« You Lost Your Job at 40 – Now What? | Main | Video: Cultural Differences between the US and China in Business »

03/15/2012

Does the Tail Wag the Dog? The Effect of Credit Default Swaps on Credit Risk


Click to Print This Page

ABSTRACT

Concerns have been raised, especially since the global financial crisis, about whether trading in credit default swaps (CDS) increases the credit risk of the reference entities. We use a unique, comprehensive sample covering 901 CDS introductions on North American corporate issuers between June 1997 and April 2009 to address this question. We present evidence that the probability of credit rating downgrade and the probability of bankruptcy both increase after the inception of CDS trading. The effect is robust to controlling for the endogeneity of CDS introduction, i.e., the possibility that firms with upcoming deterioration in creditworthiness are more likely to be selected for CDS trading. We show that the CDS-protected lenders’ reluctance to restructure is the most likely cause of the increase in credit risk. We present evidence that firms with relatively larger amounts of CDS contracts outstanding, and those with more “No Restructuring” contracts, are more likely to be adversely affected by CDS trading. We also document that CDS trading increases the level of participation of bank lenders to the firm. Our findings are broadly consistent with the predictions of the “empty creditor” model of Bolton and Oehmke (2011).

I. INTRODUCTION

Credit default swaps (CDS), which are insurance-type contracts that offer investors protection against default by a debtor, are arguably the most controversial financial innovation of the past two decades. They were praised by some market observers such as former Federal Reserve Chairman Alan Greenspan, who argued that “these increasingly complex financial instruments have contributed, especially over the recent stressful period, to the development of a far more flexible, efficient, and hence resilient financial system than existed just a quarter-century ago.”1 However, they also came in for strong criticism from several well-known market practitioners, particularly after the global financial crisis, which had its origins in July 2007. Warren Buffett, the much acclaimed investor, weighed against derivatives in general by describing them as “time bombs, for the parties that deal in them and the economic system” and went to conclude that “in my view, derivatives are financial weapons of mass destruction, carrying dangers that, while now latent, are potentially lethal.”2 In a similar vein, George Soros, a legendary hedge fund investor argued that “CDS are toxic instruments whose use ought to be strictly regulated.”3 Which of these conclusions is valid? Although one can debate this question based on theoretical arguments, the issue can only be resolved by empirical testing in specific contexts with clearly stated hypothesis that can be refuted. Our purpose in this paper is to present a careful empirical examination along these lines.

Despite the concerns expressed by regulators as well as market participants, the CDS market grew by leaps and bounds from about $0.9 trillion at the end of 2001 to a high of about $62 trillion at the end of 2007, measured by notional amount outstanding, next only to interest rate derivatives. Although the CDS market shrank considerably during the global financial crisis, it nevertheless stood at $26 trillion by December 2010. Indeed, during this period, CDS trading was introduced in countries including China and India. At the same time, CDS played a prominent role during the credit crisis of 2007-2008 and the European sovereign debt crisis of 2010-2011. In particular, the bankruptcy of Lehman Brothers and the collapse of Bear Stearns and AIG were closely related to CDS trading. In spite of misgivings about the role of CDS in potentially destabilizing markets, their role as indicators of credit quality has, in fact, expanded. CDS spreads are widely quoted by market practitioners as well as regulators, who have built them into their assessment of credit risks at both the level of each corporate debtor, as well as the aggregate level of a sector and the overall sovereign risk of a country.

Many of the issues mentioned in the context of derivatives, in general, have also been raised in the specific case of CDS. The generic arguments about the deleterious effects of derivatives, as a group, rely on market mechanisms such as the possibility of market manipulation, accounting fraud, pressure on posting collateral and their liquidity consequences, and the credit risk of counter-parties. These arguments challenge the hitherto accepted notion that derivatives are redundant securities, as assumed in most pricing and hedging models, and hence have no effect, adverse or otherwise, on the price of the underlying asset or the integrity of markets.

Apart from the above concerns that apply to all derivatives, in principle, CDS contracts are somewhat different from many other derivatives for one important reason: Buyers of CDS protection can influence the financial decisions of the reference entities, such as firms, and, indirectly, the credit risk of the claims they issue. This possibility is contrary to the “redundancy” assumption in structural models of credit risk along the lines of Merton (1974), that default risk is principally driven by leverage and asset volatility. In the spirit of that framework, CDS are regarded as “side-bets” on the value of the firm, and hence, have no effect on the credit risk associated with the individual claims issued by the firm. In particular, in such models, CDS trading does not affect the probability of bankruptcy, or indeed even the possibility of a credit downgrade.

