A segregation puzzle July 16, 2012 at 1:07 pm

From BCBS 226, the new (-ish) Basel consultative document on margin requirements for non-centrally-cleared derivatives:

Under current market practices, the exchange of two-way initial margin in bilateral trades is not universal. Accordingly, requiring the segregation or other protection of initial margin collateral may create material incremental liquidity demands and trading costs relative to current practices, as (i) firms would be required to divert significantly more liquid assets to provide initial margin to counterparties on a gross, rather than net, basis, and (ii) firms would no longer retain the ability to use initial margin collected as a source of funding, for rehypothecation or re-use, or for other discretionary purposes.

Given the potential for the net treatment of provided margin to undermine the general benefits of the proposed margin requirements, there was broad consensus among the BCBS and IOSCO that the proposed requirements should address these risks by requiring the gross exchange and the segregation or other effective protection of provided initial margin, so as to preserve its capacity to fully offset the risk of loss in the event of the default of a derivatives counterparty.

This is all very reasonable. Yes, IM can reduce credit risk for the postee, but it can also introduce both liquidity risk (because it has to be funded) and credit risk for the poster (because the margin over the MTM of the portfolio is often an unsecured claim). To mitigate this credit risk, the BCBS says

the collected margin must be subject to arrangements that fully protect the posting party in the event that the collecting party enters bankruptcy to the extent possible under applicable law.

So here’s my point. How much, roughly, of the system’s margin is currently held in arrangements that meet those criteria? My guess is not much. Moreover it is very hard to see how you get both fully segregated margin and rehypothecation. So you pay your money and you take your choice – do you want to mitigate liquidity risk or credit risk today? Pick one.

6 Responses to “A segregation puzzle”

  1. Right on David. I think it is *impossible* to see how you get full set and rehypothecation.

    And given your choice of poisons, I would choose mitigating liquidity risk. In a heartbeat.

    The head of risk for an oil super major’s trading operation told me once that Frankendodd, etc., were transforming credit risk into liquidity risk, and he would rather live with credit risk any day of the week.

    This will not end well.

  2. I might be mistaken, but I don’t quite follow you. The second except you quote appears to come from ‘Key Principle 5’. When put in the context of the whole paragraph, my reading is that the phrase ‘the collected margin’ is referring specifically to initial margin, and KP5 does not address variation margin at all.

    Banning rehypothecation only for initial margin would certainly increase demand for unencumbered liquid collateral, but I don’t quite see how it would cause liquidity *risk*. I would have thought that the initial margin is not going to change (for existing trades) once a trade is in place, so there is no uncertainty, and hence no risk. Certainly a desk that runs out of eligible unencumbered collateral might have to stop entering new trades, but this is a foreseeable limit to the number of trades a desk can open at once, not an uncontrollable risk such as the MTM variation margin calls.

    If the MTM variation margin was also required to be segregated, then I agree that this would cause a significant liquidity risk, but I don’t read it like that. Maybe they should reword this paragraph to make it clearer?

  3. You seem to be arguing that rehypo reduces credit risk. Are you sure about that? I think it’s more accurate to say that it enables leverage. My understanding is that assets that can be rehypothecated, already are rehypothecated. They aren’t just sitting around idle waiting for a liquidity drought.

  4. edit: meant to say liquidity risk in the first sentence

  5. Guest – No, I am arguing that rehypo reduces liquidity risk, because it lets the collateral flow around the system rather than forcing everyone to separately fund it, and have it sitting unused at the custodian. It _increases_ credit risk however because the return of collateral over the mark is an unsecured claim.

  6. JH – Thank you. Here’s the point. If IM is calculated as portfolio VAR, then it certainly is going to change as the VAR model is recalibrated. In a crisis, vols go up and correlations get more extreme so VAR increases. IM therefore increases causing a liquidity drain on the whole system. Nick Vause at the BIS has written cogently about this.

    Since in the world the document suggests _every_ big player will have to post seg’d IM, that IM represents a liquidity drain on the system as a whole.