection from the effects of writing CDS

protection. In addition, to control for effects from CDS participation change or position change,

we include the variable CDS_Change, which is defined as a dummy variable with a value of 1 if

Page 17: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

17

CDS participation is changed from a participation position to a non-participation or when the net

position of CDS is reversed. Moreover, Spread_Volt presents the volatility of daily CDS spreads

over a year to control for the changes in credit market conditions.

In the CDS equation (equation (4)), the coefficients of risks, are to examine how insurer’s

risk characteristics affects their decision on CDS use as either net buyers or net sellers. (Any

references or literature suggest the effects of risk on CDS participation??)

The definition of each control variable is illustrated as follows. Firm size, measured by

the natural logarithm of the book value of assets, is used to control for the firm size effect.

Leverage measures the capital structure of the insurers and is defined as the ratio of long-term

debt plus preferred stock over long-term debt plus preferred stock plus common equity to total

assets (i.e., leverage to market value of assets).9 Debt_maturity is the portion of short-term debt

measured by the ratio of short-term debt with maturity less than three years to total debt and it is

to measure the debt characteristics of the insurers. Investment opportunity is to control the

growth opportunity of the firm, which is defined as net capital expenses to assets. Rating

measures the risk class of the insurers and is based on S&P issuer credit ratings to classify firms

into investment and speculative grade to measure the credit risk and the unrated firms are

included in the speculative grade. Sale growth represents the growth of sale.

4.2.2 Empirical Results

Total Risk Model

For Life insurers, their participation in CDS transactions increases total risk at the 1%

significant level as shown in Panel A1 (for the risk equation) of Table 4. In terms of the effects

of participation positions, the total risk is positively and significantly related to the net sellers

dummy variable (0.148 with a t-value of 2.64). While buying CDS protection appears to reduce

credit risks, the effect of net-buyer positions is negative but insignificant (-0.068 with a t-value

of -1.04) on total risk, while net sellers do increase Life insurer’s risk.

In the CDs equation, CDS use and participation positions are positively and significantly

related to the total risk factor, with a coefficient of 7.504 (t-value of 3.79) for the overall result, a

coefficient value of 1.844 (t-value of 1.87) for the net buyers, and a coefficient estimate 5.398

(t-value of 3.18) for the net sellers, confirming the presence of endogeneity relations between

9 Results are robust when the leverage variable based on book asset value is used.

Page 18: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

18

CDS and risks identified in the model.

[Insert Table 4 here]

CDS_Change, a change in CDS position, does not appear to affect the total risk

significantly. Total risk is positively and significantly related to investment opportunity

measured by the capital expenditure with a coefficient value of 0.015 (t-value = 4.04) and

leverage (0.122; t-value = 4.05), but negatively and significantly related to firm size (-0.017 with

a t-value of -4.36), dividend yield (-0.023 with a t-value of -2.17) and firm rating (-0.02 with a

t-value of -2.77) for the overall sample. The results are robust across the overall sample,

net-buyer or net-seller samples.

The results for PC insurance sample shown in Panel B1 are similar to those for Life

insurance sample but net buyers or net sellers in CDS transactions show a stronger and

significant negative effect on total risk.

Idiosyncratic Risk Model

Panel A2 of Table 4 reports the results of idiosyncratic risk model for the life insurers.

Similar to the total risk results, the idiosyncratic risk (IR) is positively and significantly related to

the participation of life insurers in CDS transactions with a coefficient value of 0.056 at the 10%

significant level. Taking CDS positions as net sellers also has a positive and significant effect on

IR with a coefficient value of 0.114 at the 5% level. A higher idiosyncratic risk represents a

larger degree of information revealed to the market (Bali et al., 2005; Durnev, 2003) and thus

less information opaqueness. Therefore, the increased idiosyncratic risk through the participation

and participation positions of CDS transactions suggests that investors interpret them as a signal

to identify the higher degree of information revealed to the market by Life insurers that may not

be identified earlier by investors.

For the PC insurance sample, Panel B2 shows that idiosyncratic risk (IR) is positively and

significantly related to not only CDS participation, but also for both samples of net buyers and

net sellers, indicating that the use of CDS indeed increases IR.

For the CDS equation, results show that for PC insurers, IR increases CDS participation and

position as net buyers, but not as net sellers. On the other hand, for Life insurers, IR increases

Page 19: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

19

CDS use as well as the positions as net sellers, whereas not for the positions as net buyers.

Market Risk Model

Panels A3 and B3 of Table 4 shows the results of market risk models for Life and PC

insurers, respectively. As shown in Panel A3, Life insurers’ participation in CDS and their

participation as either net buyers or net sellers consistently and significantly increase the market

risk. However, for the PC insurers, their CDS participation has no significant effects on market

risk. Their participation positions as net-buyer show significant and positive effects on market

risk. Compared to other risks, market risk of Life insurers is more sensitive to the firm’s CDS

participation as well as participation positions regardless of net-buyer or net-seller position. Such

observations specific to Life insurers are consistent with the findings in Instefjord (2005) arguing

that the engagement in CDS transactions (either buy or sell) is harmful to Life insurer’s

systematic risks.

4.3 Performance Analysis

The next research question is to investigate whether these effects of CDS on risks can be

transferred to firm value enhancement or, on the other hand, value reduction. To this end, we

conduct a multivariate regression analysis based on equation (5) in which the dependent variable

is to measure firm value. CDS participation that increases risk (particularly the market risk)

may suggest that shareholders would require a higher cost of capital and thus triggers a lower

market value.

Our data is a pooled time-series and cross-sectional unbalanced panel data, which are likely

to firms clustering together. CDS trading volumes may also be correlated across insurers for a

given year, therefore, we also need to correct for the time effects. As a result, we follow Petersen

(2009) to adjust for insurer-clustering effects and time varying effects. In addition, the regression

model involves multiple years and thus we also include annual dummies, which are not reported.

Performance = α0 + βj ×CDSi,t + βk ×CDS_Changei,t + βl ×Spread_Volti,t + tii Z , + εi,t (5)

We use two main variables to proxy for firm value/performance measure, which are Tobin’s q

and return on asset (ROA), representing an insurer’s market-based firm value and

Page 20: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

20

accounting-based internal firm performance, respectively. CDS represents the sane

CDS-related variables, defined earlier. Similarly, firm-specific characteristics are also discussed

earlier.

Panels A1, A2, and A3 of Table 5 show the results of Tobin’s q, Market to Book value of

equity, and ROA measures for Life insurance sample, while Panels B1, B2 and B3 present the

results for the PC insurers. Each Panel also summarizes the effects of CDS utilization and

participation positions as either net buyers or net sellers on firm performance.

Panels A1, A2 and A3, for Life insurers show that CDS utilization has significant and

negative effects on market-based firm value and on ROA. Specifically, Tobin’s Q is negatively

and significantly related to CDS participation (-0.059 with a t-value of -1.99) and to CDS

net-buyer positions (-0.078 with a t-value of -2.01), while insignificantly related CDS net-seller

positions (-0.043 with a t-value of -1.45). ROA is negatively and significantly related to CDS

dummy for the overall sample (-0.013 with a t-value of -2.22), for the net-buyer sample (-0.012

with a t-value of -0.012 with a t-value of -1.75) and for the net-seller sample (-0.010 with a

t-value of -1.66).

