Mark to Market

I'm one of the very many people who know very little about hedge funds. Despite (because of?) my ignorance, I thought this article was fascinating. If you follow financial news you've probably heard that a small Bear Stearns hedge fund recently made some bad trades and one of its creditors made a margin call (I simplify), seizing its collateral and selling it at auction.

Neat. But what's so interesting? The assets in question here, CDOs, do not have clear market prices:

Unlike Treasury bonds, which are remarkably standardized financial products, CDOs are handcrafted. No two are exactly alike in the way they distribute risk, the amount of this and that they mix, the yields they offer and how they react to market stress. And unlike Treasury bonds, which trade frequently, CDOs trade infrequently. And when they do trade, they often trade privately, which provides very little price information to the public market.

Wall Street still has to value its portfolios, of course, even in the absence of public prices. To solve that problem, the investment houses that invented, packaged and marketed these infrequently traded securities and their derivatives invented mathematical models that priced their products.

The author speculates that part of the motivation of Merrill Lynch for seizing and selling the assets, rather than working out a financing deal with Bear Stearns, is to gather real market price data for these assets. It's a straightforward plan:

1. Gather market price data.
2. ???
3. Profit!

(Nah — #2 is "Improve CDO valuation models.")

There's something deliciously satisfying about this. It's obscure economic theory made visible in plain action — it's the old "economic calculation under socialism" arguments in a novel setting. Merrill knows that its models are a poor substitute for a real market, and is taking the opportunity to learn from one!

If true, this pleases me, and I hope they make a great heaping pile of money by having better information than everyone else.