Stock return distribution and predictability: Evidence from over a century of daily data on the DJIA index

This paper analyses the predictive power of the DJIA index returns, measured at different quantiles of its distribution, for future return distribution. The returns measured at quantile 0.75 have predictive power for most quantiles of future returns, except for their median. This result prevails after controlling for the predictive power of the lagged first four moments of returns and of other economic predictors used in the literature. Furthermore, this finding is stable over time. Forecasts of future mean returns based on predicted return quantiles have positive economic value, as do forecasts of future volatility, the latter especially for investors with low risk aversion. The predictive power of quantile 0.75 DJIA returns is shown to be the result of their ability to forecast shocks to future investment and consumption.