CDS و پویایی قیمت گذاری اوراق قرضه در بازارهای نوظهور

We examine the relationships between credit default swap (CDS) premiums and bond yield spreads for nine emerging market sovereign borrowers. We find that these two measures of credit risk deviate considerably in the short run, due to factors such as liquidity and contract specifications, but we estimate a stable long-term equilibrium relationship for most countries. In particular, CDS premiums tend to move more than one-for-one with yield spreads, which we show is broadly consistent with the presence of a significant “cheapest-to-deliver” (CTD) option. In addition, we find a variety of cross-sectional evidence of a CTD option being incorporated into CDS premiums. In our analysis of the short-term dynamics, we find that CDS premiums often move ahead of the bond market. However, we also find that bond spreads lead CDS premiums for emerging market sovereigns more often than has been found for investment-grade corporate credits, consistent with the CTD option impeding CDS liquidity for our riskier set of borrowers. Furthermore, the CDS market is less likely to lead for sovereigns that have issued more bonds, suggesting that the relative liquidity of the two markets is a key determinant of where price discovery occurs.

In the past decade, the credit derivatives market has experienced rapid growth, and among credit derivatives, the credit default swap (CDS) has become the most widely traded instrument for transferring credit risk. According to survey data coordinated by the Bank for International Settlements, by the end of 2005, the total notional amount of outstanding CDS contracts had surpassed $13 trillion.2 CDS contracts can help isolate credit risk from other factors affecting bond prices such as illiquidity premiums, and thus may provide more accurate pricing and cleaner measurement of credit risk than is available from the underlying debt markets. Most empirical comparisons of CDS and bond pricing, such as Hull and White (2000), Longstaff et al. (2005), Blanco et al. (2005), and Zhu (2006) have only considered investment-grade corporate names; these studies generally have concluded that arbitrage forces CDS premiums to be approximately equal to the underlying bond spreads in the absence of market frictions. Blanco et al. (2005) also report evidence that corporate CDSs seem to lead corporate bonds in reflecting changes in credit conditions. However, the number of CDS quoted for speculative-grade reference entities, while small until recent years, has since increased rapidly, and it is not obvious that investment-grade empirical regularities are necessarily applicable to riskier credits. This paper makes three key contributions. First, we analyze the “cheapest to deliver” (CTD) option that is embedded in most CDS contracts, and we show that it can be quite important to determination of hedge ratios and to pricing relationships between bonds and CDS. We present evidence of the empirical importance of the option. Second, we find somewhat different dynamic relationships between CDS and bond prices than in the previous literature. These differences may arise because of the CTD option or because borrowers in our sample are riskier, but the differences in our results are also consistent with the relative importance of public and private information being a key determinant of differences in price dynamics across assets. Third, we examine pricing relationships for sovereign credit risk, with substantial variation in credit spreads over time and across reference entities. Most prior work on CDS has focused on investment-grade corporate names. Although the number of sovereign reference entities is smaller, cross-sectional relationships are often strong enough so that we obtain statistically significant results, and the sovereign market is large in the sense of the volume of credit risk transfer. Taken together, our arguments and evidence suggest that, particularly as the CDS market extends it reach to more speculative-grade names, market participants may wish to reconsider some features of the standard CDS contract, and may need to revise pricing models, trading strategies, and hedging strategies. There are matters of considerable practical importance to a rapidly growing market. Our paper also differs from most previous work on CDS markets in its focus on the implications of the “cheapest-to-deliver” (CTD) option in CDS contracts. The CTD option often arises when the reference entity has more than one long-term debt instrument outstanding, because upon a default event, the protection buyer typically may choose to deliver virtually any long-term obligation that matches the currency and debt-seniority specified in the contract. Accordingly, there can be an incentive to deliver the (ex post) lowest-priced instrument that the protection buyer already owns or could acquire in secondary markets, even if the protection had been bought to hedge a different instrument. Most papers in the corporate CDS literature mention the CTD option in passing as a potential pricing complication, but few go further with addressing it. One exception is Packer and Zhu (2005), who consider price differences among CDS quotes for the same reference entities that differ only in how broadly a credit event is defined and in some cases, have varying restrictions on the set of deliverable instruments. 