Abstract: With the growing pollution of rural water environment in China, accurate emission reduction measurement of rural water environment management is significant for the subsequent policies. It is difficult to effectively measure the effects of different kinds of policies due to the relatively small amount of data as well as its non-stability characteristics. Therefore, a multivariate discrete grey model is introduced to measure the emission reduction effect and then an empirical research is conducted. Firstly, the policies related to rural water environment management since 1995 are sorted out, and then they are classified into different groups and some representative policies are selected respectively according to the contents and targets. Secondly, the amount of different kinds of water pollutants emission is computed using the unit investigation and evaluation method. Furthermore, different groups of policies are introduced as virtual driven variables into the model, and competitive models strategy is adopted by setting a series of emission reduction measuring models. By comparing the adaptability among those models, the most adaptive model is selected to measure the emission reduction effect of different policy groups respectively. The research results show that rural water environment management policies are effective in general. Meanwhile, there exists certain difference about the emission reduction effect among three groups of policies. Moreover, the effect displays a spatial-clustering feature at province level. Lastly, some countermeasures are proposed to improve the management of rural water environment in China. It is hoped that the study will provide some reference for measurement of policy effectwith insufficient data.

Key words: discrete grey model, water environment policy, spatial heterogeneity