Estimating Intertemporal Quadratic Adjustment Cost Models with Integrated Series.

The authors consider the estimation of parameters in Euler equations where regressand and regressors may be nonstationary, and propose a several-stage procedure requiring only knowledge of the Euler equation and the order of integration of the data. This procedure uses the information gained from pretesting for the order of integration of data series to improve specification and estimation. The authors can also offer an explanation of the frequent empirical finding that discount rates and adjustment costs are poorly estimated. Both analytical and experimental (Monte Carlo) results are provided. Copyright 1991 by Economics Department of the University of Pennsylvania and the Osaka University Institute of Social and Economic Research Association.