Estimation of structural nonlinear errors-in-variables models by simulatedleast-squares method

This article proposes a simulation approach to obtain least-squares or gene ralized least-squares estimators of structural nonlinear errors-in-variable s models. The proposed estimators are computationally attractive because th ey do not need numerical integration nor huge numbers of simulations per ob servable. In addition, the asymptotic covariance matrix of the estimator ha s a simple decomposition that may be used to guide selection of appropriate simulation sizes. The method is also useful for models with missing data o r imperfect surrogate covariates, where application of conventional least-s quares and maximum-likelihood methods is restricted by numerical multidimen sional integrations.