Where can investors find the best stock market forecasters: academia, banks, government? In the March 2012 draft of his paper entitled “On the Forecasting Quality of Professionals”, Aron Veress compares the stock market forecasting accuracies of different professional groups (academics, commercial bankers, investment bankers, government employees and non-financial professionals) who participate in the semi-annual Livingston Survey, both to each other and to quantitative predictors. He focuses on forecasts of the S&P Composite Index return at horizons of roughly one month and six months after publication of survey reports (early June/December of each year). He considers also a naive forecasting approach that invests in stocks or cash according to which has the higher preceding return. Using results from 120 semi-annual surveys and contemporaneous data for the S&P Composite Index and commonly used financial/economic U.S. stock market predictors, he finds that:
- For six-month forecasts:
- Academics and commercial bankers perform equal to or better than the survey average, while investment bankers significantly underperform.
- Compared to survey results, the naive approach based on immediately preceding return performs better (worse) regarding direction (magnitude) of future returns.
- For one-month forecasts:
- Academics are again above average, while commercial (investment) bankers drop below (rise above) average. Governmental employees perform relatively well.
- Again, compared to survey results, the naive approach performs notably better (worse) regarding direction (magnitude) of future returns.
- Investment bankers (government employees) tend to be optimistic (pessimistic), and academics the most consistent. Forecast accuracy generally degrades during economic contractions.
- Overall, bankers mostly underperform other professions in predicting U.S. stock market returns, while academicians outperform. No group convincingly outperforms the naive approach.
- Out-of-sample stock market forecasts derived from linear regressions of future returns against dividend yield, earnings-price ratio, inflation rate, change in 3-month Treasury bill yield and NBER recessions are similar to those from the Livingston surveys.
- With regard to influences:
- Commercial (investment) bankers appear most influenced by dividend yield (earnings-price ratio).
- Academics appear most influenced by the inflation rate.
- In aggregate, survey participants seem not to believe in long-run mean reversion of stock market returns.
In summary, evidence from the Livingston survey does not support belief that either qualitative stock market forecasts or those generated by simple indicators beat a naive investing approach.
Cautions regarding findings include:
- As noted in the paper, dispersion in survey response dates confounds analysis.
- Also as noted in the paper, group sample sizes are small, inhibiting confidence in findings.
For other studies of Livingston Survey forecasts, see “Professional Economists Forecasting Stock Returns” and “Should Equity Investors Hope for Good or Bad Economic Forecasts?”.