Results suggest that the market does not reward Life insurers for the hedging, replication or

income generation feature embedded in CDS utilization that are intended for diversifying credit

risk or replicating asset. Particularly, CDS protection purchase for hedging purpose does not

confer a special advantage to insurers it is likely due to the fact that well-diversified investors

can hedge on their own. That is, the credit risk exposure embedded in life insurers’ asset

holding can be diversified away by individual investors. This finding pertains to the conclusion

in Adam and Fernado (2006), Jin and Jorion (2006), and Shao and Yeager (2007) that risk

management is irrelevant to firm valuation drawn.

Life insurers writing CDS protection for taking more credit risks could be for the purpose of

income generation or asset replication, but none of these features is recognized for value creation.

Life insurers acting as CDS net sellers increase their exposure to credit risks, which increases life

insurance firms’ total risk and market risk. The combined effects of Life insurers’ CDS

participation as net sellers on firm’s risk characteristics result in an insignificant effect on their

firm value. Moreover, selling CDS protection carries the benefits of income generation and asset

replication with greater liquidity and flexibility. The insignificant effects of net-seller positions

on firm value suggest that such benefits are not explicitly transferred into firm value creation.

Page 21: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

21

[Insert Table 5 here]

The results for the sample of PC insurers are shown in Panels B1, B2 and B3. Consistent

with the results for Life sample, Tobin’s q and the ratio of market value to book value of equity

of PC are negatively related to CDS dummy across the overall sample, net-buyer and net-seller

sample. The coefficients are negatively significant at the 1% level. On the other hand, the

accounting-based internal performance measure ROA shows an insignificant result.

It is noteworthy that for PC insurers CDS position change variable appears to have

significant and negative effects on the market value of the firm and ROA, whereas shows

insignificant effects for Life insurers. The results indicate that PC insurers are not particularly

appropriate for using CDS from the perspective of the firm profitability and preference of

stockholders. (Hung-gay: Can you illustrate more on this point? I do not quite follow this.)

Taken together, for both Life and PC insurers, CDS participation consistently reduces the

firm values. The result suggests that regardless of the discrepancy of underwriting and

investment behaviors between Life and PC firms, the market consistently consider the CDS

participation and the change of CDS positions (either as net buyers or net sellers) as a

value-reduction mechanism.

5. Conclusions

This study uses a unique data set of the insurers to examine the effects of credit default swaps

(CDS)on the risk profile and financial performance of the insurance companies from 2001 to

2007. To this end, we apply a simultaneous equations model to identify the relationship between

risk and CDS use for the pooled-time-series and cross-sectional insurance data to mitigate the

bias from the endogeneity problem between risk and the use of CDS.

We empirically examine two research questions. First, we use both Life and PC insures as

research samples to examine whether CDS participation and participation positions as net buyers

or net sellers have significant effects on three types of risk: total risk, idiosyncratic risk and

systematic risk. Second, we investigate how firm performance responds to the use of CDS and its

positions. Firm performance is based on the measures of an insurer using Tobin’s q (TQ), market

value of equity to book value, and return of asset.

Our results demonstrate that for Life and PC insurers, CDS participations consistently

Page 22: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

22

increase their total risk, idiosyncratic risk, and market risk. Net CDS sellers of PC insurance

sample show a stronger and significant increase in total and idiosyncratic risks than Life net

sellers, whereas Life net sellers show a stronger increase in their market risk. For CDS net buyers

each type of PC insurance company’s risk (total risk, beta risk or idiosyncratic risk) is

significantly increased. On the other hand, CDS net-buyer positions taken by Life insurers have

marginally significant effect on increasing market risk. While have no significant effects on total

and idiosyncratic risk. As CDS protection purchase can involve both hedging and asset

replication features, our results suggest that the increase in risk through replication dominates the

risk reduction from hedging.

Moreover, we examine the effects of CDS utilization on insurers’ firm performance. Our

results suggest that CDS participation and participation positions reduce the firm values for both

Life and PC insurance samples. In particular, Life insurers acting as net buyers, their CDS

participation positions reduce market-based firm performance (Tobin’s Q and ratio of market to

book value of equity) as well as the accounting-based firm performance— return on assets. Net

sellers of Life insurers, however, do not significantly lower their firm performance. PC insurers,

on the other hand, show consistently a lower Tobin’s Q and market to book value of equity for

both net CDS buyers and sellers, whereas show no significant effects on ROA.

This study demonstrates that CDS utilization alters the risk profile of both Life and PC

insurers in increasing each dimension of risk. At the same time, the financial performance is

reduced consequently. Our findings support the effort of the National Association of Insurance

Commissioners working with the insurance regulators to monitor closely insurance companies

engaging in derivative transactions. While insurance companies are primarily regulated at the

state level, our results confirm the need of an amended regulation for derivatives from NAIC as a

national standard to be implemented by all states.

Page 23: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

23

References

Acharya, V., Johnson, T., 2007. Insider Trading in Credit Derivatives. Journal of Financial Economics 84, 110-141.

Adam, T.R., Fernando, C.S., 2006. Hedging, Speculation, and Shareholder Value. Journal of Financial Economics 81, 283-309

Allayannis, G., Weston, J., 2001. The Use of Foreign Currency Derivatives and Firm Market Value. Review of Financial Studies 14, 243-276.

Bali, T.G., Cakici, N., Yan, X., Zhang, Z., 2005. Does Idiosyncratic Risk Really Matter? Journal of Finance 60, 905-929

British Banks’ Association, 2006. Credit Derivatives Report, 2005/2006 Survey

Colquitt, L.L., Hoyt, R.E., 1997. Determinants of Corporate Hedging Behavior: Evidence from the Life Insurance Industry. Journal of Risk and Insurance 64, 649-671

Cumming, C.M., Hirtle, B.J., 2001. The Challenges of Risk Management in Diversified Financial Companies. FRBNY Economic Policy Review.

Cummins, J.D., Lewis, C., Wei, R., 2006. The Market Impact of Operational Risk Events for U.S. Banks and Insurers. Journal of Banking and Finance 30, 2605-2634.

Demarzo, P., Duffie, D., 1995. Corporate Incentives for Hedging and Hedge Accounting. Review of Financial Studies 8, 743-771.

Durnev, A., Morck, R., Yeung, B., Zarowin, P., 2003. Does Greater Firm-Specific Return Variation Mean More Or Less Informed Stock Pricing? Journal of Accounting Research 41, 797-836

Froot, K.A., Stein, J.C., 1998. Risk Management, Capital Budgeting and Capital Structure Policy for Financial Institutions: An Integrated Approach. Journal of Financial Economics 47, 55-82.

Fung, H., Sierra, G., Yau, J., Zhang, G., 2008, Are the U.S. stock market and credit default swap market related? Evidence from the CDX Indices, Journal of Alternative Investment, 2008, summer, 43–61.

Goldfried, M., 2003. CDS: An Insurance Industry Perspective. www.M-X.Ca/F_Publications_Fr/Ccad2003/Cred_Der.Pdf

Guay, W.A., 1999. The impact of derivatives on Þrm risk: An empirical examination of new

derivative users, Journal of Economics and Accounting 26, 319-351.

Hentschel, L., Kothari, S.P., 2001. Are Corporations Reducing Or Taking Risks With Derivatives? Journal of Financial and Quantitative Analysis 36, 93-118.

Hopman, J., 2007. Hedging With Credit Default Swaps: Proper Risk Management Critical. SVO Research Quarterly 6, 19-21.

Hoyt, R.E., Liebenberg, A.P., 2006. The Value of Enterprise Risk Management: Evidence from the U.S. Insurance Industry. Working Paper, University of Georgia.

