Risk, Uncertainty, and Asset-Pricing Puzzles

In conventional rational expectations expositions of the “equity premium puzzle,” “riskfree rate puzzle,” and “variability mismatch puzzle,”the subjective distribution of future growth rates is essentially made to mimic its past sample moments. This paper shows that the unobservable nature of structural growth parameters adds to expectation beliefs a permanent thick-tailed background layer of uncertainty that never converges to a stationary-ergodic rational expectations equilibrium, yet which explains the three ‘puzzles’parsimoniously by one unifying principle. The posterior data generating process of a Bayesian evolutionary-learning equilibrium with a vague prior is consistent with all three values of the ‘puzzles’observed as stylized facts. 1 Introduction: Structural Uncertainty and Asset Prices The “equity premium puzzle”refers to the spectacular failure of the standard neoclassical representative-agent model of stochastic general-equilibrium growth to explain a historical di¤erence of some six or so percentage points between the average return to a representative stock market portfolio and the average return from a representative portfolio of relatively safe stores of value. Such a large premium on risk cum uncertainty suggests either that people are perceiving more uncertainty about the future than past data would at …rst glance appear to indicate, or else that something is fundamentally wrong with the standard formulation of the (e-address: [email protected]) For helpful detailed comments on earlier drafts of this paper, but without blaming them for its remaining defects, I am grateful to Andrew Abel, Gary Chamberlain, Xavier Gabaix, Alfred Galichon, Jerry Hausman, Joseph Kadane, Angelo Melino, Jonathan Parker, Je¤ Strnad, and, especially, Evan Anderson, John Geweke, and Narayana Kocherlakota.