Eventi del

Abstract:

I examine how macroeconomic uncertainty does relate to the market variance risk premium in a standard endowment economy, in which a single agent has recursive preferences and learns in real-time the structure of economic fundamentals. In the model, macroeconomic uncertainty is defined as the dispersion of the agent's belief about the expected growth rate of consumption. Real-time Bayesian learning relates to unobservable states and parameters, acknowledging uncertainty on both. I show that rare, large and relatively transitory spikes in macroeconomic uncertainty may help explain the unconditional moments of variance risk premium and its excess returns predictability. These shocks are located at the end of the 90s (LTCM/Russian Crisis), the beginning of 2000s (dot.com bubble, 9/11, financial scandals), and at the recent great financial crisis. I provide evidences that structural learning may help capture a relevant fraction of variance risk premia unconditonal moments even with standard CRRA preferences. Besides, the model also matches higher order moments of market excess returns.