Page 24: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

24

Instefjord, N., 2005. Risk and Hedging: Do Credit Derivatives Increase Bank Risk? Journal of Banking and Finance 29, 333-345.

Jin, Y., Jorion, P., 2006. Firm Value and Hedging: Evidence from U.S. Oil and Gas Producers. Journal of Finance 61, 893-919.

Morck, R., Schleifer, A., Vishny, R., 1988. Management Ownership and Market Valuation: An Empirical Analysis. Journal of Financial Economics 20, 293-315.

Morck, R., Yeung, B., Yu, W., 2000. The Information Content of Stock Markets: Why Do Emerging Markets Have Synchronous Stock Price Movements? Journal of Financial Economics 58, 215-260.

Morrison, A.D., 2005. Credit Derivatives, Disintermediation and Investment Decisions. Journal of Business, 78, 621-648.

Myers, S., 1977. Determinants of Corporate Borrowing. Journal of Financial Economics 5, 147-75.

Nelson, J.M., Mott, J.S., Aeck-Graves, J., 2005. The Impact of Hedging On Market Value of Equity. Journal of Corporate Finance 11, 851-881.

Nicol, A., Pelizzon, L., 2008. Credit Derivatives, Capital Requirements and Opaque OTC Markets. Journal of Financial Intermediation 17, 444-463.

Petersen, M. A., 2009, Estimating standard errors in finance panel data sets: comparing approaches. Review of Financial Studies 22 (1), 435-480.

Servaes, H., 1996. The Value of Diversification during the Conglomerate Merger Wave. Journal of Finance 51, 1201-25.

Shao, Y., Yeager, T., 2007. The Effects of Credit Derivatives on U.S. Bank Risk and Return, Capital and Lending Structure. Working Paper, University of Arkansas.

Smith, C., Stulz, R.E., 1985. The Determinants of Firms’ Hedging Policies. Journal of Financial and Quantitative Analysis 20, 391-405.

Smithson, C., Simkins, B.J., 2005. Does Risk Management Add Value? A Survey of the Evidence. Journal of Applied Corporate Finance 17, 8-17.

Stulz, R.E., 1984. Optimal Hedging Policies. Journal of Financial and Quantitative Analysis 19, 127-140.

Tufano, P., 1996. Who Manages Risk? An Empirical Examination of Risk Management Practices in the Gold Mining Industry. Journal of Finance 51, 1097-1137.

Yermack, D., 1996. Higher Market Valuation of Companies with a Small Board of Directors. Journal of Financial Economics 40, 185-211.

Page 25: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

25

Table 1 Summary of CDS Transactions for Life and PC Insurers

Panel A: Life Insurers

Total

Mean: based on Firm and Year Observations

Mean: based on Number of Transactions

CDS_Transaction (#) 4889 52.57

CDS_Transaction_Buy 1458 15.68 29.82%

CDS_Transaction_Sell 3431 36.89 70.18%

CDS_Notional Amt 74,837,471 804,704 15,307

CDS_Notional Amt_Buy 46,367,856 1,783,379 31,802

CDS_Notional Amt_Sell 28,469,615 424,920 8,298

Panel B: PC Insurers

PC Total

Mean: based on Firm and Year Observations

Mean: based on Number of Transactions

CDS_Transaction (#) 1639 42.03

CDS_Transaction_Buy 1113 28.54 67.91%

CDS_Transaction_Sell 526 13.49 32.09%

CDS_Notional Amt 39,127,542 1,003,270 23,873

CDS_Notional Amt_Buy 35,329,712 1,413,188 31,743

CDS_Notional Amt_Sell 3,797,830 271,274 7,220

Page 26: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

26

Table 2 Descriptive Statistics for the Entire Sample (CDS_users & Non_CDS Users)

Idiosyncratic risk (IR) is the standard deviation of residuals from a regression of monthly returns based on capital asset pricing model: rj,t – rf,t = αj,t + βj,t×(rm,t – rf,t) + εj,t; Market risk is defined as the coefficient, β, from CAPM; Total risk is defined as the total standard deviation of monthly returns over months the past five year period. Tobin’s Q is the market value of equity plus the book value of liabilities divided by the book value of assets,

i.e.)(

)()()(

assetstotalBV

equitycommonMVequitycommonBVassetstotalBV

TQ

, where MV (common equity) is the

product of stock price and number shares outstanding; ROA is defined as return on book value asset; Investment opportunity is defined as net capital expenses to assets; Leverage is defined as the ratio of long-term debt plus preferred stock over long-term debt plus preferred stock plus common equity (i.e. leverage to market value of assets); Firm size equals the natural logarithm of book value of assets; Div_yield defined as dividend yield, Debt Maturity is the proportion of short-term debt measured by the ratio of short-term debt with maturity less than three years to total debt; Investment opportunity is defined as net capital expenses to assets, NY_Dummy is defined as a dummy variable with value 1 if insurer operates in New York State. Rating is based on S&P issuer credit ratings to classify firms into investment and speculative grade to measure the credit risk and the unrated firms are included in the speculative grade, and Sale growth (sale_g) is the growth of sale.

Panel A: Life Insurers (N = 427 firm-year)

Panel B: PC Insurers (N = 666 firm-year)

Statistics min max mean median std Statistics min max mean median std

Idiosyncratic Risk 0.030 0.357 0.099 0.087 0.054 Idiosyncratic Risk 0.027 0.340 0.095 0.084 0.050

Market Risk 0.051 2.905 0.713 0.563 0.586 Market Risk -0.145 4.867 0.679 0.578 0.639

Total Risk 0.035 0.366 0.104 0.091 0.054 Total Risk 0.035 0.353 0.100 0.087 0.051

TQ 0.943 1.663 1.150 1.039 0.234 TQ 0.895 1.561 1.099 1.065 0.151

ROA -0.003 0.126 0.037 0.020 0.042 ROA -0.042 0.131 0.034 0.035 0.045

Invest_Opp 0.000 2.549 0.541 0.157 0.782 Leverage 0.000 5.932 0.672 0.344 0.972

Leverage 0.002 0.434 0.149 0.127 0.127 Firm Size 0.000 0.561 0.150 0.121 0.137

Firm Size 6.937 14.254 9.499 9.428 2.050 Dividend Yield 4.763 13.233 8.229 8.042 2.017

Dividend Yield 0.000 0.550 0.142 0.089 0.172 Debt Maturity 0.000 1.696 0.380 0.207 0.464

Debt Maturity 0.000 0.561 0.116 0.000 0.179 NY Dummy 0.000 1.000 0.112 0.000 0.241

