Working Paper: CEPR ID: DP1663
Authors: Magnus Dahlquist; Paul Soderlind
Abstract: This paper provides evidence on the use of stochastic discount factors in the evaluation of portfolio performance. First, we discuss evaluation in this setting, and relate it to traditional mean-variance analysis. We then use Monte Carlo experiments to examine the small sample properties of generalized method of moment (GMM) estimators. Both size and power properties are characterized for various GMM approaches. Finally, we apply the methodology to Swedish-based mutual funds. We offer an evaluation allowing for passive as well as dynamic strategies. The conditional evaluation indicates that funds may have had superior performance over the sample period.
Keywords: GMM estimators; intersection and spanning tests; mean-variance analysis; mutual funds; small sample properties
JEL Codes: G11; G12; G23
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Mutual funds (G23) | Efficient benchmark portfolios (G11) |
| Efficient benchmark portfolios (G11) | Accurate comparison of returns (C52) |
| Mutual funds do not enhance investment opportunity set (G23) | Expected returns conform to SDF predictions (G17) |
| Mutual funds (G23) | Superior returns (G19) |
| Performance metrics (C52) | Expected returns (G17) |
| Significant excess returns required to reject null hypothesis of neutral performance (C12) | Causal link between performance metrics and expected returns (G17) |
| Empirical size and power properties of tests (C12) | Detect superior performance (C52) |