We first discuss performance evaluation using stochastic discount factors a
nd relate it to traditional mean-variance analysis. We then use Monte Carlo
experiments to examine the properties of various general method of moment
(GMM) estimators. The test statistics are fairly well behaved although seri
ous size distortions are found in some cases. The simulations also show tha
t a significant excess return, or a long sample, is needed to reject neutra
l performance. Finally, we offer an evaluation of Swedish-based mutual fund
s. The conditional evaluation indicates that funds have had nonneutral perf
ormance as revealed by the predictability of the unconditional performance
measure.