基于系统GMM面板模型的土地利用碳排放效应研究——以中国省际面板数据为例

Carbon Emission Effectsof Land Use Based on System GMM Panel Model--Taking ChineseProvincial Panel Data as an Example

依据2003~2015年中国省际面板数据,运用STIRPAT模型及系统GMM动态面板计量模型,分析我国不同土地利用类型下的碳排放效应,并探究我国经济发展是否与碳排放之间存在环境库兹涅茨曲线的"倒U"型关系。研究结果显示:(1)人均GDP与土地利用碳排放存在"倒U"型曲线的特征,进一步验证了我国存在土地利用碳排放的环境库兹涅茨曲线假说;(2)土地利用碳排放效应具有显著的滞后效应,基于系统GMM的动态面板计量模型的估计结果优于静态面板估计结果;(3)系统GMM估计结果显示,经济因素、能源强度、城镇化水平是碳排放最主要的影响因素。其中,人均GDP和能源消耗强度对土地利用碳排放效应存在显著正效应,城镇化率、第三产业增加值占比、农地非农化率和土地市场化水平对土地利用碳排放效应存在显著负效应。

Based on Chinese Provincial Panel Datain 2003-2015, the STIRPAT model and the system GMM dynamic panel measurement model wereused to analyze the carbon(C) emission effects under different land use types, and to explore the "inverted U" relationship in the environmental Kuznetcurve between economic development and C emissions in China. The results showed that:(1) The per capita GDP and land use C emissions presented the characteristics of "inverted U"curve, which validatedthe hypothesis forthe environmental Kuznets curve of land use C emissions in China.(2) C emission of land use showed a significant hysteresis effect, and the estimation results with the dynamic panel measurement model based on system GMM were better than those with static panel model.(3)System GMM estimation results showed that economy, energy intensity and urbanization level were the most important factors to C emissions. Among them, the per capita GDP and energy consumption intensity had significant positive effects on C emissions of land use. The urbanization rate, ratio of the tertiary industry added value, rate of the non-agriculturalization of arable land and the land marketization level had negative significant effects on land use C emissions.