半参数变系数空间误差回归模型的估计

发布时间：2018-11-04 11:46

【摘要】：研究目标:克服半参数变系数回归模型中误差项可能存在的空间相关性问题。研究方法:提出一类新的半参数变系数空间误差回归模型,并构造其截面似然估计。研究发现:在小样本条件下,模型估计量具有良好的表现,其精度随着样本容量的增加而提高;应用该方法分析我国资源禀赋与地方公共品供给之间的相互关系,进一步证实了模型较强的适用性。研究创新:证明了估计量的一致性与渐近正态性,并通过蒙特卡洛模拟考察了估计方法的小样本表现。研究价值:新方法对于其他结构的半/非参数空间计量模型理论研究具有推广价值,其估计技术在经济、管理等学科中具有应用价值。
[Abstract]:Research objective: to overcome the possible spatial correlation problem of error term in semi-parametric variable coefficient regression model. Research methods: a new semi-parametric spatial error regression model with variable coefficients is proposed and its cross-sectional likelihood estimation is constructed. It is found that the model estimator has a good performance under the condition of small sample size, and its accuracy increases with the increase of sample size. This method is used to analyze the relationship between resource endowment and local public goods supply in China, which further proves the applicability of the model. Research innovation: the consistency and asymptotic normality of estimator are proved, and the small sample performance of estimation method is investigated by Monte Carlo simulation. Research value: the new method is worth popularizing the semi-parametric spatial econometric model theory of other structures, and its estimation technology has application value in economics, management and other disciplines.
【作者单位】：福建师范大学数学与计算机科学学院;厦门大学宏观经济研究中心;中国人民银行太原中心支行;
【基金】：国家社科基金项目“半参数变系数空间自回归模型的理论及应用研究”(16BTJ018) 教育部人文社会科学重点研究基地重大项目“集聚经济下的中国地方政府财税行为研究”(15JJD790029)和教育部人文社会科学项目“空间自回归单指数模型的理论和实践”(13YJA9100002)的资助
【分类号】：F224

【相似文献】