In contrast to the redundancy argument, illustrative evidence from corporate restructuring and bankruptcy suggests that CDS positions play an important role in the case of distressed firms, especially just prior to bankruptcy. To cite one such instance, CIT Group attempted to work out its debt to avoid bankruptcy from late 2008 to mid-2009. In the event, however, some debt-holders, including Goldman Sachs (which had also bought CDS protection on the firm) rejected the firm’s exchange offer.4 CIT Group eventually filed for Chapter 11 bankruptcy on November 1, 2009.5 Hu and Black (2008) call such debt-holders whose exposures are insured with CDS “empty creditors,” meaning that they are creditors with an economic interest in the firm’s claims, but no risk alignment with the other bondholders who do not enjoy such protection.6 Along the same lines, in an op-ed piece, Henry Hu, one of the coauthors of Hu and Black (2008), named Goldman Sachs as AIG’s empty creditor shortly before becoming the director of the SEC’s Division of Risk, Strategy, and Financial Innovation.7 In a similar vein, when Delphi Corporation filed for bankruptcy on October 8, 2005, the total amount of CDS contracts outstanding was roughly 30 times the face value of the bonds outstanding and led to a “squeeze,” when the default event called for physical delivery of the bonds. It is highly likely that many creditors had become empty creditors.8

The empty creditor concern highlighted by Hu and Black (2008) is formally modeled by Bolton and Oehmke (2011).9 Their model predicts that bondholders will usually choose to “over-insure” their credit exposure by buying CDS protection, and thus, becoming empty creditors. Consequently, they have different economic interests from other bondholders and are less willing to negotiate to restructure the debt when the firm is under stress, and are even willing to push the firm into bankruptcy, since their total payoffs may be larger in that event. A similar argument applies to events that are less extreme and more common than default such as a rating downgrade: Credit rating agencies may anticipate the potential increase in credit risk and take such action. Often, a rating downgrade is the first stage of credit deterioration towards eventual default.

An alternative channel through which creditor behavior in the presence of CDS protection may adversely affect the credit risk of a firm is through the reduction of monitoring activity by lenders who are empty creditors. Such creditors may have diminished incentives to expend resources to monitor the performance of the firm; this, in turn, may lead to lower information quality, higher risk-taking, and higher bankruptcy incidence.10 It is important to distinguish between these two channels of increased credit risk emanating from the empty creditor phenomenon. The issue has great relevance to the current regulatory debate regarding CDS contracts and the relative importance of the two channels merits scrutiny. There are several obvious differences between the two channels. First, the increase in bankruptcy risk through the “restructuring” channel is positively related to the amount of CDS outstanding, but not necessarily through the “monitoring” channel. In the former case, the greater the amount of CDS outstanding, the greater is the potential for the standoff regarding restructuring, whereas in the latter case, zero monitoring is the worst possible scenario and lenders cannot do any damage below that level. Second, the “restructuring” effect of CDS trading on bankruptcy risk is expected to be more severe for CDS that exclude restructuring as credit event. This prediction is unique to the restructuring channel.11

An additional issue of interest is the ex-ante behavior of lenders to a firm, especially banks. On the one hand, the existence of CDS contracts may render a bank more willing to lend, due to the possibility of risk mitigation and enhanced bargaining power through the CDS contract. On the other hand, banks with relationships to the firm may have long-run reputation concerns about becoming empty creditors in a dynamic setting. Further, the greater the number of lenders, the more severe the problems of coordination in a stressed situation, when a workout may be necessary. To study this issue further, we explore how CDS trading affects lending relationships, and, in particular, the number of lenders, after the introduction of CDS trading. If bankruptcy risk increases with the number of lenders, this is an indirect channel for CDS trading to affect bankruptcy risk. This is also consistent with the empty creditor hypothesis as lenders tend to think they are non-pivotal in a multiple-lender structure.


