The Research Of The DEA Method Based On The Input-Output Tables

Data Envelopment Analysis (Data Envelopment Analysis) referred to DEA, is the method to evaluate the relative effectiveness of decision-making units. It has the great advantage in dealing with the comparison among the sectors (DMU) of the same nature having multi-input, multi-output. As a non-parametric statistical method, DEA has been widely used in various fields. But at present it is rare that DEA method has been applied to the input-output analysis. For input-output tables have rich, comprehensive data and can provide diffusion coefficient and other technical indicators, if input-output tables are used for Modeling DEA, the input-output analysis can be achieved with the complementary advantages of DEA , more satisfying effectiveness of Evaluation will be got.In this paper, a method of DEA based on input-output table is put forward. Three DEA models are established, respectively, taking various departments of the three major industries as decision-making units, all investing sectors as input indicators, and a total output as output indicators. At the same time, the inputs and outputs of traditional DEA models can not be zero or negative. On this issue the paper proposes a kind of solution, and introduces a method for further evaluation on DMUs which are all effective. Finally we use the established model to analyze the input-output table of Anhui Province in 2002 containing 42 departments, and then get the result that whether the sectors of three major industries are effective. At last for the non-effective DMUs, a adjustment program is put forward.Using the input-output table for Modeling DEA, not only open a new use for input-output tables, but also broaden the DEA's scope of application. This paper has conducted in-depth way on the exploration, access to a number of accomplishments and the expected results. This paper deeply explores the measurement and acquires some achievements and excepted performance.