The financial crisis of 2007–08 gave rise to widespread concerns that over-the-counter derivatives contributed to the build-up of systemic risk. As such, regulators have been quick to request and gain access to more information about activity in these markets. This information is crucial for the supervision of financial institutions and for designing more effective prudential policies (see for example Brunnermeier et al. 2013 and Peltonen et al. 2013). Equally importantly, it is invaluable for understanding the structure and dynamics of these hitherto opaque markets. Based on recent work of ours (Benos et al. 2013) we describe here some aspects of the UK single-name credit default swap (CDS) market. In particular, we focus on four policy-relevant questions:

1. What is the structure of the CDS trading network?
2. What was CDS market activity like during the financial crisis?
3. Was counterparty risk priced in the CDS market during the financial crisis?
4. Do different market participants trade at different prices in the UK CDS market?

Below we address each of them in turn.

What is the structure of the CDS trading network?

The CDS trading network is two-tiered. It consists of an inner core of major dealers and a periphery of end-users and dealer clients. Table 1 highlights the central role that dealers play in this market. The first thing to notice in Table 1 is that dealers are a counterparty to almost every transaction (they participate in about 98% of all traded volume). This means that dealers are the primary source of liquidity in this market as they facilitate the vast proportion of trades.

The second thing to notice is that the inter-dealer segment of the market accounts for about two thirds (64%) of all activity. This highlights the intermediation role that dealers play – for every client-initiated trade that a dealer accommodates, the dealer will try to find a counterparty willing to take an offsetting position, since dealers try to maintain balanced books. If such a counterparty is not available, the dealer will take an offsetting position with another dealer, and so on, until a suitable counterparty is found. Thus, each dealer-to-client trade triggers a sequence of additional and offsetting inter-dealer transactions until one or more of the dealers find an end-user willing to take the opposite side of the original trade. Using a similar dataset for the US CDS market, Shachar (2012) finds that the average dealer-to-client trade is passed between three dealers before a suitable ultimate counterparty is found. This explains the high volume of the inter-dealer CDS market segment.

Table 1. Trading network in the UK single-name CDS market (in % of total trading volume), Jan 2007–Dec 2011.

What was CDS market activity like during the financial crisis?

All over-the-counter derivatives carry counterparty risk. This is the risk that one of the two parties to a contract fails to meet their contractual obligations, which leaves their counterparty exposed to risk. Thus, concerns about counterparty risk may undermine the smooth functioning of over-the-counter derivatives markets. For instance, a hedge fund may choose not to buy CDS protection from a dealer who it believes is at risk. How did the UK CDS market perform throughout and during the peak of the financial crisis, given the overall deterioration of credit quality during this period?

Figure 1 shows that, although aggregate daily volume in the UK CDS market dropped during our sample period from about €2 billion in 2007–08 to about €1 billion after 2009, the market never really shut down. Strikingly, volumes shot up in the wake of Lehman’s default; this is surpising given that Lehman Brothers was a major dealer in the CDS market. It thus appears that the market held up well against concerns about counterparty risk. Figure 2 plots the dealer buy, sell, and net volume in relation to all other market participants. Consistent with the pattern of aggregate volumes, Figure 2 shows that dealers continued to intermediate between ultimate CDS buyers and sellers throughout and at the peak of the financial crisis.

Figure 1. Daily aggregate trading volume (€ billions) in the UK single-name, corporate, CDS market, Jan 2007-Dec 2011

Note: The bold line denotes the ten-day moving average; the vertical line marks the date of Lehman Brothers' default.

Figure 2. Total dealer buy, sell, and net volume (€ billions) traded with CDS market end-users on single-name CDS contracts written on UK reference entities, Jan 2007-Dec 2011

Was counterparty risk priced in the CDS market during the financial crisis?

Although aggregate activity was not compromised, counterparty risk may still have been reflected in the execution prices of trades. For example, dealers who were more at risk may have been only able to sell protection at lower prices relative to more creditworthy dealers. To see if that was the case, we isolate all the transactions in our sample where a G-16 dealer is on the selling side. This includes transactions where dealers sell CDS contracts both to other dealers and to end-users. For these transactions, we calculate the distance between the transaction spread and the corresponding par spread, and normalise by the par spread. Then we plot, in Figure 3, this relative distance on the vertical axis against the selling dealer’s own CDS spread on the horizontal axis.

