Published online by Cambridge University Press: 09 April 2017
A credible opposition is necessary for democratic accountability. However, in a multiparty democracy, a credible opposition may fail to emerge when it is in the strategic interest of political parties to collude rather than compete, effectively extinguishing all credible opposition. The author argues that illicit collusion among parties on a representative council is more likely when all viable parties win seats and are thus able to enter into a self-binding commitment to jointly engage in misconduct without risk of exposure. Conversely, when at least one party fails to win representation on the council, there is a credible opposition with the incentive and ability to threaten exposure of rent seeking among council members. The theory is tested using a regression discontinuity design where the electoral threshold to win a single seat is, within a narrow band, an exogenous determinant of whether or not there is an out-party or credible opposition. Exploiting the fact that Mali's decentralization produces within-country variation in both electoral and governance outcomes, the author uses data from commune council elections alongside local-level public goods provision as a measure of rent seeking. Poorer public goods provision is indeed more likely when all political parties in a district win seats on the council. To show that collusion is the mechanism driving this relationship, the author tests several observable implications in the data and uses qualitative evidence as illustration. This examination of when it is in the strategic interest of parties to engage in uncompetitive behavior contributes to the literature on when elections fail to produce democratic accountability.