Portal de Periódicos da CAPES

The effect of liquidity on the price discovery process in credit derivatives markets in times of financial distress

2011; Taylor & Francis; Volume: 17; Issue: 9-10 Linguagem: Inglês

10.1080/1351847x.2010.538529

1466-4364

Sergio Mayordomo, Juan Ignacio Peña, Juan Romo,

Global Financial Crisis and Policies

Abstract This paper analyses the role of liquidity in the price discovery process. Specifically, we focus on the credit derivatives markets in the context of the subprime crisis. We present a theoretical price discovery model for the asset swap packages (ASPs), bond and credit default swap (CDS) markets and then we test the model with data from 2005 to 2009 on Euro-denominated non-financial firms. Our empirical results show that the ASP market clearly leads the bond market in the price discovery process in all cases, while the leadership between ASPs and CDSs is very sensitive to the appearance of the subprime crisis. Before the crisis, the CDSs market leads the ASP market, but during the crisis, the ASP market leads the CDS market. The liquidity, measured as the relative number of market participants, helps to explain these results. Keywords: price discoveryvector error correction model (VECM)credit derivativescredit spreads JEL Classification : C32C51G13G14 Acknowledgements We are grateful to the Editors of this special issue: John Wilson, Barbara Casu, and David McMillan for their useful comments. We owe thanks to two anonymous referees for many astute comments that considerably improved the study. We also thank Isabel Figuerola-Ferretti, Javier Gil, Carlos González-Aguado, Jesús Gonzalo, Neil Kellard, María Rodríguez, Pedro Serrano, Javier Suárez and seminar participants at the XVII Foro de Finanzas and at the Emerging Scholars in Banking and Finance Conference at Cass Business School for all their comments and suggestions. Sergio Mayordomo and Juan Ignacio Peña acknowledge financial support from MICINN Grant Ref: ECO2009-12551. Juan Romo acknowledges financial support from MICINN Grant Ref: ECO2008-05080. The usual disclaimer applies. Notes Acharya and Schaefer (2006) Acharya, V. and Schaefer, S. 2006. "Liquidity risk and correlation risk: Implications for risk management". London Business School. Working Paper [Google Scholar] posit that the liquidity risk can be defined as unpredictable changes in transaction costs and in liquidity. These adverse liquidity shocks, systematic or idiosyncratic, are mainly due to high and negative changes in financial products' returns and reduce the amount of capital available to financial intermediaries which lowers the ability of their trading desk to provide liquidity. The value of CDSs outstanding at the end of 2004, 2005 and 2006 was $8.42, $17.1 and $34.4 trillion, respectively. The CDS market exploded over the past decade to more than $45 trillion in mid-2007 and more than $62 trillion in the second half of the same year, according to the ISDA. The size of the CDS market in mid-2007 is roughly twice the size of the US stock market (which is valued at about $22 trillion) and far exceeds the $7.1 trillion mortgage market and $4.4 trillion US treasuries market. However, the notional amount outstanding decreased to $38.6 trillion at the end of 2008. The main reason is the existence of a higher CTD option due to the higher risk of borrowers and also to liquidity reasons given that the higher the number of bonds issued, which they use as a liquidity measure, the more difficult is that the CDS market leads price discovery. According to Dötz (2007) Dötz, N. 2007. "Time-varying contributions by the corporate bond and CDS markets to credit risk price discovery". Deutsche Bundesbank. Discussion Paper Series 2: Banking and Financial Studies No 08 [Google Scholar], the relatively large contribution of the CDS market to price discovery is not necessarily tantamount to general and lasting improvement in the processing of information; the turbulence in the credit markets in the spring of 2005 was apparently handled much better by the bond market than by the CDS market. The weaknesses of the CDSs are likely to consist in the relatively high concentration and homogeneousness of its market players, whose herding behaviour, particularly in times of crisis, can strain liquidity, amplify market volatility and hamper price discovery. Chan-Lau and Kim (2004) Chan-Lau, J. A. and Kim, Y. S. 2004. Equity prices, credit default swaps, and bond spreads in emerging markets. Working Paper, IMF [Google Scholar] relate the same three financial instruments but for emerging market sovereign issuers. In most countries, they do not find any equilibrium price relationship between equity and bond markets and in terms of price discovery, it is difficult to conclude that one particular market dominates the price discovery process. As far as we know, the relationship between ASP and CDS has only been treated in De Wit (2006) De Wit, J. 2006. "Exploring the CDS-bond basis". National Bank of Belgium. Working Paper [Google Scholar]. However, the perspective adopted in De Wit (2006) De Wit, J. 2006. "Exploring the CDS-bond basis". National Bank of Belgium. Working Paper [Google Scholar] is based on the long-run equilibrium that should exist, and which the author finds, between ASP and CDS, ignoring