Free NYU Webinar


We test our hypotheses using a comprehensive data set on CDS trading since the inception of CDS market for corporate names in 1997. It should be emphasized that it is difficult to retrieve accurate data on the introduction of CDS from a single source, since CDS trading does not take place on centralized exchanges. Indeed, even the central clearing of CDS is a relatively recent phenomenon. Our identification of the launch date relies, of necessity, on multiple data sources including GFI Inc., the largest global interdealer broker with the most extensive records of CDS trades and quotes, CreditTrade, a major intermediary especially in the early stages of the CDS market, and Markit, a data disseminator and vendor, which provides daily quotes from major institutions. Our combined data set covers 901 CDS introductions from 1997 to 2009 for North American names. The list of bankruptcies for North American firms is comprehensively constructed from major data sources such as New Generation Research, the UCLA-LoPucki Bankruptcy Database, Moody’s Annual Reports on Bankruptcy and Recovery, and the Altman-NYU Salomon Center Bankruptcy List. Over the same time period, we record bankruptcy filings by 1,628 firms, of which 60 had CDS trading prior to bankruptcies. Since bankruptcy is a relatively rare event for firms, we also investigate data on credit rating downgrades, of which we find 3,863 rating downgrades in our data set. The data on credit ratings are from S&P.

Our main empirical challenge is the potential endogeneity of CDS trading due to the possibility that firms with greater future credit risk deterioration are selected for CDS trading. In other words, there could be unobserved omitted variables that drive both selection of firms for CDS trading and bankruptcy risk. We address this concern in several ways. First, we select firms by their distance-to-default from the Merton (1974) structural model and match firms with and without CDS traded on them, to examine the effect of CDS trading on this matched sample. This partially controls for the credit risk prior to CDS trading. Second, we construct a model to predict CDS trading for individual firms. This model allows us to undertake a difference-in-difference comparison and a propensity score matching analysis for firms with and without CDS trading. Third, we use the two-stage Heckman correction for the selection of firms with CDS traded. In the first stage, we run a probit model for CDS trading. In the second stage, we estimate the probability of bankruptcy subject to the likelihood of CDS trading from the first stage.

We find that the introduction of CDS on a firm increases the likelihood of both credit downgrades and bankruptcy, after controlling for variables suggested by structural models. The effect of CDS trading is both statistically significant and economically large. For our sample firms, the credit rating declines by about half a notch, on average, in the next two years after the introduction of CDS trading. In a similar vein, the probability of bankruptcy more than doubles, from 0.14% to 0.33%, once the CDS starts trading. The positive relationship between CDS trading and bankruptcy risk is significant after controlling for the propensity of CDS trading. The Heckman correction results show that the effect of CDS introduction on bankruptcy risk is robust to the selection of a firm for CDS trading. Moreover, we find that the effect of CDS trading goes beyond the influence of the rating downgrade itself.

We also distinguish empirically between the different channels through which the empty creditor phenomenon manifests itself. Specifically, our analysis separates the restructuring channel from the monitoring channel. We document that the effects of CDS trading are stronger when the the number of outstanding CDS contracts is larger, and when the CDS contract has a “No Restructuring” credit event clause. We stress that these results are less subject to the endogeneity concern. In sum, rather than insuring against borrower default, CDS can actually indirectly cause borrower default. This “tail wagging dog” effect of CDS trading is important to take into account in policy discussions of the effect of CDS trading.

The remainder of this paper is organized as follows. The next section discusses related studies in the literature and places our research in context. Section III presents the motivation for our hypotheses and states them explicitly. The construction of our data set and our empirical methods are discussed in Section IV. Section V examines closely the selection of firms for CDS trading, and incorporates this issue explicitly into our analysis of the likelihood of rating downgrades and bankruptcy filing. Section VI explores further the empty creditor problem through the restructuring and monitoring channels by which CDS trading affects bankruptcy risk. Section VII concludes.

II. LITERATURE REVIEW

Our study is related to three different strands of the literature. The first analyzes the implications of CDS trading, and more broadly, the introduction of credit risk transfer mechanisms for creditors and the firms themselves. The second related literature is on the wide array of models of bankruptcy prediction. The third examines the effects of CDS trading on the relationship between creditors and firms, including the role of monitoring and information asymmetry. Longstaff, Mithal, and Neis (2005) provide an excellent introduction to the CDS contract and market.

