Sovereign Credit Default Swap Premia∗ Patrick Augustin† ‡ Stockholm School of Economics This version: April 2012 Preliminary, comments welcome Abstract This paper provides a comprehensive overview of the young, but rapipdly growing literature on sovereign credit default swap premia, describes key statistical and stylized facts about prices, the market, its players and related trading activities and attempts to raise some thoughtprovoking questions. We thereby hope to serve as a useful starting point for anyone interested in the topic. Keywords: Credit Default Swap Spreads, Default Risk, Descriptive Statistics, Literature Review, Sovereign Debt, Term structure JEL Classification: A31, D40, E02, F30, F34, G01, G12, G13, G14, G15, O57, Y10 ∗ This version is preliminary and incomplete. Comments are welcome and appreciated. We would like to thank René Kallestrup for valuable feedback. All remaining errors are our own. † Stockholm School of Economics, Finance Department, Sveavägen 65, 6th floor, Box 6501, SE-113 83 Stockholm, Sweden. Email: Patrick.Augustin@hhs.se. ‡ 1 Introduction The late 1990s and early 2000s displayed a series of emerging market sovereign defaults. Nowadays, the popular press is covered with news about developed countries in distress, in particular the GIIPS countries: Greece, Italy, Ireland, Portugal and Spain. For most part of history, lenders to national governments were unable to insure against sovereign default and had to trust creditors that they honor their debt obligation in good faith. Our generation, on the other hand, has witnessed the creation of insurance products to protect specifically against such contingencies. These insurance products, so-called credit default swaps (henceforth CDS) have undergone two decades of spectacular growth and, covering an underlying cash market of approximately 41 trillion USD in 2011 represent an important part of the economy1 . Policy makers, regulators and investors increasingly look at sovereign CDS prices to gauge the health of our system2 . While there have been several treatments about the corporate CDS market, we feel that there is a void when it comes to a comprehensive overview of the sovereign analogue3 . The purpose of this document is to fill this gap. More specifically, the goal of this paper is to provide an integrated overview of the key stylized facts of the sovereign CDS market, its players, frictions, the determinants of prices and trading related information. This is executed by reviewing the very young but growing literature, and discussing mainly the publicly available data from the Depository Trust and Clearing Corporation (DTCC), the Bank for International Settlements (BIS) and a historical dataset provided by Markit. As the focus of this review is the synthetic fixed income market, we mainly refer in citations to academic work on sovereign CDS spreads, but refer at times to academic work on sovereign cash bonds4 . We hope that this review provides a useful starting point for anyone interested in the topic and that our exposition raises some thought-provoking questions. At the outset, we would like to note some puzzling observations. Despite the huge press coverage of credit derivatives, they represent 1 The Economist global debt clock displays an estimate of current global public debt. Source: www.economist.com. Hart and Zingales (Fall 2011) suggest the use of financial CDS to evaluate the health of banks. 3 See for example Stulz (2010) for an overview of corporate CDS and their relation to the financial crisis or Das and Hanouna (2006). 4 As our goal is specifically the treatment of sovereign CDS spreads, we apologize to any authors who feel that their work is relevant and should have been included in this review. 2 1 only a tiny fraction (4.5%) of the overall ”over-the-counter” derivatives market. Moreover, among all traded credit default swaps, sovereign contracts make up roughly 9%. Although the overall size of the CDS market is valued at 32 trillion USD, the net economic exposure is only 1.3 trillion5 . Furthermore, why is it that contracts on emerging market economies tend to trade in high numbers and small volumes, while developed economies tend to trade with a larger net economic exposure per contract and have fewer contracts outstanding? Why are hedge funds so often blamed for speculation in the sovereign CDS market, when in fact they represent only a small fraction of all counterparties? Why is there such a surge in CDS trading on the US government in 2011? What determines the shape of the sovereign yield curve when government term structures exhibit similar patterns as those of corporations, but country-specific factors play relatively weaker roles in explaining variation in CDS spread levels. And finally, why is it so important to avoid a trigger on Greek CDS when the net reported economic exposure is below 4 billion at the end of 2011? These facts and more are highlighted along the road. The remainder of this document structures as follows. The second section explains the nature and mechanism of CDS. Readers familiar with this may directly want to jump to section three, which describes the market for sovereign CDS. Section four illustrates statistical stylized facts about sovereign CDS prices. In section five, we focus on the determinants of these products and review the literature. We follow up with a discussion on the relationship between sovereign CDS and cash bonds, and CDS related frictions in section six. Finally, in section seven, we give an overview of the trading in sovereign CDS markets based on the newly available data from the global data repository DTCC, and provide some indication on how this relates to country specific fundamentals. Finally, in section eight, we conclude. 2 What Are Credit Default Swaps Engineered in the early nineties by JP Morgan to meet the growing demand to slice and dice credit risk6 , credit default swaps represent the most simple (”vanilla”) instrument among the class of credit 5 The net economic exposure can be understood as an upper bound on insurance payments upon default. Actual payments are expected to be lower, because recovery values of defaulted bonds are generally non-zero. 6 The exact year of creation is not clear. Tett (2009) refers to the first CDS in 1994, when J.P. Morgan off-loaded its credit risk exposure on Exxon by paying a fee to the European Bank for Reconstruction and Development. More 2 derivatives. Nevertheless, they remain till this date highly controversial. While the proponents would defend them as efficient vehicles to transfer and manage credit risk as well as means to widen the investment opportunity set of return chasers, opponents denounce them as weapons of mass destruction, time bombs7 , financial hydrogen bombs8 or speculative bets on government default. For the sake of this review, let us stick with the factual definition of what they really are: insurance contracts offering protection against the default of a referenced sovereign government or corporation. Technically speaking, a credit default swap is a fixed income derivative instrument, which allows a protection buyer to purchase insurance against a contingent credit event on an underlying reference entity by paying an annuity premium to the protection seller, generally referred to as the Credit Default Swap spread. The premium is usually defined as a percentage on the notional amount insured and can be paid in quarterly or bi-annual installments. Readers not familiar with this particular notion of insurance, should simply bear in mind the example of car insurance. But whereas for car insurance, the contract would be written on a specific vehicle, credit insurance could be contracted on a given brand, think of a Volkswagen for the sake of the argument. Thus, any Volkswagen involved in an accident would trigger a payout to the holder of the contract. In the language of credit derivatives, you would purchase a CDS on a company, the reference entity, and if that company fails to meet its obligations for any of a predetermined set of its debt obligations, the payout would occur. More specifically, the contract comprises usually a specific class of the firm’s capital structure, such as the senior unsecured or the junior debt portion of the company, and references a particular amount of insured debt, the notional amount. Failure to meet the debt obligations is labeled a credit event. Hence, a credit event triggers a payment by the protection seller to the insuree equal to the difference between the notional principal and the loss upon default of the underlying reference obligation, also called the mid-market value. In general, the occurrence of a credit event must be documented by public notice and notified to the investor by the protection buyer. Amid the class of qualifying credit events are Bankruptcy, Failure importantly for this review, the author also references a deal between J.P. Morgan and Citibank asset management on the credit risk of Belgian, Italian and Swedish government bonds around the same time (p.48). 7 See Warren Buffett, Berkshire Hathaway Annual Report for 2002, p.13, on line at http://www.berkshirehathaway.com/2002ar/2002ar.pdf 8 Felix Rohatyn, a Wall Street banker emplyed at Lazard Frères, quoted in Tett (2009). 3 to pay, Obligation default or acceleration, Repudiation or moratorium (for sovereign entities) and Restructuring, and represent thus a broader definition of distress than the more general form of Chapter 7 or Chapter 11 bankruptcy. The settlement of insurance contracts comes in two forms, it may be either physically or cash settled. In case of a cash settlement, the monetary exchange involves only the actual incurred losses and the claimant holds on to the physical piece of paper providing him with a claim on the capital structure of the sovereign balance sheet. On the other hand, if there is a physical settlement, the claimant may hand any reference obligation referenced in the contractual agreement to the insurer and obtains the full notional amount of the underlying contract in return. The insurer can then try to maximize its resale value in the secondary market. This implies that the claimant literally holds a Cheapest-to-Deliver option, as he may deliver the least valuable eligible reference obligation9 . This option is particularly relevant in the case of a corporate restructuring, which is why the restructuring credit event is the most critical. As a consequence, it has been adapted numerous times and comes in multiple colors and associated contractual clauses. In particular, apart from Full Restructuring, which allows the delivery of any obligation with a remaining maturity up to 30 years, a contract may simply omit this definition, in which case it is labeled No Restructuring, or it may qualify as Modified Restructuring or even Modified Modified Restructuring. The former definition is more prevalent in the U.S. and limits the remaining maturity of the deliverable reference obligation to a maximum of 30 months, whereas the latter is predominant in Europe, but restricts the maximum maturity to 60 months. Even after default, prices of bonds usually suffer from wide market fluctuations, so the question remains as to how the precise value of the insurance settlement is determined. Markets have converged to a practice where the mid-market value is obtained through a dealer poll. Whether this pricing mechanism is efficient, remains unclear. Three recent papers have taken a closer look at this issue for corporate CDS: Helwege et al. (2009), Chernov et al. (2011) and Gupta and Sundaram (2012). Credit default swaps are part of the over-the-counter market (OTC). Thus they are not trading on an organized exchange and remain to a large extent unregulated. The reason for this is a 9 See Ammer and Cai (2011) for evidence on sovereign CDS and Jankowitsch et al. (2008) for evidence on corporate CDS. Ammer and Cai (2011) also study the short-run dynamics between sovereign CDS spreads and their cash counterparts, in a similar fashion as Blanco et al. (2005) did for the corporate market. 4 provision inserted in 2000 in the Commodity Futures Modernization Act by Senator Phil Gramm, who from 1995 to 2000 presided over the Senate Banking Committee, exempting CDS from regulation by the Commodity Futures Trading Commission (CFTC)10 . Nevertheless, guidance on the legal and institutional details of CDS contracts is given by the International Swaps and Derivatives Association (ISDA)11 . They also act as a non-voting secretary for the Credit Determination Committee, which deliberates over issues involving Reference Entities traded under Transaction Types that relate to the relevant region, including Credit Events, CDS Auctions, Succession Events and other issues. Moreover, ISDA has played a significant role in the growth of the CDS market by providing a standardized contract in the 1998 (”the ISDA Master Agreement”), which was updated in 2002, and supplemented with the 2003 Credit Derivatives Definitions (”The Definitions”) and the July 2009 Supplement. The contractual details of these contracts are crucial, and as usual, the devil lies in the details, as was recently proved in the restructuring case of Greek government debt. European officials have heavily pushed towards a voluntary restructuring, which would not be binding on all bondholders. In such a case, it would not be be likely that the agreed deal would trigger payments under existing CDS contracts12 . The landscape for CDS has further altered with the implementation of the CDS Big Bang and Small Bang protocols on 8 April and 20 June 2009 for the American and European CDS markets respectively. The biggest changes relate to the standardization of the coupon structure and the settlement process. Going forward, the Dodd-Frank Act in the U.S. will likewise have a significant impact on the way these products are traded. The Inter-continental Exchange (ICE), a subsidiary of the NYSE, is already recording growing market share in the clearing process and both academic and political voices call for a move towards organized exchanges, more transparency and more orderly price dissemination13 . Before moving on to the statistical features of the market for sovereign CDS, we want to point out that we don’t describe their pricing mechanism. The interested reader may however refer to Duffie (1999), Hull and White (2000a) and Lando (2004), who provide excellent descriptions of these details. 