中文摘要：

选取2013年10月到2014年9月的PM2.5日数据,采用AR模型和马尔科夫区制转换模型研究了京津冀地区各城市雾霾污染的持续性特征,进一步基于Hsiao（1981）的格兰杰因果检验和广义脉冲响应函数考察了相邻城市间雾霾污染的空间溢出效应。实证结果表明：（1）雾霾污染存在较强的持续性,PM2.5指数表现为高度的自相关和区制转换特征,高污染状态较难向低污染状态转换,高污染状态存在粘滞效应;（2）不同城市间的雾霾污染存在空间溢出效应,格兰杰因果检验表明北京对其他周边城市均存在领先滞后关系,广义脉冲响应分析表明北京雾霾浓度的提高对周边城市均会产生正向冲击、且衰减速度较慢。政策建议是简单的产业转移并不能解决京津冀地区的雾霾污染问题,必须通过产业升级和清洁技术引进、区域综合治理、调整政绩考核方式等手段来从根本上治理京津冀地区的雾霾污染。

英文摘要：

Using data ranging from 2013,October to 2014,June,this article empirically analyzes the duration of haze of the Beijing-tianjin-hebei region based on AR model and Markov Regime Switching Model,and then investigates the spatial spillover effects of haze pollution between adjacent cities by employing Hsiao＇s（ 1981） Granger Causality test and generalized impulse response function. The results shows that：（ 1） The haze pollution has strong duration with high autocorrelation and significant regime-switching effect,besides,it＇s hard to transfer from high-polluted state to lowpolluted state,indicating that the high-polluted state has viscous effect;（ 2） Spatial spillover effects are found in haze pollution between different cities. The Granger Causality test between cities finds that there exists evident lead-lag relationship between Beijing and its adjacent cities; moreover,the generalized impulse response analysis suggests that the increase of the PM2. 5 concentration of Beijing leads to positive impact on other cities and the decaying rate is relatively low. Finally,based on the above results,we propose the policy recommendation that simple industrial transfer can hardly solve the PM2. 5 pollution of this region,thus only industrial structure upgrade,the introduction of clean technology,regionally comprehensive control and the adjustment of assessment standard for political achievement can solve the problem fundamentally.