Streetwise Professor

September 25, 2010

VaR and Margins

Filed under: Clearing,Derivatives,Economics,Exchanges,Financial crisis,Politics — The Professor @ 10:28 pm

Value at risk, VaR, is the most common means of measuring risk.  It provides useful information: it quantifies the magnitude of loss that can be incurred on a portfolio with a given probability.  But it is dangerously incomplete: it does not tell you how much you can expect to lose, conditional on the VaR boundary being breached.  VaR, in short, doesn’t tell you how bad bad can be.

This is relevant in the clearing debate.  CCPs choose margin levels in order to ensure that the probability that price moves will wipe out margin balances is sufficiently small.  This is exactly equivalent to calculating a VaR.

And hence, it is subject to the same problems as VaR.  This means that when evaluating the financial capacity of a CCP to withstand a large price shock that not only is the margin level (the VaR equivalent) that is important: the potential loss conditional on margins being blown through is important too.  If the loss conditional on margins being breached is X, but a CCP only has resources of X/2, during a crash (or a spike) that breaches margin levels–an event that will occur with positive probability–then on average the CCP will default.

This is not a purely hypothetical situation.  This important paper by David Bates and Roger Craine shows that in the days following the 1987 Crash, the loss on outstanding S&P500 futures positions conditional on breach of margin was huge, likely far beyond the financial resources available to the CME clearinghouse.

This raises the questions: do CCPs take this information into account when determining their financial resources?; will new CCPs do so; and importantly, do they have the incentive to do so?

This last question is crucial. Financial institutions have problematic incentives to take proper consideration of tail risk.  Indeed, that is arguably one of the major reasons for financial crises.  So why should CCPs, which are coalitions of self-same financial institutions, take into account the tail risks that they ignore independently?  Particularly inasmuch as collectives typically have weaker incentives (due, for instance, to free rider problems).

This is yet another, crucial issue, that has not received adequate attention amidst the paeans to clearing in the run up to and the days since the passage and Dodd-Frank; the same can be said of Europe too.

VaR was the subject of much criticism prior to the crisis: that criticism has only intensified since.  And for good reason.  That should be kept in mind when one hears assurances that CCP margins will prevent derivatives from being the source of a future financial crisis, because margin setting is in essence a VaR exercise.

Print Friendly, PDF & Email


  1. […] This post was mentioned on Twitter by John Avery, John Kiff. John Kiff said: Craig Pirrong on VaR and CCP Margins […]

    Pingback by Tweets that mention Streetwise Professor » VaR and Margins -- — September 26, 2010 @ 6:47 am

  2. Won’t price limits (or circuit breakers), when used smartly with margins, almost guarantee
    a no end-of-day loss to a CCP? Suppose the first circuit breaker is triggered, the clearing
    calls to deposit additional margins. If next limit level/circuit breaker is hit, and the
    additional margin has not been deposited, the offending portfolio can begin to be liquidated.

    Note: if the price was locked at the limit by the day’s close, it becomes the settlement
    price, limiting the loss. Ex. ICE US Cotton yesterday.

    Comment by Igor Baglay — September 28, 2010 @ 7:24 am

RSS feed for comments on this post. TrackBack URI

Leave a comment

Powered by WordPress