A. CDS Trading and Credit Risk Transfer

Duffee and Zhou (2001) provide an early discussion of the benefits of CDS contracts as risk transfer tools, but also express caution on the potential downside of CDS trading for firms. They model the impact of introduction of CDS contracts from the perspective of creditors, particularly banks. The banks’ information advantage regarding borrower credit quality can cause both adverse selection and moral hazard concerns. In particular, CDS trading may reduce other types of risk-sharing, such as secondary loan sales, with ambiguous welfare consequences. Morrison (2005) argues that CDS can cause disintermediation as banks may not have incentives to monitor borrowers as closely, once their exposures are hedged with CDS.12 Allen and Carletti (2006) show that credit risk transfer can be beneficial when banks face systematic demand for liquidity. However, when they face idiosyncratic liquidity risk and hedge this risk in the inter-bank market, credit risk transfer can be detrimental to welfare. Further, such hedging via CDS may lead to contagion between the banking and the real sectors and increase the risk of financial crises.

Several papers have investigated the impact of loan sales, an alternative tool for credit risk transfer, on the creditor’s monitoring incentive. Gorton and Pennachhi (1995) focus on the moral hazard problem after loan sales. They conclude that banks can overcome the moral hazard problem by continuing to hold a fraction of the loan, and hence, have “skin in the game.” Parlour and Plantin (2008) emphasize the impact of a liquid loan sale market on bank’s ex ante incentive to monitor the debtor firm. They provide conditions under which a liquid credit risk transfer market can be socially inefficient. Parlour and Winton (2011) focus on a bank’s decision to lay off credit risk through loan sales versus CDS protection. They explicitly present efficiency implications in terms of risk transfer and monitoring, and suggest that, overall, CDS as a risk transfer mechanism are more likely to undermine monitoring. Beyhaghi and Massoud (2011) find that banks are more likely to hedge with CDS when monitoring costs are high.

Notwithstanding the insightful theoretical work cited above, there is a lack of direct empirical evidence as to what extent CDS trading affects bankruptcy risk through the creditor’s monitoring incentives and related channels. The only existing related evidence is somewhat indirect. Ashcraft and Santos (2009) document that CDS trading does not significantly benefit firms in terms of their cost of debt, except for safe and transparent firms. Hirtle (2009) shows that CDS trading increases bank credit supply and improves credit terms for large loans. Nadauld and Weisbach (2011) find that securitized loans have lower spreads but Bord and Santos (2011) find such loans underperform non-securitized loans. Purnanandam (2011) discusses how the originate-to-distribute model reduces loan quality and increases bankruptcy risk. Arentsen, Mauer, Rosenlund, Zhang, and Zhao (2012) find supporting evidence using CDS coverage data on subprime mortgage-backed securities (MBS). Das, Kalimipalli, and Nayak (2011) find that CDS trading hurts bond market quality. After the inception of CDS trading, there is greater pricing error and lower liquidity in the bond market. Boehmer, Chava, and Tookes (2010) document a negative impact of CDS trading on equity liquidity and prices. However, Saretto and Tookes (2011) show that CDS trading affects the corporate capital structure: Firms with CDS traded on them are able to maintain greater leverage and borrow at longer maturities.

There are several recent papers discussing the CDS-bond basis. For example, Bai and Collin-Dufresne (2010) investigate cross-sectional variation in the CDS-bond basis during the crisis period. Many other works, such as Tang and Yan (2011), focus on the determinants of the CDS spread. Giglio (2011) and Huang, Zhou, and Zhu (2009) measure the impact of systemic risk based on information contained in CDS spreads. Based on the recently developed “latent liquidity” measure for corporate bonds, Nashikkar, Subrahmanyam, and Mahanti (2011) find a liquidity spillover effect from the CDS market to the corporate bond market. They also provide empirical evidence on the impact of the limits to arbitrage on the pricing of credit risk and the CDS-bond basis.

CDS spreads can sometimes be misleading and excessively high, sending out false signals about firm performance, and thus accentuating the stress faced by the firm, and buttressing the need for additional capital.13 Stanton and Wallace (2011) find that the price levels for the AAA ABX.HE index CDS (a CDS contract based on asset-backed securities with a AAA credit rating) in 2009 are inconsistent with any reasonable forecast of the future default performance of the underlying loans. Moreover, changes in the CDS spreads are only weakly related to the credit performance of the underlying loans. Their finding casts serious doubts on the practice of using the CDS for marking-to-market purposes. However, the excessively high CDS spreads are conceivably driven by the strong demand and the limited supply of credit protection, without regard to the underlying risk itself. In such cases, if the buyers’ demand is not satisfied, the CDS price spike could have feedback effects on firm value. Indeed, Hortacsu, Matvos, Syverson, and Venkataraman (2011) find that increases in GM’s CDS spreads result in a drop in the resale prices of its cars at auctions.

