Causality Channels between Corruption and Democracy: A Threshold Non-Linear Analysis

Authors: Khalid Sekkat, Fredj Fhima, Ridha Nouira

Abstract:

This paper focuses on three main limitations of the literature regarding the impact of corruption on democracy. These limitations relate to the distinction between causality and correlation, the components of democracy underlying the impact and the shape of the relationship between corruption and democracy. The study uses recent developments in panel data causality econometrics, breaks democracy down into different components, and examines the types of the relationship. The results show that Control of Corruption leads to a higher quality of democracy. Regarding the estimated coefficients of the components of democracy, they are significant at the 1% level, and their signs and levels are in accordance with expectations except in a few cases. Overall, the results add to the literature in three respects: i). corruption has a causal effect on democracy and, hence, single equation estimation may pose a problem, ii) the assumption of the linearity of the relationships between control of corruption and democracy is also possibly problematic, and iii) the channels of transmission of the effects of corruption on democracy can be diverse. Disentangling them is useful from a policy perspective.

Keywords: corruption, governance, causality, threshold models