中国GDPの3つの代理変数に関する研究 : 重回帰分析の観点から

Abstract

The aim of this study is to examine the usefulness of multiple regression analysis by conducting an analysis using Chinese GDP as an objective variable and the Li Keqiang index， namely， electricity usage by the industrial sector， railway freight quantity ， and growth of bank loans as explanatory variables. Firstly， we used correlation analysis to determine the correlation between our objective variable and explanatory variables. We found the explanatory variables electricity usage by the industrial sector and growth of bank loans to have a strong correlation with objective variable GDP. However， this correlation analysis found only weak correlation between railway freight quantity and GDP. From this study， we learned that if we add road freight quantity to railway freight quantity，using the resulting freight quantity as an explanatory variable，there was a strong correlation with GDP. We also found that when a revised version of the Li Keqiang index is used， there is a very strong correlation between these three explanatory variables and GDP. Additionally， we considered multicollinearity while conducting this study. The fact that these three new variables-specific economic activities-each showed strong correlation with China's GDP suggests that GDP statistics accurately reflect economic conditions. Therefore， Chinese GDP statistics are useful. We also believe that these three new variables are proxy variables for GDP.