At a more general level, there is evidence from the equity market that derivatives trading can affect the pricing of the underlying asset.14 However, the general conclusions drawn from the equity derivatives market may not be applicable to CDS and the underlying credit risk due to several major differences between the two types of instruments. First, CDS traders can directly influence firm operations, if they are also bond holders, especially when the firm is stressed.15 Second, the payoff from a CDS is linked to a specific corporate event (default), while that of equity options is related to the level of stock prices. Further, bankruptcy is an irreversible event, that can occur as a “jump to default” unlike the continuous movement of stock prices. Third, CDS are also traded by credit institutions that may have other devices to attenuate the impact on bankruptcy risk, such as by bailing out the stressed firm with additional junior debt. Lastly, CDS contracts are traded over-the-counter, where price transparency and discovery are less clear-cut than exchange traded markets where most equity derivatives are traded.

B. Bankruptcy Risk

The literature on bankruptcy prediction, which can be dated back to the Z-score model of Altman (1968),16 is too vast to be surveyed here. Bharath and Shumway (2008) and Campbell, Hilscher, and Szilagyi (2008) discuss the merits of simple bankruptcy prediction models over their more complicated counter-parts. On the other hand, Longstaff, Giesecke, Schaefer, and Strebulaev (2011) argue that factors suggested by structural models such as volatility and leverage predict bankruptcy better than other firm variables. Chava, Stefanescu, and Turnbull (2011) argue that the specification of the default model has a major impact on the predicted loss distribution. The literature suggests that the merits of using a large number of independent variables in bankruptcy prediction models are debatable. Hence, we extend Bharath and Shumway (2008) and use a same simple hazard specification, in the spirit of structural models, throughout our analysis.

Another aspect of the bankruptcy problem has received extensive attention in the literature is the coordination problem between creditors that increases the likelihood of bankruptcy. In an early paper, Gilson, John, and Lang (1990) show that creditor coordination failure increases bankruptcies. More recently, Brunner and Krahnen (2008) show that distress workout is less successful when there are more creditors.

C. CDS and the Lending Relationship

Several papers examine the effect of CDS trading on the incentives and behavior of lenders to firms, in general, and banks, in particular. Acharya and Johnson (2007, 2010) demonstrate evidence of insider trading activity in the CDS market. Further, they show that the intensity of insider trading is related to the number of lenders.17 Their evidence indicates that creditors often choose to become empty creditors and engage in insider trading in the CDS market. However, Minton, Stulz, and Williamson (2009) find that bank use of CDS is limited, possibly due to the lack of liquidity in CDS contracts. Moreover, Hilscher, Pollet, and Wilson (2011) provide evidence that equity returns lead returns from credit protection at daily and weekly frequencies, casting doubt on the generality of insider trading in the CDS market.

CDS could affect bankruptcy risk through two channels associated with the empty creditor problem. The first and direct channel is the effect on the willingness to restructure the debt, whereby creditors (over)insured with CDS break the link between cash flow rights and control rights. Empty creditors are unwilling to restructure the firm even if doing so is efficient for debt value as they can profit significantly from their CDS positions. Several theoretical papers model the empty creditor issue. Bolton and Oehmke (2011) emphasize the ex-ante commitment benefit of CDS trading, which relaxes the borrower’s debt constraint and decreases the probability of strategic default. However, the optimal level of CDS protection depends on the tradeoff between the ex-ante commitment benefit and the resulting intransigent overinsured creditors, who may push the firm into an inefficient bankruptcy filing. Campello and Matta (2011) show that the empty creditor problem is a pro-cyclical phenomenon. Based on their model, CDS over-insurance can minimize the moral hazard problem and maximize the probability that the firm’s investments are profitable.