3 In addition, Blanco et al. (2005), discuss a CTD option as the most likely explanation for an upward shift in CDS quotes for one firm in their sample (Fiat SpA), after it had issued an unusual bond that traded at low prices. Our empirical focus is on emerging market sovereign credit risk, which differs in several respects from the investment-grade corporate credit risk addressed in prior studies of CDSs. First, sovereigns are among the largest high-yield borrowers in the world, typically with more bonds outstanding, longer maturities, larger issues, and more liquidity than their corporate counterparts. Second, during the period we study, the outstanding debt of our sovereign borrowers was almost entirely in the form of bonds, whereas corporate bond issuers nearly always also borrow from banks. Accordingly, some of the CDS market activity for corporate reference entities likely involves hedging illiquid loan positions, which we would not expect to be the case for our sovereign sample. Third, countries in financial distress generally do not enter bankruptcy proceedings or ever liquidate their assets, so the nature of default risk is somewhat different. In practice, sovereign defaults have been followed by exchange offers or other debt restructuring mechanisms, which tend to be shaped at least in part by either explicit or implicit political forces, which would not play the same role in statutory corporate bankruptcy proceedings. Fourth, our sample spans a wide range of creditworthiness, ranging from China, with an average 5-year bond spread over the swap curve of just 26 basis points during our sample, to Uruguay, with an average spread of 1124 basis points. Finally, for sovereign credit instruments, there may be less information asymmetry between market participants than there is among traders of corporate credit risk. The relative importance of public and private information can affect the process of price discovery, and if informed traders are choosing their trading venue strategically, also affect the dynamic relationship between credit pricing in the bond and CDS markets. With respect to sovereign credit risk, much of the relevant information about the national economy and the state of government finances tends to be in the public domain, so new information might tend to be reflected in observed prices more quickly in the more liquid market, if there are any differences in timing. But if price discovery for corporate credit risk is driven more by informed trading, this might occur in the market that is institutionally less transparent, which might consequently tend to be the less liquid market. Chakravarty et al. (2004) document that relative liquidity measured by trading volume and bid-ask spreads plays an important role in capturing the relative contribution to price discovery by stocks and options. Consistent with their findings, we report evidence that bond spreads lead CDS premiums for emerging market sovereigns more often than has been found for investment-grade corporate credits, especially for sovereigns that have issued more bonds. There has been relatively little work to date on sovereign CDS markets in the literature. Packer and Suthiphongchai (2003) document the growth in the sovereign CDS market and compare average sovereign CDS premiums to corporate CDS premiums by credit rating. Pan and Singleton (2008) use data on CDS contracts of several different maturities, in an attempt to separately identify default risk and recovery risk in the context of the Duffie and Singleton (2003) credit pricing framework. Andritzky and Singh (2007) also separately consider default risk and recovery risk. For a period when Brazilian sovereign spreads were high and volatile, they find that, all else equal, the CDS premium covaries negatively with the (pre-default) price of the cheapest bond that would be eligible for delivery upon default, broadly consistent with the effect of the CTD option. Among empirical studies with an emerging-market focus, the paper that appears to be most similar to the dynamic analysis undertaken here is Chan-Lau and Kim (2004), who look for lead–lag relationships among sovereign bond indices, sovereign CDS premiums, and national stock market indexes, reporting somewhat inconclusive results. In comparison, our analysis uses more closely matched credit market data, which enables us to cast our analysis in terms of an arbitrage relation. We also cover a longer sample period and more countries. Previewing our main results, we find that sovereign CDS premiums and bond spreads do move in tandem in the long run. In the short run, when the prices deviate from their long-run equilibrium, CDS markets seem to lead bond markets in price discovery in some instances, but lag bond prices in other cases, with some evidence that the more liquid market tends to lead. We also offer a variety of evidence that CDS premiums are affected by the CTD option that arises in the standard CDS contract. Our results concerning the importance of the CTD option have a number of further implications. First, the CTD option appears to be quantitatively important enough so that it should be accounted for in any CDS pricing model. In particular, we report post-default bond quotes for one sovereign issuer (Argentina), that vary widely enough to suggest that a CTD option could increase the value CDS protection by more than 20 percent. Second, the CTD option appears to have a significant effect on optimal hedging ratios versus the underlying debt, with protection buyers often over-hedged, if they match the CDS notional amount to the face value of their debt holdings. Our dynamic pricing analysis suggests that the correct hedging ratio could be as high as 1.5 for some credits. Third, the pay-off to a CDS protection buyer of a strategic delivery choice is a factor in optimal CDS contract design, particularly in the trade-off between flexibility and specificity in settlement procedures after a covered credit event. To the extent that the least valuable bond would always be delivered, even when the protection buyer had been hedging a position in a different asset, it is not at all clear that the wide choice in deliverable instruments available under prevailing market conventions is desirable, particularly if there is ex ante uncertainty about the extent to which the CTD option will be in-the-money after a credit event. Finally, the value of the CTD option incorporated into CDS pricing can make it more difficult to draw inferences from CDS premiums about the market price of the underlying credit risk. This is a potential drawback for risk management, and for policy makers who use market prices to monitor credit risk and credit market conditions.

