Asset Pricing with Unforeseen Contingencies

Data suggest that macro asset pricing models attributing risk premia solely to covariance with aggregate consumption shocks generally perform poorly. Yet, there is no consensus over what alternative primitive pricing factors should accompany or take the place of aggregate consumption. We assume an economy where agents are like economists: they know that factors other than aggregate consumption shocks have welfare implication, but they can’t specifically and explicitly state what these factors are. Instead, agents attribute the additional source of uncertainty to ‘unforeseen contingencies’ whose behavioral impact is observationally equivalent to individual taste shocks. We investigate an economy of heterogeneous agents that cannot specify all exogenous welfarerelevant events, and characterize the appropriate equilibrium concept when securities can only trade on demandand price-contingent events. We establish the existence of an equilibrium for a class of parametric models in which aggregating unforeseen contingencies across agents can lead to non-consumption pricing factors. To fit the stylized facts, (i) non-consumption factors must dominate the pricing kernel and contribute to the variation of the wealth-consumption ratio, (ii) markets must be incomplete and the set of claims that are traded endogenously determined, (iii) agents’ preferences with respect to unforeseen contingencies must be non-expected utility, and (iv) although non-consumption pricing factors can be conditionally uncorrelated with aggregate consumption shocks, they must be correlated with shocks to expected consumption growth. JEL Classification: D51, D52, D58, G12, G13 ∗We are grateful to Jonathan Berk, Murray Carlson, Pierre Colin-Dufresne, Greg Duffee, Ilan Kremer and Martin Lettau for helpful comments and suggestions. Address all correspondence to Jacob S. Sagi, Haas School of Business, University of California at Berkeley, 545 Student Services Building, Berkeley, California, 94720-1900. email: [email protected], phone: (510)-642-3442