the price discovery process. Although we develop the case of CDS and ASP prices, the model can also be applied to study the interaction between CDS and bond spreads or between ASP and bond spreads. A popular arbitrage strategy employed by hedge funds in credit derivatives markets defines the correct price from a long-term equilibrium price based on the cointegration methodology. The arbitrageur is betting only that the spread between the two cointegrated assets will narrow, which can be understood as a permanent adjustment process towards an economic equilibrium. As Dötz (2007) Dötz, N. 2007. "Time-varying contributions by the corporate bond and CDS markets to credit risk price discovery". Deutsche Bundesbank. Discussion Paper Series 2: Banking and Financial Studies No 08 [Google Scholar] state, the market for synthetic CDO products, which as opposed to cash CDS products are not backed by bonds or loans but by CDSs, presents some advantages such as the better availability of CDSs relative to bonds or loans and the heavy demand among investors for unfunded supersenior tranches. We may also find protection sellers who hedge their positions, again, in the CDS market or individuals that participate in the CDS market in order to hedge their exposures to other institutions that are not due to debt positions. We include this type of agents in order to make the model more comprehensive, given that the nature of CDSs is to provide insurance. In fact, CDSs are usually used to manage the credit risk. As CDSs are over-the-counter (OTC) instruments, it is possible to buy a CDS contract whose maturity coincides with the bond's maturity and whose premium payments timing is agreed by the parties. As the bond's maturity date approaches, the use of CDSs with a 5-year constant maturity would lead to overhedging, given that the maturity dates of CDSs and asset swaps do not coincide. The consequence is that the investor will pay a CDS spread above the one needed to be fully hedged. Thus, we take advantage of the range of CDSs maturities to fit a CDS curve using a Piecewise Cubic Hermite Interpolating Polynomial algorithm that allows us to match asset swap and CDS maturities. This method is also used in Levin, Perli, and Zakrajšek (2005) Levin, A., Perli, R. and Zakrajšek, E. 2005. "The determinants of market frictions in the corporate market". Federal Reserve Board, Washington. Working Paper [Google Scholar]. In order to proceed in this way, we assume that the investor can borrow money at Euribor flat. If , then a profitable arbitrage opportunity exists. The investor should take long positions in both CDS and ASP and borrow the required quantity of money in order to finance the investment at 3 months Euribor. If , the inverse strategy will lead to an arbitrage opportunity. A combination of short positions on both the CDS and ASP leads to a net payment for the investor equal to the difference between CDS and ASP spreads, . The net payoff is also zero at coupon payment dates while at the bond's maturity, as in every quarterly payment, the net payoff is equal to the basis. Assumption A1 is necessary to reach a two-sided bound on the CDS rate and to guarantee that the equivalence relationship holds. An example of a statistical arbitrage analysis in credit derivatives markets according to the technique and concept introduced Hogan et al. (2004) Hogan, S., Jarrow, R. A., Teo, M. and Warachka, M. 2004. Testing market efficiency using statistical arbitrage with applications to momentum and value strategies. Journal of Financial Economics, 73(3): 525–65. [Crossref], [Web of Science ®] , [Google Scholar] can be found in Mayordomo, Peña, and Romo (2009) Mayordomo, S., Peña, J. I. and Romo, J. 2009. "Are there arbitrage opportunities in credit derivatives markets? A new test and an application to the case of CDS and ASPs". Universidad Carlos III de Madrid. Working Paper [Google Scholar]. Cossin and Lu (2005) Cossin, D. and Lu, H. 2005. "Are European corporate bonds and default swap markets segmented?". FAME. Working Paper 153 [Google Scholar] state that the liquidity premium, the CTD option and the market segmentation explain the pricing differences between bonds and CDS. The effect of the CTD option is more important as default risk increases. If the two markets price credit risk equally in the long run, then their prices should be cointegrated with cointegrating vector [1,-1,c], suggesting a stationary basis. Shorting a corporate bond or an ASP with a required maturity, even years, is not an easy task. The short sale of bonds or ASPs could be done via a repurchase agreement (repo) but as Blanco, Brennan, and Marsh (2005) Blanco, F., Brennan, S. and Marsh, I. W. 2005. An empirical analysis of the dynamic relation between investment grade bonds and credit default swaps. Journal of Finance, 60(5): 2255–81. [Crossref], [Web of Science ®] , [Google Scholar] explain, it is impossible to borrow a bond via a repo. The reason is that repo market for corporate bonds is illiquid and even if it is possible to short a bond via a repo, the tenor of the agreement would be short. Schonbucher (2003) Schonbucher, P. J. 2003. Credit derivatives pricing models: Models, pricing, implementation, Wiley Finance. [Google Scholar] states that this limitation could be solved by issuing