NY Dummy 0.000 1.000 0.073 0.000 0.260 Rating 0.000 1.000 0.099 0.000 0.299

Rating 0.000 1.000 0.353 0.000 0.479 Sale Growth 0.000 1.000 0.324 0.000 0.468

Sale Growth 0.880 1.462 1.097 1.069 0.156 Stat 0.679 1.845 1.123 1.081 0.228

Page 27: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

27

Table 3: Univariate Comparison of Mean and Median between CDS users, Non-usres, CDS Net Buyers, and Net Sellers Idiosyncratic risk (IR) is the standard deviation of residuals from a regression of monthly returns based on capital asset pricing model: rj,t – rf,t = αj,t + βj,t×(rm,t – rf,t) + εj,t; Market risk is defined as the coefficient, β, from CAPM; Total risk is defined as the total standard deviation of monthly returns over months the past five year period. Tobin’s Q is defined as the market value of equity plus the book

value of liabilities divided by the book value of assets, i.e.,)(

)()()(

assetstotalBV

equitycommonMVequitycommonBVassetstotalBV

TQ

, where MV (common

equity) is the product of stock price and number shares outstanding; Investment opportunity is defined as net capital expenses to assets, Leverage is defined as the ratio of long-term debt plus preferred stock over long-term debt plus preferred stock plus common equity to total assets (i.e., leverage to market value of assets), Firm size is measured by the natural logarithm of the book value of assets. Div_yield defined as dividend yield; Debt_maturity is the portion of short-term debt measured by the ratio of short-term debt with maturity less than three years to total debt, NY_Dummy is defined as a dummy variable with value 1 if insurer i operated in New York State. Rating is based on S&P issuer credit ratings to classify firms into investment and speculative grade to measure the credit risk and the unrated firms are included in the speculative grade, and Sale growth is defined as the growth of sale. Mean difference test between CDS users and non-users is based on t-statistics and median difference test is based on Wilcoxon statistics. ***,** and* denote significance at the 1%, 5%, and 10% levels, respectively.

Page 28: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

28

Panel A: Life Insurers (N = 427)

(1) CDS User

(2) None User

(3) Net

Buyer

(4) Net

Seller (1) - (2) (3) - (2) (4)-(2)

(1) CDS User

(2) None User

(3) Net

Buyer

(4) Net

Seller (1) - (2) (3) - (2) (4)-(2)

Statistics mean mean mean mean mean mean mean median median median median median median median

Idiosyncratic Risk 0.077 0.105 0.091 0.072 -0.028 *** -0.014 *** -0.033 *** 0.062 0.091 0.068 0.060 -0.029 *** -0.022 *** -0.030 ***

Market Risk 0.861 0.671 0.935 0.859 0.190 *** 0.264 ** 0.188 *** 0.759 0.496 0.759 0.826 0.263 *** 0.263 ** 0.331 ***

Total Risk 0.085 0.110 0.100 0.080 -0.025 *** -0.010 *** -0.029 *** 0.072 0.096 0.076 0.069 -0.025 *** -0.020 *** -0.028 ***

TQ 1.033 1.183 1.018 1.042 -0.149 *** -0.165 *** -0.141 *** 1.021 1.071 1.010 1.030 -0.050 *** -0.061 *** -0.041 ***

ROA 0.013 0.043 0.015 0.014 -0.030 *** -0.028 *** -0.029 *** 0.012 0.027 0.012 0.013 -0.015 *** -0.015 *** -0.014 ***

Invest_Opp 0.235 0.626 0.370 0.209 -0.391 *** -0.257 ** -0.418 *** 0.000 0.237 0.000 0.013 -0.237 *** -0.237 ** -0.224 ***

Leverage 0.182 0.140 0.176 0.187 0.042 *** 0.036 ** 0.047 *** 0.173 0.114 0.165 0.179 0.060 *** 0.051 ** 0.065 ***

Firm Size 11.181 9.030 10.957 11.374 2.151 *** 1.927 *** 2.343 *** 11.588 8.856 11.041 11.663 2.732 *** 2.186 *** 2.807 ***

Dividend Yield 0.133 0.145 0.114 0.147 -0.012 * -0.031 0.002 ** 0.107 0.066 0.084 0.134 0.041 * 0.018 0.068 **

Debt Maturity 0.138 0.110 0.114 0.145 0.029 *** 0.005 *** 0.035 *** 0.078 0.000 0.062 0.079 0.078 *** 0.062 *** 0.079 ***

NY Dummy 0.129 0.057 0.217 0.109 0.072 *** 0.161 *** 0.052 ** 0.000 0.000 0.000 0.000 0.000 *** 0.000 *** 0.000 **

Rating 0.452 0.327 0.250 0.492 0.125 -0.077 0.164 ** 0.000 0.000 0.000 0.000 0.000 0.000 0.000 **

Sale Growth 1.045 1.111 1.028 1.055 -0.067 *** -0.084 *** -0.057 *** 1.038 1.081 1.038 1.044 -0.043 *** -0.043 *** -0.037 ***

Panel B: PC Insurers (N = 666)

Idiosyncratic Risk 0.086 0.096 0.080 0.108 -0.010 ** -0.016 * 0.012 0.076 0.085 0.076 0.101 -0.009 ** -0.009 * 0.016 Market Risk 1.374 0.636 1.657 0.989 0.738 *** 1.021 *** 0.353 *** 0.977 0.551 1.073 0.947 0.426 *** 0.521 *** 0.395 ***

Total Risk 0.101 0.100 0.098 0.118 0.001 -0.001 0.018 0.080 0.088 0.080 0.107 -0.008 -0.008 0.019 TQ 1.014 1.104 1.003 1.011 -0.090 *** -0.101 *** -0.093 ** 1.012 1.071 1.006 1.017 -0.060 *** -0.065 *** -0.055 **

ROA 0.016 0.035 0.023 0.000 -0.019 *** -0.011 ** -0.035 *** 0.012 0.037 0.014 0.008 -0.026 *** -0.024 ** -0.030 ***

Invest_Opp 0.923 0.657 1.300 0.263 0.266 0.643 -0.394 *** 0.176 0.345 0.555 0.018 -0.170 0.210 -0.327 ***

Leverage 0.217 0.146 0.210 0.195 0.071 *** 0.064 *** 0.049 *** 0.201 0.115 0.194 0.216 0.085 *** 0.079 *** 0.100 ***

Firm Size 10.897 8.063 10.231 11.391 2.834 *** 2.168 *** 3.328 *** 11.135 7.965 10.238 12.112 3.169 *** 2.273 *** 4.147 ***

Dividend Yield 0.135 0.395 0.149 0.109 -0.260 *** -0.246 ** -0.287 0.095 0.232 0.092 0.099 -0.137 *** -0.140 ** -0.133 Debt Maturity 0.101 0.112 0.068 0.091 -0.011 *** -0.045 -0.022 0.063 0.000 0.000 0.062 0.063 *** 0.000 0.062

NY Dummy 0.154 0.096 0.000 0.182 0.058 -0.096 * 0.086 *** 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Rating 0.459 0.316 0.273 0.600 0.144 ** -0.043 0.284 * 0.000 0.000 0.000 1.000 0.000 0.000 1.000 *

Sale Growth 1.036 1.128 1.014 1.067 -0.093 *** -0.114 *** -0.062 1.038 1.083 1.017 1.095 -0.045 *** -0.065 *** 0.012

Page 29: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

29

Table 4 Risk Models

The simultaneous equations model is described as equations (3) and (4): Riski,t= α1 + β1,j ×CDSi,t, + tii Z ,,1 + β1,2 ×Div_yieldi + ε1,i,t

(3)

