The Effects of Democracy on the Quality of Governance : Evidence on the Capacity of Political Systems to Constrain Corruption

In the last decades democracy spreads in many parts of the world, and this generates several questions about its performance. The object of this work is the relationship between democracy and the “quality of governance”. The main purpose is to analyse empirically the consequences of democratization on one specific aspect of “good governance”, i.e. the efficiency and transparency of the decision making process measured with the capacity of a political system to constrain political corruption. Starting from a procedural conception of democracy the theoretical argument that guides my work claims that the transition to a democratic system introduces more competition in the political arena through the concrete possibility of government alternation. This credible threat strenghtens the relation of accountability between voters and government members inducing the latters to maintain an accountable behaviour. This should produce a general reduction of the degree of corruption among public servants. This argument generates a theoretical framework developed upon three main analytical dimension of democratization. The first dimension takes into account the presence of competitive political structure while the second analyses the actual level of democratization of a political system. Finally, the third dimension considers the longitudinal variation of democracy measuring the time exposure of different states to democratic competition. According to this analytical framework I advance five different research hypotheses concerning the role of the democratization process in constraining political corruption. To test these hypotheses I use a cross-national dataset containing data on democracy and performance indicators for 191 countries. To operationalize my dependent variable I use two indexes of perceived corruption, whereas as independent variables I use several democracy indicators. To obtain more generalizable results I test the validity of the hypothesised relations for the effects of several control variables. The results obtained seem to confirm in part my research hypotheses.