The second and indirect channel of the empty creditor mechanism is reduced monitoring by creditors who are insured by CDS, and hence, less concerned about the credit risk of the borrower. Absent monitoring activity by creditors, managers can shift risk from shareholders to creditors, since this improves shareholder value, and thereby increases the probability of bankruptcy. Parlour and Plantin (2008) show that if CDS market is liquid, lenders may initiate too many loans and reduce monitoring, ex post.18 Ashcraft and Santos (2009) also argue that such reduced monitoring may ultimately lead to a higher cost of debt. Hirtle (2009) shows that the presence of CDS does not lead to greater credit supply. Norden, Buston, and Wagner (2011) document lower loan rates for banks that use credit derivatives more intensively. The recent decline in the absolute priority deviation (APD) during bankruptcy resolution (see, for example, Bharath, Panchapagesan, and Werner (2010)) is consistent with tougher creditors and coincides with the development of the CDS market.

Another aspect of the empty creditor mechanism is the reputation effect on a bank. Relationship banks may choose not to become empty creditors. While Bolton and Oehmke (2011) use a one-period model that cannot incorporate relationship lending, Gopalan, Nanda, and Yerramilli (2011) show in a different setting that the lead arranger suffers reputation damage from borrower bankruptcies due to inadequate screening or monitoring. But since CDS encourage lending, more banks are willing to lend after introduction of CDS trading. New banks can become empty creditors. Then empty creditor problem exists even when there are relationship banks.

In contrast to the restructuring and monitoring channels that derive from the empty creditor problem associated with covered CDS positions, Che and Sethi (2011) model an alternative mechanism for the impact of “naked” CDS on economic fundamentals. They argue that CDS can crowd out debt investors, reduce the firm’s debt capacity and increase its costs of debt. They find that the permitting naked CDS positions may increase the borrower’s bankruptcy risk due to its impact on the cost of debt. Naked CDS trading induces the most optimistic investors to sell CDS protection, which channels their capital away from purchasing bonds to investing in collateral to back their naked CDS positions. The remaining less optimistic bond investors require higher returns. This increase in the cost of debt, in turn, can increase the borrowing firms’ default risk.

The changed incentives of the borrowers with regard to restructuring and monitoring as a consequence of the empty creditor problem play a critical role in the discussion in the literature on the impact of CDS trading. However, there is lack of direct empirical evidence in this regard. Even when information on the proportion of CDS insured debt for a firm is available, it is hard to distinguish between covered and naked CDS positions. Some recent research investigates the empty creditor hypothesis from an indirect perspective. Bedendo, Cathcart, and El-Jahel (2011) examine the distressed firms’ decisions regarding out-of-court restructuring and bankruptcy filing during the global financial crisis. They find that CDS contracts do not significantly increase the probability of bankruptcy when the firm is already in distress, although their relatively small sample spans a short time period. However, Danis (2012) finds that distressed firms with CDS trading are less successful in debt workouts in their sample, over the period 2006-2011. Similarly, Peristiani and Savino (2011) document the higher bankruptcy risk in the presence of CDS during 2008.

III. Theoretical Framework and Hypotheses

In this section, we present the key insights from the theoretical literature that we use to motivate the specific hypotheses for our empirical tests. The prior literature has discussed both direct and indirect mechanisms through which CDS trading affects bankruptcy risk. The direct mechanism acts to lower the success of debt restructuring due to increased coordination failures among creditors. The coordination failure can result from the creation of empty creditors or, simply a larger, more diverse group of creditors. The indirect mechanism causes an increase in firm risk due to a higher leverage ratio and higher borrowing cost, as a result of catering to a more heterogeneous group of creditors, some of whom are hedged against the credit risk of the firm. The higher leverage can result either because of more efficient risk transfer or lower monitoring by some of the creditors.19 Higher borrowing costs may arise because of potential feedback effects from the CDS market to the firm’s financing decisions: a shock to the CDS market as a whole can be transmitted to the firm’s bonds by arbitrageurs who take advantage of mispricing between the bonds and the CDS.

We use a simple example to illustrate how CDS trading by creditors affects the likelihood of bankruptcy. The example is intended to convey the basic intuition of the empty creditor problem and is based on the model of Bolton and Oehmke (2011).

