Please rotate your device to landscape mode to view the charts.

Anomalies and the Expected Market Return

Journal: Journal of Finance

Date: 20220201

Author: Dong, Xi; Li, Yan; Rapach, David E.; Zhou, Guofu

Abstract:
We provide the first systematic evidence on the link between long-short anomaly portfolio returns--a cornerstone of the cross-sectional literature--and the time-series predictability of the aggregate market excess return. Using 100 representative anomalies from the literature, we employ a variety of shrinkage techniques (including machine learning, forecast combination, and dimension reduction) to efficiently extract predictive signals in a high-dimensional setting. We find that long-short anomaly portfolio returns evince statistically and economically significant out-of-sample predictive ability for the market excess return. The predictive ability of anomaly portfolio returns appears to stem from asymmetric limits of arbitrage and overpricing correction persistence.

Link: Google Scholar

Key Findings

Predictive Power of Anomalies

Long-short anomaly portfolio returns demonstrate significant out-of-sample predictive ability for market excess returns when using appropriate forecasting strategies that guard against overfitting

Asymmetric Market Correction

Evidence shows stronger mispricing correction persistence for overpricing compared to underpricing, consistent with asymmetric limits of arbitrage

Economic Value

Strategies using anomaly portfolio returns generate substantial economic gains, with annualized utility gains of 259-638 basis points for investors

Out-of-Sample Forecast Performance

  • The ENet, C-ENet, and PLS strategies show the strongest predictive performance with R² statistics above 2%
  • All strategies designed to guard against overfitting outperform the conventional OLS approach
  • Simple combination strategy shows most conservative but still significant improvements

Market Friction Effects

  • Predictive ability increases significantly during periods of high market frictions
  • Short fees and liquidity innovations show strongest impact on predictability
  • Effect is consistent across different friction measures

Trading Position Changes

  • Increase in long-short anomaly returns leads to significant changes in arbitrageur positions
  • Short positions show stronger response than long positions
  • Changes in net positions confirm asymmetric response to mispricing signals

Contribution and Implications

  • First systematic evidence linking cross-sectional anomalies to time-series market return predictability
  • Demonstrates importance of appropriate statistical techniques for extracting predictive signals from large sets of anomalies
  • Provides new insights into the role of market frictions and limits of arbitrage in price discovery
  • Offers practical strategies for investors to improve market timing using anomaly information

Data Sources

  • Out-of-Sample Forecast Performance chart based on Table II R²OS statistics for different forecasting strategies
  • Market Friction Effects visualization derived from Table IV showing R²OS statistic differences between high and low friction regimes
  • Trading Position Changes chart constructed using standardized coefficient estimates from Table V