credit-linked notes linked to the corresponding bond and selling them to the investors in the asset swap market. This alternative presents other limitations given that the issuance of credit-linked notes takes time and implies high fixed cost. This fact implies that deviations in the equivalence relationship might not imply arbitrage opportunities whenever an asset swap short sale is needed. Thus, in some cases traders are not able to exploit price differentials when the CDS premium is higher than the asset swap spread and as Blanco, Brennan, and Marsh (2005) Blanco, F., Brennan, S. and Marsh, I. W. 2005. An empirical analysis of the dynamic relation between investment grade bonds and credit default swaps. Journal of Finance, 60(5): 2255–81. [Crossref], [Web of Science ®] , [Google Scholar] suggest, this asymmetry may affect significantly the dynamic adjustment of credit spreads. The restrictions on short-sales could be even more severe in periods of financial distress. These individual interpret that is the CDS price, given its probability of default. As the market price ¯ s t is above the reservation price, the investor is interested in selling protection in exchange of ¯ s t . Note that the total endowments can increase exogenously from period t to period t+1, for instance, by means of CDS or bond issuances. The idea is that under the assumptions employed when defining the arbitrageurs demand, the fact that the ASP spread is a good indicator, although not perfect, of the CDS spread and given that both spreads are prices of the credit risk of a given firm, it seems reasonable to assume that the elasticities are similar. The objective of this paper is not related with the literature of credit risk pricing and so, Equations (10) and (11) are not pricing equations of credit risk, such as the ones defined from the probability of default and recovery rates, but simply the ASP and CDS prices that clear both markets. We do not report the whole expression of u t in order to save space and also, because of the assumptions on residuals they are not going to appear in our analysis. Although hedgers are not important in the price discovery process, they must be included in the model because these individuals really participate in credit markets and we need them in order to make the model more comprehensive. This decline in trading during the second half of 2008 reflects a combination of significantly reduced risk appetite, expectations of stable low interest rates in major markets and lower hedge fund activity. Exchange rate movements may also have affected to this decline. Most institutions report their positions in US dollars and the euro and the pound sterling depreciated by 30% and 12%, respectively, against the US dollar between June and December 2008. During the first quarter of 2009, issuance of investment grade corporate bonds in Europe totalled a record ≠uro140bn, well above quarterly levels of less than ≠uro50bn seen in recent years. It is motivated, among other reasons, by the use of bond markets for funding and the government guarantees to aid to the bond issuance. IFSL Research (2009b) reports that the amounts outstanding on the global bond market, which includes bonds, notes and money market instruments, increased 6% in 2008 to $83 trillion. See for instance Norden and Weber (2004) Norden, L. and Weber, M. 2004. Informational efficiency of credit default swap and stock markets: The impact of credit ratings announcements. Journal of Banking and Finance, 28(11): 2813–43. [Crossref], [Web of Science ®] , [Google Scholar], Blanco, Brennan, and Marsh (2005) Blanco, F., Brennan, S. and Marsh, I. W. 2005. An empirical analysis of the dynamic relation between investment grade bonds and credit default swaps. Journal of Finance, 60(5): 2255–81. [Crossref], [Web of Science ®] , [Google Scholar], Zhu (2006), Baba and Inada (2007) Baba, N. and Inada, M. 2007. "Price discovery of credit spreads for Japanese mega-banks: Subordinated bond and CDS". Institute for Monetary and Economic Studies, Bank of Japan. Duscussion Paper 2007-E-6 [Google Scholar], Dötz (2007) Dötz, N. 2007. "Time-varying contributions by the corporate bond and CDS markets to credit risk price discovery". Deutsche Bundesbank. Discussion Paper Series 2: Banking and Financial Studies No 08 [Google Scholar], Forte and Peña (2009) Forte, S. and Peña, J. I. 2009. Credit spreads: An empirical analysis on the informational content of stocks, bonds and CDS. Journal of Banking and Finance, 33(11): 2013–25. [Google Scholar] and Coudert and Gex (2010) Coudert, V. and Gex, M. 2010. Contagion inside the credit default swaps market: The case of the GM and Ford crisis in 2005. Journal of International Financial Markets, Institutions and Money, 20(2): 109–34. [Google Scholar]. In a limit case, if N BOTH tends to infinity, the price discovery is equal to 0.5. Allocation to bonds from high-net-worth individuals increased from 27% to 29% during 2008 with equities seeing the largest decline in their share of portfolio allocation. The nominal of CDSs with respect to ASP contracts serves to show how in periods of financial distress it is much more difficult to participate in CDS than in ASP markets. The standard bond's faced value is ≠uro1000 while the CDS typical notional amount is ≠uro10–20 million