Swedroe: Rigged Horse Race

While I was performing my daily perusal of the SSRN website in order to review the latest studies on investing, I came across a paper by Atanu Saha and Alex Rinaudo of Data Science Partners, “Actively Managed Versus Passive Mutual Funds: A Horse Race of Two Portfolios,” which was written in May 2017 but only recently posted.

Given the subject, I read the abstract, which had, at least for me, a very surprising conclusion. It stated: “This paper demonstrates that the average investor would be better off by following a readily-implementable strategy of investing in a portfolio of the five largest active funds in U.S. equity, fixed income and international equity asset categories than investing in a corresponding portfolio of passive index funds. The active-fund-portfolio outperforms not only in terms of average returns, but also in risk-adjusted returns, providing far greater downside risk protection than the passive fund portfolio. This paper has important implications because its findings question the ‘wisdom’ of index investing, which has been receiving considerable attention in the financial press in the recent years.”

Lopsided Race

To create their horse race, the authors constructed portfolios across three broad asset classes: domestic stocks, domestic bonds and international stocks. They took the five largest active funds and the five largest passive funds in each category and reconstituted the portfolios annually. The study covered the period 1996 through 2015.

Saha and Rinaudo found that, in addition to the portfolios of the five largest active funds outperforming the portfolios of the five largest passive funds in each of the three broad asset classes, they also outperformed the average active fund.

Unfortunately, this methodology means they likely aren’t measuring funds on an apples-to-apples basis. In other words, the five largest active funds chosen could be very different from the five largest passive funds, even if they all are in one of the same broad asset classes that the authors used.

For example, the two Dodge & Cox funds that made the largest active funds list are value funds, and there are no value index funds among the largest passive funds selected.

In the paper’s appendix, Saha and Rinaudo list the funds used in their portfolios, and we can see that while the largest passive U.S. equity funds sometimes included midcap, small-cap and REIT index funds, the list was quite different for the active funds, which tended to be large-cap blend and growth funds. In addition, the various funds could have very large differences in geographical exposures.

For example, while the Vanguard Emerging Markets Index Fund is included among the largest passive funds for all 20 years, there are no active emerging market funds listed. Over the period of the study, the S&P 500 Index returned 8.2% per year, while the MSCI Emerging Markets Index returned just 5.5% per year, an underperformance of 2.7 percentage points per year.

This lag is reflected in the list of funds the authors used to create portfolios, and helps to create the illusion that active investors outperformed passive investors. However, we know there is far more money invested actively than passively in emerging markets, and passive emerging market strategies have outperformed active ones.

(Note that the authors did a robustness test, removing emerging markets from the data, and they did find that it closed the gap by 0.08%. That one change eliminated about one-quarter of the entire advantage of the active funds.)

Additional Caveats

Here’s another interesting observation. Vanguard’s REIT index fund appears on the list of largest passive funds for only a single year, 2013. In that year, it returned 1.2%, versus 32.4% for the S&P 500 Index. Because the authors did not check for this, we don’t know how much of an impact it might have had. On the bond side, selected funds could have had very different exposures to both term and default risk.

This is why, when studies are done on active versus passive funds, they are generally performed on an apples-to-apples basis—to account for different exposures to the factors, such as size, value, momentum and profitability/quality, that explain differences in returns.

Higher Hurdles

Another reason to be skeptical of the study’s claims is that, while some economies of scale may be available to larger funds, the larger an active fund gets, the greater the hurdles to outperformance become.

The reason is that, in order to overcome the hurdle of their higher costs (not just the expense ratio, but also all trading costs), active funds must be highly differentiated (have a high active share). As their assets increase, they’ll either maintain a highly concentrated portfolio, in which case, when they trade, their market impact costs will rise, or they will diversify further, reducing their active share and leaving them with a portfolio less differentiated from the proper benchmark. In either case, the hurdle to generating risk-adjusted alpha will increase.

That’s also one of the major problems for successful actively managed funds. If they generate alpha, cash flows follow, and the hurdles to generating alpha in the future increase. Successful active management contains the seeds of its own destruction.

Increasing competition and passive strategies’ persistently declining costs (investors, for instance, now have access to index funds with a zero expense ratio) are some of the reasons we see the results we do in the annual SPIVA tables, which regularly show not only that the vast majority of active funds underperform, but there is less persistence than is randomly expected.

Larry Swedroe is the director of research for The BAM Alliance, a community of more than 140 independent registered investment advisors throughout the country.