Is there a best measure of investor sentiment for predicting stock market returns? In his March 2016 paper entitled “Investor Sentiment and Stock Market Returns”, Lee Smales updates relationships between stock market/portfolio returns and five sentiment measures:
- CBOE Implied Volatility Index (VIX).
- Baker-Wurgler composite sentiment index (readily available only through 2012).
- American Association of Individual Investors (AAII) investor sentiment.
- University of Michigan Consumer Sentiment Index.
- Commitment of Traders (COT) reports from the Commodity Futures Trading Commission.
He controls for multiple economic and financial variables likely to be related to stock market returns (gross domestic product, industrial production, unemployment rate, consumer price index, Federal Funds target rate, term spread, credit spread and dividend yield). He also investigates economic recessions separately. Principal tests relate sentiment levels and changes in sentiment levels to S&P 500 Index and style/industry portfolio returns (from Kenneth French’s data library) at horizons of 1, 3, 6 and 12 months. Using monthly values of sentiment measures as available and monthly index/portfolio returns during January 1990 through December 2015, he finds that:
- Based on causality analyses:
- The Baker-Wurgler and COT report measures depend on lagged S&P 500 Index returns, and returns depend on lagged values of these sentiments.
- Stock market returns lead the consumer sentiment index.
- VIX leads stock market returns (for the broad market, industries and style portfolios).
- Based on predictive regressions:
- Only VIX and the Baker-Wurgler index exhibit predictive power for stock market returns, and this power persists beyond one month only for the latter.
- Results for changes in these two sentiment measures are similar.
- A one standard deviation increase in VIX (4.24) predicts a 0.70% next-month decline in the S&P 500 Index.
- Sentiment effects are stronger:
- For hard-to-value stocks (small, growth, technology and telecoms).
- During NBER recessions.
In summary, evidence suggests that level of VIX/change in VIX may be useful for predicting stock market/portfolio returns at short horizons.
Cautions regarding findings include:
- Analyses are in-sample, employing data from the entire sample period to discover predictive relationships. An investor operating in real time may have drawn different conclusions at different times.
- Regression tests assume linear relationships between sentiment measures and stock market returns. These relationships may exhibit material tail effects.
- The study tests no market timing strategies based on findings. Out-of-sample strategies may not be significantly profitable.
