我国省域碳排放量与经济发展的GWR实证研究

Local Spatial Pattern Analysis of Carbon Emissions and economic development——An Empirical Study of Provincial Low-carbon Economy Based on GWR

普通最小二乘法(OLS)仅是对变量"均值"估计,不能反映省域碳排放量在空间上的非平稳性。采用地理加权回归(GWR)技术引入空间效应,发现GWR模型比OLS模型具有明显优势:省域碳排放量与经济发展水平、产业结构、人口、外商直接投资和能源价格之前存在内生经济关系;影响碳排放量各因素在省域空间上存在明显差异。最后对模型进行验证并为低碳经济实现提出相关政策。

Ordinary least squares （OLS） is the only to estimate ＂mean＂ of variables and does not reflect the spatial non- stationary of provincial carbon emissions. The article which is introduced into space effect found geographical weighted regression （GWR） method has obvious advantages comparing to OLS model： There are endogenous economic relations among provincial carbon emissions and the level of economic development, industrial structure, the population, foreign direct investment and the price; The factors which impact of carbon emissions are differences in the provinces; there is a positive correlation among provincial carbon emissious and the level of economic development, industrial structure. The two factors which are the biggest influencing weights take on decreasing feature from east to west; But, the influence coefficients of population and foreign direct investment all exist positive and negative correlations, these mean that controlling the population and increasing foreign direct investment will reduce carbon emissions in some regions and also increase provincial carbon emissions in another regions; Energy consumption demands are not sensitive to energy price which is monopolized by the government. Finally, we test and verify models and propose policies for a low carbon economy.