Evaluation Of R&D Input-output Efficiency Based On Improved Bootstrap-DEA Model

Research and Experimental Development(R&D)is a creative scientific and technological activity in the field of science and technology in order to promote technological progress,increase knowledge theory and promote scientific and technological development.How to measure and evaluate the R&D input-output efficiency of a country or region is very important to improve the country’s scientific and technological innovation ability.Data envelopment analysis(DEA)is the most common nonparametric measurement method of R&D input-output efficiency.In the field of DEA research,the bootstrap DEA model and Malmquist total factor productivity index have attracted extensive attention of scholars.However,the traditional bootstrap DEA model can not reflect the independent changes of input-output variables among different decision-making units,and usually converges prematurely,resulting in invalid subsequent iterations.The calculation of Malmquist total factor productivity index is easy to be affected by exogenous environmental variables and random factors.Therefore,Based on the evaluation of R&D input-output efficiency,it is of great theoretical and practical significance to improve the bootstrap DEA method and calculate the Malmquist total factor productivity index.In the process of discussing the R&D input-output efficiency evaluation based on the improved bootstrap DEA model,the following work is carried out: first,carry out literature research,and put forward the problems to be solved and research ideas.Secondly,it introduces data envelopment analysis and its related concepts,as well as two classical DEA models.Third,explore the Composite DEA model combining DEA model with other models,including DEA model based on Tobit regression,bootstrap DEA model before and after improvement,and common DEA Malmquist productivity index model.Fourth,simulation and empirical analysis based on the above theory show that the R&D investment intensity has the most obvious influence among the five environmental variables.Increasing the R&D investment intensity can reduce the full-time equivalent of researchers in experimental development,applied research and basic research,and the relaxation of input variables such as external expenditure and internal expenditure of basic research;The improved bootstrap DEA model can effectively alleviate the data multicollinearity,and the variance expansion factor of the adjusted variables is significantly reduced.Compared with the results of the traditional bootstrap DEA model,the efficiency calculated by the improved bootstrap DEA model is closer to the initial DEA efficiency without resampling;From the perspective of Malmquist index decomposition,in 2019,only two of China’s total factor productivity decline areas were caused by the decline of technical efficiency and the slowdown of technological progress at the same time,and the rest were caused only by the slowdown of technological progress.Technological progress will be our first concern in the next few years.The value and significance of this study includes: first,an improved bootstrap DEA model is proposed.In order to solve the shortcomings of the traditional bootstrap DEA model(it can not reflect the independent changes of input-output variables between different decision-making units,and it is easy to converge prematurely,resulting in invalid subsequent iterations),a new idea and feasible method are proposed,which makes up for the defect that the traditional bootstrap DEA model is vulnerable to the influence of data multicollinearity,The applicable scenario of bootstrap DEA model is extended to make the nature of the model more superior.Secondly,it puts forward a new idea for the problem that Malmquist index is vulnerable to environmental variables and random shocks,which is of great significance for accurately calculating the real total factor productivity index and its decomposition results excluding external interference factors.