All investors want above-average returns, yet ironically it is those who settle for the average returns of an index fund who historically have outperformed.
No less a skilled investor than Warren Buffett plans for his heirs to benefit from a Vanguard index fund after his demise (he suggests the traditional cap-weighted S&P 500 fund, VFINX).
But while he remains active and ever alert to investing issues, the Sage of Omaha might want to consider the latest study from Research Affiliates, whose analysts Jason Hsu and Vitali Kalesnik subject index-based portfolios to the same performance evaluation typically used for actively managed funds.
The analysis by the Newport Beach, Calif., firm founded by Rob Arnott is not intended to dwell on the hoary debate between active and passive management; rather, the firm known for its smart beta index products seeks to question a cherished tenet of index investing — namely, that it produces market-based returns.
Hsu and Kalesnik find that traditional index portfolios — capitalization-weighted products like VFINX that hold the bulk of indexing assets — underperform the market average in a statistically significant way, unlike smart beta portfolios, which do deliver market returns.
In an article that is more technical than their usual newsletter articles, the two analysts describe the process for assessing active manager skill. They show this mathematically, but the basic idea is that "the expected return of a portfolio is the sum of the return for an average stock and the return due to the investor's skill."
"Covariance" with the stock market is the crucial determinant of skill. A manager whose performance co-varies, or travels together, with higher performing stocks displays skill; covariance with lower performing stocks indicates negative skill, while matching the market has zero covariance.
An effective index therefore should have zero covariance, and that is what the authors' study found for a smart-beta equal-weighted portfolio. Three other smart-beta indexes with different weighting schemes had slightly negative covariance, but the variance was not statistically significant; in other words, they were indicative of "no skill," which is what index investors are seeking.