CDSi,t= α2 + β2,1 ×Riski,t + tii Z ,,2 + β2,2 ×NY_Dummyi,t +ε2,i,t

(4)

where Risk denotes total risk, systematic risk, and idiosyncratic risk of an insurer. Total risk is defined as the total standard deviation of monthly returns over months for the past five-year period. Idiosyncratic risk (IR) is the standard deviation of residuals from a regression of monthly returns based on the capital asset pricing model: rj,t – rf,t = αj,t + βj,t×(rm,t – rf,t) + εj,t; Market risk is the coefficient, β, from CAPM. CDS represents three different dummy variables. First, CDS_Dummyi,t = 1 if insurer i participates in CDS transactions in year t and zero elsewhere; Second, Net_Buyeri,t = 1 if the aggregate notional amount of CDS buy position is greater than that of sell position for insurer i in year t and zero elsewhere. Third, Net_Selleri,t = 1 if the aggregate notional amount of CDS sell position is greater than that of buy position for insurer i in year t and zero elsewhere. Zi,t represents a set of firm characteristics variables defined as follows: CDS_Change is dummy variable with value 1 if an insurer changes its CDS position from Net Sellers (Buyers) at time t-1 to Net Buyers (Sellers) at time t or from CDS holding to zero holding. Spread_Vol is the volatility of CDS daily spread difference; CAP_EXP is defined as net capital expenses to assets, Leverage is defined as the ratio of long-term debt plus preferred stock over long-term debt plus preferred stock plus common equity to total assets (i.e., leverage to market value of assets), Firm size is measured by the natural logarithm of the book value of assets. Div_yield is defined as dividend yield. DT_Maturity is the portion of short-term debt measured by the ratio of short-term debt with maturity less than three years to total debt, Rating is based on S&P issuer credit ratings to classify firms into investment and speculative grade to measure the credit risk and the unrated firms are included in the speculative grade, and Sale growth is defined as the growth of sale. State_NY_Dummy is defined as a dummy variable with value 1 if insurer i operated in New York State.

Page 30: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

30

Panel A: Life Insurance Sample Panel A1: Total Risk Model

CDS Participation CDS Net Buyers CDS Net Sellers

Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

Risk Equation Intercept 0.268 6.57 *** Intercept 0.227 6.68 *** Intercept 0.312 5.52 ***

CDS_ Dummy 0.081 2.69 *** CDS_ Buyers -0.068 -1.04 CDS_ Sellers 0.148 2.64 ***

CDS_Change 0.016 1.15 CDS_Change 0.032 1.13 CDS_Change 0.007 0.31

Spread_Vol 0.005 1.12 Spread_Vol 0.004 1.08 Spread_Vol 0.004 0.75

CAP_EXP 0.015 4.04 *** CAP_EXP 0.012 3.62 *** CAP_EXP 0.017 3.95 ***

Leverage 0.122 4.05 *** Leverage 0.125 4.51 *** Leverage 0.133 3.53 ***

Firm Size -0.017 -4.36 *** Firm Size -0.006 -1.98 ** Firm Size -0.022 -3.77 ***

Div Yield -0.023 -2.17 ** Div Yield -0.073 -4.81 *** Div Yield -0.016 -1.70 *

DT_Maturity -0.001 -0.06 DT_Maturity 0.019 1.34 DT_Maturity -0.010 -0.46

Rating -0.020 -2.77 *** Rating -0.014 -2.60 *** Rating -0.025 -2.58 **

Sale Growth -0.020 -1.04 Sale Growth -0.057 -3.10 *** Sale Growth -0.025 -1.07

CDS Participation CDS Net Buyers CDS Net Sellers

Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

CDS Equation Intercept -2.131 -4.20 *** Intercept -0.513 -2.09 ** Intercept -1.754 -3.86 ***

Total_Risk 7.504 3.79 *** Total_Risk 1.844 1.87 * Total_Risk 5.398 3.18 ***

CDS_Change -0.114 -0.81 CDS_Change 0.228 1.23 CDS_Change -0.047 -0.31

Spread_Vol -0.038 -1.02 Spread_Vol -0.009 -0.46 Spread_Vol -0.022 -0.67

CAP_EXP -0.127 -2.78 *** CAP_EXP -0.010 -0.46 CAP_EXP -0.100 -2.81 ***

Leverage -0.958 -2.40 ** Leverage -0.321 -1.55 Leverage -0.749 -2.25 **

Firm Size 0.173 6.18 *** Firm Size 0.057 3.20 *** Firm Size 0.138 5.41 ***

DT_Maturity 0.058 0.36 DT_Maturity -0.057 -0.87 DT_Maturity 0.078 0.60

Rating 0.175 2.51 ** Rating -0.018 -0.62 Rating 0.148 2.44 **

State NY Dummy 0.118 1.50 State NY Dummy 0.283 3.04 *** State NY Dummy 0.053 0.83

Sale Growth -0.002 -0.01 Sale Growth -0.107 -1.39 Sale Growth 0.080 0.52

Page 31: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

31

Panel A2: Idiosyncratic Risk Model

CDS Participation CDS Net Buyers CDS Net Sellers

Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

Risk Equation Intercept 0.279 7.40 *** Intercept 0.246 7.58 *** Intercept 0.316 6.25 ***

CDS_ Dummy 0.056 1.78 * CDS_ Buyers -0.067 -1.00 CDS_ Sellers 0.114 2.26 **

CDS_Change 0.010 0.96 CDS_Change 0.031 1.11 CDS_Change 0.003 0.19

Spread_Vol 0.004 1.24 Spread_Vol 0.005 1.23 Spread_Vol 0.004 0.90

CAP_EXP 0.014 3.90 *** CAP_EXP 0.012 3.48 *** CAP_EXP 0.015 3.72 ***

Leverage 0.114 4.32 *** Leverage 0.119 4.43 *** Leverage 0.121 3.75 ***

Firm Size -0.017 -4.31 *** Firm Size -0.009 -2.82 *** Firm Size -0.021 -4.04 ***

Div Yield -0.034 -2.53 ** Div Yield -0.074 -5.00 *** Div Yield -0.028 -2.07 **

DT_Maturity 0.007 0.43 DT_Maturity 0.022 1.59 DT_Maturity -0.002 -0.09

Rating -0.018 -2.69 *** Rating -0.013 -2.50 ** Rating -0.022 -2.60 ***

Sale Growth -0.030 -1.68 * Sale Growth -0.057 -3.29 *** Sale Growth -0.032 -1.58

CDS Participation CDS Net Buyers CDS Net Sellers

Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

CDS Equation Intercept -2.133 -3.85 *** Intercept -0.491 -1.86 * Intercept -1.770 -3.63 ***

IR 7.033 3.34 *** IR 1.555 1.54 IR 5.143 2.91 ***

CDS_Change -0.061 -0.46 CDS_Change 0.247 1.33 CDS_Change -0.023 -0.16

Spread_Vol -0.035 -1.00 Spread_Vol -0.007 -0.33 Spread_Vol -0.021 -0.66

CAP_EXP -0.113 -2.44 ** CAP_EXP -0.004 -0.19 CAP_EXP -0.089 -2.44 **

Leverage -0.857 -2.22 ** Leverage -0.265 -1.29 Leverage -0.667 -2.08 **

Firm Size 0.181 5.79 *** Firm Size 0.057 2.89 *** Firm Size 0.146 5.22 ***

DT_Maturity 0.027 0.17 DT_Maturity -0.057 -0.84 DT_Maturity 0.057 0.44

Rating 0.156 2.29 ** Rating -0.022 -0.80 Rating 0.136 2.29 **

State NY Dummy 0.169 2.04 ** State NY Dummy 0.297 3.26 *** State NY Dummy 0.095 1.35

Sale Growth -0.022 -0.12 Sale Growth -0.109 -1.45 Sale Growth 0.056 0.37

Page 32: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

32

Panel A3: Market Risk Model

CDS Participation CDS Net Buyers CDS Net Sellers

Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

Risk Equation Intercept 1.317 2.01 ** Intercept 0.002 0.00 Intercept 2.651 2.15 **

