Research on the Economic Growth for Under-Developed Counties -- Based on Two-way Fixed-effects Model

Abstract

The county economy, especially the problem of poor counties, is currently a hot issue in China. At present, the research on the key factors for county economy is scarce, partly because it is difficult to obtain accurate and sufficient data. Because power consumption and industrial level are often highly correlated, and the power consumption is real-time data and hard to falsify, this paper used spatial, electricity consumption and economic data of 66 Chinese counties, which contains 29 poor counties, from 2009 to 2016 to find out the key factors for under-developed counties’ economic growth by applying fixed effect regression model and machine learning models. The result shows that for poor counties, though the 1st industry is still the fundamental industry for county-level economy, the development of 3rd industry has significant positive impact on local economy. However, the development of 2nd industry, including recruiting large companies, and the input of electricity resources cannot well drive the local economies, which may suffer great loss due to the elimination policy of overcapacity in recent years. And the machine learning results support the above conclusions and suggest an obvious geographical cluster of poor counties, and the location, development of 1st and 3rd industry and net income of rural residents can explain most difference between the poor and non-poverty counties. These conclusions can be helpful for the government to lead the poor counties to get rid of poverty, and the cluster of poor counties should be focused.