中国地级市GDP数据的真实性:基于NPP-VIIRS夜间灯光数据的检验

Authenticity of China’s Prefecture-level City GDP Data:Test Based on NPP-VIIRS Night Light Data

文章通过使用比DMSP-OLS灯光数据更为准确的NPP-VIIRS夜间灯光数据,从更客观的角度来分析中国地级市GDP数据的真实性。首先,从主观层面对华东地区城市灯光数据和GDP数据的散点图、1995年和2013年的灯光数据对比图进行分析,得出GDP和灯光数据之间存在高度的相关性。其次以2013-2015年全国291个地级市的面板数据为样本进行固定效应分析,得出与主观层面分析一样的结果。研究结果表明灯光数据可以作为GDP的代理变量,进而可用于研究分析GDP数据是否存在水分。通过分析各地级市2013—2015年灯光数据的增长率平均值与GDP增长率平均值差值,得出东、中、西的差值平均值依次0.56%、0.62%、0.72%。

By using the NPP-VIIRS night light data, which is more accurate than DMSP-OLS lighting data, this paper analyzes the authenticity of the GDP data of China's prefecture-level cities from a more objective perspective. Firstly from the subjective level, the paper analyzes the scatter diagram of urban lighting data and GDP data in east China and the comparison diagram of lighting data in 1995 and 2013, and it is concluded that there is a high degree of correlation between GDP and lighting data. Secondly, the panel data of 291 prefecture-level cities from 2013 to 2015 are used as samples to analyze the fixed effects, showing the same results as those obtained from the subjective level. The study results show that lighting data can be used as a proxy variable of GDP, and can be used to study and analyze whether there is water in GDP data. By analyzing the difference between the average growth rate of light data and the average GDP growth rate of each prefecture-level city from 2013 to 2015, it is concluded that the average difference of east, middle and west is 0.56%, 0.62% and 0.72% respectively.