CDS_ Dummy 2.586 8.72 *** CDS_ Buyers 4.116 1.93 * CDS_ Sellers 4.733 2.46 **

CDS_Change -0.048 -0.12 CDS_Change -0.862 -0.86 CDS_Change -0.063 -0.08

Spread_Vol -0.017 -0.24 Spread_Vol 0.015 0.18 Spread_Vol -0.022 -0.17

CAP_EXP 0.050 0.69 CAP_EXP 0.016 0.26 CAP_EXP 0.100 0.94

Leverage 1.106 1.79 * Leverage 0.982 2.00 ** Leverage 1.547 1.64

Firm Size -0.196 -3.57 *** Firm Size -0.034 -0.51 Firm Size -0.349 -2.10 **

Div Yield -0.027 -0.39 Div Yield 0.044 0.25 Div Yield 0.072 0.23

DT_Maturity -0.703 -2.02 ** DT_Maturity -0.303 -1.12 DT_Maturity -0.897 -1.53

Rating -0.299 -1.96 * Rating -0.102 -0.87 Rating -0.381 -1.33

Sale Growth 0.512 1.34 Sale Growth 0.441 1.25 Sale Growth 0.419 0.65

CDS Participation CDS Net Buyers CDS Net Sellers

Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

CDS Equation Intercept -0.506 -2.05 ** Intercept -0.008 -0.06 Intercept -0.567 -2.76 ***

Beta 0.376 10.60 *** Beta 0.222 3.99 *** Beta 0.212 2.61 ***

CDS_Change 0.027 0.18 CDS_Change 0.209 1.04 CDS_Change 0.013 0.08

Spread_Vol 0.007 0.24 Spread_Vol -0.004 -0.20 Spread_Vol 0.005 0.17

CAP_EXP -0.019 -0.71 CAP_EXP 0.001 0.08 CAP_EXP -0.020 -0.96

Leverage -0.420 -1.73 * Leverage -0.202 -1.54 Leverage -0.326 -1.42

Firm Size 0.077 4.19 *** Firm Size 0.008 0.77 Firm Size 0.074 4.45 ***

DT_Maturity 0.268 2.04 ** DT_Maturity 0.085 1.33 DT_Maturity 0.190 1.73 *

Rating 0.115 2.02 ** Rating 0.028 1.00 Rating 0.080 1.59

State NY Dummy 0.008 0.66 State NY Dummy -0.040 -0.53 State NY Dummy -0.005 -0.22

Sale Growth -0.203 -1.43 Sale Growth -0.100 -1.24 Sale Growth -0.088 -0.74

Page 33: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

33

Panel B: PC Insurance Sample

Panel B1: Total Risk Model

CDS Participation CDS Net Buyers CDS Net Sellers

Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

Risk Equation Intercept 0.394 5.14 *** Intercept 0.219 6.98 *** Intercept 0.410 5.38 ***

CDS_ Dummy 1.386 4.74 *** CDS_ Buyers 0.689 3.41 *** CDS_ Sellers 3.310 3.52 ***

CDS_Change -0.528 -1.68 * CDS_Change -0.288 -1.79 * CDS_Change -0.581 -0.83

Spread_Vol 0.022 1.15 Spread_Vol 0.014 1.79 * Spread_Vol 0.000 0.01

CAP_EXP -0.041 -1.87 * CAP_EXP -0.016 -1.58 CAP_EXP 0.007 0.70

Leverage -0.055 -0.59 Leverage 0.033 0.82 Leverage 0.094 1.57

Firm Size -0.047 -4.65 *** Firm Size -0.017 -4.41 *** Firm Size -0.040 -3.70 ***

Div Yield 0.029 1.64 Div Yield 0.000 -0.07 Div Yield 0.034 1.86 *

DT_Maturity 0.039 1.06 DT_Maturity 0.026 1.45 DT_Maturity -0.007 -0.22

Rating -0.005 -0.18 Rating -0.006 -0.47 Rating -0.041 -1.44

Sale Growth 0.043 1.15 Sale Growth 0.004 0.21 Sale Growth -0.025 -0.77

CDS Participation CDS Net Buyers CDS Net Sellers

Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

CDS Equation Intercept -0.292 -2.72 *** Intercept -0.251 -4.64 *** Intercept -0.136 -1.94 *

Total_Risk 0.827 2.05 ** Total_Risk 1.131 7.62 *** Total_Risk 0.430 1.68 *

CDS_Change 0.380 1.84 * CDS_Change 0.420 2.10 ** CDS_Change 0.172 0.83

Spread_Vol -0.017 -1.29 Spread_Vol -0.019 -1.71 * Spread_Vol -0.002 -0.25

CAP_EXP 0.030 2.04 ** CAP_EXP 0.023 1.80 * CAP_EXP -0.003 -0.83

Leverage 0.050 0.70 Leverage -0.022 -0.39 Leverage -0.030 -1.24

Firm Size 0.032 3.85 *** Firm Size 0.021 3.61 *** Firm Size 0.011 2.04 **

DT_Maturity -0.031 -1.08 DT_Maturity -0.038 -1.63 DT_Maturity 0.004 0.40

Rating -0.009 -0.42 Rating 0.004 0.22 Rating 0.008 0.86

State NY Dummy 0.001 0.06 State NY Dummy -0.002 -0.17 State NY Dummy 0.002 0.15

Sale Growth -0.025 -0.90 Sale Growth -0.011 -0.42 Sale Growth 0.013 1.26

Page 34: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

34

Panel B2: Idiosyncratic Risk Model

CDS Participation CDS Net Buyers CDS Net Sellers

Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

Risk Equation Intercept 0.392 5.21 *** Intercept 0.218 6.99 *** Intercept 0.411 5.46 ***

CDS_ Dummy 1.356 4.75 *** CDS_ Buyers 0.662 3.47 *** CDS_ Sellers 3.296 3.52 ***

CDS_Change -0.518 -1.67 * CDS_Change -0.281 -1.81 * CDS_Change -0.580 -0.83

Spread_Vol 0.023 1.22 Spread_Vol 0.014 1.94 * Spread_Vol 0.001 0.02

CAP_EXP -0.040 -1.89 * CAP_EXP -0.016 -1.61 CAP_EXP 0.007 0.66

Leverage -0.064 -0.69 Leverage 0.027 0.67 Leverage 0.089 1.52

Firm Size -0.048 -4.81 *** Firm Size -0.017 -4.67 *** Firm Size -0.041 -3.81 ***

Div Yield 0.029 1.69 * Div Yield -0.001 -0.16 Div Yield 0.034 1.88 *

DT_Maturity 0.040 1.10 DT_Maturity 0.026 1.47 DT_Maturity -0.006 -0.18

Rating -0.005 -0.18 Rating -0.005 -0.42 Rating -0.040 -1.42

Sale Growth 0.044 1.18 Sale Growth 0.005 0.27 Sale Growth -0.024 -0.76

CDS Participation CDS Net Buyers CDS Net Sellers

Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

CDS Equation Intercept -0.297 -2.49 ** Intercept -0.253 -4.79 *** Intercept -0.136 -1.83 *

