The time-varying risk–return trade-off and its explanatory and predictive factors

We analyze the intertemporal dimension of the risk–return trade-off and determine the drivers that better explain and predict its evolution. To this end, we propose a novel estimate of the relationship between return and risk where we generate time variation in the trade-off without conditioning the estimates to any state variable. We compare this dynamic approach with time-invariant or state-dependent estimates and observe that our dynamic method reasonably aligns with the constant (state dependent) methods but it offers a much broader picture of the risk–return trade-off. We also link its evolution to a set of macroeconomic, systematic and sentiment or uncertainty risk factors. We find that the risk–return relationship is positive during expansionary periods but it decreases during recessionary periods where occasionally even turns out negative. Our main conclusions hold for the consideration of hedging components, different MV-GARCH models or window lengths and several proxies of market returns and risk.