Assaying Anomalies

Abstract

We propose a protocol for testing potential cross-sectional predictors of equity returns, and describe turn-key tools for implementing this protocol. These tools are not completely exhaustive, but identify the most important issues that arise in common tests of asset pricing strategies. They go far beyond the direct inferences available from the simple tests commonly employed using standard linear factor models, paying particular attention to arbitrage limits, which can make strategies that are impossible to exploit in practice look good on paper. These tools provide a thorough, transparent analysis, along the lines of a referee report, with little more effort than pushing a button.

Keywords: Anomalies, Performance Evaluation, Trading Costs, Factor Models

JEL Classification: G11, G12, G14