国际性无形技术外溢与中国全要素生产率——基于分位数回归的实证研究

Direct International R&D Spillover and Total Factor Productivity in China: Empirical Research Based on Quantile Regression

考虑国际性技术外溢中技术外溢方和吸收方互动因素的基础上,本文测算了我国各省吸收同化的无形国际性外溢R&D存量,将其代入Coe和Helpman推导的内生增长模型中,利用我国各省的面板数据估计了自主R&D存量和无形国际性外溢R&D存量对全要素生产率的弹性系数,同时,应用了分位数回归分析了无形国际性技术外溢对TFP的贡献随着各省市TFP的分位点的变化而变化的趋势。研究结果表明,自主R&D投资显著促进全要素生产率的增长,是我国各省技术进步的主要推动力;无形的国际性技术外溢同样促进了各省的全要素生产率的提高,但其作用要小于自主R&D投资对各省全要素生产率的推动作用。分位数回归结果进一步表明,随着各省市全要素生产率的提高,无形的国际性技术外溢对各省全要素生产率的促进作用逐步降低。

In addition domestic RD investment,China has been relying on international RD spillovers from the developed country to grow its economy.International spillovers consist of international trade,foreign direct investment（FDI）,and direct（intangible） spillovers.Intangible spillovers include reverse engineering,students returning from foreign countries,dialog between academic researchers,information exchange through electronic means,academic journals,and international conferences,and so on.Researchers begin paying attention to direct RD spillovers in addition to trade and FDI. Based on Coe and Helpman＇s（1995） proposed model,we evaluate the contribution of domestic and direct foreign RD stocks to the total factor productivity（TFP） of different provinces in China.First evaluation result shows that the elasticity of TFP to the domestic RD stock is between 0.34 and 0.47.This finding implies that domestic RD investment is the major source of technological improvement in China.Second evaluation result shows that the elasticity of TFP to the direct foreign RD stock is 0.06.This finding implies that direct foreign RD spillovers have significant impact on TFP.However,its impact on TFP is weaker than that of domestic RD investment.Third evaluation result shows that the domestic and direct foreign RD stocks are complementary for TFP growth. The previous research primarily analyzes the effect of technology spillovers on OLS method.These studies have the major limitation that the impact of independent variables（IVs） on conditional dependent variable（DV） at different locations is assumed to be constant.According to the assumption,the mean estimate of OLS cannot reflect the differential impact of IVs on the conditional DV at difference locations. Is the impact of direct foreign RD spillovers on the TFP in thirty provinces constant even though so much provinces heterogeneity exists？ The quantile regression helps to answer this question.The results show that when TTP＇s quantile increases,the elasticity of TFP to direct foreign RD stocks decreases dramatically from 0.14 and 0.02.This indicates that the contribution of direct foreign RD spillovers to the TFP decreases while the TFP rises.There are two possible reasons for this negative correlation.First,imitative innovation is more difficult in provinces with higher TFP than in those with lower TFP. Second,direct international RD spillovers have some limitations.Different from technology imports and patents purchases,direct technology spillovers is not effective for developing countries,such as China,to acquire critical technologies.This finding indicates that non-critical and general technological knowledge obtained through direct technology spillovers has higher effect on the growth of TFP in provinces with lower TFP than that in provinces with higher TFP.