On the Relationship between the Conditional Mean and Volatility of Stock Returns: A Latent VAR Approach

Abstract

We model the conditional mean and volatility of stock returns as a latent vector autoregressive (VAR) process to study the contemporaneous and intertemporal relationship between expected returns and risk in a flexible statistical framework and without relying on exogenous predictors. We find a strong and robust negative correlation between the innovations to the conditional moments that leads to pronounced counter-cyclical variation in the Sharpe ratio. We document significant lead-lag correlations between the conditional moments that also appear related to business cycles. Finally, we show that although the conditional correlation between the mean and volatility is negative, the unconditional correlation is positive due to the lead-lag correlations.