Recent studies show that the standard test portfolios do not contain sufficient information to discriminate between asset pricing models. To address this issue, we develop a large-scale approach that expands the cross-section to several thousand portfolios. Our novel approach is simple, widely applicable, and allows for formal evaluation/comparison tests. Its benefits are confirmed in empirical tests of CAPM- and characteristic-based models. While these models are all misspecified, we uncover striking performance differences between them. In particular, the human capital and conditional CAPMs largely outperform the CAPM, which suggests that labor income and time-varying recession risks are primary concerns for investors.

We formally compare two versions of the market variance risk premium (VRP) measured in the equity and option markets. Both VRPs follow common patterns and respond similarly to changes in volatility and economic conditions. However, we reject the null hypothesis that they are identical and find that their difference is strongly related to measures of the financial standing of intermediaries. These results shed new light on the information content of the VRP, suggest the presence of market frictions between the two markets, and are consistent with the key role played by intermediaries in setting option prices.

This paper develops a unified approach to comprehensively analyze individual hedge fund return predictability, both in and out of sample. In sample, we find that variation in hedge fund performance across changing market conditions is widespread and economically significant. The predictability pattern is consistent with economic rationale, and largely reflects differences in key hedge fund characteristics, such as leverage or capacity constraints. Out of sample, we show that a simple strategy that combines the funds’ return forecasts obtained from individual predictors delivers superior performance. We exploit this simplicity to highlight the drivers of this performance, and find that in- and out-of-sample predictability are closely related.

We develop a flexible and bias‐adjusted approach to jointly examine skill, scalability, and value‐added across individual funds. We find that skill and scalability (i) vary substantially across funds, and (ii) are strongly related, as great investment ideas are difficult to scale up. The combination of skill and scalability produces a value‐added that (i) is positive for the majority of funds, and (ii) approaches its optimal level after an adjustment period (possibly due to investor learning). These results are consistent with theoretical models in which funds are skilled and able to extract economic rents from capital markets.

This paper develops a simple technique that controls for “false discoveries,” or mutual funds that exhibit significant alphas by luck alone. Our approach precisely separates funds into (1) unskilled, (2) zero‐alpha, and (3) skilled funds, even with dependencies in cross‐fund estimated alphas. We find that 75% of funds exhibit zero alpha (net of expenses), consistent with the Berk and Green equilibrium. Further, we find a significant proportion of skilled (positive alpha) funds prior to 1996, but almost none by 2006. We also show that controlling for false discoveries substantially improves the ability to find the few funds with persistent performance.

We develop a novel approach to separate alpha and beta under model misspecification. It comes with formal tests to identify less misspecified models and sharpen the return decomposition of individual funds. Our hedge fund analysis reveals that: (i) prominent models are as misspecified as the CAPM, (ii) several factors (time-series momentum, variance, carry) capture alternative strategies and lower performance in all investment categories, (iii) fund heterogeneity in alpha and beta is large—an important result for fund selection and models of active management, (iv) performance is increasingly similar to mutual funds, (v) fund valuation is sensitive to investor sophistication.