Identifying Beliefs from Asset Prices

Anisha Ghosh,Guillaume Roussellet

DOI: https://doi.org/10.2139/ssrn.3400005

2019-01-01

SSRN Electronic Journal

Abstract:This paper proposes a novel information-theoretic procedure to identify investors' beliefs about future macroeconomic and financial outcomes from observed asset prices. Our approach recovers price-consistent beliefs, i.e. the conditional distribution of macro and financial variables that satisfy the conditional Euler equations, given a cross-section of assets, a pricing kernel, and a conditioning set. Our procedure is non-parametric, not requiring any functional-form assumption about the dynamics of the variables, or regarding investor rationality or lack thereof. The price-consistent beliefs show strong cyclicality in the conditional mean and skewness of aggregate consumption growth, while the conditional volatility is mostly flat over the business cycle. This contrasts with the widely assumed conditionally normal dynamics in the existing literature. Looking at stock market returns, we observe that the price-consistent beliefs contain similar information as survey data on institutional investors' expectations. A comparison of these price-consistent beliefs with a non-parametric objective benchmark suggests large beliefs distortions. Investors underestimate the expected consumption growth mostly during recessions, but consistently overestimate the skewness of the consumption growth rate to a much larger extent.