Stock returns and aggregate mutual fund flows: a system approach

Abstract

To investigate if the mutual fund flows have been a driving factor in the US stock market at the macro level, we combine information from the stock market with information from bond and money markets in a system method. The empirical evidence from Seemingly Unrelated Regression Error Correction Model (SURECM) and Granger and Sims causality tests in a system indicates that the fund flows are weakly exogenous and stock performance causes fund flows, implying that investors move their money to the securities that yield higher returns to rebalance their investment portfolios in the US market.

Acknowledgements

We are grateful to the editor, an anonymous referee, Bong-Soo Lee and seminar participants at the ASSA/KAEA meeting in Boston, WEA meeting in Seattle, 2007 FEMES meeting in Taiwan and FMA in Orlando for their helpful comments and discussions. I also thank Xi Tian for updating the data. All remaining errors are our own.

Notes

¹ See Ippolito (1992) and Grinblatt et al. (1995) for details.

² See Edelen and Warner (2001), Goetzmann and Massa (2003) and Boyer and Zheng (2004) for details.

³ Price pressure theory implies that when inflows in mutual funds increase, this stimulates a demand to hold stocks, and causes stock prices to go up. Information revelation hypothesis shows that based on the assumption that the market is to react to the available information instead of responding to the fund flows, well-informed investors’ purchases may signal to other less-informed investors to buy mutual fund, thus cause stock prices to rise. Further, investor sentiment is also considered as one of the important factor affecting mutual fund market. See Warther (1995) and Boyer and Zheng (2004) for details.

⁴ See Fortune (1998) and Potter and Schneeweis (1998) for details.

⁵ For unit root tests, we employed Augmented Dickey–Fuller (ADF) and Kwiatkowski–Phillips–Schmidt–Shin (KPSS) (1992) tests. The descriptions of the tests will be available upon requests.

⁶ See Kim (2004) for details.

⁷ The λ_{0 i} measures how much Y_t responds to a deviation from the long-run equilibrium in the previous period and we use the half-life for this. See Engle and Granger (1987) for details.

⁸ Mutual funds played a smaller role in the pre-1984 markets, we focus on the post-1984 markets.

⁹ We compute net flows as total sales minus redemptions, plus exchange sales minus exchange redemptions.

¹⁰ These are the actual rates of return experienced during the month, including both cash income and capital value changes.