So Bear Stearns announces it's going to save one of its funds ($3.2 billion loan) because, in my opinion, it's afraid that if their fund sold that much paper quickly to close shop, many other funds would have to write down a lot of their books, and this would cause a contagion that would induce so many of its clients to close as well. And though they could afford to lose this fund, if a lot of their clients went down, that would be very bad for Bear Stearns. You see, you can mark securities at par (eg, 100) as long as the stuff doesn't trade, even though you know it would trade at 60 (not best practices, but a clever risk manager can help you rationalize this). This is because many firms use "matrix' pricing, where the matrix refers to things that have traded. Untraded bonds can therefore be unaffected by their market price, just as you can consider your house worth what you bought it for regardless of the market, as long as you don't sell it. Further, it is true that the market is highly variegated, so you can say with a straight face that the problems for Market X means little about Market Y.

What kills me about the Bear Stearns fund is that it appears they targeted some 20% returns in bonds (I've read elsewhere it was a 1% edge, levered 10 times, but the point is similar). That's asking for trouble. Bonds have very little edge, they are for widows and orphans. If you want to put lipstick on them so you can make them part of a hedge fund--where investors expect a 10%+ return--you have to lever them so much, or take so much risk, it's a bad deal. I've looked at a lot of debt, and figure a very good model (or, by extension, human) could add at most 100 basis points to High Yield strategy unlevered. But with the volatility of High Yield bonds, their lack of liquidity, and expensive hedges, I don't see it as a good Sharpe ratio, that is, 1+ Sharpe. Can't do it over the cycle. Debt markets are characterized by 8 good years (nothing happens) and 2 lousy ones, this trade will work great if you start it at the right part of the business cycle.

It's a classic mode-mean trade: the mode is great but the mean sucks. Further, the variance is severely understated in the good years, so from a sharpe ratio perspective its very subject to abuse: when it works, the numerator (return) and denominator (volatility) look great, and you rationalize this to investors with a tale about 'focusing on value' or someone's connections at Fannie Mae. Then, when it fails, conveniently both the numerator and denominator go bad. This is convenient because if you are going to lose your job anyway, you might as well fail in two dimensions as opposed to one. For a trader using someone else's capital, it's heads we win, tails you lose.

I know cycles repeat themselves, but I thought they should be more different, the way the 1990 recession centered on real estate while the 2001 recession centered on technology. Now I feel pretty old, because I thought everyone knew that you only get 15% expected returns in debt by either 1) taking on warrants, equity or 2) taking on too much risk.