10 See Nouriel Roubini and Stephen Mihm, Crisis Economics, pp.199. See www.isda.org. 12 See Greek Sovereign Debt Q&A, October 31 2011, www.isda.org 13 Another growing clearing platform for CDS is provided by the CME. 11 5 3 The Market for Sovereign Credit Default Swaps The last decade was accompanied with a spectacular growth in the market for credit default swaps, nearing three-digit growth rates year by year. The Bank for International Settlements (BIS) publishes bi-annual statistics on the notional amounts outstanding and gross market values of OTC derivatives14 , and also includes credit default swaps in their reporting since 2004. As can be seen in Table (1), the market grew from approximately 6 trillion USD of notional amount outstanding to a peak of 58 trillion USD in the second term of 2007. Subsequent surveys show a significant drop in trading with figures down to 32 trillion USD in June 2011, which is mostly linked due to a netting of outstanding positions as regulators’ increasing concerns about counterparty risk and calls for higher transparency have led to portfolio compressions, but also partly due to the fact that credit derivatives were central to the 2007-2009 financial crisis15 . While contracts written on a single underlying reference made up for the bulk of the trading volume in 2004 with a market share of 80%, multi-name instruments have increased more rapidly due to the practice of synthetic securitization and represented roughly 44% of the market in 2011. While these numbers sound impressive, the reader should bear in mind that credit default swaps still account for only a small fraction of the overall OTC market, as is illustrated in Table (2). With a total trading volume of roughly 32 trillion USD in June 2011, credit default swaps represent merely 4.58% of the total amount of notional oustanding in all OTC derivatives, which has grown to 707.57 trillion USD. The largest fraction comes from interest rate derivatives, which is estimated to have a size 553.88 trillion USD, or equivalently 78.28% of the market. Nominal or notional amounts outstanding provide a measure of market size and a reference from which contractual payments are determined in derivatives markets. However, such amounts are generally not those truly at risk. The amounts at risk in derivatives contracts are a function of the price level and/or volatility of the financial reference index used in the determination of contract payments, the duration and liquidity of contracts, and the creditworthiness of counterparties. They 14 The notional amounts are likely to slightly underestimate the total value of the market, as until the end of 2011, only 11 countries are reporting their OTC derivative statistics to the BIS. Belgium, Canada, France, Germany, Italy, Japan, the Netherlands, Sweden, Switzerland, the United Kingdom and the United States. From December 2011, Australia and Spain are expected to contribute to the semiannual survey, bringing to 13 the number of reporting countries. 15 Portfolio compression refers to the process by which two counterparties maintain the same risk profile, but reduce the number of contracts and gross notional value held by participants. 6 are also a function of whether an exchange of notional principal takes place between counterparties. A measure commonly reported along the lines is gross market value, which provides a more accurate measure of the scale of financial risk transfer taking place in derivatives markets. As is reported in the BIS statistical notes, ”gross market values are defined as the sums of the absolute values of all open contracts with either positive or negative replacement values evaluated at market prices prevailing on the reporting date. Thus, the gross positive market value of a dealers outstanding contracts is the sum of the replacement values of all contracts that are in a current gain position to the reporter at current market prices (and therefore, if they were settled immediately, would represent claims on counterparties). The gross negative market value is the sum of the values of all contracts that have a negative value on the reporting date (i.e. those that are in a current loss position and therefore, if they were settled immediately, would represent liabilities of the dealer to its counterparties). The fact that these numbers are gross merely reflects the fact that contractual positions of positive and negative values with the same counterparty are not netted out”. Taking a look at the final column of Table (2), it is possible to see that the gross market value of credit default swaps is 1.35 trillion USD during the first half of 2010, which is still a sizable market, but this number is much less breathtaking than the total notional amount outstanding of 32 trillion USD. Going a step further, gross credit exposure, which takes account of all legally enforceable netting agreements, makes this number shrink further to a value of 375 billion USD, as can be seen in the last column of Table (3). This document is however about sovereign CDS, so we may wonder what these numbers look like for insurance contracts written on government debt. Statistics on sovereign CDS have until recently been much less detailed, but what is clear from Table (4) is that the amount of trading in the sovereign CDS market with 2.91 trillion USD in 2011 represents approximately 9% of the overall market for OTC credit derivatives. In contrast to corporate CDS, they are on the other hand largely concentrated around single-name products, who reflect a trading volume of 2.75 trillion USD, or respectively a fraction of 95%. While more detailed statistics for gross credit exposures of sovereign CDS are not publicly available on the BIS website, we can get a noisy estimate if we make the simplifying assumption that the fraction of gross credit exposure to total notional amount traded is the same for sovereign reference entities and the overall market. Under this assumption, 7 the gross market value of all oustanding contracts is approximately 121 billion USD, while the total net exposure reduces to roughly 33.6 billion USD. This is still an economically meaningful number, but admittedly too small to justify interference in the market to avoid a payout trigger, unless it is for other indirect reasons. So who trades these products. Hedge funds are often blamed for their speculative one-sided bets, but if we look at the counterparties involved in sovereign CDS trading in Table (5), we see that it is to a large extent the reporting dealers, who operate in this market. Their reported trading volume is approximately 1.8 trillion USD, reflecting a market share of 66.81%, followed second by Banks and Security Firms with a market share of 21.54% or an underlying notional amount outstanding of 592 billion USD. Hedge funds are only third in line, with a total volume of 146 billion USD. While this is no evidence, it suggests at least that sovereign CDS are used primarily for hedging purposes, rather than for speculative bets16 . We should however be cautious with such fast conclusions, as their trading volume has almost doubled from 2010 to 2011, which reflect arbitrage opportunities linked to the European sovereign debt crisis, wheras banks and security firms have reduced their exposure from 828 billion USD to 592 billion USD over the same time period. In contrast to cash bonds, CDS have the attractive feature that they are constant maturity products. This means that on every day, a market participant can buy an insurance contract with an identical remaining maturity. Bonds, on the other hand, are issued only seldomly, and their maturity decreases as time goes by. In addition, an investor has access to a full set of maturities, as CDS trade usually from maturities of 6 months up to about 30 years. However, liquidity is still mostly concentrated around five years17 . Table (6) shows that in 2011, the total volume of notional amount outstanding with maturities above one and up to five years, was 23.19 trillion USD, representing a total market share of 71.57%. This fraction has fluctuated closely around this level since the BIS has started publishing their reports about CDS. The remaining contracts are well balanced among maturities below one year, with a total volume 3.92 trillion USD or 12.11%, or above five years, with a total volume of 5.29 trillion USD or 16.32%18 . 16 Bongaerts et al. (2011) also cite evidence that banks are primarily net buyers of (corporate) CDS, while insurance companies and funds are mainly net sellers. 17 The following numbers are for the overall CDS market, and not specifically for sovereign CDS. 18 In fact, we should emphasize that for sovereign reference entities, trading is not as much concentraed in the 5-year contract as it is the case for corporate reference entities. Evidence is also provided in Pan and Singleton (2008) and Packer and Suthiphongchai (2003). 8 Last, but not least, Table (7) illustrates that the CDS market is mostly a transnational market. 81.70% trade with a counterparty based in a foreign country, in contrast to only 18.30% of all contracts that are handled with a counterparty of the home country. The international nature of the trading raises questions about the usefulness of unilateral policy decisions affecting the market, such as the ban of naked CDS by Germany in May 2010. 4 Statistical Facts about Sovereign CDS Beyond knowing the size of the market, which parties trade and with whom, a comprehensive overview should also provide some stylized statistical facts about the probability distributions of sovereign CDS. The following statistics are based on a sample of daily mid composite USD denominated CDS quotes for 38 sovereign countries taken from Markit over the sample period 9 May 2003 until 19 August 2010, where all contracts contain the full restructuring clause. The sample covers prices for a wide selection of the term structure, including 1, 2, 3, 5, 7 and 10-year contracts and spans 5 geographical regions, including the Americas, Europe, Africa, the Middle East and Asia. A first set of summary statistics is provided in Table (8), where we have grouped the countries in rating categories according to their average credit rating on external debt over the sample period. The average term structure is always upward sloping, going from 15 basis points at the 1-year horizon to 26 basis points at the ten-year maturity for AAA rated entities, and from 433 to 599 basis points for the B rated countries. A mean upward sloping term structure was also shown for a set of three emerging countries in Pan and Singleton (2008). As the countries become less creditworthy, the volatility of their CDS spreads jumps up as well. For example, the average standard deviation of the 5-year CDS spread is 34 basis points for the most creditworthy country, and 328 basis points for the least creditworthy. Moreover, the level of CDS spreads exhibit positive skewness, with a value usually around 2, and they have very fat tails. But the excess kurtosis generally decreases with the asset horizon. Finally, CDS spreads are also very persistent. Among all statistics, the lowest and highest first-order autocorrelation coefficients are 0.9901 and 0.9979 respectively for daily observations. 9 While this measures provide a rough overview of the statistical stylized facts, a common theme of interest among credit instruments is the slope of the term structure. A structural credit model of Merton (1974) predicts that the credit curve is increasing for high-quality credit levels, humpshaped for intermediate credit quality and decreasing for low levels of creditworthiness. Lando and Mortensen (2005) have confirmed these predictions for corporate CDS. For their analysis, they rank their observations based on the average five-year spread level. We attempt to do the same exercise here for our sovereign CDS sample. Hence for each country on each day, we look at the 5-year CDS spread and group the observations into ranges of 100 basis points and then report within each category, the average spread, as well as the difference between the 10-year and the 1-year spread, the 10-year and the 3-year spread and the 3-year and the 1-year spread. The average values are calculated across time and entities. The results, which are shown in Table (9), yield similar conclusions as for corporate CDS. The overall slope, defined as the difference between the 10-year and 1-year spread, first increases as the credit quality weakens, and then decreases before becoming negative at very high spread observations. Yet, if we look separately at the gap between the 1 to 3-year, as well as the 3 to 10-year spread, it is evident that the former turns negative only much later, confirming the humped shape curve for intermediate groups. Another way to look at this is by computing the average slope levels in the same three segments, as well as the fraction of negative values for sliding groups of 400 curves. This yields the outcome plotted in Figure (1), from which we can draw the same conclusion. At first glance, such results make us feel comfortable that statistical models of credit curves can be applied equally well to corporate and sovereign reference contracts. From an empirical perspective, however, it raises the question about the determinants of the shape of the sovereign credit curve, as country-specific factors seem to have little explanatory power for the time-series variation of sovereign CDS, at least at higher sampling frequencies. This contrasts with the predictions of a Merton (Merton 1974) type world. We will address this issue in the following section. 10 5 Determinants of Sovereign CDS spreads Up to this point, it remains to a large extent unclear what is driving variations in sovereigns CDS spreads. Let us again go back to the classical Merton model to frame our thoughts. According to this set-up, the credit spread of a company should be determined by the asset volatility of the firm’s balance sheet. Or in other terms, characteristics specific to each firm should influence its likelihood of default. To make the analogy with sovereigns, we would also intuitively expect the fluctuations in default probabilities to be driven by country-specific fundamentals. Yet, it turns out that a major fraction of the variation in sovereign CDS spreads is determined by a few global variables, at least at higher trading frequencies. Moreover, most studies reference some sort of risk originating in the United States. Where the camps of thought are divided, is whether these global factors are related to fundamental macroeconomic risk, or rather to financial risk. There is a long strand of literature, which has tried to understand bond spreads. Here, apart a few minor exceptions, we will only focus on derivative contracts written on government debt. On average, sovereign CDS spreads generally have a downward trend, with sharp spikes every time there is a run-up in risk aversion due to a global risk-related event. Figure (2b) illustrates this more clearly by plotting the average five-year CDS spread for the sample of 38 countries, which we have discussed in the previous section. Indeed, the average spread quickly rises each time the global economy suffers a major shock, such as changes in U.S. monetary policy following falling growth rates, which have led traders to unwind their carry trades, major corporate bankruptcies of the usual suspects Bear Stearns and Lehman Brothers, or when the European governments underwent major political bailouts. Of course, it might be that the risk premium embedded in spreads, compensating investors for unpredictable variations in future default rates, is driven by one set of determinants, while expected losses are a function of others. But even for this point, the literature is at odds. It might be useful to review the papers who have analyzed this question. Sovereign CDS spreads comove significantly over time, as can be seen in graph (2a) for the 38 countries studied previously. Performing a principal component analysis on a sample of sovereign CDS spreads generally reveals that the first principal component explains a very large fraction of the variation, much higher than is the case for equities. The precise figure depends on sample 11 frequency and countries studied, but can be as high as 96% at the daily level or 64% at a monthly decision interval as shown by Pan and Singleton (2008) and Longstaff et al. (2010) respectively19 . This strong common component motivates the holy grail journey of what this common factor ought to be. Among those studies leaning more towards an explanation of global factors is for example Longstaff et al. (2010). The authors carefully study monthly 5-year CDS of 26 countries over the time period October 2000 to January 2010 and consider variables related to the local economy, global financial variables, global risk premia, global investment flow measures and regional spreads. They conclude that sovereign CDS can be explained to a large extent by U.S. equity, volatility, and bond market risk premia. A tight link between sovereign CDS spreads is also reported by Pan and Singleton (2008), who study the term-structure of daily CDS spreads of Korea, Mexico and Turkey from March 2001 to August 2006. Their main interest lies in disentangling the risk neutral default and recovery processes implicit in the term structure. With this goal in mind, they develop a theoretical pricing model to decompose the spread into the part related to expected losses, and the risk premium. Risk premia appear to comove strongly over time and are strongly related to the CBOE VIX option volatility index, the spread between the 10-year return on US BB-rated industrial corporate bonds and the 6-month US Treasury bill rate, and the volatility in the owncurrency options market. These findings corroborate the intuition of time-varying risk premia in the sovereign CDS market. Ang and Longstaff (2011) conduct a comparative analysis of CDS spreads written on sovereign states in the United States and European countries. Rather than extracting the risk premium, they extract the dependence of spreads on a common component and define this part as systemic risk. The authors find evidence that this systemic risk component is influenced by global financial factors and dismiss links with macroeconomic fundamentals. Related to the paper by Pan and Singleton (2008) is the work of Zhang (2003), who develops a CDS pricing model and applies it to Argentina to infer differences in expectations about expected recovery rates upon default and expected default probabilities. In addition, the author also documents that the wedge between risk-adjusted and historical default probabilities are associated with changes in the business cycle, both the U.S. and 19 Pan and Singleton (2008) study 3 countries. Augustin and Tédongap (2011) report that the first principal component explains 78% of the daily variation for a set of 38 countries. 12 Argentine credit conditions as well as the overall local economy. Augustin and Tédongap (2011) provide empirical evidence that the first two common components among a set of 38 countries spanning a wide geographical region are strongly associated with expected consumption growth and macroeconomic uncertainty in the United States. The authors study daily quotes over the time period April 2003 to August 2010 and document that this link is robust against the influence of global financial market variables such as the CBOE volatility index or the Variance Risk Premium. In addition, they rationalize this new finding in a structural asset pricing model with recursive preferences and a long-run risk economy where the default rate process is driven by the global longrun expectations of future consumption growth and macroeconomic uncertainty. Further evidence of influence from the United States on sovereign CDS premia is provided by Dooley and Hutchison (2009), who document how the subprime crisis channeled through to a sample of 14 geographically dispersed countries, based on a series of both positive and negative real and financial news over the sample period January 2007 to February 2009. While they document that a range of news had statistically and economically significant effects, they find that in particular the Lehman event and the expansion of Federal Reserve swap lines with the central banks of industrial and emerging countries uniformly moved all country spreads (in the sample)20 . Finally, Wang and Moore (2012) study the dynamic correlations between the sovereign CDS spreads of 38 emerging and developed economies with the U.S. from January 2007 to December 2009. Results indicate that in particular developed economies have become much more integrated with the United States since the Lehman shock. This tighter link with the U.S. seems to be driven mainly by the U.S. interest rate channel. A good transition between the ”pro-global” and ”pro-local” camps is provided by Remolona et al. (2008). Based on the intuition that financial asset prices are driven by country-specific fundamentals and investor’s appetite for risk, Remolona et al. (2008) decompose monthly 5-year emerging markets sovereign CDS spreads over the period January 2002 to May 2006 into a Market-based measure of expected loss and a risk premium to analyze how each of these two elements is linked to measures of country-specific risk and measure of global risk aversion/risk appetite. Fundamental variables include inflation, industrial production, GDP growth consensus forecasts, and foreign 20 On 13 October 2008, the Fed removed its USD swap limits to industrial countries, and on 29 October 2008, the FOMC established swap lines with the central banks of Brazil, Mexico, Korea and Singapore for up to $30 billion each. 13 exchange reserves. Proxies for global risk aversion are taken as the Tsatsaronis, Karamptatos (2003) effective risk appetite indicator, the VIX and a Risk Tolerance Index by JP Morgan Chase. They find empirical evidence that global risk aversion is the dominant determinant of sovereign risk premium, while country-specific fundamentals and market liquidity matter more for sovereign risk. Both components thus behave differently. Carlos Caceres and Segoviano (2010) analyze how much of the rising spreads can be explained by shifts in global risk aversion, country-specific risks, directly from worsening fundamentals, or indirectly from spillovers originating in other sovereigns. The authors argue that the widening of spreads during the early period of the crisis was essentially driven by changes in a self-computed measure of risk aversion, but later in the crisis, countryspecific factors identified by each countrys stock of public debt and budget deficit as a share of GDP played the dominant role. Carr and Wu (2007) present a joint valuation framework for sovereign CDS and currency options written on the same economy and execute an empirical test for Mexico and Brazil over the time period January 2, 2002 to March 2, 2005. They document that CDS spreads show strong positive contemporaneous correlations with both the currency option implied volatility and the slope of the implied volatility curve in moneyness21 . This analysis confirms their conjecture that economic or political instability often leads to both worsened sovereign credit quality and aggravated currency return volatility in a sovereign country. Interestingly, their results suggest that there are additional systematic movements in the credit spreads that the estimated model fails to capture. Hui and Fong (2011) reverses the analysis and documents information flow from the sovereign CDS market to the dollar-yen currency option market during the sovereign debt crisis from September 2009 to August 2011. in a similar spirit, Pu (2012) exploits pricing differences between USD and EUR denominated sovereign CDS spreads to show that the dual-currency spread significantly affects the bilateral exchange rate for ten Eurozone countries from January 2008 to December 2010 and has predictive power up to a perdio of ten days. Gray and Bodie (2007) apply the contingent claims analysis to price sovereign credit risk by using the government’s balance sheet and exchange rate volatility as inputs to their model and compare the results to observed CDS spreads. More recently, Plank (2010) also suggest a structural model for sovereign credit risk based on macroeconomic fundamentals. In 21 More specifically, the authors refer to the delta-neutral straddle implied volatilities and the risk reversals. 14 this model, the probability of default reflected in the CDS spread depends on a country’s foreign exchange reserves, as well as its exports and imports. Ismailescu and Kazemi (2010) consider the possibility of rating changes affecting the sovereign CDS spreads of 22 emerging economies over the time period January 2, 2001 to April 22, 200922 . Using classic event study analysis, the authors conclude that the CDS of investment grade countries respond mainly to negative credit rating announcements, while the spreads of speculative grade countries respond largely to positive announcements. In addition, the information content of negative credit events is anticipated and already reflected in CDS spreads by the time the credit rating change is announced. Alternatively, CDS spreads do not fully anticipate upcoming positive credit rating events, which seem to contain new information. Furthermore, the paper documents evidence of spillover effects arising mainly from rating upgrades, with a one notch rise in the rating of an event country increasing the average CDS spread of a non-event country by 1.18%. These effects are asymmetric, high grade countries reacting stronger to rating upgrades and low-grade countries reacting stronger to rating downgrades. Although controlling for previous rating downgrades reduces the magnitude of the spillover effect, it is increased if countries have common a common creditor, which seems to be one channel of transmission. However, competition effects arise if two countries have similar levels of trade flow correlation with the U.S. There is also a class papers documenting a link between sovereign credit risk implicit in the CDS price and the health of the financial system. One such example is Acharya et al. (2011), who illustrate the two-way feedback effect between sovereign risk and the financial sector, where government bailouts negatively impact the financial strength of the sovereign, which in turn reduces the value of the government guarantees implicit in the financial institutions. As a consequence, the CDS spreads of sovereign countries and financial companies co-move strongly once the government has committed excessively to financial guarantees. Altman and Rijken (2011) propose a ”bottomup” approach to evaluate sovereign default probabilities by adapting the well-know credit-scoring method to countries. This suggests that the local real economy significantly impacts default risk. The outcomes are compared to implied CDS rankings provided by the financial market. Dieckmann and Plank (2011) study eighteen advanced western economies on a weekly basis during the 22 For early evidence on the relationship between credit ratings and pricing information, see Cossin and Jung (2005), who show that ratings become more informative after a crisis event. 15 2007/2008 financial crisis23 . They find that both the state of a country’s financial system and of the world financial system have strong explanatory power for the behavior of CDS spreads, while the magnitude of this impact seems to depend on the importance of a country’s financial system pre-crisis. In addition, they provide evidence that Economic and Monetary Union member countries exhibit higher sensitivities to the health of the financial system. While Acharya et al. (2011) stress the two-way feedback effect, the authors here underscore the unilateral private-to-public risk transfer through which market participants incorporate their expectations about financial industry bailouts. Evidence that contingent liabilities of sovereigns arising from implicit or explicit guarantees of the banking system influence sovereign CDS premia is also provided in Kallestrup et al. (2011). Also Kallestrup (2011) shows that there are strong interactions between sovereign credit risk and macrofinancial risk indicators calculated based on bank balance sheet variables. Sgherri and Zoli (2009) conduct an analysis of the co-movement between sovereign CDS and those of the financial sector for ten European countries from January 1999 to April 2009. While they confirm the dominance of a common time-varying factor, they find that concerns about the solvency of the national banking systems have become increasingly important. Finally, Brutti and Sauré (2012) show how financial shocks to Greece spill over to eleven other European economies. The magnitude of contagion depends on the cross-country bank exposures to sovereign debt. All else being equal, the transmission rate to the country with the greatest exposure to Greece (1.22 percent of GDP) has been roughly 46 percent higher than the rate to the country with the least exposure (0.08 percent of GDP). 6 24 . The relationship between sovereign CDS and bonds and Frictions Duffie (1999) illustrates theoretically that the spread on a par floating rate note over a risk-free benchmark should essentially be equal to the CDS spread25 . As such, the CDS-Bond basis, defined as the difference between the CDS and bond spread, should be zero. However, empirically observed 23 The exact time period is January 2007 to 2010. Further evidence on the relationship between sovereign and bank CDS is provided by Ejsing and Lemke (2011), Alter and Schuler (2011) and Aktug and Vasconcellos (2010) 25 Other references are Lando (2004) and Hull and White (2000a). 24 16 imperfections of this theoretical arbitrage relationship has led researches to investigate whether the CDS market is more informationally efficient than the cash bond market and whether it reflects new information more quickly, or vice-versa. Alternatively, people have looked at the determinants of the basis, which relates explicitly to limits of arbitrage and frictions in one of the two markets, or both. Regarding price discovery for the sovereign asset class, there has recently been a proliferation of papers addressing the question of which market is more efficient because of the turbulences in the sovereign debt market. Abstracting from technical details, these papers generally have the common approach to test whether the two markets are co-integrated, that is whether they are characterized by a long-run stationary relationship, and then to look at short-term deviations from equilibrium to verify which market adapts to the other one. This is executed using the techniques of vector error correction modeling, computing Hasbroucks’s and Gonazalo-Granger’s information measures, or by analyzing statistical causality using Granger’s method. While for corporate reference entities, there is more or less consent about the CDS market being more efficient, results for sovereign reference entities are very mixed and ambiguous. These discrepancies are certainly related to the different samples, time periods, sampling frequency and data sources. While some conclude in favor of the bond market and others in favor of the CDS market, our interpretation of the literature is that there is increasing price discovery in the credit derivative market. Arce et al. (2011) make the sensible point that price discovery is state dependent and related to the relative liquidity in both markets. As the CDS market has become more mature over time, this would also explain why its relative informational efficiency has increased. A similar argument is also made in Ammer and Cai (2011), who provide some evidence that the relatively more liquid market tends to lead the other, as their measure of relative CDS price leadership is correlated positively with the ratio of the bond bid-ask spread to the CDS bid-ask spread and negatively with the number of bonds outstanding. Our further overall interpretation is that a positive basis is usually more persistent, due to the difficulty of shorting bonds. Moreover, except for some minor exceptions, almost all studies focus exclusively on the most liquid 5-year segment of the market. Without spending too much time on the details, we provide a shopping list of the references in the following paragraph. Palladini and Portes (2011) study 6 developed economies in Europe from 30 January 2004 17 through 11 March 2011 using CMA quote data and conclude that CDS prices lead in the price discovery process26 . A similar conclusion is drawn by Varga (2009), who takes a special look at Hungary from 6 February 2004 to 18 June 2008 using the CMA database27 . Using likewise the CMA data, O’Kane (2012) studies cross-correlations and performs Granger causality tests for 6 European economies from 1 January 2008 to 1 September 201128 . He finds that for Greece and Spain, Granger causality flows from the CDS to to the bond market, wile the results is reversed for Italy and France. For Portugal and Ireland, there seems to be two-way causality. Studying 18 sovereign developed and emerging economies from January 2007 to March 2010, Coudert and Gex (2011) conclude that bonds seem to lead for developed European economies, while the derivative market wins the horse race in emerging economies29 . The authors also put forth a differential liquidity argument and argue that the role of CDS has intensified during the crisis, as bond market players with a bearish view will simply stay out of the markets, while CDS buyers will stay in and purchase protection. This contrasts with Arce et al. (2011), who argue that the role of bond markets has increased for European Economies during the financial crisis. They study 11 European economies from January 2004 to September 2010 using CMA data30 . Fontana and Scheicher (2010) look at 10 Euro area countries from January 2006 to June 2010 using again CMA data31 . In contrast to most other studies, which use the 5-year spread, the authors study 10-year spreads. In line with the previous mixed results, in half of the sample countries, price discovery takes place in the CDS market and in the other half, price discovery is observed in the bond market. Further evidence is provided by M. Kabir Hassan and Suk-Yu (2011), who study 7 emerging economies from January 2004 to October 2009 using data from JP Morgan Chase32 . They find that bond markets lead for Brazil and the Philippines, CDS markets lead for Argentina, and there is no clear dominance for Mexico, China, and South Africa. The additional contribution is to link pricing efficiency to 26 The 6 countries are Austria, Belgium, Greece, Ireland, Italy, and Portugal. While Varga focuses on Hungary, some additional analysis is also provided for 14 emerging countries from 3 January 2005 to 30 May 2008. These countries are Brazil, Bulgaria, Czech Republic, South Africa, Estonia, Croatia, Poland, Latvia, Lituania, Russia, Romania, Slovkia, Turkey and Ukraine. 28 The 6 countries are France, Greece, Italy, Ireland, Portugal and Spain. 29 See also Coudert and Gex (2010). The sample includes Austria, Belgium, Denmark, Finland, France, Netherlands, Greece, Ireland, Italy, Portugal and Spain for the advanced European countries; Argentina, Brazil, Mexico, Lithuania, Poland, Turkey and Philippines for the emerging countries; in addition the authors study 17 financial companies. 30 The 11 EMU countries are Austria, Belgium, Finland, France, Germany, Greece, Ireland, Italy, the Netherlands, Portugal and Spain. 31 Austria, Belgium, France, Germany, Greece, Ireland, Italy, Netherlands, Portugal and Spain 32 The countries are Argentina, Brazil, China, Colombia, Mexico, Philippines and South Africa. 27 18 financial integration. It is argued that sovereigns, for which bond markets lead the price discovery process, are more financially integrated with the world economy. Li and Huang (2011) witness an increasing contribution of sovereign CDS rates to credit risk discovery based on a sample of 22 emerging countries using Thomson reuters data from 1 January 2004 to 31 July 200833 . C. Emre Alper and Gerard (2012), studying CDS premia in relation to asset swap spreads, conclude in favor of price discovery in the derivative market given a sample of 21 advanced economies from January 2008 to January 201134 . Their data comes both from Datastream and Markit. Aktug and Bae (2009) study 30 emerging markets at a monthly sampling frequency from January 2001 to November 2007 using data from Markit35 . They show that bond markets lead CDS markets in general, but that they lag CDS spreads in some cases, and that the markets have become more integrated over time. Support for the bond markets is also found in Ammer and Cai (2011). The latter investigate 9 emerging economies from 26 February 2001 to 31 March 2005 using Markit data36 . Overall, they find that the bond market leads the CDS market more often. In an early paper, Chan-Lau and Kim (2004) have a hard time concluding in favor of any market using CreditTrade data for 8 emerging economies from 19 March 2001 to 29 may 2003.37 . Delis and Mylonidis (2011) look at Greece, Italy, Portugal and Spain using CMA data from 9 July 2004 to 25 May 25 2010. Based on 10-year spreads, they argue that CDS almost uniformly Granger cause bond spreads. Feedback causality is, however, detected during times of intense financial and economic turbulences. Finally, additional elements of analysis can be found in In et al. (2007), Boone et al. (2010), Li (2009), Bowe (2009), Carboni (2010) and Adler and Song (2010). As a conclusion, results are thus very inconclusive, but differential liquidity seems to matter. Based on this intuition, Calice et al. (2011) investigate liquidity spillovers between the two markets 33 The sample is composed of Argentina, Brazil, Chile, China, Colombia, Czech Republic, Egypt, Hungary, Indonesia, Israel, Korea, Malaysia, Mexico, Morocco, Pakistan, Peru, Philippines, Poland, Russia, South Africa, Turkey, Venezuela. 34 The 22 countries are Australia, Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Japan, the Netherlands, Norway, Portugal, Spain, Sweden, the UK and the US. 35 They study Brazil, Bulgaria, Chile, China, Colombia, Croatia, Czech Republic, Dominican Republic, Ecuador, Egypt, El Salvador, Hungary, Korea, Lebanon, Malaysia, Mexico, Morocco, Pakistan, Panama, Peru, Philippines, Poland, Russia, South Africa, Thailand, Turkey, Ukraine, Venezuela and Vietnam. 36 Brazil, China, Colombia, Mexico, the Philippines, Russia, Turkey, Uruguay, and Venezuela. 37 The following countries are reviewed: Brazil, Bulgaria, Colombia, Mexico, the Philippines, Russia, Turkey, and Venezuela. 19 and find substantial variation in the patterns of the transmission effect between maturities and across countries. For several countries, including Greece, Ireland and Portugal the liquidity of the sovereign CDS market has a substantial time varying influence on sovereign bond credit spreads. In et al. (2007) study the volatility transmission among the two markets across emerging economies. While the dynamics between the cash and derivative market raise interesting questions, another angle to analyze is the determinants of the short-term deviations from the equilibrium relationship. Differential liquidity may naturally explain the seemingly arbitrage opportunity. Early studies on this topic assumed that the credit derivative market was perfectly liquid, as the underlying traded instruments are by nature simple contractual agreements, in contrast to the cash market, where hard money is actually exchanged when the asset is purchased. In this spirit, Longstaff et al. (2005) postulated CDS spreads to be pure indicators of default risk in order to infer liquidity characteristics of the bond market. Academic work on liquidity effects in the CDS market is very young, even for corporate reference contracts, but the literature is rapidly growing. Two papers that study liquidity characteristics in the synthetic credit derivative market for corporate reference entities are for example Tang and Yan (2007) and Bongaerts et al. (2011). The former show that both liquidity characteristics and liquidity risk explain about 20% of the time series variation in CDS spreads, while the latter provide strong evidence for an expected liquidity premium earned by the protection seller. The premium is however economically small. For sovereign CDS, Pan and Singleton (2008), who study the term structure of recovery rates of Mexico, Brazil and Turkey, report anecdotal evidence of liquidity effects from discussion with practitioners. Regarding the determinants of the CDS-bond basis, a liquidity story is supported by Arce et al. (2011), who show that differential liquidity between the bond and the CDS partially explains the difference. Specifically, differential liquidity is proxied by the ratio of the bid-ask to the mid spread in the CDS market divided by the bid-ask to the mid spread in the bond market. In addition, the authors show evidence for the role of counterparty risk, which is the risk that a CDS counterparty may not be able to honor its obligations from the insurance contract. This risk becomes especially important when the default of the counterparty occurs at the same time as the default event of the underlying reference entity. Arora et al. (2012) show that counterparty risk is significantly priced in the corporate market, but that it has economically a small magnitude. On average, counterparty credit risk needs to increase 20 by 646 basis points to reduce the insurance price by 1 basis point. This small effect is likely the result of collateralization (and maybe overcollateralization) agreements. In a similar spirit, Levy (2009) also links the pricing discrepancies to both counterparty risk and liquidity. Kucuk (2010) finds that the basis is strongly associated with various bond and CDS liquidity variables such as (for bonds and CDS) the average sample bid-ask spread and percentage bid-ask spread, and (for bonds) lagged trading volume, notional amount outstanding, age and time to maturity. Similar to the debate on global versus local factors for sovereign CDS spreads, Fontana and Scheicher (2010) relate the basis mainly to common factors. Another potential source of friction, which can create a wedge between the CDS and bond spread is the Cheapest-to-Deliver option embedded in CDS contracts. As we have already explained, upon the occurrence of a credit event, the holder of the insurance contract has the right to deliver the cheapest among a set of defaulted debt obligations to his counterparty in return for the insurance settlement. This option makes the contract riskier from the perspective of the CDS seller, and should be relatively more important, the closer a country is to default. Ammer and Cai (2011) provide a treatment of the sovereign case, while Jankowitsch et al. (2008) provide positive evidence for corporate CDS. For sovereigns, some early analysis is also available in Singh (2003). Fisher (2010) studies theoretically how heterogenuous investors with different beliefs about the probability of default of a sovereign borrower affects the sovereign CDS-Bond basis. The results of the model depend on the assumption that a fixed proportion of investors is unable to sell CDS. The latter constraint becomes binding in bad times, when a relatively bigger proportion of investors is pessimistic and wants to sell of credit risk (buying CDS). But as the supply is limited, this constraint creates a wedge between the bond and CDS prices and induces a positive basis. Thus, the time-varying equilibrium basis depends on the time-varying proportion of pessimistic investors. A second mechanism, that induces the (positive basis) in his model is the prospect of lending fees from the bonds in bad times. As supply in the CDS market is restricted, some pessimistic investors, who want to take a short position in sovereign credit risk need to short-sell the physical bond. This allows those, who hold on to the bond to charge higher fees by lending it out to short-sellers. This fee raises bond prices, lowers spreads, and creates a positive basis. Finally, Adler and Song (2010) provide evidence that short-selling costs were partly responsible for persist- 21 ing positive bases for sovereign reference entities. In addition, the authors provide a more general theoretical pricing framework for the basis, which corrects for biases arising from acrrued spread or coupon payments and bond prices away from their par value38 . In fact, the model illustrates how accrued payments and bond prices below (above) par may mechanically introduce a negative and positive (negative) basis respectively. It is exactly this latter effect, which may explain the well documented basis smile. The average credit quality of a portfolio of bonds depends on the statistical distribution of historical rating transitions. Hence, the joint consideration of price and credit quality will determine the shape of the basis smile, which could very well be a ”frown”. Another ”hot” topic about sovereign CDS is whether speculators manage to (negatively) impact the market. In May 2010, BaFin, the German financial regulator went ahead with an outright ban on naked sovereign CDS (and bond) positions, despite an official report failing to provide conclusive evidence (See Criado et al. (2010).). The academic opinions don’t seem to agree on this idea. Richard Portes (see Portes (2010)) has advocated a ban, arguing that naked CDS buying (equivalent to shorting the bond) artificially drives up prices, thereby increasing a country’s borrowing costs. This opinion seems to be shared in a study by Delatte et al. (2012). Darrell Duffie on the other hand (Duffie (April 29 2010a, 2010b)) dismisses the argument and shows some empirical evidence that there is no such evidence. He argues that a ban on naked sovereign CDS trading may instead increase execution costs and lower the quality of price information. In contrast, he argues that holders of both the derivative and the underlying cash bond have in fact no more incentive to monitor the borrower, which may induce the borrower to invest less efficiently and thus increases the probability of default. Bolton and Oehmke (2011) study this ”empty creditor” problem theoretically for corporate CDS, while Subrahmanyam et al. (2012) provide empirical evidence39 . An early theoretical treatment of the effect of the existence of CDS on the incentives of sovereign borrowers was also made available by Goderis and Wagner (2011). More specifically, the authors show that without insurance, the sovereign country has no incentive to optimally invest (or exert effort) as there is the possibility of offering a debt restructuring in the bad states of the world, 38 The pricing framework considered in Adler and Song (2010) is a useful extension to the treatment in Duffie (1999). 39 The concept of empty creditor refers to the separation of cash flow and voting rights through the use of credit derivatives. A debtholder, whose exposure is insured with a CDS, keeps an economic interest in the firm’s claims, but has no more risk alignment with other bondholders, who enjoy no such proection. 22 which the lender cannot credibly commit to reject. When the lender can purchase insurance on the bond, the threat of rejecting the restructuring offer becomes credible and the debtor will have to internalize more of the costs in the bad states. This provides incentive to invest more efficiently ex-ante, which in turn reduces default. Strategic purchase by a (single) bondholder to increase the bargaining power in debt renegotiation deal is thus beneficial from an ex-ante perspective. However, in the presence of many bondholders, there may be coordination failures. This may lead to a situation, where lenders buy more insurance than is socially optimal, which leads to an increased probability of default. Sambalaibat (2011) studies the issue of naked CDS theoretically and shows that the effect depends on the infrastructure of the insurance market. The overall outcome depends on the parameter values, and naked CDS buyers may either induce over-investment, leading to lower borrowing costs or to under-investment with higher borrowing costs. For a legal treatment of how credit derivatives may affect the sovereign debt restructuring process, see Verdier (2004). 7 Trading in the Sovereign CDS Market Very little is known about the trading patterns in the sovereign CDS market. Based on a threemonth sample of detailed transaction-level data over the time period May 1 to July 31 2010 generously received from a set of dealers, the Federal Reserve Bank of New York spoiled us a with a more detailed picture of the CDS trading behavior in their staff report No. 51740 . Keeping in mind the caveat that the data set stems from a period, where overall trading activity was below historical average, the results indicate that CDS trade not that often despite their highly standardized nature. For the 29,146 single name sovereign CDS transactions recorded for 74 reference entities, the most actively traded reference entities traded on average 30 times per day, the less frequently traded 15 times per day on average and the infrequently traded reference entities traded only 2 times per day on average. Another way to say this, is that a long tail of reference entities trade less than once a day. In terms of size, the USD denominated trades were mostly traded in 5 million USD tickets, with a median and mean trade of 10 and 16.74 million USD respectively. For EUR denominated trades (which amounted to only 574 transactions), the mode was 10 million USD, with a median 40 See Chen et al. (2011). 23 trade of 5 million and a mean trade of 12.53 million. For matters of comparison, sovereign single name CDS trades were on average twice as large as their corporate counterparts. Although the number of trades seems rather low, market participation for sovereign reference entities was rather active. On average, 50 market participants are reported to have traded at least once a day, 200 traded at least once a week and 340 traded at least once a month. In addition, dealers were more likely to be sellers of protection and the four most active dealers were involved in 45% of the buying and selling of all CDS trades and in 50% of the overall notional amount. Yet, according to regulatory definitions of the Department of Justice, the Herfindahl Hirshman concentration index with levels ranging between 885 and 965 is considered to be low41 . Finally, the market is confirmed to be highly standardized, as 92% of all single-name CDS contracts within the sample had a fixed coupon and 97% had fixed quarterly payment dates42 . Another more exhaustive source of information on CDS trading behavior is the the Depository Trust and Clearing Corporation (DTCC). In October 2008, the DTCC Trade Information Warehouse started to publish detailed weekly reports on stocks and volumes in CDS trading. More specifically, current and historical positions are shown on an aggregate level and for the 1,000 mostly traded reference contracts. This move towards higher transparency is greatly welcome. Oehmke and Zawadowski (2011) have used this data to study the determinants of corporate trading and find that higher amounts of (corporate) CDS outstanding are associated with companies, which have more assets on their balance sheet, more bonds outstanding, that are investment grade firms or firms that have just recently lost their investment grade status, as well as with higher analysts’ forecast dispersion. Duffie (April 29 2010a) illustrates a few statistics on sovereign CDS from this database in his testimony to the United House of Representatives, but we are not aware of any formal treatment at this stage. Here, we provide a summarized overview of the main statistical facts for government credit derivatives. Tables (10) and (11) illustrate the average gross and net notional amounts outstanding in (mil41 Note that Giglio (2011) emphasizes the high concentration of the CDS market, thereby referring to industry reports, which state that in 2006, the top 10 counterparties (all broker/dealers) accounted for about 89% of the total protection sold. See footnote 8. 42 Using transactions and quote data from CreditTrade on 77 sovereign reference entities from January 1997 to June 2003, Packer and Suthiphongchai (2003) emphasized the low trading volume by indicating that (in 2002), only 6% of all quotes resulted in transactions, with a strong concentration in a few reference entities. The top five names accounted for more than 40% of all quotes. These countries were Brazil, Mexico, Japan, the Philippines and South Africa. 24 lion) USD equivalents on all sovereign CDS contracts (excluding sovereign states in the US) among the 1,000 mostly traded contracts over the time period 31 October 2008 through 11 November 201143 . In addition, there is information on the average number of contracts over that period, the ratios of gross to net notional amount outstanding and the net notional amount outstanding to the number of contracts. The countries are classified into five geographical regions following the practice of Markit: Americas, Asia Ex-Japan, Australia and New Zealand, Europe-Middle East-Africa (EMEA) and Japan. While the weekly total gross notional volume recorded for sovereign countries in the data repository is approximately 2.1 trillion USD, the net exposure, which accounts for offsetting effects between buyers and sellers and represents therefore a more economically meaningful measure, reflects with 203.6 billion USD about 9.86% of that amount. Keeping in mind that this number does not report the values for sovereign US states and other non-government supranational bodies, this amount represents approximately 75% of the sovereign single name CDS outstanding as reported by BIS in the first semester of 2011 (see Table (4)). The total number over all country-specific averages of traded contracts is 147,258, the mean ratio of gross to net notional amount is 10.74 and on average, the net credit exposure per contract is 1.8 million USD. There is however a significant cross-sectional dispersion. Further analysis also reveals that four of the five GIIPS countries rank among the top ten countries with the highest amount of net notional exposure. Another striking feature we observe is that emerging economies tend to have higher numbers of traded contracts, but smaller net exposures per contract on average, while developed economies trade in bigger bulks, but have less contracts outstanding. The world’s two biggest reference bond markets, the US and Germany, top the list with 5.16 and 6.66 million USD per contract respectively. At the bottom of the list are the Philippines with an average of 360,000 USD per traded contract. Moreover, the top ten list of countries with the highest amount of gross volume outstanding is dominated by emerging market economies, while developed distressed economies cluster in the top ten list with the highest net notional exposure. Tables (12) and (13) provide similar statistics at a yearly frequency since the data has been made available by DTCC. From these numbers, it is clear to see that average weekly gross and net volumes, as well as ratios of gross to net exposure have fluctuated only modestly, with the ratio 43 Tables (10) and (11) provide essentially the same information, but Table (10) provides a sort in descending order for each column to help the exposition. The reader may find one or the other table more reader-friendly. 25 of gross to net exposure being roughly equal to a factor of 10 over time. The average number of contracts reported in the warehouse has changed slightly more over time, which explains the bigger fluctuations in the net notional exposure per contract, which has come down from 3.6 million USD in 2008 to a mean 1.4 million in 2011. Figures (3) and (4) provide some more intuition about how the CDS trading is related to individual government characteristics for EMEA and Asian countries (ex-Japan). For these regions, it is clear that within each year, there seems to be a statistical relationship between the net notional amount of CDS outstanding and the nominal amount of government debt. This relationship appears much weaker if debt levels are compared against gross exposures. Duffie (April 29 2010a, 2010b) showed empirical evidence that the volume of CDS is unrelated to the level of spreads. This raises once more interesting questions. We have seen that government characteristics and CDS volumes are unrelated to to spreads, but characteristics are linked to volumes. Table (14) separately reports statistics for the United States government and the sovereign US states, which are registered in the DTCC database. In contrast to the statistics on other sovereign governments, all yearly average values for CDS on US Treasuries exhibit a sharp increase in 2011. In fact, the mean gross notional amount jumps by 90% from 2011 to 2011, going from 13.2 to 25 billion USD. That’s a huge surge. The same observation holds for net exposure, which rises by 84% over the same time period, going from 2.4 to 4.5 billion USD, while the number of live contracts increases from 479 million to approximately 1.1 billion contracts, which corresponds to an increase of 120%. These spikes raise the provoking thought that investors may actually have seriously considered the possibility of US government default during the lock-out period in summer 2011. In particular, as no such sharp changes are recorded for the individual states. There the evidence is mixed. Gross exposures and number of traded contracts unilaterally increased, while net notional exposure increased mainly for Illinois, while it dropped for instance for New York City, Texas or Florida. 26 8 Conclusion Sovereign credit risk is undoubtedly a topic of crucial importance. The supranational bailouts of Greece, Ireland, Portugal and the vast amount of European Union summits ever since the start of the sovereign debt crisis eliminate any doubt about the relevance of this topic44 . As we speak, trying to avoid information on sovereign debt issues is like walking through a mine field in North Korea. This paper tries to provide a comprehensive overview of the young, but steadily growing literature on sovereign CDS, describes key statistical facts and features of prices, about the market, its players and related trading activities and attempts to raise some thought-provoking questions. We thereby hope to provide a good starting point for anyone interested in the topic. We had fun writing this document, we hope you enjoyed it too. 44 We stopped counting at 16. 27 References Acharya, V. V., Drechsler, I. and Schnabl, P. (2011). A pyrrhic victory? - bank bailouts and sovereign credit risk, NBER Working Paper 17136 . Adler, M. and Song, J. (2010). The behavior of emerging market sovereigns’ credit default swap premiums and bond yield spreads, International Journal of Finance & Economics 15(1): 31–58. Aktug, E. and Vasconcellos, G. (2010). Is sovereign risk related to the banking sector?, Working paper, Lehigh University . Aktug, Erdem, G. V. and Bae, Y. (2009). The dynamics of sovereign credit default swap and bond markets: Empirical evidence from the 2001-2007 period, WP . Alter, A. and Schuler, Y. S. (2011). Credit spread interdependencies of european states and banks during the financial crisis, SSRN eLibrary . Altman, E. I. and Rijken, H. A. (2011). Toward a bottom-up approach to assessing sovereign default risk, Journal of Applied Corporate Finance 23(1): 20–31. Ammer, J. and Cai, F. (2011). Sovereign cds and bond pricing dynamics in emerging markets: Does the cheapest-to-deliver option matter?, Journal of International Financial Markets, Institutions and Money 21(3): 369 – 387. Ang, A. and Longstaff, F. A. (2011). Systemic sovereign credit risk: Lessons from the u.s. and europe, Working Paper 16982, National Bureau of Economic Research. Arce, O., Mayordomo, S. and Pena, J. I. (2011). Do sovereign cds and bond markets share the same information to price credit risk? an empirical application to the european monetary union case, Working Paper . Arora, N., Gandhi, P. and Longstaff, F. A. (2012). Counterparty credit risk and the credit default swap market, Journal of Financial Economics 103(2): 280 – 293. Augustin, P. and Tédongap, R. (2011). Sovereign credit risk: Consumption strikes again, Working Paper Stockholm School of Economics . Blanco, R., Brennan, S. and Marsh, I. W. (2005). An empirical analysis of the dynamic relation between investment-grade bonds and credit default swaps, The Journal of Finance 60(5): 2255– 2281. Bolton, P. and Oehmke, M. (2011). Credit default swaps and the empty creditor problem, Review of Financial Studies 24(8): 2617–2655. Bongaerts, D., De Jong, F. and Driessen, J., J. (2011). Derivative pricing with liquidity risk: Theory and evidence from the credit default swap market, The Journal of Finance 66(1): 203–240. Boone, L., Fransolet, L. and Willemann, S. (2010). Euro public debt and the markets: sovereign fundamentals and cds market dynamics, Banque de France Financial Stability Review - Derivatives, Financial innovation and stability No 14: Barclays Capital. Bowe, M., K. A. T. A. (2009). Information transmission and price discovery in emerging sovereign credit risk markets, Working Paper . 