Bank Underground

Trimming the Hedge: How can CCPs efficiently manage a default?

Fernando Cerezetti, Emmanouil Karimalis, Ujwal Shreyas and Anannit Sumawong

When a trade is executed and cleared though a central counterparty (CCP), the CCP legally becomes a buyer for every seller and a seller for every buyer. When a CCP member defaults, the need to establish a matched book for cleared positions means the defaulter’s portfolio needs to be closed out. The CCP then faces a central question: what hedges should be executed before the portfolio is liquidated so as to minimize the costs of closeout?  In a recent paper, we investigate how distinct hedging strategies may expose a CCP to different sets of risks and costs during the closeout period. Our analysis suggests that CCPs should carefully take into account these strategies when designing their default management processes.

Closeout procedures and the importance for a CCP

In the event of a member defaulting, closeout procedures play an important role in the sustainability of CCPs, as they allow CCPs to resume normal operations under a matched book of clearing obligations. Typically, closeout procedures entail three main steps: splitting, hedging and liquidation. Although some degree of variability exists depending on the type of product cleared (i.e. exchange traded vs over-the-counter (OTC)), a CCP usually splits a defaulting portfolio by currency, hedges each one of the currency sub-portfolios to minimise its market risk exposure, and finishes the closeout phase by liquidating the hedged portfolios to non-defaulting clearing members.

When implementing these procedures, hedging plays a pivotal role, given that different hedging strategies may affect distinctly how losses accumulate during the default management phase (see Cerezetti et al. (2017) for details). For instance, consider the following two extreme scenarios. One in which the CCP does not perform any hedge, and a second one where the CCP hedges perfectly the defaulter’s portfolio. While the former scenario implies that the CCP faces market risk until liquidation is complete, but not paying any transaction costs to hedge, the latter scenario implies higher transaction costs but reduces exposure to adverse market movements. A full spectrum of strategies is possible between these two extremes cases, and identifying the one that efficiently minimises the total cost for the CCP is the ultimate goal of the present work.

Losses, costs, and hedging design

Using EMIR trade repository data on sterling denominated OTC interest rate swaps (IRS), we carried out two empirical exercises. The first one focused on the period between 15 and 22 January 2015. This corresponds to the period when the Swiss National Bank unpegged the Swiss Franc against the Euro. We assumed that a particular clearing member defaulted at the end of January 15, with closeout procedures completed over the next 5 days. Actual trades cleared by the 5 largest clearing members over this period, covering over 20 distinct maturity subgroups (silos), were used as reference to design a number of hedging strategies for the assumed defaulting portfolio. In the second exercise, we considered the worst loss in the defaulting portfolio observed over the past 10 years. Similar to a historical simulation approach, the objective of this exercise is to provide forecasting measurements, converting simple realisations from the closeout period into risk metrics.

To measure the marginal effects of different hedging strategies, distinct metrics were employed (e.g. profits & loses (PnL), DV01 , permanent loss, transient loss, transaction costs, etc.). Following Vicente et al. (2015), permanent loss is defined as the portfolio’s change in market value over the closeout horizon, and the transient loss as the worst accumulated loss exceeding the permanent value. The transaction cost of a new hedging trade was approximated by its end-of-day marked-to-market (MtM) value on the day of the trade.

First exercise

Figure 1 and 2 show the market value and the cumulative PnL of the defaulting portfolio over the closeout period analysed (i.e., from 15 to 22th of January 2015) in the absence of any hedging trades. The total end-of-day market value of the defaulter’s portfolio on January 15 (T + 0), the last MtM date before the theoretical default, was approximately £6.31 billion. On January 22 (T + 5), the end of the closeout period, the portfolio value was equal to £6.36 billion.

Figure 1 – Market Value of the outstanding portfolio over the closeout period

Figure 2 – Cumulative PnL over the closeout period

Even though no permanent losses are observed during this period, in order to assess the impact of the hedging process the defaulter’s portfolio DV01 was analysed, as displayed in Figure 3. At T + 0, the unhedged portfolio DV01 was close to -£40 million. As the hedging process becomes more precise, moving from a single macro hedge to a set of more granular hedges, each targeting different maturity silos, the DV01 reduces progressively. Not all DV01 can be neutralised, as we restricted the total hedging amount to be less than or equal to the observed traded volumes on each day of the closeout horizon. Moreover, other hedging instruments that could be used to hedge risk exposures (i.e. interest rate futures, FRAs etc.) are not considered in the study. The marginal efficiency of the incorporation of more silos into the hedging process also decreases with the number of silos, and the later the hedging process starts (i.e.,T + 1,…, T + 5 ), the less efficient it is – as the portfolio is left exposed for a longer period of time.

As shown in Figure 4, the execution of the hedging strategies does not come at zero cost. Specifically, the larger the number of maturity silos considered in the hedging process, the more effective is the mitigation of market risk (in DV01 terms), but also the more expensive is the strategy. Importantly, the surge in costs are not just from increased trading, but also from trading less liquid instruments with relatively wider bid-ask spreads. Although the above pattern is relatively stable with respect to the starting date of the hedging process, total costs are lower for late starts due to the reduction in the total number of hedging trades that could be executed.

Figure 3 – DV01 for different configurations of the total number of maturity silos (0 to 20) and first date to start the hedging process (T+1 to T+5)

Figure 4 – Transaction costs for different configurations of the total number of maturity silos (0 to 20) and first date to start the hedging process (T+1 to T+5)

Second exercise

The descriptive evaluation of the marginal effects of hedging strategies for a single reference date, as analysed in the previous subsection, although informative, exhibits limitations as a risk management tool. As such, an assessment of 10 years of data for risk factors moves was performed. Over this period, the total worst-case potential loss (i.e. worst-case permanent loss plus transient loss) of the defaulter’s portfolio over a five day closeout period was £1.5 billion.

Similarly to the first exercise, hedging strategies are introduced to assess the sensitiveness of the portfolio to risk neutralisation mechanisms. As displayed in Figure 5, hedging can efficiently reduce potential losses. In particular, when the most efficient hedging strategy is considered (i.e., using 16 different maturity silos and starting hedging immediately after the default, T + 1) the total worst-case potential loss is reduced from £1.5 billion to £734 million. However, once again, the loss reducing benefits of hedging do not come at zero cost, and total transaction charges associated with that strategy would sum to £756 million. The surge in costs arise not only from the fact that more trades are executed, but also from trading at less liquid instruments with a relatively wider bid-ask spreads.

At the above configuration, hedging is certainly not a viable option, as the margin benefits from the loss reduction are entirely offset by its costs. Therefore, hedging strategies defined taking into account only the effects over risk measurements may not be economically viable, suggesting that transaction charges embedded into them should also be considered. Under this expanded perspective, a new efficient hedging strategy can be obtained (i.e., performing a macro hedge with only one maturity silo), as displayed in Figure 6. Specifically, the implementation of this particular hedging strategy would cost the CCP £113 million, while the reduction in total worst-case potential loss would be approximately equal to £343 million. Any attempt to deviate from this hedging configuration would leave the CCP worse-off, either with higher losses or with extra transaction costs.

Figure 5 – Total worst-case potential loss (TL) for different configurations of the total number of maturity silos (0 to 20) and first date to start the hedging process (T+1 to T+5).

Figure 6 – Total worst-case potential loss (TL), transaction costs (TC), and total worst-case potential losses plus transaction costs (TL.TC) for different configurations of the total number of maturity silos (0 to 20) and first date to start the hedging process equal to T+1.

Final remarks

Our analysis provided evidence that transaction costs is a significant factor and it should be taken into account when designing a hedging strategy. It was shown that the loss reducing benefits arising from more delicate hedging strategies may introduce higher transaction costs and therefore change optimal strategies.

From a regulatory perspective, the evaluation of how distinct hedging strategies expose a CCP to different sets of losses and costs could contribute to the enhancement of the CCPs’ resilience. Furthermore, the framework could helpfully inform default management planning, while serving as a complementary tool to fire drills to test the sufficiency of the financial resources available to manage a default.

Fernando Cerezetti and Emmanouil Karimalis both work in the Bank’s Risk, Research and CCP Policy Division. Ujwal Shreyas works in the Bank’s Advanced Analytics Division, Research & Statistics. Anannit Sumawong works at the University of Sussex.

If you want to get in touch, please email us at bankunderground@bankofengland.co.uk or leave a comment below.

Comments will only appear once approved by a moderator, and are only published where a full name is supplied.

Bank Underground is a blog for Bank of England staff to share views that challenge – or support – prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees.