Stock and fund return predictability are partially driven by predictable price pressure, which is the result of capital flows from retail investors to mutual funds and from mutual funds to individual stocks. This flow-driven return effect can fully account for both mutual fund performance persistence and the "smart money effect" and can partially explain stock price momentum.
Mutual fund managers typically expand their existing holdings with capital flows and liquidate their holdings to pay for redemptions. This flow-induced trading can have a significant impact on individual stock returns.
The author attempts to apply this flow-based mechanism to well-known findings on return predictability. Flow-induced purchases and sales can drive winning funds to continue to outperform losing funds—a pattern often thought to demonstrate manager ability. The mechanism also gives rise to the "smart money effect," whereby flows predict next-quarter performance.
How Is This Research Useful to Practitioners?
A pertinent question for practitioners to ask is whether using expected capital flows (versus past flows) can result in a better forecast of future stock returns. Using past abnormal fund performance to forecast future fund flows, the author shows that expected flow–induced trading positively predicts stock returns in the short run but negatively predicts stock returns in the long run because of the dissipation of flow price pressure. Furthermore, he finds that outperformance in the first year is completely reversed by the end of the third year. This finding is consistent with the notion that arbitrageurs have limited capacity to absorb demand shocks in the equity market. Typically, practitioners have the objective of minimizing portfolio volatility; this may be more difficult to achieve in funds with high expected capital flow. The research results suggest that capital flows play a role in causing excess stock return co-movement. The author finds that after he controls for common risk factors, stocks that are expected to receive inflow-induced funds tend to move together. The same is true for stocks that are expected to experience outflow-induced sales.
Although past abnormal fund returns and fund flows are significant predictors of future fund performance, both are subsumed by expected flow–induced trading, which the author refers to as E[FIT]. He concludes that the observed patterns of mutual fund performance and the smart money effect are likely to be manifestations of price effects induced by predictable flows. These effects are important to consider when investment professionals are in the process of choosing active managers and approving performance incentive fees. Once the author controls for E[FIT], stock price momentum is no longer statistically significant. This flow-based mechanism explains momentum’s persistence over time and robustness across stocks.
How Did the Author Conduct This Research?
The author uses quarterly mutual fund holdings data from the CDA/Spectrum database for 1980–2006. This database includes mandatory U.S. SEC filings and disclosures. Mutual fund net asset returns and other characteristics are from the CRSP survivor bias–free database, and stock liquidity data are from Joel Hasbrouck. The author restricts the sample to domestic diversified equity funds, excluding balanced and sector funds.
To construct the measure of expected flow–induced trading, or E[FIT], he first estimates the part of mutual fund trading that is associated with capital flows. He finds that managers sell their holdings dollar-for-dollar to meet redemptions but invest only 62 cents for every dollar of inflow. Then, he computes a measure of flow-induced trading across all mutual funds for every stock in every quarter. Finally, he computes E[FIT] by replacing realized capital flows with expected flows.
The author begins his analysis by sorting all mutual funds into deciles based on alpha and holding the resulting portfolios for the next 12 quarters. He finds continuation in abnormal mutual fund performance and reports that more than half of the return spread is a result of the continued outperformance of past winning funds. The author distinguishes between a flow-based explanation and a manager ability explanation and conducts performance analysis. The results suggest that the pattern of mutual fund persistence is likely to be a result of forecastable, flow-induced trading.
The author suggests that performance-chasing investment flows from retail investors can cause institutional managers to follow momentum strategies; this flow-induced herding can have asset-pricing implications. It would be interesting to see the results for such other asset classes as international equities or precious metal funds. In addition, updated research that includes 2007–2012 would be useful.