Betting the Business

Reforming the financial system involves not only the grand public battles over legislation and rulemaking, but also the substantial trenchwork that falls to staffers in the many agencies responsible for carrying out the mandates. And it is heartening to observe this work advance. One of the many interesting analyses being produced en route is a study by Daniel Heller and Nicholas Vause at the Bank for International Settlements (BIS), the international organization of central banks. The purpose of the report is to produce an estimate of the financial resources that Central Counterparties (CCPs) would need to safely clear interest rate and credit default swaps. Central clearing of derivative trades is one of the major mandates of the Dodd-Frank Financial Reform Act in the U.S. and of comparable reforms in Europe.

In a number of our previous posts, we have emphasized that end-users have much to gain from the central clearing mandate—see here, here, here and here. Central clearing creates the possibility to reduce the total amount of credit risk in the system, lowering the overall costs to the various parties using derivatives, including non-financial companies seeking to hedge their commercial risk. One reflection of how risks can be minimized, depending upon how the mandates are implemented, shows up in Graph 6 of this BIS study, reproduced below.

The graph shows the impact of integrating the clearing of different types of derivatives at a single clearinghouse. Suppose, for example, that one clearinghouse clears all single-name credit default swaps, and another clearinghouse clears all multi-name credit default swaps. The report calculates the potential losses at each clearinghouse and then sums the two. This establishes a benchmark level of credit risk. Now, as an alternative, suppose that instead of two separate clearinghouses for the two types of swaps, there is a single clearinghouse that handles both single-name and multi-name credit default swaps. This report calculates the potential losses at this integrated clearinghouse. The red line in the Graph shows the size of losses at this single, integrated clearinghouse as a fraction of the sum of losses at the two separate clearinghouses. Potential losses at the integrated clearinghouse are always less than 75 percent of the total potential losses at the two separate clearinghouses, regardless of the percentile chosen to define the severity of losses included. As we move to the right in the graph, looking at only the largest sized losses (defined by the higher percentile cutoff), the potential losses at the integrated clearinghouse drop to less than 70 percent of the total potential losses at the two separate clearinghouses.

Total credit risk is not fixed. The market structure matters. Some structures leave the system with more total credit risk and some leave the system with less.

The OTC derivatives market is based on derivative dealers setting themselves up as decentralized mini-exchanges. The events of 2008 exposed a fundamental flaw in that structure. It is highly profitable to the individual banks, but as a whole contains too much risk. As a result, the OTC derivatives markets exacerbated the recent financial crisis. Individual dealer banks don’t take into account the effects of their actions on the system as a whole, and ignore the viral risk of contagion.

Central clearing, if done properly, can lower the total credit risk in the system. The reason for the lower losses in CCPs shown in the graph above come primarily from the diversification of credit risks that stem from clearing different contracts whose values do not reflect the same underlying risk factors. Another reason for lower losses in CCPs more generally comes from greater risk cancelation in multilateral netting (versus bilateral netting), and the lower risk of losses cascading through a chain of brokers in a decentralized clearing system of yet interconnected intermediaries.

A system with lower potential losses is a system with lower credit risk, and therefore lower total collateral requirement.