Probe into the targeted poverty mitigation policy in China based on causal inference: Evidence from Chongqing in the Three Gorges Reservoir region

A quantitative analysis of the panel data of 21 districts in Chongqing City in the Three Gorges Reservoir region from 1998 to 2015 was conducted to determine the influence of China's focused poverty reduction policy in terms of causal inference. Specifically, the effects of this policy at the initial phase of execution through break-point regression; the impacts of this policy during the implementation period by using an instrumental variable panel regression; and comparative studies of the policy impacts by using the generalized synthetic control method. The outcomes showed that: (1) China's state policy for specific poverty mitigation has a substantial impact on reducing poverty, which has been observed from the start of the enactment, namely, peasants' net income per capita from 2011 to 2013 increased by 13.9%, as compared to that during 2008-2010. Meanwhile, the net earnings of each farmer grew by 22% from 2011 to 2015 throughout the implementation phase of the policy, as opposed to that in 1998-2010. (2) Economic behavior in the marketplace and agricultural output figure of impoverished regions has significantly improved farmers' individual net income. (3) The comparative study further proves that the poverty mitigation policy has the effect of reducing poverty. Based on these findings, it is proposed that China should increase investment to consolidate its poverty alleviation policy in the most poverty-stricken areas, activate rural markets and develop agricultural production at the local level to accelerate poverty reduction in the region. The method of causal inference was utilized as an innovative method to study China's focused poverty reduction policy for both immediate and permanent effects. It is the first time to use the generalized integrated control method, which is a causal inference frontier method, to further verify the effectiveness of poverty reduction policies.