Efficient estimation of linear asset-pricing models with moving average errors

This article studies alternative methods for estimating parameters from multiperiod conditional moment restrictions. Our discussion is couched in the context of a multivariate linear time series model, and we use the log-linear intertemporal asset-pricing model as a prototype when comparing alternative econometric methods. We propose a generalized method of moments estimator that is scale invariant and is asymptotically equivalent to one used previously in empirical work on asset pricing. We then show how to improve the efficiency of this estimator. Finally, we apply these methods in an empirical investigation of the log-linear intertemporal asset-pricing model.