Discussion And Improvement Of Analysis Methods Of Returns-to-Scale In DEA

Returns-to-Scale analysis is a very important research field of data envelopment analysis.At present,this field has proposed six methods to analyze returns-to-scale.These methods can correctly determine the returns-to-scale status of effective units,but often gain inconsistent conclusions for non-effective unit calculations,which is due to different oriented models and different projection methods used in the above methods.Therefore,the key theoretical problems that need to be solved urgently are which oriented models to choose and which projection methods to adopt in this field.This paper will solve the mentioned problems derived from various oriented models and multiple projection methods in order that different methods can obtain consistent calculation results.In addition,this paper will use the existing method of returns-to-scale analysis to conduct in-depth analysis and evaluation of energy use efficiency in Anhui Province.The main tasks are as follows:Firstly,as for the situation that many returns-to-scale analysis methods can adopt in both intput-oriented forms and output-oriented forms,this paper takes the Z&S-scale efficiency index method and the reference set method as examples,and demonstrates that the analysis procedure should be based on output-oriented models from three aspects of the comparison of theoretical results,theoretical analysis and economic interpretation.Secondly,as for the situation that many returns-to-scale analysis methods use multiple projection methods to determine the returns-to-scale status of non-effective units,this paper takes the GDEA-WYL method and the BT-ZSP improvement method as examples,and based on the comparison of theoretical results and theoretical analysis,it argues that the analysis procedure should use a radial projection method that keeps the input constant,i.e.transforms non-effective units into weak effective units instead of effective units.Finally,the paper discusses the implementation of scale efficiency decomposition methods and other methods,and then applies the improvement of scale efficiency decomposition method to the empirical study of energy utilization efficiency in 16 cities in Anhui Province,which provides scientific decision-making advice for the rational use of energy in Anhui Province.The above tasks not only solve the problem of selecting models' orientations and projection methods,but also obtains important conclusions such as the input-oriented BCC model and the output-oriented BCC model are not equivalent,and the input-oriented CCR model and the output-oriented CCR model is equivalent.Based on these,various methods can obtain consistent analysis results,which theoretically guarantee the equivalence and correctness of various methods,so this work has important theoretical value and practical significance.