But just how good are analysts at predicting the directions of stocks? A decent paper (though dated) is available here which discusses the difficulties in truly evaluating analysts. However, at the very least, if analysts were at all useful, they can at the very least tell the difference between a company that has value and one that is going under. Yet these two very different realities in a company's status continue to baffle these "experts".
Consider the latest examples of this: Fannie Mae and Freddie Mac. Before going down (where shareholders lost pretty much everything) two days ago, they each had 13 analysts following them. A full 40% of analysts rated these two companies as either "Strong Buy" or "Buy", compared to just 30% on the sell side. If we can't trust analysts to tell the difference between a company worth something and a company worth nothing, what can they be used for? What do you think many "analysts" said about Bear Sterns right before it went down?
In fairness to analysts (and conversely in unfairness to investors), analysts are often confronted with conflicts of interest that may be preventing them from extolling their true opinions of a stock. So analysts might actually be better at stock prediction than we think. However, this makes them no more useful for investors, who should put little stock in what analysts say, and should instead do their own homework, in the same way as they would for any investment outside of the stock market! Why should stocks be treated any differently as an investment than a house, or a private business? You wouldn't let an appraiser pick your house would you?