Job security is of major importance for most people – and fund managers are no exception. In their influential Quarterly Journal of Economics paper “Career Concerns of Mutual Fund Managers” from 1999, Chevalier and Ellison explore whether career concerns affect the investment behavior of mutual fund managers. Their main hypothesis: past performance affects a manager’s termination risk and therefore drives managers’ investment decisions.
Age and termination: Fund families trust younger managers less
Chevalier and Ellison use data on mutual funds from 1992 to 1994 for their analysis. They first run a regression of whether fund families fire a manager on the manager’s past performance, her age relative to average age of the population – and the interaction of the two factors. Table 1 shows that the loading on this interaction term is positive and significant. So, a younger manager who underperforms has significantly higher termination risk than an older manager!
|Performance x (Manager Age – Av. Age)||0.259|
Table 1: Impact of the interaction term of performance with manager age (in excess of the average age) on a manager’s termination risk in the next year. Source: Table 3, Chevalier and Ellison (1999). Bold font indicates significance at the 1% level. Fund- and manager-level control variables are included in the regression.
The economic magnitude of the effect is also substantial. If two managers with an age difference of 10 years have the same performance, the younger’s termination risk is higher by 7 percentage points!
Better run with the herd!
How does this difference affect managers? Younger managers are “punished” more for underperformance. Apparently, fund families trust younger managers less. This could give them an incentive to “herd” more than older managers: for them, the consequences of not beating the market are worse. To prevent termination, they should therefore behave exactly like other managers. Chevalier and Ellison empirically document exactly this strategy: younger managers deviate less from others in their sector weighting and in the unsystematic and systematic risk they take – see Table 2.
|Sector Deviation||Unsystematic Risk Deviation||Systematic Risk Deviation|
|Manager Age (/100)||0.21||0.02||0.20|
Table 2: Impact of the manager age on the manager’s deviation from the average sector weights, unsystematic risk, and systematic risk. Source: Table 6, Chevalier and Ellison (1999). Bold font indicates significance at the 1% level. Fund- and manager-level control variables are included in the regression.
But what about the upside?
Up until now, we have focused only on the downside of risk taking – the higher termination probability when a manager underperforms. But taking risks and behaving different from the herd can of course also lead to outperformance. So, do fund families reward outperformance with promotion? Chevalier and Ellison also analyze this question. And though the data basis is quite small, they can show that older and younger managers have the same chance of promotion – given the same performance. And, bad news for managers of all ages: even very high performance hardly increases the promotion probability! In summary, young managers have all the downside risk, but no upside potential!
The focus of research on delegated portfolio management has historically been on the classical principal/agent relationship: between the fund family and the investor. Chevalier and Ellison show that the fund manager plays a crucial role in this relationship. Hence, their paper is among the first to link delegated portfolio management research to labor economics research.