Several people have written me about my Incerto 2 posting yesterday, unsettled perhaps by the observation that professional fund managers, let alone amateur stock pickers, might be doing no better than random. There have been several studies about this. See, for example, the 2014 New York Times article entitled Heads or Tails? Either Way, You Might Beat a Stock Picker. Or the study the S&P itself did (referenced in the article) entitled Does Past Performance Matter: The Persistence Scorecard. If you search, there are also a number of academic studies as well, all of which more or less confirm the conclusion that on average, managed funds and managed stock portfolios do no better than random stock picks over the long run, which is why unmanaged index funds, with much lower fees, have become popular. And then the fact that the market trend has been upward for years now in the U.S., despite the occasional dip, means that almost everyone has made some money in the market long-term.
Of course the price a stock sells for at a particular moment isn’t random. It is determined not only by the real fundamentals of the company it represents, but also by thousands of other factors, from the market rumors, correct or incorrect, circulating about it, the momentary demand for that stock, the momentary mood and individual strategies of thousands of players, big and small, in the market, external worldwide economic and political factors, and even the actions of high-speed trading algorithms detecting the offer and buying ahead to push the price up. So although the price isn’t really random within limits, it is subject to so many unmeasurable momentary forces that for all practical purposes its short-term movement is largely random.
Given these conditions, it is no wonder that funds that do well in one year often do poorly the next. There is a claim by one statistician that anything less than an 11-year history of a fund is meaningless for determining if it is really doing better than average (I don’t know how he got to exactly 11 years, but you see the point). There is even a market strategy that says to sell the highest performing funds each year and buy the lowest performing ones, because of the statistical tendency for them to regress to the mean (meaning that on average the poorer performing funds ought to do better next year, and the better performing ones ought to do worse next year.)
It is a little hard to actually model this statistically, because funds and fund managers that do poorly for a few years, even if by chance, tend to get eliminated (managers get fired, funds close or are merged with other funds), so there is what is called an “absorbing state” on one tail of the distribution, which muddies the waters a bit statistically.
In any case, given all of this, Taleb’s observation that there is a “survivorship bias” operating when “winning” fund managers claim their strategy is responsible for their success seems reasonable. And of course much the same logic applies to “winning” CEOs who write books about their successes.
Posts
