Scholarlink

<p>This paper introduces a novel random effects semiparametric additive spatial autoregressive panel model with spatiotemporal correlated errors. By using local linearization method to approximate additive components and establishing linear and quadratic moments, generalized method of moments (GMM) estimators of unknown parameters and additive components are proposed. Under some weak conditions, we prove that GMM estimators are consistent and asymptotically normal. Under normality of individual random effects and remaining errors, we derive the asymptotically efficient best GMM (BGMM) estimators by constructing the best instrumental variables. Monte Carlo simulation results show that both GMM and BGMM estimators have good small sample performance, and BGMM estimators have better estimation accuracy than GMM estimators. The empirical analysis shows that the proposed model and estimation approach have broad application prospects.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

GMM estimation of random effects semiparametric additive SAR panel model with spatiotemporal correlated errors

  • 期刊论文

  • 发表日期：

    2025年3月26日

查看全文
  • Jianbao Chen,
  • Bogui Li,
  • Shuangshuang Li

摘要

This paper introduces a novel random effects semiparametric additive spatial autoregressive panel model with spatiotemporal correlated errors. By using local linearization method to approximate additive components and establishing linear and quadratic moments, generalized method of moments (GMM) estimators of unknown parameters and additive components are proposed. Under some weak conditions, we prove that GMM estimators are consistent and asymptotically normal. Under normality of individual random effects and remaining errors, we derive the asymptotically efficient best GMM (BGMM) estimators by constructing the best instrumental variables. Monte Carlo simulation results show that both GMM and BGMM estimators have good small sample performance, and BGMM estimators have better estimation accuracy than GMM estimators. The empirical analysis shows that the proposed model and estimation approach have broad application prospects.