First consider the case where there is no CDS traded on the firm. Assume that creditors lend X to a firm. If the firm is in financial distress and consequently declares bankruptcy, creditors will recover r x X, where r is the recovery rate in bankruptcy. Consider, on the other hand, that the creditors allow the firm to restructure the debt, since the recovery value of the assets in bankruptcy is less than its value as a going concern. Suppose the firm offers the creditors part of the difference between the going concern value and the recovery value of the assets in bankruptcy, and agree to pay them say R x X, with R > r. Clearly, the creditors would consider such a restructuring and try to avoid bankruptcy.20 In general, restructuring would dominate bankruptcy.

Suppose next that the creditors can also buy CDS protection against the firm’s credit events. Clearly, bankruptcy would always be defined as a credit event. However, restructuring may or may not be defined as a credit event, as per the clauses of the CDS contract. If restructuring is included as a credit event, we define the contract to be a Full Restructuring (FR) CDS. If it is not, we defined it as a No Restructuring (NR) CDS.21

We first consider the case of FR CDS. Assume that the CDS premium (price) is F in present value terms at the time of default. Suppose the creditors buy CDS against Y of face value of the CDS. Therefore, if the firm defaults, the creditors’ total payoff with CDS protection is [rxX+(1-r-F)xY], in the event of bankruptcy, and [RxX+(1-R-F)xY] if the debt is restructured. Again, the creditors are better off with bankruptcy than with restructuring, if

[rxX+(1-r-F)xY]>[RxX+(1-R-F)xY],

i.e., when Y > X, since R > r. Hence, for FR CDS, bankruptcy dominates restructuring as a choice for empty creditors for whom the amount of CDS purchased exceeds the bonds held.

Now consider the case of NR CDS. Assume that the CDS premium (price) in this case is f in present value terms, where f < F. Suppose again that the creditors buy CDS against Y of face value of the CDS. Therefore, if the firm defaults, the creditors’ total payoff with CDS protection is [r x X+(1–rf)xY], in the event of bankruptcy, and [RxX-fxY] if the debt is restructured. Bankruptcy is a preferred outcome for the creditors if

[rxX+(1-r-f)xY]>[RxX-fxY],

or when

1

which is true even when Y < X, since R < 1. Thus, for NR CDS, bankruptcy is preferred when even a relatively small amount of CDS are purchased; hence, bankruptcy is the preferred alternative for a larger range of holdings of CDS by the creditors.

It is also easy to see that buying CDS protection with NR CDS contracts is a better choice in bankruptcy than restructuring without CDS protection, so long as

[rxX+(1-r-f)xY]>RxX,

which is equivalent to saying that:22

2

This condition is met when Y > X as long as R < 1 - f which is almost always true as the cost of CDS protection should not be higher than the loss in the event of restructuring. As before, it is likely to be true, even if Y < X, for reasonable values of R and f. Further, the greater the difference between Y and X, the greater the incentive to push the firm into bankruptcy. Hence, our example shows that a) creditors have an incentive to over-insure and push the firm into bankruptcy, b) this incentive increases with the difference between Y and X, i.e., the amount of CDS contracts outstanding relative to the firm’s debt, and c) the probability of bankruptcy occurring is greater for NR CDS contracts. This analysis provides the intuition for our first three hypotheses:

1 | 2 | 3 | Next Page >

NOTES

1Greenspan, Alan. 2004. “Economic Flexibility.” Speech given to the Her Majesty’s Treasury Enterprise Conference, London, January 26, 2004.

2Berkshire Hathaway Annual Report for 2002.

3Wall Street Journal, March 24, 2009.

4See, for example, “Goldman Purchase Puts CDS in Focus,” Financial Times, October 4, 2009. “Goldman Sachs May Reap $1 Billion in CIT Bankruptcy”, Bloomberg, October 5, 2009.

5Appendix A lists several other cases of a similar nature, demonstrating that the example cited is not that unique.

6The use of equity derivatives such as options or swaps in the context of equities creates the analogous issue of “empty voters” who enjoy voting rights in the firm, but without any financial risk, by breaking the link between cash flow rights and control rights.

7Wall Street Journal, April 10, 2009.

8Following this episode, the New York Fed launched its first round of regulatory actions on CDS in September 2005. It requested major CDS dealers to clear the backlog of unsettled contracts.

9Other studies such as Duffie (2007), Stulz (2010), and Jarrow (2011) also offer related discussions.

10See, for example, Ashcraft and Santos (2009) and Parlour and Plantin (2008), for this line of argument.

