Identifying the elasticity of substitution with biased technical change: a structural panel GMM estimator
提出一种结构面板GMM估计量，用于估计资本与劳动间的替代弹性，无需外部工具变量，且适用于有偏技术变化。蒙特卡洛研究表明该估计量在时间期数不太小时几乎无偏，应用于挪威制造业企业发现替代弹性为1.9，而固定效应估计为1.0，忽略联立性可能导致误判要素为互补或柯布-道格拉斯技术。
Summary: This paper provides a structural panel GMM (P-GMM) estimator of the elasticity of substitution between capital and labour that does not depend on external instruments, and which can be applied in the presence of biased technical change. We identify the conditions under which P-GMM is a consistent estimator and compare it to a fixed effects estimator. Using a Monte Carlo study, we find that the P-GMM estimator is nearly unbiased provided the number of time periods (T) is not too small. We show analytically how the small-T bias is related to metrics of weak identification. In an application on manufacturing firms in Norway, we estimate the elasticity of substitution to be 1.9 using the P-GMM and 1.0 using the fixed effects estimator. Neglecting simultaneity may thus lead to the conclusion that capital and labour are complements or can be described by Cobb–Douglas technology, when, in fact, they are substitutes.