An Empirical Research On R&D Spillover Effects On Productivity Based On The Perspective Of Different Spillover Source

Endogenous growth theory regards R&D spillovers as an engine to overcome diminishing returns to scale and maintain economic growth. A large number of literature carried out empirical research on R&D spillover effects on productivity using data of different levels from different countries such as time series data, cross-section data or panel data. However, systematicness was missing in pre-existing research on intrinsic mechanism of R&D spillovers, empirical study mostly focused on developed economies, empirical research on developing countries as the sample has not attracted enough attention. This article carries out systematical study on the depth and mechanism of R&D spillovers using the data from Chinese enterprise, industries and regions, specifically through the channels of intra-industry, inter-industy and space based on the perspective of spillover source.In theory, this article sorts out the extensive literature combined with selected research perspective, and we sum up two ways to R&D spillover effects on productivity:on the one hand, R&D spillovers have effects on productivity directly, on the other hand, R&D spillovers have effects on productivity through effects on factor inputs such as R&D. Based on the model of quality stage of Barro Sala-I-Martin (2004),the article describes mechanism on the R&D spillover effects on productivity from direct and indirect paths,considering the R&D spillovers.Through the comparative static analysis of R&D spillover steady state solution,we finds that R&D spillovers is the key variable to productivity; Determinant model of R&D effort shows that R&D spillovers is the important variable to affect R&D effort and that R&D spillovers promote R&D effort by reducing the cost of R&D.In the analysis of experience, firstly, the article examines not only manufacturing inter-industry, service inter-industry but also the inter-industry spillovers between manufacturing and service sectors in China based on the framework of mechanism analysis. Different from existing research which neglected considering heterogeneity of R&D spillovers, this article distinguishes between rent spillovers and pure knowledge spillovers.Unlike existing research with disadvatage of bias on using OLS to estimate to explain the endogenous variables, we accurately estimate inter-industry R&D spillover effects by FMOLS. The empirical results show that there exists inter-industry R&D spillovers in China, which is asymmetrical,that is to say there exists manufacturing inter-industry R&D spillover effects, service sector inter-industry R&D spillover effects and manufacturing to service sector interindustry R&D spillover effects, not existing interindustry R&D spillover effects on manufacturing from service sector. The analysis on inter-industry R&D spillovers impacting R&D effort shows that there exists the complementary effect between the inter-industry R&D spillovers in manufacturing and R&D effort in manufacturing, so does the service industry.Secondly, this article further investigates R&D spillover effects on productivity using micro data of14,227enterprises of Chinese manufacturing industry from the perspective of production cost, which is a new perspective different from previous study. To avoid the disadvantage caused by setting functional form in advance, we use the logarithm function model to accurately estimate them. The results show that intra-industry and inter-industry R&D spillovers reduce the cost of production and actively promote the R&D effort increased and have positive direct and indirect effects on productivity.Then, this article examines the effect of international inter-industry and intra-industry R&D spillovers on productivity based on the media of international trade in the open economic conditions. Different from the previous study with disadvantage of selecting relative data to measure indicators, we accurately estimate international intra-industry and inter-industry R&D spillover effect on productivity promotion through absolute data. We finds that our access to international R&D spillovers mainly comes from inter-industry R&D spillovers, the international intra-industry spillovers don’t have significant effects on TFP,TC and EC. The analysis on R&D spillovers impacting R&D effort shows that there exists the complementary effects between international interindustry R&D spillovers and R&D effort, there exists the substitution effects between international intraindustry R&D spillovers and R&D effort. International interindustry R&D spillovers have a positive direct and indirect effects on productivity. However, international intraindustry R&D spillovers have a negative direct and indirect effects on productivity.Further, considering spatial factor and modifying MRW model, this article constructs the spatial panel model and spatial simultaneous equations and estimates accurately the R&D spatial effects, which avoids partial defects because of using cross-section data and the omission of spatial variables, which is a new expansion to previous study. Results of spatial panel model show that R&D of spatial adjacent areas of region has positive spillover effects on economy growth, but spillover effects mainly come from R&D spillovers of enterprises. The results from Spatial simultaneous equations between productivity and R&D from Multinational companies in China show that R&D spillovers from Multinational companies by neighboring areas has positive spillover effects on productivity. Regional R&D spatial spillover, R&D spatial spillovers of Multinational companies actively promote own R&D effort. What’s more, R&D spatial spillovers have a positive direct and indirect spillover effects on productivity.Finally, according to the assumption that the absorption capacity has effects on R&D spillover, this article establishes dynamic threshold panel data model and firstly examines the threshold value of different R&D spillovers. Using some variables such as openness, market-oriented, R&D effort, industry growth,the article measures the integrated absorption capacity unlike most existing literature using a single variable such as human capital to measure. We finds that different elements play different role in R&D spillovers in model of different spillover sources and that the threshold in different spillover sources has certain disparity. The above conclusion provides a quantitative basis with choice of appropriate technology.