Friday, January 6, 2017

Bias Against Those From Less Wealthy Families: The Example of Mutual Fund Managers

When you are thinking about investing in a mutual fund , here's an impolite but seemingly relevant question to ask: How wealthy were the parents of the manager of the fund?  Oleg Chuprinin and Denis Sosyura raise this question in "Family Descent as a Signal of Managerial Quality: Evidence from Mutual Funds" (August 2016, NBER Working Paper No. 22517). The paper is not freely available online, but Jay Fitzgerald offers a short accessible overview in the December 2016 NBER Digest. As Fitzgerald writes:

"This study relies on hand-collected data from individual U.S. Census records on the wealth and income of managers' parents. The researchers also identified and verified fund managers via Morningstar, Nelson's Directory of Investment Managers, and LexisNexis Public Records. They ultimately identified hundreds of fund managers, most born in the mid-1940s, whose parents' Census records were in the public domain. They then examined the performance of hundreds of actively managed mutual funds focused on U.S. equities between the years 1975 and 2012.
"The researchers find that mutual fund managers from wealthier backgrounds delivered `significantly weaker performance than managers descending from less wealthy families.' Managers from families in the top quintile of wealth underperformed managers in the bottom quintile by over one percent per year on a risk-adjusted basis." 
Here's a figure to illustrate the findings. The results show monthly returns in "basis points," which means that 20 is actually .2%. Fund managers whose parents came from the highest wealth quintile performed the worst, while fund managers whose parents came from the lowest wealth quintile performed the best.

What's going on here? It seems unlikely that skill levels are distributed in this way across families. Instead, the plausible explanation offered by the authors is that it's a lot easier for people from high-wealth families to become fund managers, and a lot harder for people from low-wealth families to do so. Because those who start in low-wealth families face more barriers in becoming a fund manager, only the most very highly suited and skilled actually work their way into such a job. Fitzgerald notes some other differences, too:
"Indeed, in tracking career trajectories of mutual fund managers, they find that the promotions of managers from well-to-do families are less sensitive to their performance. In other words, managers who are born rich are more likely to be promoted for reasons unrelated to performance. In contrast, those born into poor families are fewer in number and are promoted only if they outperform. They also find that fund managers from less-affluent families who do make it into top ranks are more active on their job: they are more likely to trade and deviate from the market, whereas those born rich are more likely to follow benchmark indexes." 
The study is interesting to me on both conceptual and practical grounds. On conceptual grounds, it's difficult to measure the extent of favoritism to those from more-wealthy families, or equivalently, the extent of bias against those from less-wealthy families. After all, those who come from families with greater wealth may often have an advantage because their families were in a better position to invest in their human capital, or else because of social favoritism, and it's not easy to find data that will distinguish between the two. But in the example of mutual funds, the measure of job performance is very simple--it's just the return on the fund. So if those from high-wealth families don't (on average) measure up as managers of mutual funds, it certainly smacks of favoritism toward them

On practical grounds, there's no particular reason to believe that the underlying takeaway from the paper applies only to mutual funds. When those who face  higher barriers to success manage to overcome those barriers in any occupation, it may often be a sign that their competence level is not just high, but exceptionally high.