More evidence that passive funds are superior to active

By LARRY SWEDROE

The publication of research (such as the Standard & Poor’s Indices versus Active Scorecards) on the failure of actively managed funds to demonstrate risk-adjusted outperformance has fuelled the trend toward passive investing. Yet, despite the overwhelming body of evidence, at the end of 2018 actively managed funds still controlled the majority of both equity and bond assets in publicly available vehicles (mutual funds and ETFs).

Edwin Elton, Martin Gruber and Andre de Souza contribute to the literature on the debate about whether active or passive strategies are superior with their study, Are Passive Funds Really Superior Investments: An Investor Perspective, which appeared in the Third Quarter 2019 issue of the Financial Analysts Journal. The authors sought to determine whether a simple implementable passive strategy could be expected to outperform active funds.

They began by examining whether a combination of a small set of ETFs could be found which, while matching the risk of each active fund in one period, would outperform active funds in the following period. They used ETFs because not only are they live vehicles (not benchmark indices, which do not have expenses), they can be held as either long or short vehicles. They wanted a small set so that the strategy could be easily implemented by individual investors. They found they only needed five ETFs to explain more than 90 per cent of the variation in returns of the active funds — a market ETF, a Russell 1000 Growth ETF, a Russell 1000 Value ETF, a Russell 2000 Growth ETF, and a Russell Mid-Cap Value ETF. They examined risk-matching and future performance for two cases: when short sales are allowed, and when short sales are not allowed (many investors will not engage in shorting). Their sample included 883 active funds.

Elton, Gruber and de Souza used the five ETFs to match the risk of active equity funds every December starting in 2004 using two years of previous monthly data. This provided them with 15 years of data (through 2018). They then compared the return of each active fund with its risk-matching ETF portfolio in the subsequent year. Following is a summary of their findings:

The risk-matching ETFs outperformed the active mutual fund 77 per cent of the time with unlimited short sales and 78 per cent of the time with no short sales.

When loads and transaction costs were considered, ETFs had higher returns more than 90 per cent of the time.

In the no short sales case, the matched portfolio outperformed the active portfolio on average in 12 and 15 years. In the short sales allowed case, the outperformance occurred in 13 out of 15 years. Underperformance occurred around the crash (2007 and 2009) in both cases.

In both the unlimited short sale case and the no short sale case, the average extra return earned was about 1.4 per cent per year — statistically significant at the 1 per cent confidence level.

The risk-matching ETFs on average had a lower standard deviation of returns. In the no short sales case, the matching ETFs has a monthly standard deviation that was lower by 0.15, and in the short sales allowed case, the standard deviation was lower by 0.10.

Four out of the five ETFs had extremely small bid-ask spreads. The one exception was the low-cost ETF matching the Russell 2000 Growth Index — the maximum bid-ask spread observed on this index was still less than three-tenths of a basis point.

The matching portfolios outperformed in each of the nine Morningstar style categories. For the short sales allowed case, the difference in performance was statistically significant at the .01 level in all categories. In the no short sales case, eight out of nine Morningstar category, the standard deviation of the risk-matching portfolio was smaller than the standard deviation of the active fund.

Elton, Gruber and de Souza also examined whether common suggested methods of selecting actively managed funds would have outperformed their ETF replication strategy. They considered screens such as expense ratio, past turnover, R-squared (measure of active share), R-squared plus alpha and Morningstar ratings. They found that for long-only portfolios (which would be the case for most investors), none produced superior results. However, when unlimited short sales were allowed, the R-squared screen outperformed the ETF portfolio. Note that because short sales are very expensive, the costs would likely destroy any benefit. Thus, the no short sales (long only) case, which does not produce positive performance, is the more relevant.

They next examined whether the strategy of simply investing in an ETF that matched the active fund’s prospective benchmark would outperform. They found that the “ETF outperformed the active fund 72% of the time with an excess return of 1.01% per year.” They also found that “the five ETF model outperformed the prospectus benchmark model at a statistically significant level.”

Summary

Elton, Gruber and de Souza demonstrated the typical long-only individual investor can outperform active funds by simply investing in the low-cost ETF that matches each fund’s benchmark. However, they also showed that investors can do even better by employing the five ETF model they develop in their paper.

Are Passive Funds Really Superior Investments: An Investor Perspective adds to the body of evidence demonstrating that active management is a loser’s game. While it’s possible to win, the odds of doing so are so poor that it’s not prudent to try. And finally, as you consider the results, remember that for taxable investors, Elton, Gruber and de Souza’s results are far more damning because the largest expense for actively managed funds in taxable accounts is typically taxes — and ETFs are generally the most tax-efficient vehicles.