Portfolio Managers: Better Born With Wooden or Silver Spoons?

Class Wars A new academic paper claims that mutual fund managers who grew up relatively poor outperform those who came from wealthy families. The early press reports have taken the paper at face value, but I don't believe that you should.

To be sure, their effort is impressive. The authors--Oleg Chuprinin of New South Wales and Denis Sosyura of Michigan--slogged through a morass of data, tracking down information about each portfolio manager's childhood household in the National Census Archives. They recorded the income, occupations, and educational backgrounds of the parents, as well as the home value. They also contacted the registrars at the universities that the managers attended to confirm what degrees they earned and their fields of study.

They then ran a whole lot of numbers. For performance comparisons, they used neither raw total returns, nor risk-adjusted total returns, but instead the academic standard of the four-factor model, which measures the exposure a fund has to value, small-company, and high-momentum stocks and adjusts its performance accordingly.

(The idea of using the four-factor model is to sift out the major systemic effects on a fund's returns--its beta exposures, to use investment-industry jargon. That way, managers are neither penalized nor rewarded by the calculation for decisions that are out of their hands, such as owning mostly small-company stocks while running a small-company fund. What remains after applying the model are, in theory, the alphas--the managers' individual contributions.)

The key finding: Portfolio managers from middle-class or lower-middle-class families (very few fund managers come from outright poverty) show statistically significantly better performance than the golf-and-prep-school set. Managers who come from families in the bottom quintile of wealth, relative to other managers' families, outdid those from the top quintile by about 2 percentage points per year.

That is what has been reported, and it's true enough as far as it goes. But there is more to be said.

Upon Closer Examination For one, there is the usual caveat that the results apply to U.S. diversified stock funds only--not sector funds, not bond funds, not balanced funds, not international-stock funds. Most fund research does the same, because U.S diversified equity funds come with the richest data set and are the easiest to benchmark (for example, there's no four-factor model for bond funds). So, the sin can scarcely be pinned on Chuprinin and Sosyura. But really, that's a minor problem.

The bigger one--a whale of a problem, in fact--is that I suspect that the findings are accidental.

The danger of confusing data accidentals with reality affects all experimental sciences. Professors get published by discovering the positive relationship between two things. If they run enough tests, they will surely uncover some positives, as surely as buying a fistful of instant lottery tickets will lead to some payouts. But are those true, lasting relationships, or just how the darts happened to land this time around? Will they be repeated with the next data set?

(For more on the subject of false positives, see "Why Most Published Research Findings are False," by Stanford's John Ioannidis.)

What gives me particular concern about this paper is a side discovery that male portfolio managers substantially outperform women. At 4 percentage points per year, this little-discussed gender effect is considerably larger than the family-wealth effect. The gender effect also carries a higher t-statistic than does the family-wealth effect. Thus, from the numbers alone, one would expect the paper's primary claim to be the superiority of male managers rather than the manager's family background.

You can guess why the paper headed in the direction that it did. Arguing for the existence of the "weaker sex" would be unfashionable, to put the matter mildly, and would subject the paper to fierce attack. Critics would point out that the paper's sample size is quite small, being less than 300; that the number of females in the sample was much smaller yet; and that the material is irrelevant for current investors, because the study's youngest manager was born in 1945. If you wish to make grand and sweeping conclusions, they would say, you'd better bring more to the party than that.

Aside from the point about there being particularly few females, those same objections apply to this version of the paper. The sample size is still less than 300. Because the authors did not have access to census data from later than 1940, owing to privacy rules, the study covers the previous generation's mutual-fund managers as opposed to those working today. (They also could measure only solo managers rather than those who operated in a team structure. Solo managers were once common, but no longer; even if one accepts the paper's assertion that lower-income managers are better as lone operators, there is an additional step needed to accept that they also excel while working in a team structure. Probably, but not necessarily.)

What's more, this paper's thesis is every bit as sweeping and grand as would be a gender thesis. It claims that, as outsiders, those born into less-wealthy families must be more skilled than those who have the inside track. The authors write, "We argue that managers born poor face higher entry barriers into asset management, and only the most skilled succeed. Consistent with this view, managers born rich are more likely to be promoted, while those born poor are promoted only if they outperform."

How the authors reconcile such a belief with the weaker performance of female managers--who faced similar if not higher entry barriers as managers from less-wealthy families--I do not know. Suffice to say that, from the perspective of the authors' thesis, the gender figures contradict what the family-wealth figures provide.

A Bridge Too Far The authors then postulate that this pattern extends throughout the nation's professions, not just investment management. "We believe our findings have implications that extend beyond asset management. Our evidence suggests that an individual's social status at birth may serve as an important signal of quality in other industries with high barriers to entry, such as corporate management or professional services."

That’s quite a stretch, considering that they haven't done any research into those other fields. After all, if they wished, the authors could instead have argued that their statistically significant finding for the superiority of male fund managers could be generalized across the workforce. The data table's numbers would support such a statement, just as they support the paper's existing statements.

Might the family-wealth effect exist? Might investors wish to seek out portfolio managers from hardscrabble backgrounds? Maybe. Could be. But if so, this paper represents only a start for that proof. An ending, it is not.

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.