Academic studies have found that Wall Street analyst stock recommendations trail the market and do so with more volatility. As a result, investors who use sell side research should be careful to pay attention to certain data points that analysts have spent hours putting together, but to completely ignore price targets and ratings and instead coming up with their own opinion on the ultimate value of a stock.

The latest example that illustrates this point is the call we got out of Merrill Lynch today. ML's auto analyst downgraded shares of General Motors (GM) from "buy" to "sell" and slashed the price target from $28 to $7 per share. That's right, this analyst thinks GM is worth 75% less than it was 24 hours ago.

As for how to value GM, I think it is simply too difficult to do so. It is nearly impossible to estimate future legacy costs, and trying to figure out what a reasonable profit margin on cars should be is simply a guess because they are not even making money at all and their competitive position has deteriorated since they were last in the black.

Besides, if an analyst can tweak their model and get $28 one day and $7 the next, that is a pretty clear signal to me that valuing GM right now is just not something anyone can do with a large degree of confidence.

Anybody think GM is a buy at 10 bucks? If so, why?

Full Disclosure: No position in GM at the time of writing