28 Brutti, F. and Sauré, P. (2012). Transmission of sovereign risk in the euro crisis, Working Paper . C. Emre Alper, L. F. and Gerard, M. (2012). Pricing of sovereign credit risk: Evidence from advanced economies during the financial crisis, IMF Working Paper 12/24 . Calice, G., Chen, J. and Williams, J. M. (2011). Liquidity spillovers in sovereign bond and cds markets: An analysis of the eurozone sovereign debt crisis, SSRN Working Paper 1898596 . Carboni, A. (2010). The sovereign credit default swap market: price discovery, volumes and links with banks’ risk premia, Working Paper Bank of Italy . Carlos Caceres, V. G. and Segoviano, M. (2010). Sovereign spreads: Global risk aversion, contagion or fundamentals?, IMF Working Paper 10/120 . Carr, P. and Wu, L. (2007). Theory and evidence on the dynamic interactions between sovereign credit default swaps and currency options, Journal of Banking & Finance 31(8): 2383–2403. Chan-Lau, J. A. and Kim, Y. S. (2004). Equity prices, credit default swaps, and bond spreads in emerging markets, IMF Working Paper 04/27 . Chen, K., Fleming, M., Jackson, J. and Sarkar, A. (2011). An analysis of cds transactions: Implications for public reporting, Federal Reserve Bank of New York Staff Report No. 517 . Chernov, M., Gorbenko, A. S. and Mak, I. (2011). Cds auctions, Working Paper London School of Economics . Cossin, D. and Jung, G. (2005). Do major financial crises provide information on sovereign risk to the rest of the world? a look at credit default swap markets, FAME Research Paper N.134 . Coudert, V. and Gex, M. (2010). Credit default swap and bond markets: which leads the other?, Banque de France Financial Stability Review - Derivatives, Financial innovation and stability . Coudert, V. and Gex, M. (2011). The interactions between the credit default swap and the bond markets in financial turmoil, Centre d’Etudes Prospectives Et d’Informations Internationales Document de Travail 2011/02 . Criado, S., Degabriel, L., Lewandowska, M., Linden, S. and Ritter, P. (2010). Sovereign cds report, Official Report DG COMP, DG ECFIN and DG MARKT . URL: http://www.ft.com/intl/cdsreport Das, S. and Hanouna, P. (2006). Survey of the literature: Credit default swap spreads, Journal of Investment Management 4: 93–105. Delatte, A.-L., Gex, M. and Lpez-Villavicencio, A. (2012). Has the cds market influenced the borrowing cost of european countries during the sovereign crisis?, Journal of International Money and Finance 31(3): 481 – 497. ¡ce:title¿Financial Stress in the Eurozone¡/ce:title¿. Delis, M. D. and Mylonidis, N. (2011). The chicken or the egg? a note on the dynamic interrelation between government bond spreads and credit default swaps, Finance Research Letters 8(3): 163 – 170. Dieckmann, S. and Plank, T. (2011). Default risk of advanced economies: An empirical analysis of credit default swaps during the financial crisis, Review of Finance . 29 Dooley, M. and Hutchison, M. (2009). Transmission of the u.s. subprime crisis to emerging markets: Evidence on the decoupling-recoupling hypothesis, Journal of International Money and Finance 28(8): 1331 – 1349. Duffie, D. (1999). Credit swap valuation, Financial Analysts Journal 55(1): 73–87. Duffie, D. (2010b). Is there a case for banning short speculation in sovereign bond markets?, Banque de France Financial Stability Review - Derivatives, Financial innovation and stability No. 14. Duffie, D. (April 29 2010a). Credit default swaps on government debt: potential implications of the greek debt crisis, Testimony to the United States House of Representatives, Subcommittee on Capital Markets, Insurance, and Government Sponsored Enterprises . Ejsing, J. W. and Lemke, W. (2011). The janus-headed salvation sovereign and bank credit risk premia during 2008-09, Economics Letters 110: 28–31. Fisher, N.-F. (2010). Explaining sovereign bond-cds arbitrage violations during the financial crisis 2008-09, Working Paper London School of Economics . Fontana, A. and Scheicher, M. (2010). An analysis of euro area sovereign cds and their relation with government bonds, SSRN eLibrary . Giglio, S. (2011). Credit default swap spreads and systemic financial risk, Working Paper Harvard University . Goderis, B. and Wagner, W. (2011). Credit derivatives and sovereign debt crises, Unpublished Paper, Tilburg University. . Gray, Dale . F., R. C. M. and Bodie, Z. (2007). Contingent claims approach to measuring and managing sovereign credit risk, Journal of Investment Management 5 4: 5–28. Gupta, S. and Sundaram, R. K. (2012). Cds auctions and informative biases in cds recovery rates, Working Paper New York University . Hart, O. and Zingales, L. (Fall 2011). A new capital regulation for large financial institutions, American Law and Economics Review 13(2): 453–490. Helwege, J., Maurer, S., Sarkar, A. and Wang, H., Y. (2009). Credit default swap auctions and price discovery., Journal of Fixed Income 19(2): 34 – 42. Hui, C.-H. and Fong, T. (2011). Information flow between sovereign cds and dollar-yen currency option markets in the sovereign debt crisis of 2009-2011, Working Papers 402011, Hong Kong Institute for Monetary Research. Hull, J. C. and White, A. (2000a). Valuing credit default swaps i: No counterparty default risk, SSRN eLibrary . In, F., Kang, B. U. and Kim, T. S. (2007). Sovereign credit default swaps, sovereign debt and volatility transmission across emerging markets, Working Paper . Ismailescu, I. and Kazemi, H. (2010). The reaction of emerging market credit default swap spreads to sovereign credit rating changes, Journal of Banking and Finance 34(12): 2861 – 2873. 30 Jankowitsch, R., Pullirsch, R. and Veza, T. (2008). The delivery option in credit default swaps, Journal of Banking and Finance 32(7): 1269 – 1285. Kallestrup, R. (2011). Bank balance sheets and the dynamics of bank and sovereign credit spreads, Working Paper Copenhagen Business School, . Kallestrup, R., Lando, D. and Murgoci, A. (2011). Financial sector linkages and the dynamics of bank and sovereign credit spreads, Working Paper Copenhagen Business School, . Kucuk, U. N. (2010). Non-default component of sovereign emerging market yield spreads and its determinants: Evidence from the credit default swap market., Journal of Fixed Income 19(4): 44 – 66. Lando, D. (2004). Credit Risk Modeling, Princeton University Press. Lando, D. and Mortensen, A. (2005). Revisiting the slope of the credit spread curve, Journal of Investment Management, Vol. 3, No. 4, Fourth Quarter 2005 3(4): 1–27. Levy, A. (2009). The cds bond basis spread in emerging markets: Liquidity and counterparty risk effects, Working Paper . Li, N. (2009). The price discovery process in credit derivative market: evidence from sovereign cds market, American Journal of Finance and Accounting 1(4): 393 – 407. Li, N. and Huang, A. Y. (2011). Price discovery between sovereign credit default swaps and bond yield spreads of emerging markets, Journal of Emerging Market Finance 10(2): 197–225. Longstaff, F. A., Mithal, S. and Neis, E. (2005). Corporate yield spreads: Default risk or liquidity? new evidence from the credit default swap market, The Journal of Finance 60(5): 2213–2253. Longstaff, F. A., Pan, J., Pedersen, L. H. and Singleton, K. J. (2010). How sovereign is sovereign credit risk?, American Economic Journal: Macroeconomics (13658). M. Kabir Hassan, G. M. N. and Suk-Yu, J. (2011). Credit default swaps and sovereign debt markets, Working Paper . Merton, R. C. (1974). On the pricing of corporate debt: The risk structure of interest rates, The Journal of Finance 29(2, Papers and Proceedings of the Thirty-Second Annual Meeting of the American Finance Association, New York, New York, December 28-30, 1973): 449–470. Oehmke, M. and Zawadowski, A. (2011). Positioning in the cds market: Evidence from the dtcc data, Working Paper Columbia University . O’Kane, D. (2012). The link between eurozone sovereign debt and cds prices, Working Paper Edhec Business School . Packer, F. and Suthiphongchai, C. (2003). Sovereign credit default swaps, BIS Quarterly Review . Palladini, G. and Portes, R. (2011). Sovereign cds and bond pricing dynamics in the euro-area, Working Paper London Business School . Pan, J. and Singleton, K. J. (2008). Default and recovery implicit in the term structure of sovereign cds spreads, The Journal of Finance 63(5): 2345–2384. 31 Plank, T. J. (2010). Do macro-economic fundamentals price sovereign cds spreads of emerging economies?, Working Paper . Portes, R. (2010). Ban naked cds, Euro Intelligence . Pu, X. (2012). The information content of dual-currency sovereign cds for exchange rate returns, Working Paper . Remolona, E., Scatigna, M. and Wu, E. (2008). The dynamic pricing of sovereign risk in emerging markets: Fundamentals and risk aversion, Journal of Fixed Income 17(4): 57 – 71. Sambalaibat, B. (2011). Credit default swaps and sovereign debt with moral hazard and debt renegotiation, Working Paper Carnegie Mellon University . Sgherri, S. and Zoli, E. (2009). Euro area sovereign risk during the crisis, IMF Working Papers 09/222, International Monetary Fund. Singh, M. (2003). Are credit default swap spreads high in emerging markets? an alternative methodology for proxying recovery value, IMF Working Paper 03/242 . Stulz, R. M. (2010). Credit default swaps and the credit crisis, Journal of Economic Perspectives 24(1): 73–92. Subrahmanyam, M. G., Tang, D. Y. and Wang, S. Q. (2012). Does the tail wag the dog? the effect of credit default swaps on credit risk, SSRN eLibrary . Tang, D. Y. and Yan, H. (2007). Liquidity, liquidity spillover, and credit default swap spreads, Working Paper . Tett, G. (2009). Fool’s Gold: How the Bold Dream of a Small Tribe at J.P. Morgan Was Corrupted by Wall Street Greed and Unleashed a Catastrophe, Free Press. Varga, L. (2009). The information content of hungarian sovereign cds spreads, Working Paper Central Bank of Hungary . Verdier, P.-H. (2004). Credit derivatives and the sovereign debt restructuring process, Working Paper Harvard Law School . Wang, P. and Moore, T. (2012). The integration of the credit default swap markets during the us subprime crisis: Dynamic correlation analysis, Journal of International Financial Markets, Institutions and Money 22(1): 1 – 15. Zhang, F. X. (2003). What did the credit market expect of argentina default? evidence from default swap data, SSRN eLibrary . 32 Table 1: Credit Default Swaps: Notional Amount outstanding (’000,000) This table reports the total notional amount outstanding in million USD of credit default swaps, its growth (in %) and the market share (in %) for the overall market, single-name and multi-name instruments. Data on total notional amount outstanding are shown on a net basis, that is transactions between reporting dealers are reported only once. Source: BIS. Period 2004-H2 2005-H1 2005-H2 2006-H1 2006-H2 2007-H1 2007-H2 2008-H1 2008-H2 2009-H1 2009-H2 2010-H1 2010-H2 2011-H1 Total 6 395 744 10 211 378 13 908 285 20 352 307 28 650 265 42 580 546 58 243 721 57 402 760 41 882 674 36 098 169 32 692 683 30 260 923 29 897 578 32 409 444 Growth 0.00 59.66 36.20 46.33 40.77 48.62 36.78 -1.44 -27.04 -13.81 -9.43 -7.44 -1.20 8.40 Single-Name 5 116 765 7 310 289 10 432 038 13 873 445 17 879 280 24 239 478 32 486 272 33 412 115 25 739 929 24 165 086 21 917 050 18 493 632 18 144 580 18 104 619 Growth – 42.87 42.70 32.99 28.87 35.57 34.02 2.85 -22.96 -6.12 -9.30 -15.62 -1.89 -0.22 33 Fraction 80.00 71.59 75.01 68.17 62.41 56.93 55.78 58.21 61.46 66.94 67.04 61.11 60.69 55.86 Multi-Name 1 278 979 2 901 090 3 476 247 6 478 863 10 770 985 18 341 068 25 757 449 23 990 644 16 142 745 11 933 083 10 775 633 11 767 291 11 752 998 14 304 825 Growth – 126.83 19.83 86.38 66.25 70.28 40.44 -6.86 -32.71 -26.08 -9.70 9.20 -0.12 21.71 Fraction 20.00 28.41 24.99 31.83 37.59 43.07 44.22 41.79 38.54 33.06 32.96 38.89 39.31 44.14 34 Total contracts Foreign exchange contracts (Fraction) Forwards and forex swaps Currency swaps Options Interest rate contracts (Fraction) Forward rate agreements Interest rate swaps Options Equity-linked contracts (Fraction) Forwards and swaps Options Commodity contracts (Fraction) Gold Other commodities Forwards and swaps Options Credit default swaps (Fraction) Single-name instruments Multi-name instruments of which index products Unallocated (Fraction) 2005-H1 282 665 31 081 11.00 15 801 8 236 7 045 204 795 72.45 13 973 163 749 27 072 4 551 1.61 1 086 3 465 2 940 1.04 288 2 652 1 748 904 10 211 3.61 7 310 2 901 – 29 086 10.29 Notional amounts outstanding (’000,000,000) 2006-H1 2007-H1 2008-H1 2009-H1 2010-H1 372 513 507 907 672 558 594 553 582 685 38 127 48 645 62 983 48 732 53 153 10.24 9.58 9.36 8.20 9.12 19 407 24 530 31 966 23 105 25 624 9 696 12 312 16 307 15 072 16 360 9 024 11 804 14 710 10 555 11 170 262 526 347 312 458 304 437 228 451 831 70.47 68.38 68.14 73.54 77.54 18 117 22 809 39 370 46 812 56 242 207 588 272 216 356 772 341 903 347 508 36 821 52 288 62 162 48 513 48 081 6 782 8 590 10 177 6 584 6 260 1.82 1.69 1.51 1.11 1.07 1 430 2 470 2 657 1 678 1 754 5 351 6 119 7 521 4 906 4 506 6 394 7 567 13 229 3 619 2 852 1.72 1.49 1.97 0.61 0.49 456 426 649 425 417 5 938 7 141 12 580 3 194 2 434 2 188 3 447 7 561 1 715 1 551 3 750 3 694 5 019 1 479 883 20 352 42 581 57 403 36 098 30 261 5.46 8.38 8.53 6.07 5.19 13 873 24 239 33 412 24 165 18 494 6 479 18 341 23 991 11 933 11 767 – – – – 7 499 38 332 53 213 70 463 62 291 38 329 10.29 10.48 10.48 10.48 6.58 2011-H1 707 569 64 698 9.14 31 113 22 228 11 358 553 880 78.28 55 842 441 615 56 423 6 841 0.97 2 029 4 813 3 197 0.45 468 2 729 1 846 883 32 409 4.58 18 105 14 305 12 473 46 543 6.58 2007-H1 11 118 1 345 12.10 492 619 235 6 063 54.53 43 5 321 700 1 116 10.03 240 876 636 5.72 47 589 – – 721 6.48 406 315 – 1 237 11.13 Gross market values (’000,000,000) 2008-H1 2009-H1 2010-H1 20 340 25 298 24 697 2 262 2 470 2 544 11.12 9.77 10.3 802 870 930 1 071 1 211 1 201 388 389 413 9 263 15 478 17 533 45.54 61.18 70.99 88 130 81 8 056 13 934 15 951 1 120 1 414 1 501 1 146 879 706 5.63 3.48 2.86 283 225 189 863 654 518 2 213 682 458 10.88 2.69 1.85 72 43 45 2 141 638 413 – – – – – – 3 192 2 973 1 666 15.69 11.75 6.75 1 901 1 950 993 1 291 1 023 673 – – – 2 264 2 816 1 789 11.13 11.13 7.25 2011-H1 19 518 2 336 11.97 777 1 227 332 13 244 67.85 60 11 864 1 319 708 3.63 176 532 471 2.41 50 421 – – 1 345 6.89 854 490 – 1 414 7.25 This table reports the total notional amounts outstanding in billion USD of Over-the-counter derivatives, their gross market values and the market share (in %), broken down by risk category and type of instrument. Data on total notional amount outstanding are shown on a net basis, that is transactions between reporting dealers are reported only once. Source: BIS. Table 2: OTC Derivatives: Notional Amount outstanding (’000,000,000) Table 3: Exposures of Credit Default Swaps: Notional Amount Outstanding (’000,000) This table reports the total notional amount outstanding in million USD of credit default swaps as well as the gross and net market values. Data on total notional amount outstanding are shown on a net basis, that is transactions between reporting dealers are reported only once. Source: BIS. Notional Amounts outstanding Gross Market Values Gross Credit Exposure 2005-H1 10 211 378 187 932 – 2006-H1 20 352 307 294 332 – 2007-H1 42 580 546 720 725 – 2008-H1 57 402 759 3 191 868 – 2009-H1 36 098 169 2 973 449 – 2010-H1 30 260 923 1 665 944 – 2011-H1 32 409 444 1 344 731 374 525 Table 4: Sovereign Credit Default Swaps: Notional Amount outstanding (’000,000) This table reports the total notional amount outstanding in million USD of Sovereign credit default swaps and the market share (in %) by type of product. Data on total notional amount outstanding are shown on a net basis, that is transactions between reporting dealers are reported only once. Source: BIS. Period 2004-H2 2005-H1 2005-H2 2006-H1 2006-H2 2007-H1 2007-H2 2008-H1 2008-H2 2009-H1 2009-H2 2010-H1 2010-H2 2011-H1 Total 6 395 744 10 211 378 13 908 285 20 352 307 28 650 265 42 580 546 58 243 721 57 402 760 41 882 674 36 098 169 32 692 683 30 260 923 29 897 578 32 409 444 Sov.Single-Name –– –– 1 258 139 ( 9.05 %) 640 971 ( 3.15 %) 1 100 990 ( 3.84 %) 1 489 655 ( 3.50 %) 1 837 873 ( 3.16 %) 2 177 364 ( 3.79 %) 1 650 743 ( 3.94 %) 1 761 083 ( 4.88 %) 1 943 005 ( 5.94 %) 2 392 580 ( 7.91 %) 2 542 331 ( 8.50 %) 2 749 259 ( 8.48 %) Sov.Multi-Name –– –– –– –– –– –– –– –– –– –– –– –– –– 158 516 (0.49 %) Sov.All –– –– –– –– –– –– –– –– –– –– –– –– –– 2 907 775 (8.97 %) Table 5: Counterparties of Single-Name Sovereign Credit Default Swaps This table reports the