Real-Time Profitability of Published Anomalies: An Out-of-Sample Test

Abstract:     
Empirical evidence on the out-of-sample performance of various asset-pricing anomalies in equity is mixed so far and arguably is often subject to data-snooping bias. This paper conducts a performance analysis of real-time strategies that can significantly reduce data-snooping bias and that involve only asset-pricing anomalies published prior to that time and non-forward-looking filters. In particular, we consider a simply long-only strategy that each year recursively picks the best past-performer among published anomalies. We find that this strategy can outperform the equity market even after taking transaction costs into account. This outperformance tends to be stronger when we pick performers based on their performances between the past two and five years and when we include more published anomalies. We also find that the performance of published anomalies --- either relative to standard benchmarks or their relative ranks among contemporaneous anomalies --- do not decrease over time, indicating a relatively stable performance once being published. Overall, our results suggest that published anomalies persist even after controlling for data-snooping bias.

Number of Pages in PDF File: 40

Keywords: Data-snooping Bias, Asset-pricing Anomalies, Out-of-sample Test, Real-time Simulation

JEL Classification: G11, G14, D83

working papers series