中文摘要：

在模型的部分协变量为内生性协变量的情况下，考虑广义变系数模型的一类估计问题．通过结合基函数逼近和一些辅助变量信息，提出了一个基于工具变量的估计过程．并得到了估计的相合性和收敛速度等渐近性质．所提出的估计方法可以有效地消除协变量的内生性对估计精度的影响，并且具有较好的有限样本性质．

英文摘要：

This paper considers the estimation for a class of generalized varying coefficient models with endogenous covariates. By combining basis function approximations with some auxiliary variables, an instrumental variable based estimation procedure is proposed. The asymptotic results of the estimator such as the consistency and the convergence rate are obtained. The proposed procedure can attenuate the effect of endogeneity in covariates, and has proved workable for finite samples.