Measuring Risk Appetite from Financial Assets' Excess Returns

Abstract

We introduce a new meaure of risk appetite in financial markets, based on the cross sectional behavior of excess returns. Turning them into probabilities through a Markov Switching model, we define one global risk appetite measure as the cross-sectional average of the individual probabilities for each asset to be in a "risk appetite" regime. Given the probabilistic approach that comes naturally with this Markov Switching framework, we present various tests to gauge the interest of the risk appetite measure that is presented here. Using these tests we show that our index behaves well vs. various competitors, especially in out-of-sample results. We test for the information content of various assets and find that a core of asset allocation-related assets provide the best possible choice over various competing specifications.