Working Paper: NBER ID: w12261
Authors: James J. Choi; David Laibson; Brigitte C. Madrian
Abstract: Experimental subjects allocate $10,000 across four S&P 500 index funds. Subject rewards depend on the chosen portfolio's subsequent return. Because the investments are not actually intermediated by the fund companies, portfolio returns are unbundled from non-portfolio services. The optimal portfolio therefore invests 100% in the lowest-cost fund. Nonetheless, subjects overwhelmingly fail to minimize fees. When we make fees transparent and salient, portfolios shift towards cheaper funds, but fees are still not minimized. Instead, subjects place high weight on normatively irrelevant historical returns. Subjects who choose high-cost index funds are relatively much less confident about their asset allocation choices.
Keywords: index mutual funds; law of one price; investor behavior; mutual fund fees
JEL Codes: D14; D18; D43; D83
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Making fee information transparent and salient (D49) | Reduces allocations to high-cost funds (G23) |
| High weight on normatively irrelevant historical returns (G41) | Increased allocations to funds with higher past returns (G11) |
| Transparent fee information (G19) | Investors do not fully minimize fees (G19) |
| Choosing high-cost funds (G23) | Lower confidence in investment decisions (G11) |