Entry and exit in OTC derivatives markets

Andy Atkeson, Andrea Eisfeldt, Pierre-Olivier Weill

17 September 2014



Several observations regarding entry and trading patterns in over-the-counter (OTC) markets for interest rate and credit derivatives have recently received considerable attention from policymakers and the public alike. 

  • OTC derivatives markets are large. For example, the total gross notional of derivatives contracts on US bank balance sheets is over 17 times the size of their assets. 
  • The large volume of bilateral trades at varied prices creates an intricate liability structure between participating banks. These liability linkages are thought to potentially create systemic risk (see Stultz 2010 for a detailed discussion in the context of CDS markets).
  • OTC derivatives gross notionals are concentrated on the balance sheets of the largest dealer banks. In the US, on average, about 95% of OTC derivatives gross notional held on bank balance sheets was concentrated in the top five banks (Atkeson et al. 2013 describe the stylised facts characterising CDS markets).  
  • These large banks intermediate a large volume of trade. They trade many contracts in both directions, long and short, and they have very large gross positions relative to their net positions. 
  • In contrast, medium-sized banks act as ‘customers’ – they trade mostly in one direction, either long or short, so their gross and net positions are close to each other.

To put a few of these facts into context, Figure 1 plots the net to gross notional for index credit default swap (CDS) contracts reported by the Bank for International Settlements (BIS) in the first half of 2013. The figure shows that reporting dealer banks and central clearing parties tend to net almost all of their long and short trades, while other institutions are either long or short on net.  Gross notional is not plotted, but, as we will show below, reporting dealers trade the majority of notional value, followed closely by central clearing parties.

Figure 1. Net/gross trades H2 2013 (CDS bought-CDS sold)/(CDS bought+CDSsold)

Figure 1 also shows that only the hedge fund sector sold CDS index contracts on net. In the second half of 2013, $8.75 trillion in total CDS index notional was traded. This seems a very large gross volume of trade in order to move credit risk from those who bear too much (loan making institutions and corporate bond holders) to those who may hold too little (at least ex-ante).

The OTC trading patterns outlined above have raised a number of interrelated questions. Why do OTC markets create such a large demand for dealers’ intermediation services? Why are much of these intermediation services provided by large dealer banks? Do these dealer banks provide too few or too many intermediation services relative to what might be socially optimal? Does the concentration of intermediation activity in these large banks make OTC markets excessively vulnerable to negative shocks?

In response to some of these concerns, considerable regulatory efforts have been made to improve disclosure (through trade reporting) to learn more about the operation of these markets, and to mitigate counterparty risk (through central clearing and margin requirements, see FSB 2014). Correspondingly, the academic community has focused much of its theoretical and empirical effort on central clearing and its effects on collateral needs and counterparty risk (see Duffie and Zhu 2011, Duffie et al. 2014).

Figure 2. Share of CDS bought + CDS sold H1 2011

However, despite these efforts, little is understood about the social role of dealers in OTC derivatives markets. In a recent article, ISDA states that dealer banks are a segment “that is often portrayed as purely speculative and contributing little to the economy at large.” They disagree with this view and argue instead that dealer banks provide socially valuable intermediation services.1 While ISDA’s article must, of course, be read with a critical eye, it shows that in order to assess the desirability of the many regulations and policy interventions that will shape the future structure of OTC markets, it is critical to study the private and social incentives of dealers to participate in these markets and to provide intermediation services.

Banks trading in OTC derivatives markets

To study this question, in recent research we develop a formal analysis of banks’ incentives to enter into, and trade in, OTC derivatives markets. Our conceptual framework highlights the role of entry and exit incentives in shaping the structure and ultimate resilience of OTC markets (Atkenson et al. 2014). In our model, all banks have two motives for participating and trading, namely to hedge by changing their underlying risk exposure, or to make intermediation profits. Due to the frictions in OTC markets, risk sharing is imperfect and participants have heterogeneous valuations for the hedging derivative contract. Because trades are bilateral, this creates price dispersion and incentives to make intermediation profits. We show that the entry incentives associated with intermediation profits are smaller than the entry incentives associated with hedging. Hence, given fixed entry costs, banks require a large enough scale to enter as a pure intermediary. In equilibrium, only large banks enter to become dealers, and middle-sized banks only enter as customers.

