Remember collateral support default November 4, 2011 at 9:23 am
I am reluctant to use the phrase ‘What Went Down at MF Global‘ as I think events at the broker have a somewhat different tenor to a drug bust. Let’s ignore the title, then, and look at what Brad DeLong has written about the risk that brought MF Global down:
- In stage zero MF Global sets up the financing with its counterparty and buys southern Europe’s bonds.
- In stage one we learn whether the market is tolerant or intolerant of southern Europe risk: if the market is tolerant the bond prices stay high; if the market is intolerant the bond prices collapse.
- In stage two southern Europe either pays off its bonds or defaults.
In other words, DeLong (correctly) says that MF Global faced two risks in its repo-to-maturity trade: risk that repo haircuts increased during the trade (risk in what DeLong calls stage one); and default risk (stage two risk).
DeLong gives approximately* this diagram:
The key point, I think, is that while MF focusses on the default risk – witness the fact that they hedged against a Europe wide crisis by shorting France – they didn’t focus on the possibility of large increases in repo haircuts. This failure to understand the liquidity impact of collateral calls is exactly the risk that caused AIG to fail.
Thus I think DeLong is not precisely right about this:
MF Global’s bet is (i) highly leveraged…
MF Global’s bet is attractive to… (a) rogue traders (and rogue CEOs) speculating with other people’s money, (b) those who are highly confident in their ability to switch from highly-leveraged speculators to patient well-capitalized investors in fundamentals if necessary, and (c) those who don’t believe that there are shocks to risk tolerance that are orthogonal to shocks to fundamentals.
The bet is only highly leveraged in the tail. For most states of the world, it produces low, stable returns; this makes it look much more benign than many leveraged positions. If you didn’t know about the collateral support risk, then you would not have to have fallen into DeLong’s (a)-(c) to do the trade. You would just have had to have thought that default was unlikely and the impact of it, even if it happened, was within your risk tolerance; a perfectly reasonable conjecture.
It seems much more likely to me that Corzine et al. didn’t understand the repo haircut risk – a modern phenomenon – than that they were rogues. Or, to use my own slang, this is a cockup not a conspiracy.
*Specifically, I fixed a typo in DeLong’s diagram.
Update. My thanks are due to the Streetwise Professor for a very kind commentary on this post.
Unkind update. Normally I don’t go in for selective quotation, but this, from Matt Levine, is too good to miss:
It is normally a good and noble endeavor to make fun of Gretchen Morgenson, because she really really doesn’t care at all if what she writes about Big Evil Banks is true or not, so it’s mostly not.
Oh, OK then.