What risks does collateral transformation bring to buy-side institutions, in meeting the margin requirements of a CCP?
As the OTC market heads towards clearing in 2012, a wide population of buy-side firms will begin posting assets via Futures Commission Merchants (FCMs) in the US or Direct Clearing Organisations (DCOs) outside the US, against their liabilities at clearing houses. At present Central Counter Parties (CCPs) accept cash, government securities and sometimes gold as eligible collateral, but don’t accept corporate bonds or asset backed securities. For a buy-side firm to meet their liabilities, they need to pledge cash or eligible securities, and bear any costs of doing so. Buy-side firms tend to have higher funding costs for uncollateralized cash, and have fewer government bonds as assets due to their relatively low performance.
This creates a new funding cost for the buy-side which in turn creates an opportunity for FCMs (and DCOs) to step in. FCMs have conceived of a service to accept securities from buy-side firms and ‘transform’ them into cash for delivery to CCPs. The mechanism to achieve this transformation is the Repo market, the securities provided to the FCM are pledged as collateral in the Repo market in exchange for cash, which can then be provided to the CCP.
It isn’t too far back to remember that Lehman Brothers were brought down via a margin call they couldn’t settle, and that funding via the money markets can be risky. Whilst this new idea seems to offer a solution to meeting CCP margin liabilities, this incurs a whole set of new risks, connecting what is supposed to be a risk reducing activity (central clearing) to a more complex risk transfer process, potentially tightly linking ineligible and illiquid securities to the safety and stability of the capital markets.
In theory a Fund will reduce it’s costs through using the Repo market (indirectly), by providing collateral, to cover the credit exposure of the cash borrowing. Using an FCM to provide access to the Repo market offloads the need for the Fund to have the capability in-house, but, this isn’t as simple as it seems, there are a whole variety of risks in using the Repo market which the Fund needs to be aware of, and their FCM managing fully.
Credit risk: The FCM now becomes an active investor in the Repo markets, requiring a full suite of tools to monitor the Repo business including counterparty credit risks with limits setting and monitoring. Each new counterparty will require full on-boarding including master agreements, KYC checks and settlement instructions.
Valuations: The securities provided to the FCM must always cover the value of the cash provided to the CCP, this means monitoring the value of securities (requiring FX rates, bonds prices, interest rates and curves) at least once per day, if not intraday. As an example SwapClear recalculate the Initial Margin requirements for all their customers up to seven times per day – potentially making margin calls to the FCM as a result.
Liquidity risk: What isn’t clear, is how an FCM will set it’s own eligibility criteria for acceptable securities, and whether liquidity is a key determinant, after all, if the FCM is ‘marking to model’ an asset back security – the amount of systemic risk in the capital markets will go up, rather than down. This area is outside any planned regulations, so the link between ineligible assets and CCPs is a hidden risk without any central supervision.
Concentration risk: The FCM will need to monitor the proportion of securities provided in each issue, having a high proportion of a single issuer amplifies the loss in a default, rather than limiting concentration to an issuer and bearing lower losses in a default. Additional monitoring criteria for concentration risk could include currency, country or market sector.
Collateral management: In addition the FCM will need to use the valuations to make and receive margin calls to their Repo counterparties, and where necessary mirror those calls to the Fund. Should the securities drop in value, the Fund will need to deliver additional securities to the FCM, or additional cash, to re-balance any gap in value. One way of reducing any volatility is for the FCM to apply a haircut to the securities (in a reputable salon of course), therefore requiring the Fund to over-collateralise their CCP liabilities, adding to the cost.
Repo settlement: The bond market has developed a number of solutions to ensuring the safe and efficient settlement of bond transactions with simultaneous movements of cash and securities. Settling repo trades is best done via a tri-party repo service such as those offered by DTCC, Euroclear, or the RepoClear service from LCH. All these services rely on being able to value the bonds daily, and specify the securities must be liquid to be accepted. Any security which is ineligible for these services probably ought not be used for this transformation process.
