Size, determinants, and consequences of corruption in China's provinces : the MIMIC approach

This paper uses a multiple indicators and multiple causes (MIMIC) model and estimates the extent of corruption in 30 Chinese provinces from 1995 to 2015. Treating corruption as an unobserved latent variable, the MIMIC results show that both government size and public investment have significant positive effects on corruption, while fiscal decentralization, citizen education level, average public sector wages, intensity of law enforcement, media supervision, political control and FDI all have significant negative effects on corruption. Among them, education level, size of public investment, intensity of law enforcement and political control are the most important determinants of China's corruption. Additionally, we find that corruption decreases GDP and residents' income significantly. In the 30 provinces the corruption index shows a negative trend from 1995 to 2015. Comparing the extent of corruption in the eastern, central and western provinces, we also find that the more developed the region, the lower the extent of corruption.