The Efficient Market Theorem
There are three forms of the efficient market theorem: Weak Form, which holds only that there is not a correlation between price movements over time of a stock. I agree with this (essentially negating technical analysis) and believe that many statistical studies have shown this to be true. Then there is semi strong form which holds that the price of any stock will already reflect all public information regarding that stock and strong form which holds that the price of a stock will already reflect all public and private information regarding that stock.
For both semi-strong and strong form I believe there is a good argument against these theorems in the price behavior of stocks after an analyst upgrade or downgrade. In this instance the public information concerning the stock has not changed, but the price, typically goes up with an upgrade and down with a downgrade. It seems to me that this price movement tends to disprove both the semi strong and strong theorems. Either the price was correct before the rating change, or correct after the rating change, but I don't see how the price could be correct at both times, when the public information has not changed, but the price has changed far more than could be explained by price movement of the broad market.
There is a defense against my argument in the mosaic theory of analyst opinions, which holds that such an opinion may legitimately reflect a mosaic of material public information and nonmaterial nonpublic information. Accordingly, at least for the semi strong theory, some nonpublic information, immaterial by itself, is reflected in the analysts' opinions. I just have a hard time believing that information that is nonmaterial in itself could grow in importance by being mixed with the material public information in the analysts mind, so that when an opinion is released reflecting that information it justifies a 2 or 3% movement in stock price (which does happen). This defense does not apply to strong form efficient market theory, because in that theory even the nonpublic information should already be factored into the stock's price, before the release of the analyst's opinion.
Also, semi-strong form and strong form do not take account of the human being's ability to have a world view that may be more accurate than the world view of others, to form a unique understanding of the same set of facts and to draw connections that others miss. Yes, if investing decisions could be broken down to the application of broadly accepted theory to the same set of facts, semi-strong form and strong form would work. But it is in seeing that broadly accepted theory is wrong in some cases, and that the world understanding of others is limited, that some analysts do better than others.
Of course it is difficult to tell the difference between skill and luck, when there are only 40 quarters in 10 years, unlike the situation in baseball, where there are more than 100 games in a single season. If each baseball team played a single game per quarter, it would be very difficult to tell if the win/loss results were the result of differing levels of skill or mere chance. After all, someone has to win and someone has to lose. In like manner, if 100 different people chose 100 portfolios by randomly selecting sectors in which to concentrate and then randomly selecting companies in those sectors, there would be a good spread at the end of each quarter. But statistical analysis shows that the distribution of results from portfolio managers displays leptokurtosis, or fat-tails, indicating that there is indeed some portion of the distribution that is unlikely to be the result of mere random distribution.
I've studied for and passed the first CFA exam, and have reviewed some of the material for the second, and I believe I can say that to be an analyst one need not know the population of Brazil, or the history of the great depression, or the South Sea bubble or the 1965 to 1980 market stagnation. Perhaps in preparation for the third exam CFA applicants learn that if one investment philosophy grows very popular it is likely to warp the market in a manner from which others, who do not fully accept the philosophy, can profit. When the market is filled with private equity players, waving about their sharp axes that they want to use to chop away the deadwood that is kept around due to the misplaced sentimentality of corporate officers, it tends to greatly reduce the amount of deadwood, making the private equity game that much more difficult. All of this information is important to know for portfolio managers to know. Every portfolio manager should know the world and economic history, and the basic philosophy of investment philosophies. And I'm sure that the great majority do. But you don't have to know that information to set out in the industry. And that creates at least some opportunity for the knowledgeable and insightful.
Of course, in one respect the efficient market theorem is correct. This is in the sense that the average investor really will have a very difficult time advantaging himself by studying the market (others may well do a better job of it), or looking for an actively managed fund that will beat the market (it is just as hard to pick pickers as it is to pick stocks), or by finding some great portfolio manager picker to pick the portfolio managers (it is just as hard to pick the picker picker as it is to pick the stock). So if you are reading this and are thinking what a crazy, crazy thing Marketocracy is, well, hey, it is no crazier (and me be far more sane) then the more usual way of doing things.