March 2, 2010
Analysts’ reports now say this manager is too volatile to be a core holding. Jim’s own financial analyst ran the numbers and now concludes that Manager X’s recent risk-adjusted performance is poor. Jim wonders where he went wrong.
Jim’s experience highlights the critical question of persistence in manager performance – whether a manager’s past performance is predictive of his or her future performance. Certainly, considering the avalanche of media articles on top winning funds and the endless sales pitches to investors trumpeting “best in class” managers, one would assume that there is some reasonable level of persistence in performance.
Fortunately, we can garner insights based on empirical evidence, not puffery. Over the past half a century, there have been over 100 academic studies on the question of persistence in managed money performance. In 2003, the Fund Management Research Centre undertook a sweeping review of 49 of the most recent or robust of these studies from the U.S., U.K. and Australia in a report1 to the Australian Securities and Investment Commission.
The report’s major conclusions provide serious investors with some clear answers:
- Good past performance is, at best, an unreliable and weak predictor of future good performance over the medium to long-run. Approximately 50 percent of the studies found no correlation at all between good past performance and good future performance. Where persistence was found, it tended to be short–term, i.e. only one to two years.
- In those studies that found some level of persistence in positive performance, the outperformance tended to be small and in many cases, would be swamped by the cost of swapping between funds.
The report’s authors hypothesized some reasons for the lack of persistence in past performance – style cyclicality; the erosion of competitive advantage as managers battle it out for better staff and methods, and; the negative impact of large capital inflows on outperforming managers.
The implications for investors are clear. An analysis of past performance alone is not sufficient for the selection of an investment manager. The chance of a given outperforming manager repeating this performance is almost random. An investor might as well use a dartboard if he or she is selecting managers solely on past return numbers.
If active managers are to be used in a portfolio, extensive investigation far beyond a simple review of past performance is required. A recent study2, for example, suggests that analysis of a manager’s portfolio holdings and the extent of their deviation from the benchmark as well as historic returns might point the way to managers who are more likely to exhibit positive performance persistence. However, once adjusted for style, size and momentum factors, much of this positive performance disappears and hence, more research is needed to validate these findings.
Finally, since managers as a group underperform the market by their fees and costs, the absence of positive performance persistence by active managers in general suggests that low cost, tax efficient index funds should form the core of a portfolio and that active managers, if included, should be used in a satellite role. Jim Smith wishes he had taken this approach in 2000.
Tacita Capital Inc. is based in Toronto, Canada and is a private, independent family office and investment counselling firm that specializes in providing integrated wealth advisory and portfolio management services to families of affluence. We understand the challenges of affluence and apply the leading research and best practices of top financial academics and industry practitioners in assisting our clients to reach their goals.
1 Allen, D., T. Brailsford, R. Bird, and R. Faff. 2003. A review of the research on the past performance of managed funds. ASIC REP 22, Australian Securities and Investment Commission.
2 K.J. Martijn Cremers, Antti Petajisto, How Active is Your Fund Manager? A New Measure That Predicts Performance, Yale School of Management, October 3, 2007.
Display article as PDF for printing.
Would you like to send this article to a friend?
Remember, if you have a question or comment, send it to .
