中文摘要：

中国地方政府特殊的竞争行为和政绩考核机制使环境成为“逐底竞争”的牺牲品．跨境河流极易形成污染的行政“边界效应”。本文基于2004-2013年9大水系重点监测断面周数据，运用OLS和最邻近匹配（Nearest Neighbor Matching，NNM），考察了行政边界对河流污染程度的影响：利用双重差分（Differencein Difference，DID）检验了“十一五”和“十二五”期间的减排政策对“边界效应”的改善作用。研究结果显示，不论是OLS还是NNM都证实确实存在较为明显的“边界效应”．发现边界监测点的污染水平比非边界监测点的污染水平分别高出16％（净pH值）、105．02％（化学需氧量）和90．02％（氨氮）。五年规划的节能减排政策确实有利于缓解严重的“边界效应”．但仅限于纳入考核指标的化学需氧量和氨氮。研究对于逐步弱化经济考核、规范地方考核办法、鼓励区域间环境合作治理以及出台合理的生态补偿方案都具有重要的政策含义。

英文摘要：

China＇s local governments＇ special competition behavior and performance evaluation system lead to the environment become the victim of ＂race to the bottom＂. Transboundary rivers are very inclined to administrative＂boundary effect＂. Based on the key monitoring sections＇ weekly data of nine river systems from 2004 to 2013, we apply OLS regression and Nearest Neighbor Matching method, and investigate the effect of administrative boundaries on the rivers＇ pollution level. Then we research the influence of the emissions mitigation policy on ＂boundary effect＂ with the method of difference in difference（DID）. The results show that the ＂boundary effect＂ significantly exists with both the method of OLS and NNM. It turns out that the average pollution indices at boundary monitoring sites are respectively 16%（NpH）, 105.02%（COD） and 90.02%（NH3-N）, and higher than those at non-boundary monitoring sites. Emissions mitigation policy of the five-year plan can help ease serious ＂boundary effect＂, but just for COD and NH3-N, which are integrated into performance evaluation system. The research is of great practical significance to understand the domestic cross administrative region of river pollution, break the economy assessment mechanism and focus on cooperating cross river environment problems etc.