If counterparty risk were reflected in execution prices, the relationship between the distance of the transaction and par spreads on one hand, and the dealer CDS spread on the other would be negative, and the linear fit of the observations in Figure 3 would be downward-sloping. However, the linear fit in Figure 3 (blue line) is almost indistinguishable from zero. This is consistent with earlier findings in the academic literature (e.g. Arora et al. 2012) and also with the fact that there was no disruption in the UK CDS market during the financial crisis. Thus, it appears that the collateral and netting arrangements in place, by and large, mitigated counterparty credit risk concerns in the CDS market during this period of market stress.

Figure 3. Scatter-plot of the unconditional relationship between the relative spreads (%) at which the G-16 dealers sell UK CDS contracts and the selling dealer CDS spread, Jan 2007–June 2009

Do different market participants trade at different prices in the UK CDS market?

Finally, we take a look at execution prices in the UK CDS market. The CDS market is generally opaque – dealers do not observe each other’s quoted prices to end-users. Moreover, throughout our sample time there was no publicly-available record of CDS transactions that would allow end-users to infer the prevailing CDS valuations from recently concluded transactions. This opacity of the CDS market could in principle lead to considerable costs for end-users, who need to search for the best execution price available from a relatively small group of dealers.

To see if and how the lack of transparency manifests itself in the UK CDS market, we calculate again the relative distance between transaction spreads and end-of-day par spreads separately for inter-dealer and dealer to end-user transactions. By doing this, we can infer the impact of the CDS market transparency regime on end-users’ trading costs. In other words, if, under the prevailing levels of transparency, end-users face higher execution costs, then we would expect end-users’ relative distance measure to be substantially higher than that of dealers.

In Figure 4, we plot the median of the relative distance between transaction spreads and par spreads, separately for dealer-to-dealer, dealer-bank, dealer-asset manager and dealer-hedge fund trades, focusing on the most heavily traded five-year tenor. In addition to the counterparty breakdown, we report the median trading cost measures separately by size – ‘small trades’ (<= 5 million) and ‘large trades’ (> 5 million notional amount) – and seniority (senior vs. subordinated). We also augment the median by the interquartile range in order to facilitate comparison across the different break-downs.

Figure 4 shows that there are relatively small variations in our trading cost measure across inter-dealer and dealer to end-user trades. While the dispersion of transaction prices around par spreads is generally lower for inter-dealer trades, the differences are not economically significant – the medians of the relative distances are no more than 1% apart from each other across breakdowns. For a contract with a par spread of 200 basis points and a notional amount of €5 million, a relative distance of 1% would correspond to an additional trading cost of €1,000 per annum.1

Thus, our findings suggest that the relative opacity of the UK CDS market does not seem to cause end-users to trade at much inferior prices relative to the dealers. This may seem puzzling, but one potential explanation is that most CDS market participants are sophisticated and presumably well informed; there are no retail investors in this market. In Benos et al. (2013) we discuss some additional potential explanations for this lack of discrepancy in execution prices.

Figure 4. Medians and interquartile ranges of the relative (to the par spread) distance between transaction spreads and end-of-day par spreads for five-year senior (top panel) and subordinated (bottom panel) CDS contracts, Jan 2007–June 2009


Benos E, A Wetherilt and F Zikes (2013), “The structure and dynamics of the UK credit default swap market”, Bank of England Financial Stability Paper No. 25, November.

Brunnermeier M, L Clerc, Y El Omari, S Gabrieli, S Kern, C Memmel, T Peltonen, N Podlich, M Scheicher, and G Vuillemey (2013), “Assessing contagion risks from the CDS market”, European Systemic Risk Board Occasional Paper No. 4, September.

Arora, N, P Gandhi, and F Longstaff (2012), “Counterparty credit risk and the credit default swap market”, Journal of Financial Economics, 103: 280–293.

Peltonen T, M Scheicher, and G Vuillemey (2013), “The network structure of the CDS market and its determinants”, European Central Bank Working Paper No. 1583, August.

Shachar, O (2012), “Exposing the exposed: intermediation capacity in the credit default swap market”, New York University, mimeo.

1 1% x 200 bps x €5 million = €1,000

2,732 Reads