In James Montier's book 'The Little Book of Behavioral Finance', he paraphrases Graham's quote that “Analysis should be penetrating not prophetic". Essentially, Montier suggests that all investors should devote themselves to understanding the nature of the business and its intrinsic worth, rather than wasting their
time trying to guess the unknowable future. Several research studies show that few analysts achieve better results than by tossing a coin, and many in fact are much worse.
Friends of mine who have worked in or interned in buy side firms, are often visited by the sell side equity research analysts. Contrary to what we may think, the buy side analyst does not care about the final recommendation, but delves deeper into the assumptions/thought process/logic of the sell side analyst. After all, the buy side guy essentially aggregates the views/data/logic, but finally condenses them to form his own view on how the stock will move. And that is what investor should do-rely on analyst reports to make them more informed about the company, but not to attribute any superior forecasting skills to the analyst.
Joint probability informs us as to WHY so many forecasts just get it wrong. Even an elementary DCF model has several point estimates about cash flows, taxes, discount rate, inflation, costs etc. Even assume 80% probability of getting each one right, the chance of getting ALL 5 right is just (0.9^5)=0.33. Now, one may argue that some errors may cancel out(some overstated and some understated) and so the overall forecast may still hold. In my view, analysts being the eternally optimistic beings that they are, the errors are likely to be one sided as in upward bias.
Much as we would like an all knowing oracle, that is not gonna happen anytime soon, atleast not in the public domain. Hence, we as investors should take explicitly take responsibility for our forecasts, instead of blaming analysts for their wrong ones.