total notional amount outstanding in million USD of Single-Name sovereign credit default swaps by type of counterparty as well as their relative market share (in %). Data on total notional amount outstanding are shown on a net basis, that is transactions between reporting dealers are reported only once. Central Counterparty (CCP) is defined as an entity that interposes itself between counterparties to contracts traded in one or more financial markets, becoming the buyer to every seller and the seller to every buyer. Both contracts post-novation are captured. Insurance and Financial Guaranty firms include reinsurance and pension funds. Source: BIS. All Counterparties Reporting Dealers Other Financial Institutions Central Counterparties Banks and Security Firms Insurance and Financial Guaranty Firms SPVs, SPCs and SPEs Hedge Funds Other Financial Customers Non-Financial Institutions 2006-H1 640 971 301 320 (47.01) 292 907 (45.7) – 168 174 2 110 712 18 386 103 525 46 745 (7.29) 2007-H1 1 489 655 781 382 (52.45) 644 922 (43.29) – 216 677 5 293 2 375 27 603 392 976 63 347 (4.25) 35 2008-H1 2 177 364 1 123 078 (51.58) 985 117 (45.24) – 569 346 8 356 323 13 519 393 573 69 169 (3.18) 2009-H1 1 761 083 877 496 (49.83) 813 241 (46.18) – 647 135 4 371 418 9 437 151 880 70 357 (4,00) 2010-H1 2 392 580 1 320 039 (55.17) 1 017 122 (42.51) 828 023 8 858 7 905 88 519 83 817 55 429 (2.32) 2011-H1 2 749 259 1 836 889 (66.81) 891 483 (32.43) 2 075 592 194 14 688 18 106 145 178 119 241 20 885 (0.76) Table 6: Maturity Structure of Credit Default Swaps: Notional Amount Outstanding (’000,000) This table reports the total notional amount outstanding in million USD of credit default swaps by remaining maturity as well as their respective market shares. Data on total notional amount outstanding are shown on a net basis, that is transactions between reporting dealers are reported only once. The remaining contract maturity is determined by the difference between the reporting date and the expiry date of the contract and not by the date of execution of the deal. Source: BIS. Period 2004-H2 2005-H1 2005-H2 2006-H1 2006-H2 2007-H1 2007-H2 2008-H1 2008-H2 2009-H1 2009-H2 2010-H1 2010-H2 2011-H1 Total 6 395 744 10 211 378 13 908 285 20 352 307 28 650 265 42 580 546 58 243 721 57 402 760 41 882 674 36 098 169 32 692 683 30 260 923 29 897 578 32 409 444 1y or less 491 270 670 627 861 871 1 573 542 2 336 401 2 866 737 3 376 160 3 968 090 2 978 334 3 843 679 3 432 429 3 328 461 3 181 610 3 924 650 Fraction 7.68 6.57 6.20 7.73 8.15 6.73 5.80 6.91 7.11 10.65 10.50 11.00 10.64 12.11 Over 1y up to 5y 4 659 521 7 139 210 9 821 293 13 019 148 16 876 503 24 353 126 36 037 715 36 923 426 26 724 773 23 191 972 21 307 787 20 786 997 21 480 952 23 194 893 Fraction 72.85 69.91 70.61 63.97 58.91 57.19 61.87 64.32 63.81 64.25 65.18 68.69 71.85 71.57 Over 5 years 1 243 778 2 400 398 3 225 106 5 759 336 9 437 359 15 360 681 18 829 848 16 511 244 12 179 538 9 062 498 7 952 441 6 145 442 5 235 000 5 289 922 Fraction 19.45 23.51 23.19 28.30 32.94 36.07 32.33 28.76 29.08 25.11 24.32 20.31 17.51 16.32 Table 7: Credit Default Swaps by location of counterparty This table reports the total notional amount outstanding in million USD of credit default swaps by location of counterparty as of June 2011, as well as their respective market shares. Data on total notional amount outstanding are shown on a net basis, that is transactions between reporting dealers are reported only once. The notional amounts outstanding is allocated to one of the locations listed in the table on an ultimate risk basis, ie according to the nationality of the counterparty. Home country means country of incorporation of the reporters head office. Based on the data reported by 10 countries. Source: BIS. June 2011 Total Total 32 409 444 Home Country 5 931 892 36 Fraction 18.30 Abroad 26 477 552 Fraction 81.70 37 AAA Mean Median Minimum Maximum Volatility Skewness Kurtosis AC1 A Mean Median Minimum Maximum Volatility Skewness Kurtosis AC1 BB Mean Median Minimum Maximum Volatility Skewness Kurtosis AC1 1 15 2 0 304 26 2 6 0.9940 1 49 12 1 1235 69 2 6 0.9967 1 110 78 15 822 92 2 7 0.9979 2 17 2 0 301 28 2 6 0.9949 2 55 18 2 1172 72 2 5 0.9970 2 157 115 7 845 111 1 4 0.9977 3 19 3 1 293 30 2 5 0.9970 3 61 24 3 1127 74 2 5 0.9971 3 196 146 1 903 125 1 3 0.9976 5 23 4 1 274 34 2 5 0.9975 5 71 34 5 1015 77 2 5 0.9974 5 255 197 1 1032 138 1 3 0.9974 7 24 5 2 270 34 2 4 0.9975 7 76 41 5 952 75 2 5 0.9973 7 281 225 74 1036 138 1 3 0.9972 10 26 7 2 266 33 2 4 0.9974 10 81 49 6 893 73 2 5 0.9972 10 305 250 72 1039 138 1 3 0.9971 category. AC1 is the first-order autocorrelation coefficient. Source: Markit AA Mean Median Minimum Maximum Volatility Skewness Kurtosis AC1 BBB Mean Median Minimum Maximum Volatility Skewness Kurtosis AC1 B Mean Median Minimum Maximum Volatility Skewness Kurtosis AC1 1 24 3 0 557 36 2 6 0.9969 1 77 33 3 1190 105 3 9 0.9976 1 433 328 19 3654 371 2 7 0.9901 2 27 4 1 536 38 2 6 0.9972 2 95 54 3 1100 106 2 8 0.9974 2 484 403 34 3504 362 2 6 0.9916 3 31 5 1 499 40 2 5 0.9975 3 109 74 5 1110 105 2 7 0.9970 3 517 455 55 3400 350 2 6 0.9926 5 38 8 2 461 44 2 5 0.9978 5 132 106 8 1106 103 2 7 0.9967 5 564 517 117 3234 328 2 6 0.9927 7 41 12 2 433 44 2 4 0.9978 7 143 123 11 1096 99 2 6 0.9966 7 574 538 140 3111 303 2 7 0.9889 10 45 16 3 410 43 2 4 0.9977 10 155 139 14 1081 96 2 6 0.9964 10 599 562 186 3053 286 2 6 0.9928 by taking the mean. Similarly, the standard deviation is calculated as the standard deviation of the data series aggregated at each date within a given rating Mean (Median) spread is calculated as the historical mean (median) spread, where at each date, all observations within a given rating category are aggregated each date to each country. The equally weighted historical average is then rounded to the nearest integer, which is used as the final rating categorization. The CDS prices are mid composite quotes and USD denominated. Rating classification is achieved by assigning an integer value ranging from 1 (AAA) to 21 (C) at The table reports summary statistics for the CDS term structure of 38 sovereign countries over the sample period May 9th, 2003 until August 19th, 2010. All Table 8: Summary Statistics Table 9: Slope of the CDS Spread Curve Summary statistics by groups of the 5-year spread level. For each group and maturity, the average spread level and slope has been calculcated across time and entities. The groups are sorted according to the 5-year spread level. Source: Markit 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Group (101-200 bps) (201-300 bps) (301-400 bps) (401-500 bps) (501-600 bps) (601-700 bps) (701-800 bps) (801-900 bps) (901-1000 bps) (1001-1100 bps) (1101-1200 bps) (1201-1300 bps) (1301-1400 bps) (1401-1500 bps) (1501-3234 bps) 1y 66 139 209 273 362 503 617 622 724 1 008 1 075 1 207 1 347 1 357 2 369 Average CDS premium (bps) 2y 3y 5y 7y 88 107 143 160 172 201 245 263 256 295 347 367 334 383 447 468 433 484 542 557 567 603 642 633 693 725 745 713 759 812 850 846 845 894 940 932 1 051 1 052 1 043 1 002 1 142 1 149 1 145 1 095 1 256 1 257 1 251 1 195 1 382 1 376 1 363 1 309 1 410 1 425 1 444 1 395 2 362 2 334 2 281 2 213 38 10y 177 279 386 490 572 661 741 840 919 964 1 052 1 148 1 260 1 349 2 146 Average Slope (bps) 10y-1y 3y-1y 10y-3y 111 41 69 141 62 79 176 86 91 217 111 106 210 122 88 158 100 58 125 108 16 218 189 29 195 170 25 -44 44 -87 -23 74 -98 -60 50 -109 -87 29 -116 -8 67 -75 -223 -35 -188 #curves 14 012 5 995 3 711 2 444 1 137 645 346 230 148 96 70 48 33 69 155 #entities 35 34 27 24 21 17 10 8 6 4 2 1 1 1 1 Table 10: Trade Information Warehouse Data This table reports the average gross (Column 3) and net (Column 4) notional amount (in million USD) on CDS contracts outstanding in USD equivalents of the sovereign reference entities among the 1,000 mostly traded contracts over the time period 31 October 2008 through 11 November 2011 (excluding sovereign US states). All numbers are reported in million USD. Column 5 indicates the average number of contracts live in the Depository Trust & Clearing Corporation’s (DTCC) Trade Information Warehouse (Warehouse) over the same time period. Column 6 reports the ratio of gross to net notional amount outstanding and column 7 is the average ratio of net notional amount outstanding to number of contracts outstanding. The notional values are represented as US dollar equivalents using the prevailing foreign exchange rates. The final row labeled Total reports the sum over all rows for column 3, 4 and 5, and the average for columns 6 and 7. The values are sorted in descending order. (1) Country Italy Turkey Brazil Spain Mexico Russia Germany Greece France Philippines Korea Portugal Hungary Argentina Venezuela Ukraine UK Austria SouthAfrica Indonesia Ireland China Belgium Colombia Poland Japan Kazakhstan Peru Malaysia Thailand Bulgaria Netherlands Romania United States Sweden Finland Denmark Australia Czech Rep. Slovak Rep. Latvia Iceland Vietnam Israel Panama Croatia Qatar Norway Lithuania Slovenia Chile Egypt Estonia NewZealand SaudiaArabia Lebanon Tunisia HongKong Total (2) Gross Notional 232 651 152 667 147 673 111 477 103 758 103 099 70 774 66 072 64 073 63 228 57 731 55 638 54 333 50 781 49 909 46 002 41 957 40 389 37 797 34 006 34 003 33 142 32 153 29 762 28 989 28 935 21 914 21 441 18 575 18 148 17 412 15 908 15 732 14 574 14 560 11 385 10 718 10 499 9 070 8 881 8 241 8 186 7 855 7 550 6 912 6 730 5 864 5 275 5 149 3 994 3 974 3 205 2 968 2 700 2 188 2 027 1 972 1 705 2 066 309 (3) Country Italy Spain Germany Brazil France UK Austria Portugal Greece Mexico Turkey Belgium Russia Japan Ireland Korea China Hungary Sweden United States Netherlands Philippines Indonesia Denmark SouthAfrica Poland Colombia Finland Argentina Australia Venezuela Peru Ukraine Romania Bulgaria Malaysia Thailand Kazakhstan Czech Rep. Slovak Rep. Iceland Norway Israel Slovenia Egypt Latvia Lithuania Panama Croatia Vietnam HongKong Chile NewZealand Qatar SaudiaArabia Lebanon Estonia Tunisia Total (4) Net Notional 23 287 14 846 13 665 13 307 12 687 7 740 7 120 7 102 6 926 6 845 5 748 5 642 4 973 4 709 4 594 4 054 3 952 3 702 2 907 2 804 2 762 2 677 2 261 2 239 2 190 2 144 2 108 2 031 2 004 1 979 1 953 1 806 1 697 1 255 1 213 1 199 1 136 1 086 1 002 941 903 881 808 805 726 726 710 699 663 602 586 539 517 514 483 459 457 289 203 661 Source: http://www.dtcc.com 39 (5) Country Brazil Turkey Mexico Russia Philippines Korea Italy Argentina Venezuela Spain Hungary Indonesia SouthAfrica Ukraine China Greece Colombia France Japan UK Poland Portugal Thailand Malaysia Peru Germany Kazakhstan Bulgaria Ireland Austria Romania Belgium Vietnam Iceland Latvia Australia Panama Croatia Israel Sweden Czech Rep. Netherlands Qatar Egypt Slovak Rep. Denmark Lithuania United States Finland Chile Estonia Slovenia Lebanon Tunisia NewZealand SaudiaArabia Norway HongKong Total (6) # Contracts 10 955 10 087 8 760 7 401 7 364 6 156 6 132 5 413 5 033 4 606 4 603 4 294 4 168 3 764 3 341 3 215 3 107 2 736 2 693 2 683 2 627 2 519 2 437 2 291 2 212 2 051 1 938 1 771 1 749 1 733 1 638 1 533 1 145 1 127 1 020 963 957 918 831 781 756 737 730 718 689 617 598 544 443 420 391 334 324 300 297 264 220 125 147 258 (7) Country Germany United States HongKong France Finland Austria Norway Italy Netherlands Sweden Belgium Denmark Spain UK Portugal Ireland Slovenia Greece Australia SaudiaArabia Japan NewZealand Lebanon Slovak Rep. Czech Rep. Chile Brazil Lithuania China Estonia Egypt Israel Tunisia Poland Peru Hungary Iceland Mexico Romania Panama Croatia Latvia Qatar Bulgaria Colombia Russia Korea Turkey Kazakhstan Indonesia SouthAfrica Vietnam Malaysia Thailand Ukraine Venezuela Argentina Philippines Average (8) Net/Contract 6.66 5.16 4.69 4.64 4.58 4.11 4.01 3.8 3.75 3.72 3.68 3.63 3.22 2.89 2.82 2.63 2.41 2.15 2.05 1.83 1.75 1.74 1.42 1.37 1.32 1.29 1.21 1.19 1.18 1.17 1.01 0.97 0.97 0.82 0.82 0.8 0.8 0.78 0.77 0.73 0.72 0.71 0.7 0.69 0.68 0.67 0.66 0.57 0.56 0.53 0.53 0.53 0.52 0.47 0.45 0.39 0.37 0.36 1.8 Table 11: Trade Information Warehouse Data This table reports the average gross (Column 3) and net (Column 4) notional amount (in million USD) on CDS contracts outstanding in USD equivalents of the sovereign reference entities among the 1,000 mostly traded contracts over the time period 31 October 2008 through 11 November 2011. All numbers are reported in million USD. Column 5 indicates the average number of contracts live in the Depository Trust & Clearing Corporation’s (DTCC) Trade Information Warehouse (Warehouse) over the same time period. Column 6 reports the ratio of gross to net notional amount outstanding and column 7 is the average ratio of net notional amount outstanding to number of contracts outstanding. The notional values are represented as US dollar equivalents using the prevailing foreign exchange rates. The final row labeled Total reports the sum over all rows for column 3, 4 and 5, and the average for columns 6 and 7. The values are reported in descending order according to the net notional amount outstanding. The countries are grouped in five regions: Americas, Asia ex-Japan, Australia and New Zealand, Europe/Middle East and Africa (EMEA) and Japan. (1) Country Italy Spain Germany Brazil France UnitedKingdom Austria Portugal Greece Mexico Turkey Belgium RussianFederation Japan Ireland Korea China Hungary Sweden United States Netherlands Philippines Indonesia Denmark SouthAfrica Poland Colombia Finland Argentina Australia Venezuela Peru Ukraine Romania Bulgaria Malaysia Thailand Kazakhstan CzechRepublic SlovakRepublic Iceland Norway Israel Slovenia Egypt Latvia Lithuania Panama Croatia Vietnam HongKong Chile NewZealand Qatar SaudiaArabia Lebanon Estonia Tunisia Total/Average(∗) (2) DC Region EMEA EMEA EMEA Americas EMEA EMEA EMEA EMEA EMEA Americas EMEA EMEA EMEA Japan EMEA Asia Ex-Japan Asia Ex-Japan EMEA EMEA EMEA EMEA Asia Ex-Japan Asia Ex-Japan EMEA EMEA EMEA Americas EMEA Americas Australia NZ Americas Americas EMEA EMEA EMEA Asia Ex-Japan Asia Ex-Japan EMEA EMEA EMEA EMEA EMEA EMEA EMEA EMEA EMEA EMEA Americas EMEA Asia Ex-Japan Asia Ex-Japan Americas Australia NZ EMEA EMEA EMEA EMEA EMEA (3) Gross Notional 232 651 111 477 70 774 147 673 64 073 41 957 40 389 55 638 66 072 103 758 152 667 32 153 103 099 28 935 34 003 57 731 33 142 54 333 14 560 14 574 15 908 63 228 34 006 10 718 37 797 28 989 29 762 11 385 50 781 10 499 49 909 21 441 46 002 15 732 17 412 18 575 18 148 21 914 9 070 8 881 8 186 5 275 7 550 3 994 3 205 8 241 5 149 6 912 6 730 7 855 1 705 3 974 2 700 5 864 2 188 2 027 2 968 1 972 2 066 309 (4) Net Notional 23 287 14 846 13 665 13 307 12 687 7 740 7 120 7 102 6 926 6 845 5 748 5 642 4 973 4 709 4 594 4 054 3 952 3 702 2 907 2 804 2 762 2 677 2 261 2 239 2 190 2 144 2 108 2 031 2 004 1 979 1 953 1 806 1 697 1 255 1 213 1 199 1 136 1 086 1 002 941 903 881 808 805 726 726 710 699 663 602 586 539 517 514 483 459 457 289 203 661 Source: http://www.dtcc.com 40 (5) # Contracts 6 132 4 606 2 051 10 955 2 736 2 683 1 733 2 519 3 215 8 760 10 087 1 533 7 401 2 693 1 749 6 156 3 341 4 603 781 544 737 7 364 4 294 617 4 168 2 627 3 107 443 5 413 963 5 033 2 212 3 764 1 638 1 771 2 291 2 437 1 938 756 689 1 127 220 831 334 718 1 020 598 957 918 1 145 125 420 297 730 264 324 391 300 147 258 (6) Gross/Net 9.99 7.51 5.18 11.10 5.05 5.42 5.67 7.83 9.54 15.16 26.56 5.70 20.73 6.14 7.40 14.24 8.39 14.68 5.01 5.20 5.76 23.62 15.04 4.79 17.26 13.52 14.12 5.60 25.33 5.30 25.56 11.87 27.11 12.53 14.35 15.49 15.98 20.18 9.05 9.44 9.06 5.99 9.35 4.96 4.41 11.35 7.25 9.89 10.14 13.04 2.91 7.37 5.22 11.40 4.53 4.41 6.50 6.82 10.74∗ (7) Net/Contract 3.80 3.22 6.66 1.21 4.64 2.89 4.11 2.82 2.15 0.78 0.57 3.68 0.67 1.75 2.63 0.66 1.18 0.80 3.72 5.16 3.75 0.36 0.53 3.63 0.53 0.82 0.68 4.58 0.37 2.05 0.39 0.82 0.45 0.77 0.69 0.52 0.47 0.56 1.32 1.37 0.80 4.01 0.97 2.41 1.01 0.71 1.19 0.73 0.72 0.53 4.69 1.29 1.74 0.70 1.83 1.42 1.17 0.97 1.80∗ 41 2008 . 48 945 . 17 169 12 964 136 964 15 442 2 897 19 814 28 208 4 326 4 929 4 866 . . 2 240 3 736 22 091 38 051 35 990 . 33 875 8 864 30 813 17 589 5 149 153 732 7 296 22 479 50 311 6 440 . 3 302 16 178 75 514 4 904 . 2 313 6 475 19 707 66 345 17 152 25 066 4 033 11 839 105 708 . 5 581 . 31 681 64 649 6 245 16 620 . 179 469 59 556 4 537 12 843 46 073 5 997 30 539 Panel A: Gross Notional (’000,000) 2009 2010 2011 2 493 4 032 6 683 45 810 51 213 56 383 2 722 8 025 19 634 30 413 44 938 51 203 21 035 30 670 50 583 123 478 148 786 176 463 15 647 17 676 19 535 3 262 4 134 4 825 24 200 33 615 45 583 27 278 29 422 33 344 5 775 6 880 8 136 7 389 10 047 10 691 7 995 10 966 14 742 . 5 649 7 264 . 2 236 4 174 2 762 3 280 2 984 6 545 13 011 16 591 37 464 62 020 105 638 49 896 74 005 97 639 48 128 78 189 78 552 . 1 705 . 43 582 57 916 66 628 8 839 7 974 7 544 31 737 34 112 37 142 24 300 37 482 44 400 6 115 7 225 10 069 187 481 238 376 293 888 12 034 29 874 51 688 23 375 22 225 19 745 56 473 59 615 58 449 7 359 8 724 9 053 1 675 1 949 2 127 4 476 5 394 6 006 18 174 18 961 19 063 88 742 106 557 123 461 10 125 17 825 22 533 1 659 2 393 3 086 2 959 5 435 8 251 6 375 6 792 7 761 18 658 19 760 26 984 67 258 63 381 57 767 22 602 30 165 37 352 41 192 62 951 69 833 4 789 5 540 7 855 13 878 16 542 17 700 101 719 103 458 103 749 . . 2 188 7 666 9 561 10 145 2 945 4 138 4 943 34 253 37 776 43 140 77 492 111 699 159 851 11 093 15 493 19 132 17 873 19 266 17 456 1 932 1 841 2 141 156 877 151 600 143 699 46 198 46 450 42 535 8 716 13 156 25 020 21 323 47 368 65 251 42 703 50 787 57 968 6 246 8 281 9 582 27 732 32 192 38 228 Source: http://www.dtcc.com Country AbuDhabi Argentina Australia Austria Belgium Brazil Bulgaria Chile China Colombia Croatia CzechRepublic Denmark Dubai Egypt Estonia Finland France Germany Greece HongKong Hungary Iceland Indonesia Ireland Israel Italy Japan Kazakhstan Korea Latvia Lebanon Lithuania Malaysia Mexico Netherlands NewZealand Norway Panama Peru Philippines Poland Portugal Qatar Romania RussianFederation SaudiaArabia SlovakRepublic Slovenia SouthAfrica Spain Sweden Thailand Tunisia Turkey Ukraine United States UnitedKingdom Venezuela Vietnam Average Average 4 403 50 588 10 127 35 931 28 813 146 423 17 075 3 780 30 803 29 563 6 279 8 264 9 642 6 457 3 205 2 816 9 971 56 803 64 898 60 215 1 705 50 500 8 305 33 451 30 943 7 140 218 369 25 223 21 956 56 212 7 894 1 917 4 795 18 094 98 568 13 847 2 379 4 740 6 851 21 277 63 688 26 818 49 760 5 554 14 990 103 659 2 188 8 238 4 009 36 713 103 423 12 991 17 804 1 971 157 911 48 685 12 857 36 696 49 383 7 527 32 376 2008 . 2 815 . 4 883 3 871 11 160 1 717 670 2 099 2 442 680 1 139 1 706 . . 549 1 413 5 533 9 905 7 459 . 4 388 1 193 2 194 4 485 595 17 388 1 741 2 471 5 326 891 . 709 1 485 4 529 1 588 . 651 708 2 158 2 903 2 083 5 030 456 1 561 7 188 . 1 017 . 2 484 13 895 1 829 1 187 . 5 816 2 777 1 296 2 368 2 175 627 3 305 (14)). The values are reported in descending order according to the country name. Panel B: Net Notional (’000,000) 2009 2010 2011 Average 468 834 1 379 894 1 817 1 755 2 353 2 185 438 1 427 3 862 1 909 6 933 8 325 6 364 6 626 4 836 5 670 6 893 5 317 10 204 13 811 16 730 12 976 1 331 1 167 1 031 1 311 482 548 568 567 1 938 3 723 6 919 3 670 2 152 1 983 2 136 2 178 611 674 709 668 1 145 946 874 1 026 1 887 2 330 2 646 2 142 . 519 545 532 . 518 935 726 466 456 428 475 1 648 2 272 2 314 1 912 7 587 11 744 21 123 11 497 10 936 13 988 17 192 13 005 7 959 7 533 4 911 6 966 . 586 . 586 4 052 3 569 3 316 3 831 976 815 864 962 1 860 1 974 3 075 2 276 4 741 4 868 4 123 4 554 576 662 1 289 781 20 972 25 306 24 762 22 107 2 061 4 843 8 206 4 213 1 271 855 867 1 366 3 720 3 772 4 517 4 334 792 720 623 757 412 441 482 445 785 703 632 707 1 149 1 016 1 416 1 266 5 769 6 727 8 690 6 429 2 524 3 117 2 852 2 520 481 495 544 507 684 975 1 035 836 646 649 818 705 1 705 1 612 2 083 1 889 2 569 2 506 2 957 2 734 1 936 2 237 2 288 2 136 6 659 8 355 6 554 6 649 341 434 821 513 1 377 1 189 1 130 1 314 5 951 4 110 4 415 5 416 . . 483 483 1 042 888 870 954 687 815 918 807 2 079 2 106 2 357 2 256 12 095 15 087 17 932 14 752 2 842 3 086 2 987 2 686 1 051 1 036 1 341 1 154 281 268 317 289 5 396 5 809 6 067 5 772 2 062 1 541 1 242 1 905 1 987 2 441 4 478 2 550 3 361 9 185 12 174 6 772 1 630 1 884 2 363 2 013 423 602 806 614 2 818 3 145 3 674 3 239 2008 . 17.4 . 3.5 3.3 12.3 9.0 4.3 9.4 11.6 6.4 4.3 2.9 . . 4.1 2.6 4.0 3.8 4.8 . 7.7 7.4 14.0 3.9 8.6 8.8 4.2 9.1 9.4 7.2 . 4.7 10.9 16.7 3.1 . 3.6 9.1 9.1 22.9 8.2 5.0 8.8 7.6 14.7 . 5.5 . 12.8 4.7 3.4 14.0 . 30.9 21.4 3.5 5.4 21.2 9.6 8.8 Panel C: Ratio Gross/Net 2009 2010 2011 Average 5.3 4.8 4.8 5.0 25.2 29.2 24.0 23.9 6.2 5.6 5.1 5.6 4.4 5.4 8.0 5.3 4.3 5.4 7.3 5.1 12.1 10.8 10.5 11.4 11.8 15.1 19.0 13.7 6.8 7.5 8.5 6.8 12.5 9.0 6.6 9.4 12.7 14.8 15.6 13.7 9.5 10.2 11.5 9.4 6.5 10.6 12.2 8.4 4.2 4.7 5.6 4.3 . 10.9 13.3 12.1 . 4.3 4.5 4.4 5.9 7.2 7.0 6.0 4.0 5.7 7.2 4.9 4.9 5.3 5.0 4.8 4.6 5.3 5.7 4.8 6.0 10.4 16.0 9.3 . 2.9 . 2.9 10.8 16.2 20.1 13.7 9.1 9.8 8.7 8.7 17.1 17.3 12.1 15.1 5.1 7.7 10.8 6.9 10.6 10.9 7.8 9.5 8.9 9.4 11.9 9.8 5.8 6.2 6.3 5.6 18.4 26.0 22.8 19.1 15.2 15.8 12.9 13.3 9.3 12.1 14.5 10.8 4.1 4.4 4.4 4.3 5.7 7.7 9.5 6.9 15.8 18.7 13.5 14.7 15.4 15.8 14.2 15.5 4.0 5.7 7.9 5.2 3.4 4.8 5.7 4.7 4.3 5.6 8.0 5.4 9.9 10.5 9.5 9.7 10.9 12.3 13.0 11.3 26.2 25.3 19.5 23.5 11.7 13.5 16.3 12.4 6.2 7.5 10.7 7.3 14.0 12.8 9.6 11.3 10.1 13.9 15.7 11.8 17.1 25.2 23.5 20.1 . . 4.5 4.5 7.4 10.8 11.7 8.8 4.3 5.1 5.4 4.9 16.5 17.9 18.3 16.4 6.4 7.4 8.9 6.8 3.9 5.0 6.4 4.7 17.0 18.6 13.0 15.7 6.9 6.9 6.7 6.8 29.1 26.1 23.7 27.4 22.4 30.1 34.2 27.1 4.4 5.4 5.6 4.7 6.3 5.2 5.4 5.6 26.2 27.0 24.5 24.7 14.8 13.8 11.9 12.5 10.0 11.2 11.6 10.5 ends on 11 November 2011. The overall statistics in the final row labeled Average include information for the U.S. states, which are reported in a separate table (Table outstanding. The notional values are represented as US dollar equivalents using the prevailing foreign exchange rates. The sample period starts on 31 October 2008 and among the 1,000 mostly traded contracts as of 11th November 2011. All numbers are reported in million USD. Panel C indicates the ratio of gross to net notional amount This table reports the average yearly gross (Panel A) and net (Panel B) notional amount on CDS contracts outstanding in USD equivalents of the sovereign reference entities Table 12: Trade Information Warehouse Data 42 2008 . 5 011 . 686 437 11 097 1 515 302 1 736 3 168 588 415 236 . . 304 92 547 760 1 131 . 3 126 1 209 3 973 660 546 3 390 295 2 097 4 612 857 . 423 1 929 6 560 157 . 58 895 2 168 8 037 1 589 797 416 1 294 7 483 . 451 . 3 639 2 008 276 2 061 . 13 325 5 402 124 433 4 781 866 2 279 Panel A: # Contracts 2009 2010 2011 427 866 1 510 4 916 5 478 5 990 270 684 1 847 1 338 1 905 2 196 793 1 441 2 715 9 818 11 097 12 074 1 618 1 782 1 985 359 432 498 2 177 3 458 4 869 3 112 3 024 3 186 791 937 1 109 649 823 870 399 498 1 087 . 691 941 . 424 1 012 375 424 389 242 503 675 1 005 2 490 5 464 1 185 2 097 3 254 1 701 3 944 4 524 . 125 . 3 857 4 788 5 542 1 111 1 149 1 105 3 967 4 314 4 713 1 000 1 858 2 704 665 743 1 183 4 122 6 401 8 685 655 2 840 5 354 2 143 1 854 1 769 5 630 6 590 6 561 972 1 068 1 053 244 305 349 541 612 682 2 195 2 367 2 386 8 061 8 979 9 748 413 818 1 134 145 256 348 90 228 387 896 924 1 078 2 031 1 969 2 715 8 014 7 378 6 464 2 136 2 725 3 285 1 446 2 834 3 732 525 638 1 139 1 463 1 672 1 869 7 352 7 394 7 447 . . 264 618 734 765 274 328 407 3 916 4 122 4 620 2 429 4 699 7 533 594 795 1 083 2 267 2 569 2 553 269 283 332 11 088 9 580 8 880 3 958 3 587 3 420 241 479 1 054 881 3 120 4 699 4 464 5 093 5 669 900 1 206 1 413 1 994 2 273 2 736 Source: http://www.dtcc.com Country AbuDhabi Argentina Australia Austria Belgium Brazil Bulgaria Chile China Colombia Croatia CzechRepublic Denmark Dubai Egypt Estonia Finland France Germany Greece HongKong Hungary Iceland Indonesia Ireland Israel Italy Japan Kazakhstan Korea Latvia Lebanon Lithuania Malaysia Mexico Netherlands NewZealand Norway Panama Peru Philippines Poland Portugal Qatar Romania RussianFederation SaudiaArabia SlovakRepublic Slovenia SouthAfrica Spain Sweden Thailand Tunisia Turkey Ukraine United States UnitedKingdom Venezuela Vietnam Average Total 934 5 349 934 1 531 1 346 11 021 1 725 398 3 060 3 123 856 689 555 816 718 373 378 2 377 1 824 2 825 125 4 328 1 143 4 242 1 556 784 5 649 2 286 1 966 5 848 987 299 565 2 219 8 337 630 250 191 948 2 221 7 473 2 434 2 202 680 1 574 7 419 264 642 336 4 074 4 167 687 2 363 295 10 718 4 092 475 2 283 5 002 1 096 2 330 Panel B: Ratio Net Notional/Contracts (’000,000) 2008 2009 2010 2011 Total . 1.1 1.0 0.9 1.0 0.6 0.4 0.3 0.4 0.4 . 1.6 2.1 2.1 1.9 7.1 5.2 4.4 2.9 4.9 8.9 6.1 3.9 2.5 5.4 1.0 1.0 1.2 1.4 1.2 1.1 0.8 0.7 0.5 0.8 2.2 1.3 1.3 1.1 1.5 1.2 0.9 1.1 1.4 1.1 0.8 0.7 0.7 0.7 0.7 1.2 0.8 0.7 0.6 0.8 2.7 1.8 1.1 1.0 1.7 7.2 4.7 4.7 2.4 4.8 . . 0.8 0.6 0.7 . . 1.2 0.9 1.1 1.8 1.2 1.1 1.1 1.3 15.3 6.8 4.5 3.4 7.5 10.1 7.6 4.7 3.9 6.6 13.0 9.2 6.7 5.3 8.6 6.6 4.7 1.9 1.1 3.6 . . 4.7 . 4.7 1.4 1.1 0.7 0.6 0.9 1.0 0.9 0.7 0.8 0.8 0.6 0.5 0.5 0.7 0.5 6.8 4.7 2.6 1.5 3.9 1.1 0.9 0.9 1.1 1.0 5.1 5.1 4.0 2.9 4.3 5.9 3.1 1.7 1.5 3.1 1.2 0.6 0.5 0.5 0.7 1.2 0.7 0.6 0.7 0.8 1.0 0.8 0.7 0.6 0.8 . 1.7 1.4 1.4 1.5 1.7 1.5 1.1 0.9 1.3 0.8 0.5 0.4 0.6 0.6 0.7 0.7 0.7 0.9 0.8 10.1 6.1 3.8 2.5 5.6 . 3.3 1.9 1.6 2.3 11.2 7.6 4.3 2.7 6.5 0.8 0.7 0.7 0.8 0.7 1.0 0.8 0.8 0.8 0.9 0.4 0.3 0.3 0.5 0.4 1.3 0.9 0.8 0.7 0.9 6.3 4.6 2.9 1.8 3.9 1.1 0.6 0.7 0.7 0.8 1.2 0.9 0.7 0.6 0.9 1.0 0.8 0.6 0.6 0.7 . . . 1.8 1.8 2.3 1.7 1.2 1.1 1.6 . 2.5 2.5 2.3 2.4 0.7 0.5 0.5 0.5 0.6 6.9 5.0 3.2 2.4 4.4 6.6 4.8 3.9 2.8 4.5 0.6 0.5 0.4 0.5 0.5 . 1.0 0.9 1.0 1.0 0.4 0.5 0.6 0.7 0.6 0.5 0.5 0.4 0.4 0.5 10.4 8.2 5.1 4.2 7.0 5.5 3.8 2.9 2.6 3.7 0.5 0.4 0.4 0.4 0.4 0.7 0.5 0.5 0.6 0.6 3.6 2.5 1.8 1.4 2.2 include information for the U.S. states, which are reported in a separate table. The values are reported in descending order according to the country name. prevailing foreign exchange rates. The sample period starts on 31 October 2008 and ends on 11 November 2011. The overall statistics in the final row labeled Average sovereign reference entities among the 1,000 mostly traded contracts as of 11th November 2011. The notional values are represented as US dollar equivalents using the A) and the average ratio (Panel B) of net notional amount on CDS contracts outstanding in million USD equivalents to the average yearly number of contracts of the This table reports the average yearly number of contracts live in the Depository Trust & Clearing Corporation’s (DTCC) Trade Information Warehouse (Warehouse) (Panel Table 13: Trade Information Warehouse Data 43 2008 . . . . . . . . 4 537 Panel A: Gross Notional (’000,000) 2009 2010 2011 Average 4 253 7 928 9 981 7 387 2 688 3 855 4 636 3 726 . 2 061 3 088 2 575 . 1 682 1 707 1 695 1 989 3 075 4 358 3 141 2 333 2 922 3 318 2 857 3 004 4 423 5 655 4 361 1 937 2 363 2 719 2 340 8 716 13 156 25 020 12 857 Source: http://www.dtcc.com Country California Florida Illinois Masschusetts NewJersey NewYork NewYorkCity Texas United States 2008 . . . . . . . . 1 296 Panel B: Net Notional (’000,000) 2009 2010 2011 Average 698 885 986 856 298 296 267 287 . 329 608 468 . 171 174 172 516 440 475 477 282 348 437 355 756 528 386 557 394 221 186 267 1 987 2 441 4 478 2 550 2008 . . . . . . . . 3.5 Panel D: Ratio Gross/Net 2009 2010 2011 Total 6.1 9.0 10.1 8.4 9.0 13.0 17.4 13.2 . 6.3 5.1 5.7 . 9.9 9.8 9.8 3.9 7.0 9.2 6.7 8.3 8.4 7.6 8.1 4.0 8.4 14.7 9.0 4.9 10.7 14.6 10.1 4.4 5.4 5.6 4.7 ends on 11 November 2011. The values are reported in descending order according to the sovereign’s name. 2008 . . . . . . . . 124 Panel 2009 232 161 . . 120 104 171 81 241 C: # Contracts 2010 2011 599 989 274 318 152 403 112 116 215 401 194 266 262 356 96 155 479 1 054 Average 607 251 277 114 245 188 263 111 475 (Warehouse). The notional values are represented as US dollar equivalents using the prevailing foreign exchange rates. The sample period starts on 31 October 2008 and amount outstanding and Panel D lists the average yearly number of contracts live in the Depository Trust & Clearing Corporation’s (DTCC) Trade Information Warehouse entities among the 1,000 mostly traded contracts as of 11th November 2011. All numbers are reported in million USD. Panel C indicates the ratio of gross to net notional This table reports the average yearly gross (Panel A) and net (Panel B) notional amount on CDS contracts outstanding in USD equivalents of the U.S. sovereign reference Table 14: Trade Information Warehouse Data 44 (a) windows of 400 curves. Source: Markit. (b) up to the right-most point representing curves 71,801 to 72,200. Figure 1b shows the average slope level in basis points as a function of CDS level, in sliding against the average CDS premium among the 400 curves with the lowest 5-year spread level. The next point does the same for curves 2 to 401, and so on, 10y-1y. The 72,200 curves are sorted by the 5-year premium. Source: Markit. For instance, the left-most points plots the fraction of decreasing CDS pairs Graph (1a) plots the fraction of decreasing CDS pairs as a function of the CDS level in sliding windows of 400 basis points for the segments 3y-1y, 10y-3y and Figure 1: The Slope of the Credit Spread Curve 45 Figure 2: Co-Movement of Sovereign CDS Spreads and Risk Aversion 0 05/2003 200 400 600 800 1000 1200 02/2004 11/2004 08/2005 06/2006 03/2007 Time 12/2007 5-year Sovereign CDS Spreads (Source: Markit) (a) 09/2008 06/2009 04/2010 inspired by the illustration in Pan and Singleton (2008) Source: Markit. (b) 19th, 2010. Graph (2b) plots the historical average 5-year CDS spread of the 38 countries in the sample over the same time period. The second plot was Graph (2a) illustrates the historical 5-year CDS spread for 38 countries spanning five geographical regions over the time period May 9th, 2003 until August Spread Figure 3: Trade Information Warehouse Data The following graphs plot the average yearly net notional amount of CDS outstanding (in million USD) by the yearly amount of Government debt (in million USD) for the EMEA and Asia ex-Japan countries among the 1,000 mostly traded reference entities. The graphs are separated by year. Graph (3a), (3b) and (3c) refer to 2008, 2009 and 2010 respectively. The notional values are represented as US dollar equivalents using the prevailing foreign exchange rates. Source: www.dtcc.com. (a) (b) (c) 46 Figure 4: Trade Information Warehouse Data The following graphs plot the average yearly gross notional amount of CDS outstanding (in million USD) by the yearly amount of Government debt (in million USD) for the EMEA and Asia ex-Japan countries among the 1,000 mostly traded reference entities. The graphs are separated by year. Graph (4a), (4b) and (4c) refer to 2008, 2009 and 2010 respectively. The notional values are represented as US dollar equivalents using the prevailing foreign exchange rates. Source: www.dtcc.com. (a) (b) (c) 47