Bond settlement: The delivery of bonds from the Fund to the FCM is less complex but any delay in receiving the bonds by the FCM would expose the FCM to a timing risk, being obliged to settle the margin call at the CCP without delay. An FCM would need to put in place back-stop funding lines and charge the Fund an intra-day overdraft charge, or else put the Fund into default, should there be no credible delivery of assets to back the margin calls on the Fund.
FCM default: Once the Fund has provided the securities to the FCM, the FCM must also have a right to re-hypothecate the bonds, without which they cannot be traded in the Repo market. At that point the Fund won’t know who is the current holder of their bonds. If the FCM were to default, as with Lehman Brothers, the Fund may find it hard to recover their original securities, whilst the FCM is wound up. As of today the administrators at Lehman are still working through the wind-up process 3 years later, without being able to return assets to hedge funds who used their prime broker service.
Back-loading: Both SwapClear & CME (for example) provide offline tools in which a firm can simulate their initial margin requirements for a given portfolio. These tools will show how much Initial Margin would be required using the historic VaR which has become prevalent for interest rate markets. At the moment neither Dodd Frank nor EMIR mandate back-loading, meaning trades executed before the new rules become active, can remain outside clearing. But, given that new regulations are likely to require punitive amounts of margin for trades held outside clearing – there will be strong interest in comparing the margin calculations within the CCP with those of un-cleared but eligible trades outside the CCP. FCMs will need to be prepared to support their buy-side clients in performing these analyses and potentially supporting large amounts of back-loading and avoiding margin swings in the process.
Substitution: And just to add one more thing, the bonds pledged by the Fund will pay coupons, which the Fund will want to receive. Depending on the approach to settling the Repo trades, some of the above services will keep track of the beneficial owner of the securities, and ensure that coupons are paid back to the original owner, the Fund. In other cases it isn’t practical to track the coupon payment all the way back to the Fund, market practice is to return the original securities to the Fund prior to the ex-div date, and substitute other securities to cover their positions. The timing of these substitutions is critical to ensure continuous coverage of the margin liabilities.
Should you feel stressed?
A stressed market may involve sudden drops in value of illiquid / higher risk assets, such as those held by Funds, with a corresponding flight to high quality government securities. At the same time the CCPs will issue more frequent margin calls driven by volatility in the underlying collateral assets, and/or the portfolio of OTC trades. For an FCM to survive such markets, being at the confluence of value flows for many Funds, the FCM should consider creating a stress testing program.
The goal of this would be to explore the outcome of stressed markets on Fund portfolios, including the valuations of collateral provided, and generate a map of value swings and funding flows which the FCM would need to manage. FCMs would need to invest in simulation tools such as a full historic or Monte Carlo VaR model, covering the portfolios and their collateral assets, including the Repo trades.
Most of the existing or planned CCPs such as SwapClear, CME, IDCG, SGX, HKEX, TSE, plus Korea, Indonesia, Brasil, Canada and Australia intend to provide access to clearing for clients, and will then be linked to DCOs offering collateral transformation. Whilst it isn’t their responsibility to manage the risks of providing such a service, should there be a regulatory requirement for a CCP or the DCO to carry out stress tests and aggregate the results to look for systemic risks where a single DCO is clearing for a client at multiple CCPs? Or even to look for concentration risk at each DCO with large swings in valuations of client portfolios and margin assets? Should there be a regulatory framework to define how a DCO manages the many risks in being an intermediary where they go beyond passing margin assets back and forth between the CCP and clients, and move into active risk transformation?
Firms moving into the DCO space include traditional custodians such as State Street and BoNY, plus banks like JP Morgan, Deutsche Bank and others. The latter will already have a mature and established capability to run a Repo business, but for firms unused to being a principal rather than an agent, this will be a new departure. Given the CFTC decision to lower the capital requirement for a US FCM to $50m, we should see a new wave of smaller firms offering services in this space, will they be able to offer transformation at a level of sophistication that covers all the risks? Somehow it seems that risks in the capital markets never disappear, they just get transferred to someone else.
Stress testing collateral assets at CCPs by Stephen Loosely / Catalyst, derivsource.com
Kevin Samborn and Rajan Alexander of the Valuation and Risk Analytics practice at Sapient for their review input.