Using conditional moments of asset payoffs to infer the volatility of intertemporal marginal rates of substitution (Q921792)

A strategy for efficient utilization of conditioning information is developed. The strategy is implemented empirically by the seminonparametric methodology [\textit{A. Gallant} and \textit{G. Tauchen}, Econometrica 57, No.5, 1091-1120 (1989; Zbl 0679.62096)] to estimate the conditional distribution of a vector of monthly asset payoffs. Conditional moments of asset payoffs are used to deduce volatility bounds on the intertemporal marginal rates of substitution (IMRS) implied by asset market data. The authors use fitted conditional distributions to calculate both conditional and unconditional standard deviation bounds for the IMRS, conditional on information available to economic agents. Empirical estimations of the conditional densities and the volatility bounds are reported, underlying data with a long time series (1926-1987) of 744 monthly observations on the ex post real returns on stocks and T-bills, and a time series (1959-1984) on ex post real returns on the same two assets together with consumption data. The authors consider conditioning as a better control for the impact of outlier events on the moments of the asset payoffs, and as a substitute of splitting a sample into subsamples.