Middle-income trap and corruption: Evidence from a dynamic panel data analysis

This paper empirically tests the income trap phenomenon by analyzing the convergence rate of a growth model that incorporates corruption explicitly as a rent extraction out of capital accumulation. Our model shows that the countries are ‘trapped’ in a middle-income group because they are corrupt. Since they failed to have a less corrupt economy, then they would be less productive and do not have the sufficient and necessary capital to develop. By applying a simultaneous equations dynamic panel data model to 76 middle-income economies from 1998 to 2020 and using the iterated GLS estimation method, the results show that countries with high corruption (i.e., negative control of corruption index) are closer to their own steady state than countries with low corruption (i.e., positive control of corruption index) are. This implies the existence of a middle-income trap, which we define in this paper, for some of the countries in the middle-income group. We find the adverse effect of corruption on real GDP per capita in these economies, and in determining the different steady states among middle-income countries.