外商直接投资对中国的溢出效应:基于Meta回归分析方法的再分析

The Spillover Effect of FDI in China:A Re-analysis Based on Meta Regression Analysis

本文应用Meta回归分析方法研究了国内关于外商直接投资溢出效应的32篇论文,结果表明,若实证研究选择面板数据作为样本,会比截面数据或时间序列数据得出更高的溢出效应;若选择省际层面或行业层面的数据,则比企业层面的微观数据估计的溢出效应更高;若样本数据涉及的年限越长,估计出的溢出效应也越大;普通最小二乘法比其他的估计方法(如固定效应、随机效应、动态化模型)得到的溢出效应更高;若实证研究采用工具变量法控制了外商直接投资的内生性问题,那么比没有考虑内生性得出的溢出效应更低。

This paper applies Meta regression to analyze 32 domestic papers on the spillover effect of FDI,and it concludes that if the empirical research chooses panel data as its sample,it will get higher spillover effect than cross - section data or time series data; if it uses the province level or industry level data,the spillover effect could be higher than the firm level micro data; the longer is the year span of the sample,the greater spillover effect could be estimated; OLS estimates higher spillover effect than other methods, such as fixed effect,random effect,dynamic model; if the empirical research controls the endogeneity problem of FDI by instrument variables, then it would tend to reDort lower sDillover effects than those without controlling endogenity of FDI.