The Perils of Past Performance, Part 1
In November, legendary equity investor Bill Miller stood down as portfolio manager of the Legg Mason Capital Management Value Trust, after underperforming in four of the past five years. Does this mean Miller had lost his touch?
Maybe, but as even good managers underperform, I thought I would try and take a more objective look.
The market environment has been particularly hostile for bottom-up stock pickers, such as Miller, over the past few years. Movements in share prices have become highly correlated, as marginal investors have increasingly bought and sold “the market” using futures, exchange-traded funds or derivatives thereof to put risk on or take risk off. This high correlation implies that investors are not differentiating much between stocks, making it tough for a stock picker to add value.
I thought it would also be interesting to look at the probability—regardless of market conditions—of a very good manager like Miller underperforming. Starting in 1991, Miller had outperformed the S&P 500 Index for 15 consecutive years before his performance slumped.
In the real world, of course, investors can’t know if their managers are going to be good, but presumably they invest with them in the hope or belief that they will be. For this analysis I removed all doubt and defined a “good” manager as having an information ratio of 0.5, meaning the manager will add one unit of return per two units of risk. A manager who could add one unit of return per two units of risk would therefore rank among the very best.
Statistically speaking, to have 95% confidence that our manager with an information ratio of 0.5 has skill, one would have to monitor performance over at least 11 years. But in the real world, few investors measure managers over such lengthy time horizons. A three-year window is much more the norm.
Therefore, I assessed the probability of our manager underperforming the benchmark over the next three years. Specifically, I wanted to know the probability that the manager would underperform the benchmark in one or two of the next three years.
As shown in the Display below, there is a two-thirds probability that our good manager will underperform the benchmark in one of the next three years; that is, it is much more likely than not. And there is a still surprisingly high 10% probability that a good manager will underperform over two of the next three years.
So, simple statistics suggest that even a very good manager, such as Bill Miller, is likely to underperform at times. Indeed, as we argued in a prior research paper, simply relying on recent performance to judge a fund manager can be dangerous.
In my next post, I will examine the track records of other good managers to see how they’ve actually done. In the meantime, let me take this opportunity to wish Mr. Miller farewell and good luck with his next endeavor.
The views expressed herein do not constitute research, investment advice or trade recommendations and do not necessarily represent the views of all AllianceBernstein portfolio-management teams.
Patrick Rudden is Head of Blend Strategies at AllianceBernstein.