Time-varying stock return correlation, news shocks, and business cycles

Abstract

The cross-sectional average of the pairwise correlations between U.S. stock returns is considered as a measure of risk to aggregate wealth priced by the stock market. We show that this measure predicts future U.S. output growth at a horizon of one to four years. A stronger average correlation of stock returns foreshadows significantly lower future output growth, even when controlling for some other widely used financial predictors. An innovation to average correlation gives rise to macroeconomic dynamics that resemble negative news about future total factor productivity (TFP) in a vector autoregression. TFP news shocks thus appear to be a key source of aggregate risk priced into stocks.