空间动态短面板的估计及中国城市经济增长收敛性的分析

Estimation of Short Spatial Dynamic Panel Data Models: With Application to Growth Convergence Among Chinese Cities

本文研究了在样本时期长度T不大的情况下,如何对空间动态面板模型进行估计和统计推断。通过对初始值的设定,本文研究了在固定效应和随机效应模型设定下如何进行准最大似然估计。通过使用该计量方法对中国城市增长收敛性进行研究,我们发现中国城市经济增长呈条件收敛,且外溢效应显著。相较于动态面板模型下的估计,利用空间动态面板模型考虑外溢效应后,城市经济增长收敛速度显著提高。

This paper investigates the estimation and statistical inference of spatial dynamic panel data model when sample period T is short.By specifying initial conditions,we explore the quasi-maximum likelihood(QML)estimation under both fixed effects and random effects specifications.We then investigate the growth convergence among Chinese cities using this econometric technique.We find that the spillover effect is significant factors.Comparing to estimation using dynamic panel data model,the convergence rate is much higher after allowing the spillover effect under the spatial dynamic panel data model.