Mood Beta and Seasonalities in Stock Return.

Hirshleifer, D., Jiang, D. and Meng, Y., 2018. Mood betas and seasonalities in stock returns (No. w24676). National Bureau of Economic Research.

Research Question and General Results

This paper proposes an explanation for aggregate and cross-sectional return seasonalities based on investor mood. Authors find that historical seasonal return of a security relative to other counterparts can predict its future seasonal patterns. They also propose a mood beta that measures a security’s sensitivity to mood variations and strongly forecasted seasonal stock return in future periods.

Hypothesis Development

Prior studies in psychology and behavioral finance suggest that investors are susceptible to seasonal affective disorders and will possess different but predictable moods in different periods. For instance, investor mood will be improving in January and March, and deteriorating in September and October. Similarly, investing population have a downbeat mood on Monday and an upbeat mood on Friday.

Based on the evidence of mood variations, this paper hypothesizes that seasonal mood variations lead to seasonal factor mispricing because investors are usually optimistic or pessimistic about future expected payoffs. Specifically, investors’ seasonal mood shifts render a factor to get overpriced or underpriced. Assets sensitive to investor moods will earn higher or lower return, respectively.

Thus, this paper proposes two main hypotheses:

Empirical Results

Mood Recurrence and Reversal Effects:

Consistent with the first hypothesis, authors find that securities’ return in current mood periods is positively associated with that of the prespecified mood and realized mood periods during congruent-mood periods and have inverse relationships during noncongruent-mood periods. The result is robust to both monthly and weekday levels. For instance, stock returns in current Januarys and Marchs have a positive relationship with stock returns in historical Januarys and Marchs, and a negative relationship with stock returns in historical Septembers and Octobers.

Mood Beta:

Authors estimate the mood beta by regressing each asset’s historical excess returns earned during prespecified and realized mood periods on contemporaneous CRSP excess return. It measures how an asset’s average return change responds to aggregate return change in identified mood periods. Consistent with the second hypothesis, mood beta can forecast stock return’s seasonalities. Specifically, stocks with high mood beta perform better in high mood periods (January, March or Friday) and worse in low mood periods (September, October and Monday). Besides, authors suggest that mood beta provides additional explanatory power other than market beta and sentiment beta.