Improving the Odds in Active Management

Finding Better Large-Company Stock Funds
Last year, historically quiet American Funds stepped into the active-passive fray with a report entitled The Active Advantage. The contents did not match the wrapper. The study convincingly showed that the firm's own stock funds had excelled, decade after decade. However, there was no apparent advantage to investing in other companies' actively managed funds.

American Funds has returned with a sturdier effort. Released last week, the company's follow-up (blandly titled Expect More From the Core) delivers what the first paper did not--a case for selecting an active stock-fund manager over an index fund. At which point your suspicions are probably aroused. Rest assured, while American Funds most certainly is interested in spreading the gospel of active management, I am not. If the numbers work, they work. If they don't, they don't.

The study evaluates large-company stock funds, both domestic and foreign, for the 20-year period of 1994 through 2013. The comparison indexes are the S&P 500 for U.S. stock funds and the MSCI ACWI ex USA benchmark for foreign-stock funds. As is customary for mutual fund studies, the fund returns are net of annual expenses (but not initial sales charges), while the indexes are unadjusted for costs.

Low Costs, High Ownership
Among these funds, the study sorts for those with a) low costs and b) high manager ownership.

Low cost is defined as landing in the cheapest quartile for annual expenses as of the fund's most recent reporting date.

The manager ownership data point is more complex. It is assigned at the company level, not the fund level. That is, if a fund company's managers tend to invest heavily in the funds that they run, then all the company's funds register as high manager ownership. As it turns out, designating the data point to companies rather than individual funds is a useful, important decision.

But more on that shortly. On to the numbers. Over the 20-year time horizon, American Funds measured how often a fund category (adjusted for survivorship bias) outperformed its relevant index, looking at rolling five- and 10-year periods. Separating into U.S. and international funds, it then sifted each of those groups into four pools:

1) All actively managed funds.

2) Only funds landing in the lowest cost quartile.

3) Only funds from families within the highest quartile for fund ownership.

4) Only funds that satisfied both the second and third points, that is, funds that showed low costs and high manager ownership.

The rolling five-year figures:

That cheaper was better is no surprise, nor that more manager ownership was beneficial. It also is no shock that funds that combined the two attributes of low cost and high ownership fared best of all. However, I would not have predicted that manager ownership would be the stronger of the two effects. Merely tilting toward organizations that had high manager ownership, without regard to any other factor, moved active international funds to be even with the costless index and pushed active U.S. stock funds moderately ahead of their benchmark.

The same pattern held true for the rolling 10-year periods, only more so:

When the time horizon extended to a decade, the low cost/high ownership funds outperformed almost every time. This isn't quite the achievement that it might first seem, as there are only two independent periods in the study, but nonetheless, it is a powerful result. Would you have believed that screening actively run large-company stock funds with two readily available factors (the manager ownership data point is a Morningstar statistic) would lead to a pool of funds that reliably beat the costless indexes?

Probably not. Nor, I think, would you have suspected that in the rolling five-year periods, the low cost/high ownership funds gained an average of 7.47% per year, as opposed to a blended 5.69% return for the two largest exchange-traded funds that invest in the S&P 500 and the MSCI ACWI indexes. That makes 106 actively managed mutual funds, selected through two basic screens, handily beating their ETF rivals.

Cautionary Notes
Two caveats, one minor and the other major.

The minor caveat is the issue of data mining. (As the saying goes, torture the data enough and it will confess.) While a common scourge with investment research, data mining does not look to be a problem here. American Funds immediately identified cost as the first factor for its screen, and then tested only a handful of stewardship-related items as the second factor. This study didn't do much fishing.

The major caveat is that the data come from the most recent reporting date, rather than at the time of the measured performance.* This is of little consequence for the cost factor, as fund expenses are remarkably steady over time. (Although the fund industry promises to pass along economics of scale, in reality, fund costs are largely independent of asset growth.) It is of much greater importance for fund ownership. Obviously, managers of successful funds will be more highly paid, and therefore have more assets to invest in their funds. It is a virtuous circle.

* American Funds could have laboriously created year-by-year expense tables so as to assign cost quartiles at the time of performance, but it had no choice but to use relatively recent information for manager ownership, as those figures were unavailable during the first half of the 20-year time period. 

This is where American Funds' decision to measure manager ownership by company comes in handy. Whereas manager ownership is clearly circular for individual funds, it is much less so when applied at the company level. Hot funds rarely change the data point. As measured in this study, manager ownership is driven mostly by corporate culture, as opposed to what occurs at this fund or that fund.

That hypothesis is strengthened by the list of fund companies that made the cut, which is light on large organizations and heavy on boutiques, many of which are partnerships. According to American Funds, which conducted sensitivity testing with a data set from 2006, the list of today's high-ownership firms looks much like the list from eight years ago.

Of course, the stability of manager ownership remains to be proved. Thus, the study as it stands, while provocative and promising, cannot be considered as complete. To finish the task, American Funds must conduct its tests fully ex-ante. I think there's a good chance that the final numbers will pan out, though. The arguments for the data points' stability are solid. 

Ironically, the study commissioned by the king of load funds echoes advice long given by the king of no-load funds. Jack Bogle's advocacy of low costs is well known. Less known is Bogle's argument that investors are best served by fund companies that are not publicly traded, nor part of a larger conglomerate. To a large extent, the companies on American Funds' list fit that description. The two parties arrived there in different ways, one from big data and the other from experience and observation, but the conclusions are similar: Owning cheap funds from good stewards is the way to go. 

John Rekenthaler has been researching the fund industry since 1988. He is now a columnist for Morningstar.com and a member of Morningstar's investment research department. John is quick to point out that while Morningstar typically agrees with the views of the Rekenthaler Report, his views are his own.