How Is This Research Useful to Practitioners?
Addressing the relationship between fund flows and fund performance, the authors show that funds that experience outflows in response to poor performance (i.e., suffer “attrition”) are left with relatively performance-insensitive investors. This attrition concept conceptually resembles the “burnout” concept in mortgage-security prepayment sensitivity analysis. A period of low interest rates separates heterogeneous homeowners in a mortgage pool into two types: homeowners with relatively high interest rate sensitivity who will quickly refinance and leave the mortgage pool and homeowners with relatively low interest rate sensitivity who are less likely to refinance.
Similarly, attrition in the mutual fund space can be used to measure the predominant type of investor in a fund. Investors are heterogeneous; they differ in their willingness to remain invested in a fund given the fund’s performance. Some are relatively more performance sensitive, whereas others are relatively less performance insensitive. There is no separation of investor types when performance is good because everyone is content to stay or enter but the fact that a fund has lost investors because of poor performance (i.e., attrition) reveals something about the investors who remain in the fund. Funds that experience high attrition after poor performance exhibit a significantly flatter flow–performance relationship in the future because mostly performance-insensitive investors remain.
Investors and analysts can use a fund’s attrition measure, which is easily calculated, to help estimate the stability of the fund’s investor base. For example, firms buying asset management firms can—somewhat counterintuitively—benefit from purchasing client relationships that have experienced poor recent performance because these assets might be relatively “sticky.” Funds with higher attrition rates exhibit lower net flows, higher fees, and lower overall flow–performance sensitivities compared with funds experiencing less attrition.
How Did the Authors Conduct This Research?
The authors first measure attrition for US domestic equity funds. A fund’s attrition measure at the end of a year is calculated as 1 – (Current fund size at year-end/Historical maximum fund size over all months prior to year-end). The authors use data from the CRSP Survivor-Bias-Free US Mutual Fund Database from 1998 to 2012.
They then regress a fund’s next-year flows on the fund’s ranked return this year, its attrition dummy (i.e., either above or below the median), an interaction term of the ranked return and attrition dummy, age, lagged flows, expenses, and several other control variables.
The authors find that although this year’s ranked return helps explain next year’s fund flows (as expected), the coefficient on the interaction term is negative and statistically significant. In other words, funds with above-median attrition have a lower sensitivity of fund flows to prior returns. The authors also show that attrition affects the price sensitivities of inflows as well as outflows.
Prior researchers of the fund flow–performance relationship have highlighted the role of a fund’s age in dampening this connection. The authors show that fund age and attrition are positively correlated and that the role of age as a predictor of the flow–performance relationship is much diminished once the regression explicitly includes attrition. If attrition is caused by a nonperformance characteristic (e.g., a service or governance issue)—and investors leave for reasons unrelated to performance—attrition does not sort investors by performance sensitivity.