Swedroe: Hard To Know Who’ll Outperform

This article was originally published on ETF.com

The efficient market hypothesis asserts that financial markets are "informationally efficient"; that is, investors shouldn't expect to consistently achieve returns in excess of average market returns on a risk-adjusted basis, given the information available at the time the investment is made. However, we know that the market isn't perfectly efficient.

In fact, as I explained in my Seeking Alpha series on the subject, it doesn't hold for any of the three forms of market efficiency: strong; semi-strong; or weak. However, there's a large body of evidence demonstrating it succeeds in the only way that really matters: There are fewer active managers that outperform appropriate risk-adjusted benchmarks, after expenses, than would be randomly expected. In addition, there's little to no evidence of persistence of performance beyond the randomly expected.

It's important to understand that this doesn't rule out the possibility that some investors will outperform. In fact, given the large numbers of investors engaged in the effort, randomly we should expect to see many outperform their appropriate risk-adjusted benchmarks purely by chance. However, we should expect that as the investment horizon increases, the percentage that outperforms will decrease.

As Bradford Cornell of the California Institute of Technology put it, "Successful investing, like most activities in life, is based on a combination of skill and serendipity. Distinguishing between the two is critical for forward-looking decision-making because skill is relatively permanent while serendipity, or luck, by definition is not. An investment manager who's skilful this year, presumably will be skilful next year. An investment manager who was lucky this year is no more likely to be lucky next year than any other manager. The problem is that skill and luck are not independently observable."

Since skill and luck are not directly observable we are left with observing performance. However, we can apply standard statistical analysis to help differentiate the two.

Brad Cornell's 2009 study, "Luck, Skill and Investment Performance," used Morningstar's 2004 database of mutual fund performance to analyse a homogenous sample of 1,034 funds that invest in large-cap value stocks. Cornell found that the great majority (92 percent) of the cross-sectional variation in fund performance is due to random noise. This result demonstrates that "most of the annual variation in performance is due to luck, not skill."

He concluded: "The analysis also provides further support for the view that annual rankings of fund performance provide almost no information regarding management skill."

Professors Eugene Fama and Kenneth French also studied this issue in their March 2009 paper "Luck versus Skill in the Cross Section of Mutual Fund Alpha Estimates." They found that active managers as a group have not added any value over appropriate passive benchmarks. They also found that few funds produce benchmark-adjusted expected returns sufficient to cover their costs.

A new study on the Swedish stock market provides us with further evidence. Harry Flam and Roine Vestman studied the performance of 115 actively managed Swedish equity funds for the period 1999-2009. The Swedish stock market is of particular interest because the market participation rate is especially high. Outside of the Swedish pension system, the participation rate is 63 percent at the household level (higher than the 53 percent in the U.S.)

The following is a summary of their findings:

• Actively managed Swedish equity mutual funds generated an average positive 4-factor alpha of 0.9 percent per year before expenses, but a negative alpha of 0.5 percent per year after expenses.
• There was practically no persistence in returns.
• When funds are ranked on past performance, returns converge to the mean in about two years.
• There was practically no evidence of true stock-picking skills among active managers. Returns at both ends of the distribution can be obtained by good and bad luck.
• The actual four-factor alphas of most funds, including those with the highest alphas, was statistically indifferent from zero.
• The relationship between net returns and expenses was negative. Costs matter.

The finding that fund managers produced gross alphas of 0.9 percent demonstrates that there's evidence of some skill in stock picking. However, the fund managers received the benefits in terms of the expense ratio, while investors earned negative alphas. This is consistent with both the U.S. evidence and Jonathan Berk's premise that since the ability to generate alpha is the scarce resource, it should earn what's referred to as the "economic rent."

The bottom line is the study on Swedish funds contributes to the literature that demonstrates that while some active managers appear to have stock-picking skills, because it's difficult to separate skill from luck, it's hard to identify them in advance. In addition, even if you were able to identify them in advance, it's the manager who's likely to be the beneficiary of the excess returns, not the investor.

Larry Swedroe is director of Research for the BAM Alliance, a community of more than 130 independent registered investment advisors throughout the country.