Figure 3. Share of CDS bought + CDS sold H2 2013

We show that these large dealers do in fact play a socially valuable role in mitigating OTC market frictions. In our analysis, we assume that, because of risk management concerns, individual trade sizes are limited and thus dealers can help to ‘clear’ contracts between customers. However, these dealer banks also tend to provide too many intermediation services relative to what would be socially optimal. This is because dealer banks have a business stealing motive for entry and trade in OTC markets. They do not internalise whether the trades they create lead to new hedging opportunities for final customers, or simply crowd out intermediary services provided by other dealer banks. In our modelling framework, a policy of taxing the entry of dealer banks and subsidising the entry of final customers into the OTC market would improve the social allocation of risk.

Our work also speaks to concerns about large dealers’ contribution to systemic risk in OTC markets. Because these dealer banks derive value mainly from intermediating many trades, we find that they indeed have the weakest incentive to remain active in the market in the face of negative market shocks. Moreover, the exit of these dealer banks has externalities on other market participants by reducing trading opportunities and thus reducing other banks’ incentives to remain active in the market. Hence, the exit of dealers can trigger the exit of other market participants. Whether these large dealer banks exit more than is socially optimal in a face of a negative market shock, however, depends crucially on market resilience. In particular, we show that, for it to be optimal to subsidise the continued participation of some large dealers, OTC market participants must find it sufficiently difficult to engage with new counterparties following the exit of dealers from the market. 

Concluding remarks

In conclusion, as evidenced by ISDA (2014), market participants and regulators are still in disagreement about whether the private profits that large dealer banks gain from intermediating trade in OTC markets encourage socially optimal market structures. Tables 1.1 and 2.1 in FSB (2014) summarise the global progress of reforms to trade reporting, central clearing, exchange trading, capital requirements, and margin requirements. While efforts to promote centralisation and transparency in OTC markets should help to shed light on the surplus that is created and its distribution across participants, they will also change this surplus in ways that may not be anticipated. Many of the proposed and implemented regulations effectively reduce the surplus accruing to dealers, and so reduce their incentives to participate in OTC markets. Clearly, this hurts dealers; however, the overall change in value may either increase or decrease, depending on whether there was previously too much or too little dealer activity. Our model shows that typically in equilibrium some reduction in dealer rents is warranted. However, one should be careful not to reduce rents so much as to discourage socially valuable intermediation. Our work also suggests that particular care should be given not to reduce rents when dealer participation costs are high, such as during times when negative shocks drive exit. This is because such times may coincide with less than socially optimal dealer participation. In short, we hope that our work will help to highlight participation incentives in order to preserve the ability to share risk in OTC markets, while limiting the inefficiencies associated with excess intermediation and dependence on large dealers.


Atkeson, A, A Eisfeldt, and P O Weill (2012), “The market for OTC credit derivatives”, Working paper, UCLA.

Atkeson, A, A Eisfeldt, and P O Weill (2014), “Entry and Exit in OTC Deriviatives Markets”, Working paper, UCLA.

BIS Statistics http://www.bis.org/statistics/dt21.csv

Duffie, D, M Scheicher, G Vuillemey (2014), “Central Clearing and Collateral Demand” Working paper.

Duffie, D and H Zhu (2011), “Does a Central Clearing Counterparty Reduce Counterparty Risk?”, Review of Asset Pricing Studies.

FSB (201), “OTC Derivatives Market Reforms Seventh Progress Report on Implementation”, 8 April.

ISDA Research Study (2014), “Dispelling Myths: End-User Activity in OTC Derivatives”, August.

OCC (2014), “Quarterly Report on Bank Trading and Derivatives Activities”.

Stultz, R M (2010), “Credit Default Swaps and the Credit Crisis”, Journal of Economic Perspectives, Volume 24, Number 1, Pages 73–92


1 See ISDA (2014), page17.



Topics:  Financial markets International finance

Tags:  OTC derivatives markets, trading, banks in OTC markets

Stanley M Zimmerman Professor of Finance and Economics and Director of the Business Economics Program, UCLA

Professor of Finance at the Anderson School of Management, UCLA

Associate Professor of Economics, UCLA; Research Affiliate, CEPR