Testing Whether Investor Sentiment Predicts Stock Anomalies Spuriously

Summary

The document reviews research on whether investor sentiment helps explain stock return anomalies or appears predictive because of spurious regression. Persistent predictors can seem to forecast returns in finite samples when expected returns also vary over time, so the authors compare sentiment’s results across multiple anomaly portfolios with results from randomly generated persistent predictors. The analysis draws on 11 anomalies and tests three proposed relationships involving long-short returns, short-side returns, and long-side returns.

The document reports that researchers generated 200 million autoregressive predictor series and re-estimated 36 regressions. No simulated series matched sentiment’s combined support for all three hypotheses; matching the first two together was also rare. This cross-anomaly consistency is presented as evidence against a chance explanation. The exercise does not prove that sentiment causes mispricing or guarantee that the findings generalize beyond the studied portfolios, samples, and simulation assumptions. The article also emphasizes that the proposed effect is concentrated on anomaly portfolios’ short side, where short-sale constraints may matter.

Key ideas

• Persistent predictors can appear to forecast returns spuriously when expected returns vary over time.
• Testing a predictor across several anomaly portfolios makes chance consistency harder to obtain.
• The study compares investor sentiment with 200 million simulated autoregressive predictors across 36 regressions.
• The reported evidence is strongest for sentiment’s association with anomaly short-side returns.
• Simulation results challenge a chance explanation but do not establish causality or universal validity.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.