for investment grade credits and ≠uro2–5 million for high yield credits. Ammer and Cai (2007) Ammer, J. and Cai, F. 2007. "Sovereign CDS and bond pricing dynamics in emerging markets: Does the cheapest-to-deliver option matter?". Board of Governors of the Federal Reserve System. International Finance Discussion Papers Number 912 [Google Scholar] state that the main reason to support that bond spreads lead CDS premiums is the existence of a higher CTD option and for liquidity reasons. In particular they employ as a liquidity measure the number of bonds outstanding by a given firm and find that the higher the number of bonds the less likely it is that CDSs will lead price discovery. Under illiquid scenarios, the CTD option embedded in CDS becomes more valuable and the liquidity premium to bear liquidity risk increases. The GFI FENICS Credit curve gives preference to real trades and quoted mid- points where available, and in their absence, it is based on the calculation of a running point level using the John Hull and Alan White methodology to ensure a credit curve always exists for each reference entity. This curve is a good approximation for CDSs at any maturity as several error analyses reveal. The median of the absolute difference in basis points between five years CDS premiums as defined from credit curve and the actual quotes or transaction prices for the period between April 2001 and May 2002, is equal to 1.16, 2.01 and 3.82 basis points for AAA/AA, A and BBB ratings for a total of 2659, 9585 and 8170 companies, respectively. Moreover, market CDS spread could be different from what we are assuming to be the true CDS spread by as much as 3.725 bps. on average. This limit is set in order to avoid the selection of bonds with a small volume which could require higher transaction costs due to their reduced liquidity. Our initial sample was 285 corporate bond issuers. We found a total of 116≠uro denominated bonds that mature before February 2012 but only 67 of them include information on 5-year bid/ask CDS spreads, asset swap spreads and Fenics Curve for at least 90 trading days. Of these, two bonds have been discarded because the issued amount does not exceed 300 millions of Euros, another four bonds were discarded because they were not investment grade bonds throughout the whole sample period. Another four bonds were discarded because their asset swap spreads were persistently negative and, finally, seven bonds were discarded because prices were too far from par. We do not report these results because they are similar to the ones presented in Table 3. However, these results are available upon request. Cointegration test detailed results for ASP and bond spreads are available upon request. The detailed results are available upon request. Results of this analysis are available upon request. These results are available upon request. The t-statistics presented in Table 5 are obtained using the GG metrics reported in Table 4. Moreover, we employ the average GG metric obtained across the 1000 bootstrap repetitions and obtain similar results. Amato and Remolona (2003) state that when it turns out to be very costly to undertake transactions in a given instrument, the investors must be compensated for it. This compensation is reinforced in the presence of uncertainty about the liquidity (or illiquidity) of an ASP, bond or CDS at a given time, and thus the investors could require a premium to bear this risk. This liquidity premium has been proved to exist both in CDS and bond markets. Longstaff, Mithal and Neis (2005) Longstaff, F. A., Mithal, S. and Neis, E. 2005. Corporate yield spreads: Default risk or liquidity? New evidence from the credit default swap market. Journal of Finance, 60(5): 2213–53. [Crossref], [Web of Science ®] , [Google Scholar] and Tang and Yan (2007) Tang, D. Y. and Yan, H. 2007. "Liquidity and credit default swap spreads". Kennesaw State University. Working Paper [Google Scholar] among others support the presence of a liquidity premium in CDS spreads. Collin-Dufresne, Goldstein, and Martin (2001) Collin-Dufresne, P., Goldstein, R. and Martin, S. 2001. The determinants of credit spread changes. Journal of Finance, 56(6): 2177–207. [Crossref], [Web of Science ®] , [Google Scholar], Perraudin and Taylor (2003) Perraudin, W. and Taylor, A. 2003. "Liquidity and bond market spreads". Bank of England. Working Paper [Google Scholar], Elton et al. (2001) Elton, E., Gruber, D., Agrawal, D. and Mann, C. 2001. Explaining the rate spread on corporate bonds. Journal of Finance, 56(1): 247–77. [Crossref], [Web of Science ®] , [Google Scholar], Delianedis and Geske (2001) Delianedis, G. and Geske, R. 2001. "The components of corporate credit spreads: Default, recovery, tax, jumps, liquidity, and market factors". University of California. Working Paper [Google Scholar] and Chen, Lesmond, and Wei (2007) Chen, L., Lesmond, D. A. and Wei, J. 2007. Corporate yield spreads and bond liquidity. The Journal of Finance, 62(1): 119–49. [Crossref], [Web of Science ®] , [Google Scholar] among others find that liquidity is an additional factor to credit risk which is present in bond spreads. We do not report the whole expression of u t in order to save space and also, because of the assumptions on residuals they are not going to appear in our analysis.