IR 0.846 1.84 * IR 1.146 8.13 *** IR 0.429 1.58

CDS_Change 0.379 1.82 * CDS_Change 0.424 2.10 ** CDS_Change 0.173 0.83

Spread_Vol -0.019 -1.39 Spread_Vol -0.021 -1.80 * Spread_Vol -0.002 -0.26

CAP_EXP 0.030 2.07 ** CAP_EXP 0.023 1.79 * CAP_EXP -0.003 -0.77

Leverage 0.060 0.83 Leverage -0.016 -0.28 Leverage -0.028 -1.15

Firm Size 0.033 3.68 *** Firm Size 0.022 3.79 *** Firm Size 0.011 1.98 **

DT_Maturity -0.033 -1.13 DT_Maturity -0.038 -1.64 DT_Maturity 0.003 0.33

Rating -0.009 -0.42 Rating 0.003 0.20 Rating 0.007 0.80

State NY Dummy 0.001 0.06 State NY Dummy -0.003 -0.21 State NY Dummy 0.002 0.15

Sale Growth -0.027 -0.96 Sale Growth -0.013 -0.50 Sale Growth 0.013 1.22

Page 35: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

35

Panel B3: Market Risk Model

CDS Participation CDS Net Buyers CDS Net Sellers

Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

Risk Equation Intercept 0.644 1.69 * Intercept 0.459 2.13 ** Intercept -0.037 -0.10

CDS_ Dummy 1.041 0.55 CDS_ Buyers 3.616 3.15 *** CDS_ Sellers -5.452 -1.53

CDS_Change -0.188 -0.26 CDS_Change -1.279 -1.59 CDS_Change 0.645 0.45

Spread_Vol -0.038 -1.09 Spread_Vol -0.002 -0.04 Spread_Vol -0.029 -0.52

CAP_EXP 0.031 0.47 CAP_EXP -0.023 -0.38 CAP_EXP 0.009 0.27

Leverage 0.754 2.84 *** Leverage 0.859 3.82 *** Leverage 0.650 2.95 ***

Firm Size -0.024 -0.36 Firm Size -0.022 -0.88 Firm Size 0.068 1.40

Div Yield 0.029 0.18 Div Yield 0.034 0.59 Div Yield -0.072 -0.74

DT_Maturity -0.167 -1.82 * DT_Maturity -0.080 -0.72 DT_Maturity -0.158 -1.47

Rating -0.219 -2.55 ** Rating -0.165 -2.27 ** Rating -0.164 -1.80 *

Sale Growth -0.044 -0.25 Sale Growth 0.045 0.28 Sale Growth 0.005 0.03

CDS Participation CDS Net Buyers CDS Net Sellers

Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

CDS Equation Intercept -0.375 -1.79 * Intercept -0.051 -1.41 Intercept -0.032 -1.39

Beta 0.678 1.09 Beta 0.127 2.73 *** Beta -0.039 -0.93

CDS_Change 0.237 1.19 CDS_Change 0.345 1.79 * CDS_Change 0.037 0.16

Spread_Vol 0.019 0.46 Spread_Vol -0.011 -0.94 Spread_Vol -0.002 -0.24

CAP_EXP -0.021 -0.45 CAP_EXP 0.015 1.30 CAP_EXP 0.001 0.19

Leverage -0.560 -0.82 Leverage -0.067 -0.92 Leverage 0.032 0.91

Firm Size 0.010 0.47 Firm Size 0.007 1.34 Firm Size 0.007 2.12

DT_Maturity 0.107 0.92 DT_Maturity -0.003 -0.14 DT_Maturity -0.013 -1.03 **

Rating 0.129 1.01 Rating 0.007 0.34 Rating -0.022 -1.97

State NY Dummy 0.118 0.96 State NY Dummy -0.015 -0.81 State NY Dummy 0.028 2.79 **

Sale Growth 0.024 0.25 Sale Growth -0.028 -1.06 Sale Growth -0.001 -0.13 ***

Page 36: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

36

Table 5 : Regression Model on Firm Performance The regression model is: Performance

=

α0 + βj ×CDSi,t + tii Z , + εi,t

(5)

Three main variables are used to proxy for firm value/performance measure: Tobin’s Q, ratio of Market value of equity to book-value equity, MV(Eqty)/BV(Eqty), and return on asset (ROA). Tobin’s Q is defined as the market value of equity plus the book value of liabilities divided by the book value of assets,

i.e.,)(

)()()(

assetstotalBV

equitycommonMVequitycommonBVassetstotalBVTQ

, where MV (common equity) is the product of stock

price and number shares outstanding; )(

)()(_)(

equitycommonBV

equitycommonMVEqtyBVEqtyMV ; ROA is return on book value asset.

CDS represents those CDS-related variables, namely, CDS is the participation dummy (CDS_Dummy) and net participation positions (Net_Seller and Net_Buyer). CDS_Dummyi,t = 1 if insurer i participate in CDS transactions in year t and zero elsewhere; Net_Buyeri,t = 1 if the aggregate notional amount of CDS buy position is greater than that of sell position for insurer i in year t and zero elsewhere; Net_Selleri,t = 1 if the aggregate notional amount of CDS sell position is greater than that of buy position for insurer i in year t and zero elsewhere. Zi,t represents a set of firm characteristics variables defined as follows: CDS_Change is dummy variable with value 1 if an insurer changes CDS position from Net Sellers (Buyers) at time t-1 to Net Buyers (Sellers) at time t or changes participation position from CDS holding to no holding. Spread_Vol is the volatility of CDS daily spread difference; CAP_EXP is defined as net capital expenses to assets to measure investment opportunity; Leverage is defined as the ratio of long-term debt plus preferred stock over long-term debt plus preferred stock plus common equity to total assets (i.e., leverage to market value of assets), Firm size is measured by the natural logarithm of the book value of assets. Div_yield defined as dividend yield, DT_Maturity is the portion of short-term debt measured by the ratio of short-term debt with maturity less than three years to total debt, Rating is based on S&P issuer credit ratings to classify firms into investment and speculative grade to measure the credit risk and the unrated firms are included in the speculative grade, and Sale growth is defined as the growth of sale.

Page 37: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

37

Panel A: Life Insurance Sample Panel A1: Tobin's Q Model

CDS Participation CDS Net Buyers CDS Net Sellers

Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

Tobin's Q Intercept 1.070 7.43 *** Intercept 1.065 6.75 *** Intercept 1.072 7.24 ***

CDS_ Dummy -0.059 -1.99 * CDS_ Buyers -0.078 -2.01 ** CDS_ Sellers -0.043 -1.45 CDS_Change -0.021 -0.83 CDS_Change -0.012 -0.41 CDS_Change -0.004 -0.15

Spread_Vol -0.035 -1.72 * Spread_Vol -0.041 -1.85 * Spread_Vol -0.039 -1.82 *

CAP_EXP 0.129 5.18 *** CAP_EXP 0.128 4.87 *** CAP_EXP 0.133 5.14 ***

Leverage -0.577 -5.35 *** Leverage -0.569 -4.74 *** Leverage -0.584 -5.32 ***

Firm Size -0.002 -0.26 Firm Size -0.004 -0.40 Firm Size -0.003 -0.37

Div Yield -0.254 -2.91 *** Div Yield -0.265 -2.90 *** Div Yield -0.245 -2.80 ***

DT_Maturity -0.110 -1.43 DT_Maturity -0.123 -1.45 DT_Maturity -0.111 -1.39

Rating 0.076 2.08 ** Rating 0.087 1.95 * Rating 0.077 2.05 **

Sale Growth 0.202 2.19 ** Sale Growth 0.225 2.23 ** Sale Growth 0.208 2.23 **

Panel A2: Market to Book Equity Value Model

CDS Participation CDS Net Buyers CDS Net Sellers

Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

MV_BV Intercept 1.272 1.29 Intercept 1.239 1.15 Intercept 1.289 1.27 CDS_ Dummy -0.184 -1.02 CDS_ Buyers -0.453 -1.86 * CDS_ Sellers -0.039 -0.21 CDS_Change -0.110 -0.75 CDS_Change 0.017 0.10 CDS_Change -0.017 -0.11

