大多数学者研究环境库兹尼茨曲线(EKC)时,都隐含地假定一个地区人均污染与相邻的地区不存在空间相关性,因而其得到的估计结果可能出现偏误,进而将结论应用到环境政策方面也令人难以信服。为克服上述缺陷,利用中国28个省域1993~2005年的面板数据,应用考虑空间效应的空间面板数据模型,对中国的环境库兹尼茨曲线进行实证分析。结果发现:(1)我国的区域污染存在空间相关性,因此空间面板数据模型更适于研究EKC曲线;(2)区域污染排放不仅受本地区人均收入影响,而且相邻区域的污染物排放也对本地区有重大影响,即污染物排放具有空间溢出作用;(3)我国居民环境质量-收入弹性仍然处于较低的水平,环境质量在居民消费需求中比重并不高。

In the past 20 years, the Environmental Kuznets Curve has been regarded as one of the most powerful tools in studying the relationship between economic growth and environmental pollution. Among the conventional study in Environmental Kuznets Curve, most researchers implicitly assume that regional pollution is not affected by the neighboring regions. So, from analytical viewpoint and empirical behavior, the traditional methods used to examine the Environmental Kuznets Curve are suspected for neglecting spatial dependence among different regions. Consequently, the environmental policies based on the conclusions are hard to be convinced. With a view to the above shortcoming in describing the Environmental Kuznets Curve with the traditional methods, this paper investigates the relationship between environmental pollution and economic growth in China using panel data models which takes the spatial dependence into account. In this study, we select five types of pollution including waste water emission, waste gas emission, sulfur dioxide mission, dust emission and smoke dust emission as the environmental indexes and per capita GDP as the income index with 28 provincial data from 1993 to 2005. The results show that: (1) Pollution has spatial dependence in China and spatial panel data model is more suitable for studying the Environmental Kuznets Curve; (2) Considering spatial effect, the estimate result seems to be more robust than OLS estimate. Because the emissions of pollution are not only effected by the per capita income in the same region, but also largely impacted by the emissions from neighboring regions. (3) Resident environment quality-income is relatively low and does not take up a greater proportion in consumption demand.