In this paper, we analyze the relationship between sovereign CDS premiums and bond yield spreads for nine emerging-market sovereign credits over a four-year period. For most countries in our sample, we find that sovereign CDS premiums and bond spreads are linked by a stable linear long-run equilibrium relation. The two prices of credit risk, however, often diverge from the equilibrium in the short run, with these temporary deviations typically dissipating at a rate of 5–13 percent per day. This gradual convergence in prices implies that there is some predictability of relative price changes between the two markets, which is sustained by some combination of illiquidity in at least one of the two markets and risk that arises from the imperfections in the underlying arbitrage relation. CDS markets seem to lead bond markets in price discovery in some instances, but lag bond prices in other cases. We find some evidence that the relatively more liquid market tends to lead the other. In particular, across our sovereign borrowers, we calculate that the GG measure of relative CDS price leadership is correlated (i) positively with the ratio of the bond bid-ask spread to the CDS bid-ask spread and (ii) negatively with the number of bonds outstanding (a proxy for collective bond market liquidity), with a statistically significant coefficient in the second case. Price leadership in the more liquid market is consistent with sovereign credit risk pricing being driven mostly by public information. Overall, the bond market leads more often for our emerging-market sovereign borrowers than has been reported by Blanco et al. (2005) for investment-grade corporate names. Accordingly, the CDS market may not be as clear-cut a choice as a benchmark for credit pricing for sovereign bond issuers as it is for corporate borrowers. We also derive the pricing implications of the CTD option in CDS contracts, and report a variety of evidence that the CTD option is quantitatively important in our sample, and in particular that the value of the CTD option is an important determinant of the CDS basis – the difference in credit pricing between the CDS and bond markets. For example, we find that the CDS basis tends to be higher for a borrower for which the ex post CTD option is likely to be larger, because it has issued at least one bond with a below-market coupon, long maturity, and/or low pre-default market value, as a fraction of the face value that would be insured by CDS protection. We also find that the CDS basis tends to be higher for riskier reference entities with lower credit ratings and higher bond yield spreads. This is consistent with the presence of the CTD option, because all else equal, the option is more valuable ex ante when the probability of a default event is higher. Similarly, in our dynamic analysis, we find that CDS premiums tend to move more than one-for-one with bond yield spreads; all else equal, a rising yield spread is associated with an increasing probability of a credit event. In addition, we report a statistically significant negative correlation across sovereign borrowers between the estimated long-run slope and the cheapest bond price, further evidence that CDS premiums move more than one-for-one with bond yield spreads specifically because of the CTD option. We also report some weaker evidence that suggests the estimated long-run slope is negatively associated with CDS liquidity, consistent with the CTD option creating a market friction that is an impediment to liquidity. Our results concerning the importance of the CTD option have a number of further implications. First, the CTD option appears to be quantitatively important enough so that it should be accounted for in any CDS pricing model. Second, the CTD option appears to have a significant effect on optimal hedging ratios versus the underlying debt, with protection buyers often over-hedged, if they match the CDS notional amount to the face value of their debt holdings. Third, the pay-off to a CDS protection buyer of a strategic delivery choice is a factor in optimal CDS contract design, particularly in the trade-off between flexibility and specificity in settlement procedures after a covered credit event. To the extent that the least valuable bond would always be delivered, even when the protection buyer had been hedging a position in a different asset, it is not at all clear that the wide choice in deliverable instruments available under prevailing market conventions is desirable, particularly if there is ex ante uncertainty about the extent to which the CTD option will be in-the-money after a credit event. Finally, the value of the CTD option incorporated into CDS pricing can make it more difficult to draw inferences from CDS premiums about the market price of the underlying credit risk. This is a potential drawback for risk management, and to policy makers who use market prices to monitor credit risk and credit market conditions.