Building on the work of Barras, Scaillet and Wermers (BSW, 2010), we propose a modified approach to inferring performance for a cross-section of investment funds. Our model assumes that funds belong to groups of different abnormal performance or alpha. Using the structure of the probability model, we simultaneously estimate the alpha locations and the fractions of funds for each group, taking multiple testing into account. Our approach allows for tests with imperfect power that may falsely classify good funds as bad, and vice versa. Examining both mutual funds and hedge funds, we find smaller fractions of zero-alpha funds and more funds with abnormal performance, compared with the BSW approach. We also use the model as prior information about the cross-section of funds to evaluate and predict fund performance.

We develop methods for testing factor models when the weights in portfolios of factors and test assets can vary with lagged information. We derive and evaluate consistent standard errors and finite sample bias adjustments for unconditional maximum squared Sharpe ratios and their differences. Bias adjustment using a second-order approximation performs well. We derive optimal zero-beta rates for models with dynamically trading portfolios. Factor models’ Sharpe ratios are larger but standard test asset portfolios’ maximum Sharpe ratios are larger still when there is dynamic trading. As a result, most of the popular factor models are rejected.

Portfolio performance measures using holdings data are panel regressions. The returns of a fund’s stocks are regressed on its lagged portfolio weights. Stock fixed effects isolate average performance from time-series predictive ability. Control variables condition for fund performance on the characteristics of the stocks held. The long-term performance of average holdings drives some of the classical measures, while predictive ability drives others. A “buy-and-hold drift,” where portfolio weights increase over time in the higher alpha stocks, affects performance measures. Investor flows respond to average performance net of the buy-and-hold drift. (JEL G11, G14, G23, G29). Received September 3, 2020; editorial decision January 25, 2021 by Thierry Foucault. Authors have furnished an Internet Appendix, which is available on the Oxford University Press Web site next to the link to the final published paper online.

Junbo L. Wang Louisiana State University January 14, 2019 PRELIMINARY Abstract We provide asymptotic distributions for tests of asset pricing models and factor model comparisons with conditioning information in the form of lagged instruments, including closed-form asymptotic variance estimators for squared Sharpe ratios and their normalized differences. We provide results for both traded and non-traded factor models. We evaluate the asymptotic standard errors with simulations and provide applications to asset pricing model tests and factor model comparisons. We find that the incremental performance of the FF5 model over the FF3 and the CAPM is greater when dynamic trading is allowed. A dynamic consumption hedging portfolio contributes to the performance of most of the models, as does a dynamic liquidity hedging factor.

We provide asymptotic standard errors for tests of asset pricing models and factor model comparisons with dynamic trading using conditioning information in the form of lagged instruments. The tests are based on comparing squared Sharpe ratios or their normalized differences. We provide results for both traded and non-traded factor models, and we study the optimal choice of the zero beta rate. We evaluate the asymptotic standard errors with simulations and provide applications to asset pricing model tests and factor model comparisons. We find that the incremental performance improvement of the FF5 model over the FF3 and the FF3 over the CAPM is greater when dynamic trading is allowed, as is the effect of a momentum factor. Dynamically trading consumption and liquidity hedging portfolios contribute significantly to the performance of most of the models.