Read Full Paper

Mutual funds are important investment vehicles for many households. When these households invest in a fund, their capital is pooled and entrusted to a professional fund manager; this manager is responsible for growing his clients’ money by investing in stocks, bonds and other assets. Understandably, therefore, the choice of fund manager is crucial. There are large differences among fund managers in terms of skill, so how do potential investors gain knowledge that can assist them in picking a manager?

Darwin Choi, Bige Kahraman and Abhiroop Mukherjee provide evidence that investors learn about mutual fund managers in a relatively sophisticated manner; but there is still room for improvement. Their paper studied managers who manage more than one mutual funds simultaneously, and examined whether investors learn about managerial ability from past performance in all funds managed by the same person, as they should.

People often decide whether and how much to invest in a fund by looking at the fund’s past performance. If they are doing so to rationally figure out whether the manager has investment skill, they should decide how much to invest in a fund based not only on that fund’s past performance, but also on the performance of other funds managed by the same manager. But does the average investor do this? The paper’s findings generally find that they do – capital flows into a fund are also predicted by past performance in the manager’s other funds.

However, capital allocation does not seem to be fully rational and complete. If investors allocated funds exactly as they should in theory, fund performance should not have been predictable based on past records. This is because when an investor realizes that a particular fund manager has skill, he should invest more money with that manager, as should all other investors. When the manager’s funds grow bigger, he will likely not be able to repeat his past success – if he had 5 great stock trading ideas managing a $50 million fund, getting an additional $50 million does not mean that he will find 5 more ideas to invest the extra money in! On the other hand, if investors did not correctly allocate as much more money as they should have to a skilled manager, he could still be able to do well in the future—his original 5 trading ideas may still be good enough to yield high future returns on a $55 million fund.

From their performance predictability tests, the researchers found evidence of positive cross-fund “return predictability”: underperformance in one fund managed by a multi-fund manager today predicts poor performance of the manager’s other fund in the next quarter. This suggests that investors do not withdraw enough capital in response to poor performance in the manager’s other funds, and hence incur losses. These losses come mostly from funds managed by multi-fund managers who underperformed in other funds that they managed in the past. The paper concluded that although investors seem to respond in the right direction to their manager’s performance in the whole portfolio of funds he manages and not just focus on one fund in isolation, their response to performance of these other funds is insufficient.

This conclusion is different from prior literature on investor learning about mutual funds, and has implications for why some mutual fund investors continue to invest in poorly performing funds: frictions in learning about managerial skill may be one mechanism driving this phenomenon.

This research complements other studies on cross-fund learning, and contributes to the debate on whether capital allocation into mutual funds. The researchers suggests that we, as investors, will be better off if we pay closer attention to the performance – particularly underperformance – in the whole portfolio of funds our manager was in charge of, instead of narrowly focusing only on the fund that we had invested in.