Threshold Regression with Errors-in-Variables
提出一种新的广义矩估计方法，用于处理解释变量存在测量误差的线性阈值回归模型，无需外部工具变量，并通过蒙特卡洛模拟和公司投资方程应用验证了方法的有效性。
We propose a novel generalized method of moments (GMM) estimator for linear threshold regression models with a mismeasured regressor and an exogenous threshold variable. The estimator addresses a key challenge in empirical applications: identifying model parameters in the absence of valid external instruments for the mismeasured regressor. We establish the consistency and asymptotic normality of the proposed estimator and develop test statistics for detecting threshold effects. Monte Carlo simulations show strong finite-sample performance, and an empirical application to firm investment equations highlights the practical relevance of the method.