Research On Technology Efficiency Estimation And Influencing Factors Based On Spatial Stochastic Frontier Model

Stochastic frontier analysis,as a representative parameter method for technical efficiency measurement and total factor productivity calculation,was first proposed since 1977.Both theoretical and empirical research have developed rapidly,from the random frontier of cross-section data to the random frontier of panel data.As for the random frontier which considers the time-varying technology efficiency and the non-time-varying technology efficiency,a series of theoretical innovations and empirical application results on stochastic frontier analysis are constantly appearing.With the rise of regional economics,information and computational science,and new economic geography in recent years,scholars interested in economic growth have increasingly paid more attention to the spatial information contained in data.In the actual production process,it is impossible for different production units to be independent of each other.The widespread spatial effect between production units will cause various economic phenomena such as spillover effects and industrial agglomeration between production units.A stochastic frontier model is used.When the efficiency measurement is not taken into consideration,the parameter estimation of the model will be problematic.For example,if the parameter estimation does not satisfy the consistency,and the technical efficiency measurement will be biased.Therefore,some scholars introduced the spatial effect into the stochastic frontier model and constructed the spatial stochastic frontier model to deal with the spatial intersection effect in the economic activities,providing a new research perspective and analysis method for the analysis of modern economic activities.Based on the existing literature,this paper combines different forms of spatial econometric models with stochastic frontier models to build different forms of spatial stochastic frontier models,and derives the corresponding maximum likelihood functions and techniques based on the JLMS method.Efficiency calculation method.In terms of empirical application,based on data from 30 provinces and cities in China from 2000 to 2017,this paper measured the regional technical efficiency in China and compared it with the technical efficiency value obtained by the ordinary stochastic frontier model.The spatial correlation tests based on the spatial adjacency matrix and geographic distance matrix both show that there is a positive spatial correlation in annual GDP between 2003 and 2017,and the positive spatial correlation is weakened.The SARSFA model based on cross-section data shows that both the spatial autoregressive coefficient ? and the error autocorrelation coefficient ? are highly significant,which indicates that the spatial stochastic frontier model is superior to the ordinary stochastic frontier model,and there are positive spatial effects between provinces,but in 2016 China There is a negative spatial correlation between province output.The results of the SSFA model based on cross-section data show that the model reduces the effect of spatial effects on the model by introducing spatial effects in the technical invalid term,making the estimation of the model's random error term and technical invalid term closer to reality.The coefficients of the model show that capital investment still occupies a dominant position in the growth of China's economic aggregate.In terms of technical efficiency,the technological efficiency of all provinces and cities is between 0.45 and 0.98.The high-tech efficiency level area and the high-tech efficiency level area show adjacent or set phenomena,which also confirms the conclusion of the previous spatial correlation test.The analysis of the influencing factors of technical efficiency shows that the degree of economic outgoing has a significant positive impact.The structural adjustment of the secondary and tertiary industries also has a positive impact on technical efficiency,and the investment in science and technology has a significant negative impact on technical efficiency.Convergence analysis shows that the overall technical efficiency shows "overall divergence,partial convergence",there is a conditional ? convergence,and there is absolute ? convergence in the eastern region.