Government size, democracy, and corruption: An empirical investigation

Previous studies on the effect of government size on corruption have produced mixed results. In an attempt to explain these ambiguous results, this study investigates the effect of government size on corruption by taking into account the role of the democracy in each country. Using annual data of 82 countries between 1995 and 2008, the estimation results indicate that an increase in government size can lead to a decrease in corruption if the democracy level is sufficiently high and, in contrast, can lead to an increase in corruption if it is too low. As robustness checks, the estimations using a different index of corruption and a different proxy for government size are also conducted. The results reveal that our main findings are robust. Furthermore, to address endogeneity problems, we conduct the instrumental variables estimation and the system generalized method of moments estimation, the results of which also support our primary findings. These results provide some important implications for policymakers seeking to perform government interventions without aggravating corruption. (C) 2012 Elsevier B.V. All rights reserved.