11In results not reported here to conserve space, we also examined the validity of an alternative mechanism through which CDS trading may affect the credit risk of a firm. If a shock causes the CDS to become overpriced relative to the bonds issued by the firm, this overpricing may spill over to the bond market, increasing the cost of debt for the firm. This may affect the ability of the firm to refinance its debt by increasing its cost and in extreme cases, affecting its ability to pay off the bondholders. We did not find supportive evidence for this overpricing feedback mechanism.

12Arping (2004) shows that credit risk transfer alters the incentives of lenders and borrowers. With the shelter of the credit protection, lenders may be less willing to monitor the borrowers. The problem can be mitigated by setting the length of CDS protection less than the maturity of the project. Thompson (2007) extends the Duffee and Zhou (2001) formulation by allowing for informational asymmetry in the CDS market and relaxing the maturity mismatch assumption. Then, it is unclear whether the use of CDS to transfer credit risk would be beneficial, since it would depend on the nature of the moral hazard problem, the relationship between the bank and the borrower, the cost of loan sales and the cost of capital.

13The Department of Justice investigated Markit, the data aggregator and vendor for price manipulation in July 2009.

14See, for example, an early survey by Damodaran and Subrahmanyam (1992), and Sorescu (2000), an example of such studies.

15Although bond holders can also buy equity derivatives, CDS provide a direct protection on their exposure. Moreover, given the maturity of equity derivatives, it is harder for bond holders to hedge their exposure with equity derivatives.

16This model and its variants have been widely used to measure of bankruptcy risk. Recent additions to the literature include Duffie, Saita, and Wang (2007), who propose a reduced form model with good out-of-sample default prediction, Das, Duffie, Kapadia, and Saita (2007) who find that defaults are more clustered than would be implied by conventional credit risk models, and Duffie, Eckner, Horel, and Saita (2009) who propose a frailty model, which solves the omitted variable bias. Other follow-up studies to identify and include important risk factors in bankruptcy risk models include Lando and Nielsen (2010) and Jorion and Zhang (2009).

17Several other studies also find that lenders exploit their information advantage. Hale and Santos (2009) show that if banks exploit their information advantage, firms respond by expanding their borrowing base to include lenders in the public bond market or adding more bank lenders. Massoud, Nandy, Saunders, and Song (2011) and Ivashina and Sun (2011) find that institutional investors trade on their private information from syndicated loan lending relationships. Gormley, Gupta, and Jha (2011) show that creditor incentives to monitor borrowers and recover loans affect bankruptcy outcomes.

18DeMarzo and Duffie (1999) model security design with risk transfer in a similar setting.

19Another channel through which the empty creditor problem may manifest itself is through a reduction in monitoring by the empty creditors, who no longer derive any benefit from such activity. Thus, reduced monitoring is the attenuated case of the empty creditor problem, and contributes to an increase in the credit risk of the firm, and in turn, leads to a higher probability of bankruptcy.

20The precise size of R would be determined in a bargaining process between the creditors and the shareholders of the firm.

21Other types of contracts also exist, but are not relevant for purposes of this simple illustration. See Appendix C for details.

22The calculation for the FR CDS is the same, except that the fee is replaced by F instead of f. The precise range of values for Y relative to X would be smaller than for the NR CDS, as argued above.

Related Posts Plugin for WordPress, Blogger...

Comments

The comments to this entry are closed.

UConn-ad

NYSSA Job Center Search Results

To sign up for the jobs feed, click here.


UPCOMING EVENT
MF16

NYSSA Market Forecast™: Investing In Turbulent Times
January 7, 2016

Join NYSSA to enjoy free member events and other benefits. You don't need to be a CFA charterholder to join!


CFA® EXAM PREP

CFA® Level I 4-Day Boot Camp
Midtown

Thursday November 12, 2015
Instructor: O. Nathan Ronen, CFA

CFA® Level II Weekly Review - Session A Monday
Midtown

Monday January 11, 2016
Instructor: O. Nathan Ronen, CFA

CFA® Level III Weekly Review - Session A Wednesday
Midtown

Wednesday January 13, 2016
Instructor: O. Nathan Ronen, CFA

CFA® Level III Weekly Review - Session B Thursday
Thursday January 21, 2016
Instructor: O. Nathan Ronen, CFA

CFA® Level II Weekly Review - Session B Tuesday
Thursday January 26, 2016
Instructor: O. Nathan Ronen, CFA