Sunday, October 4, 2009

Lehman Again

Richard Robb has an interesting counter to the argument that the impact of Lehman's failure was overblown. His main point is that the looking only at Libor and the overnight swap market was misleading at that time:
The latest contribution by John Cochrane and Luigi Zingales, like others before them, rests partly on misunderstanding of the data. The authors deduce that Lehman wasn’t the main cause of last autumn’s turmoil by inspecting the daily movements in the spread between Overnight Interest Rate Swaps and three-month Libor, which they define as “the rate at which banks can borrow unsecured for three months.”

But a better definition of Libor under the circumstances was “the rate at which banks said they can borrow”. Libor is the result of a survey, not a measure of actual transactions. In the week of September 15 last year, big banks refused to settle foreign exchange with each other. They were not lending interbank for three month terms, so Libor during that week tells us little.

We could say the same thing for OIS. Volume was light to nonexistent in the week of September 15 last year. What we do know is that three-month T-bills traded at 0.04 per cent on September 17, down from 1.47 per cent on Friday September 12. These are real data that ought to impress the professors that the market was breaking down as fast as it knows how.

In addition to this argument there is the additional point that Lehman's collapse led to the run on money markets because the Reserve Primary Fund held lots of Lehman debt. This cause them to break the buck, and this led to a run on other money market funds. Since money market funds are the main purchasers of commercial paper, this spread the crisis to the rest of the economy. This is the amplification mechanism that was crucial in generating such a large impact from the failure of a relatively small bank failure.

No comments: