Are investment banks' earnings a black box?

I should be hesitant to admit this on a financial blog, but there's really only one reason that I've never purchased shares of an investment bank on my own, outside of those that I own through an index fund: The financials are often very complex, and difficult for me to understand with the level of depth that I like to have before I make an investment.

Apparently Fortune's Peter Eavis agrees with me, at least on the most recent numbers. Lehman Brothers Holdings Inc. (NYSE: LEH), The Bear Stearns Companies Inc. (NYSE: BSC), and Goldman Sachs Group, Inc. (NYSE: GS) all booked substantial gains from hedging bets/shorting mortgage bonds. Eavis asks thought-provoking questions, one that I somehow doubt many of the investors who have been pushing shares of these stocks up in recent weeks have bothered asking:

How, for instance, can it be that the three firms were able to rack up large gains by betting in the same direction? Were these bets made in liquid markets where prices are dependable and positions can be sold quickly? Or were they made in illiquid markets where brokers have to make their own estimates about what the bets are worth - and where it may be difficult to exit?

Very interesting. For now, the earnings for Bear Stearns appear to be somewhat of a "black box", exactly the way one analyst described Enron in March of 2001. At the time, Goldman Sachs analyst David Fleischer disagreed: "Enron is no black box. That's like calling Michael Jordan a black box just because you don't know what he's going to score every quarter."

Of course, I'm not suggesting that any of these firms are involved in any kind of fraud. But one of the most important, and most neglected, questions to ask before investing in a stock is "Do I really understand the earnings?"

In the case of these firms, I sure as heck don't, and so I'll stay on the sidelines.