CORRUPTION AND INNOVATION: LINEAR AND NONLINEAR INVESTIGATIONS OF OECD COUNTRIES

Abstract

Employing annual data over the period 1996–2013 for 29 OECD countries, this paper explores the impact of corruption on domestic innovative activity, measured by the number of patent and trademark applications, via a linear panel fixed effect model and a nonlinear panel smooth transition regression with all lagged explanatory variables as instrumental variables and under the consideration of potential endogeneity biases. The results indicate several important findings. First, there exists a strong threshold effect between the control of corruption and levels of innovative activity across nations. Second, we note that corruption only has a substantial positive impact on innovation when it is over the threshold level, but not when a country has a seriously corrupt government with low bureaucratic quality, no matter for patent or trademark applications. Hence, heterogeneous beliefs about low transition speed show that OECD countries may not take actions instantly and identically to pursue better bureaucratic quality. Finally, we discover that an improvement over corruption presents greater impacts on patent applications than on trademark applications. Taken together, we confirm that corruption plays a fundamental role in determining innovation activities in OECD countries, offering meaningful policy implications for those policymakers and industries in accordance with our findings.