Spread_Vol -0.325 -3.00 *** Spread_Vol -0.328 -2.77 *** Spread_Vol -0.321 -2.87 ***

CAP_EXP 0.667 3.49 *** CAP_EXP 0.671 3.36 *** CAP_EXP 0.698 3.50 ***

Leverage -3.098 -5.42 *** Leverage -2.926 -4.73 *** Leverage -3.101 -5.29 ***

Firm Size 0.090 1.62 Firm Size 0.079 1.32 Firm Size 0.080 1.36

Div Yield -1.108 -2.71 *** Div Yield -1.122 -2.64 ** Div Yield -1.040 -2.57 **

DT_Maturity -0.406 -1.02 DT_Maturity -0.503 -1.14 DT_Maturity -0.393 -0.94

Rating 0.407 2.10 ** Rating 0.482 2.05 ** Rating 0.409 2.05 **

Sale Growth 0.483 0.99 Sale Growth 0.560 1.05 Sale Growth 0.491 0.98

Page 38: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

38

Panel A3: ROA Model

CDS Participation CDS Net Buyers CDS Net Sellers

Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

ROA Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

Intercept 0.033 1.10 Intercept 0.031 0.95 Intercept 0.033 1.05

CDS_ Dummy -0.013 -2.22 ** CDS_ Buyers -0.012 -1.75 * CDS_ Sellers -0.010 -1.66 CDS_Change -0.001 -0.28 CDS_Change -0.002 -0.42 CDS_Change 0.001 0.28 Spread_Vol 0.006 1.44 Spread_Vol 0.006 1.41 Spread_Vol 0.005 1.22

CAP_EXP 0.024 5.76 *** CAP_EXP 0.024 5.48 *** CAP_EXP 0.025 5.72 ***

Leverage -0.058 -3.19 *** Leverage -0.056 -2.84 *** Leverage -0.059 -3.21 ***

Firm Size -0.003 -1.46 Firm Size -0.003 -1.55 Firm Size -0.003 -1.57

Div Yield -0.030 -1.84 * Div Yield -0.033 -2.01 ** Div Yield -0.027 -1.67 * DT_Maturity -0.001 -0.05 DT_Maturity -0.002 -0.13 DT_Maturity 0.000 -0.03

Rating 0.012 1.81 * Rating 0.014 1.77 * Rating 0.012 1.81 *

Sale Growth 0.012 0.65 Sale Growth 0.015 0.75 Sale Growth 0.014 0.75

Page 39: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

39

Panel B: PC Insurance Sample

Panel B1: Tobin's Q Model

CDS Participation CDS Net Buyers CDS Net Sellers

Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

Tobin's Q Intercept 0.946 16.23 *** Intercept 0.947 16.21 *** Intercept 0.949 16.10 ***

CDS_ Dummy -0.096 -4.29 *** CDS_ Buyers -0.093 -4.67 *** CDS_ Sellers -0.099 -3.19 ***

CDS_Change -0.065 -3.24 *** CDS_Change -0.059 -2.43 ** CDS_Change -0.056 -2.11 **

Spread_Vol -0.044 -4.07 *** Spread_Vol -0.042 -3.79 *** Spread_Vol -0.043 -3.93 ***

CAP_EXP 0.016 1.77 * CAP_EXP 0.015 1.72 * CAP_EXP 0.015 1.52

Leverage -0.229 -2.89 *** Leverage -0.233 -2.90 *** Leverage -0.236 -2.87 ***

Firm Size 0.010 1.78 * Firm Size 0.009 1.67 * Firm Size 0.009 1.68 *

Div Yield 0.015 0.52 Div Yield 0.016 0.56 Div Yield 0.016 0.55 DT_Maturity -0.042 -1.53 DT_Maturity -0.043 -1.55 DT_Maturity -0.044 -1.58

Rating 0.072 2.95 *** Rating 0.072 2.85 *** Rating 0.072 2.81 ***

Sale Growth 0.121 3.73 *** Sale Growth 0.121 3.71 *** Sale Growth 0.123 3.74 ***

Panel B2: Market to Book Equity Value Model

CDS Participation CDS Net Buyers CDS Net Sellers

Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

MV_BV Intercept 0.436 2.12 ** Intercept 0.440 2.20 ** Intercept 0.442 2.18 **

CDS_ Dummy -0.330 -3.32 *** CDS_ Buyers -0.396 -4.24 *** CDS_ Sellers -0.352 -2.79 ***

CDS_Change -0.222 -2.19 ** CDS_Change -0.168 -1.44 CDS_Change -0.250 -2.17 **

Spread_Vol -0.272 -5.83 *** Spread_Vol -0.264 -5.63 *** Spread_Vol -0.267 -5.80 ***

CAP_EXP 0.030 1.03 CAP_EXP 0.033 1.10 CAP_EXP 0.029 0.85 Leverage -0.682 -1.53 Leverage -0.689 -1.54 Leverage -0.697 -1.51

Firm Size 0.077 3.22 *** Firm Size 0.075 3.25 *** Firm Size 0.076 3.25 ***

Div Yield 0.011 0.11 Div Yield 0.013 0.13 Div Yield 0.012 0.12 DT_Maturity -0.087 -0.80 DT_Maturity -0.097 -0.91 DT_Maturity -0.100 -0.92

Rating 0.257 2.66 *** Rating 0.251 2.57 ** Rating 0.250 2.51 **

Sale Growth 0.626 4.06 *** Sale Growth 0.618 4.01 *** Sale Growth 0.622 4.02 ***

Page 40: Will the Use of Credit Default Swaps Affect Risks and Firm ...conference/conference2010/proceedings...1 Will the Use of Credit Default Swaps Affect Risks and Firm Value? Evidence from

40

Panel B3: ROA Model

CDS Participation CDS Net Buyers CDS Net Sellers

Parameter Coef. tValue Parameter Coef. tValue Parameter Coef. tValue

ROA Intercept -0.025 -1.22 Intercept -0.025 -1.21 Intercept -0.024 -1.15 CDS_ Dummy -0.010 -1.36 CDS_ Buyers -0.005 -0.71 CDS_ Sellers -0.014 -1.57

CDS_Change -0.018 -2.35 ** CDS_Change -0.021 -2.24 ** CDS_Change -0.015 -1.29

Spread_Vol 0.009 3.00 *** Spread_Vol 0.009 2.98 *** Spread_Vol 0.008 2.92 ***

CAP_EXP 0.007 3.09 *** CAP_EXP 0.007 3.01 *** CAP_EXP 0.006 2.40 **

Leverage -0.034 -1.80 * Leverage -0.035 -1.83 * Leverage -0.035 -1.81 *

Firm Size -0.002 -1.11 Firm Size -0.002 -1.19 Firm Size -0.002 -1.15 Div Yield 0.008 1.19 Div Yield 0.008 1.21 Div Yield 0.008 1.13 DT_Maturity 0.011 0.81 DT_Maturity 0.011 0.82 DT_Maturity 0.012 0.88

Rating 0.017 2.46 ** Rating 0.017 2.41 ** Rating 0.018 2.50 **

Sale Growth 0.028 2.89 *** Sale Growth 0.028 2.89 *** Sale Growth 0.028 2.88 ***