Working Paper: NBER ID: w11468
Authors: John Y. Campbell; Samuel B. Thompson
Abstract: A number of variables are correlated with subsequent returns on the aggregate US stock market in the 20th Century. Some of these variables are stock market valuation ratios, others reflect patterns in corporate finance or the levels of short- and long-term interest rates. Amit Goyal and Ivo Welch (2004) have argued that in-sample correlations conceal a systematic failure of these variables out of sample: None are able to beat a simple forecast based on the historical average stock return. In this note we show that forecasting variables with significant forecasting power in-sample generally have a better out-of-sample performance than a forecast based on the historical average return, once sensible restrictions are imposed on the \nsigns of coefficients and return forecasts. The out-of-sample predictive power is small, but we find that it is economically meaningful. We also show that a variable is quite likely to have poor out-of-sample performance for an extended period of time even when the variable genuinely predicts returns with a stable coefficient.
Keywords: equity premium; forecasting; historical average
JEL Codes: G1
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| valuation ratios (G32) | stock returns (G12) |
| interest rates (E43) | stock returns (G12) |
| restrictions on coefficients (C51) | out-of-sample performance (C52) |
| historical average return (G17) | predictive models (C52) |
| predictor variables (C29) | stock returns (G12) |