Pricing Intertemporal Risk When Investment Opportunities Are Unobservable

Abstract

The intertemporal capital asset pricing model (ICAPM) predicts that an unobservable factor capturing changes in expected market returns should be priced in the cross section. My Bayesian framework accounts for uncertainty in the intertemporal risk factor and gauges the effects of prior information about investment opportunities on model inferences. Whereas an uninformative prior specification produces weak evidence that intertemporal risk is priced, incorporating prior information about market-return predictability generates a large space of ex ante reasonable priors in which the estimated intertemporal risk factor is positively priced. Overall, the cross-sectional tests reject the capital asset pricing model (CAPM) and indicate support for the ICAPM.