This paper combines the Cobb-Douglas (CD)with the method of Data Envelopment Analysis and Analytic Hierarchy Process(DEA-AHP) and applies it to the input efficiency evaluation of fiscal and taxation policies. Based on the input of the number of fiscal policy items and the output model design of transformation and upgrading, the overall efficiency of fiscal policy is calculated by the method of Data Envelopment Analysis and Analytic Hierarchy Process. It turns out that the overall efficiency of fiscal and taxation policies is 0.397, and the efficiency of pure policy factors is 0.616. There is a difference in input efficiency between preferential tax policies and fiscal subsidy policies. Tax preferential policy has insufficient investment, and financial subsidy has investment redundancy. Furthermore, the absolute value of tax incentives and financial subsidies are introduced into the CD Function to calculate the contribution rate, and the government and enterprise factors that are not effective in the efficiency of fiscal and taxation policies are attempted to be separated. As a result, it is found that the contribution rate of financial subsidies is weakly negatively correlated. It shows that the conversion rate of enterprises to the financial subsidy policy is low, and the tax preference policy has a better conversion rate. The results of the study reveal that the scope and intensity of tax optimization needs to be increased at the government level, and the utilization of fiscal subsidy policies needs to be improved at the enterprise level. The government should reduce financial subsidies and establish a universal tax preferential policy system.

本文将柯布-道格拉斯(CD)与数据包络分析与层次分析法(DEA-AHP)相结合，将其应用于财税政策的投入效率评价。基于财政政策项目数量的输入和转型升级的输出模型设计，采用数据包络分析和层次分析法计算财政政策的整体效率。结果表明，财税政策的整体效率为0.397，纯政策因素的效率为0.616。税收优惠政策与财政补贴政策在投入效率上存在差异。税收优惠政策投资不足，财政补贴投资冗余。此外，在CD函数中引入税收优惠和财政补贴的绝对值来计算贡献率，试图分离对财税政策效率无效的政府和企业因素。结果发现财政补贴的贡献率呈弱负相关。说明企业对财政补贴政策的转化率较低，税收优惠政策的转化率较好。研究结果表明，政府层面需要加大税收优化的范围和力度，企业层面需要提高财政补贴政策的运用力度。