Economics of corruption: game-theoretic modelling of traffic police bribery in transition countries

Abstract

Jain argued in 2001 that the focus of future corruption studies should be (1) to “build a comprehensive models of how corruption works at the micro as well as at the macro levels” (p.101) and (2) subject the models to empirical tests. Moreover, he noted, “research on effective mechanisms to solve corruption is even scarcer. Although we have a good idea of what approaches are possible, we do not have more than anecdotal information on which approaches work.” (p.102). This study of traffic police bribery does just that. Modelling of utility payoffs has been derived from specifically-designed qualitative study to present determinants behind strategy selection at the micro-level, building a more comprehensive model. The three theoretical model variations of traffic police bribery were subjected to two empirical tests of anti-corruption reforms in the case of Georgia and in the case of Ukraine. As a result, this provides non-anecdotal information on which approaches work—such as the case of Georgia—and which do not. All in all, this thesis also accomplishes the original study aim—development of a game-theoretic framework for modelling of traffic police bribery—by building a model that produces results in line with the empirical cases. Besides filling the gap in research on the issues outlined by Jain (2001), this thesis also incorporates a regional dimension of transition countries to the above modelling. Conceptually, this research generates additional support to both the wage theory of anti-corruption, as well as to the ‘big bang’ theory: affecting factors like wages, probability of detection, ticket-issuing procedures, simplifying rules and regulations are important components of anti-corruption programmes and need to be tackled simultaneously, so as to generate a rapid shift from a corruption to a no-corruption equilibrium. The insights from interviews revealed that respondents utilised all three types of rationalities—instrumental, bounded and expressive rationalities, suggesting that utility-maximising elements of the models have to be further augmented with incorporation of social norms and beliefs about behaviour of others. Hence, this research can and needs to be taken further. The modelling analysis does not include role of past experience and perception about others, while the case studies suggest that past experience is an important determinant, and that replacement of old police officers with new police officers is an effective tool. Suggestions for future research include incorporating dynamics of evolutionary game theory, based on Young (1998), and Mishra (2006) analysis relating to persistence of corruption. Alternatively or concurrently, the aims of the state (that implements the anti-corruption measures) could also be incorporated into the game, taking it away from a micro-level interaction, similarly to the work of Yao (1997). Experimental research where players role-play drivers and traffic policemen, given set variables and objectives, could be considered for further testing of the theoretical framework. Lastly, current methodological shortcomings could be improved by including increasing the sample size beyond the forty-two cases of driver-traffic policeman interaction and incorporating additional empirical cases.