In my paper, a coin-flipping analogy is used to cast doubt on skill implied by past performance.  Ken Kam, in his letter, strongly suggests the conclusion reached in the paper was "you cannot tell anything about a manager's future performance from his past track record."  Not quite.  The point made in the paper is it is difficult to distinguish luck from skill when comparing past track records.  In a given group of managers, some are bound to have "outperformed" simply by pure chance.

Mr. Kam hints that the paper might have attempted to pull the wool over readers' eyes by using a coin-flipping analogy, since "once you assume that everyone is tossing a fair coin you've already conceded the argument because no one can be 'better' at tossing a fair coin than anyone else."  On the contrary, I believe the coin-flipping analogy to be particularly relevant precisely because skill is not a factor in coin-tossing.  The example, therefore, isolates the role chance can play in any outcome.  Simply put, it is possible to be lucky long enough to appear "skilled."

In quoting well-known behavioral economist Hersh Shefrin, in addition to describing his own elaborate coin-flipping experiment, it appears Mr. Kam is implying past performance does indeed have predictive value.  Specifically, he writes "restricting your choices to those with a track record of throwing 10 heads in a row greatly increases your odds of selecting a gold coin holder who can then be expected to throw 6 heads in the next ten tosses and thus beat the average..."  Sounds easy enough, but does it work?

Unfortunately, scores of academic studies indicate past performance is not a reliable indicator of future performance.  In the paper, I reference one such study which examined the subsequent performance of the twenty best-performing equity mutual funds from 1983-1993.  Of those past winners, only four maintained their stellar record in the next decade.  Sixteen failed to even match the market return.  While holding a portfolio of yesterday's winners might feel comfortable today, it hardly seems a recipe for investment success tomorrow.

Perhaps the more relevant question for investors (and their advisors) to ask is not whether skillful managers exist or whether such managers can be reliably identified in advance, but rather how will investors and their advisors respond to good or bad manager performance when it inevitably occurs?  Again, study after study suggests investors, to their detriment, too readily buy high and sell low (whether via stocks, funds, managers or asset classes).

I will readily admit human emotions are difficult to overcome, even in a simple coin-flipping experiment.  In conducting a test in which we tossed a fair coin 100 consecutive times, I too, like many investors, fell prey to the whims of chance.
As a colleague tossed the coin and I sat by tabulating the results, I knew that by the end of the trial we should have flipped somewhere near 50 heads.  But, as we recorded head after head--at one point flipping a string of fourteen in a row--I turned to him incredulously and asked, "what are you doing differently?"  Of course, he wasn't doing anything differently.  It was simply luck.  But with the odds of recording a string of 14 consecutive heads in a trial of 100 tosses at 1:370, had he not replied that he wasn't doing anything differently, I probably would have believed whatever creative explanation he could have concocted. 

Dougal T. Williams, CFA
Vista Capital Partners, Inc.

Display article as PDF for printing.

Would you like to send this article to a friend?