Sowing Seeds of Support: Elections, Bureaucrats, and Agricultural Clientelism in Africa

This dissertation asks when politicians in African countries will subvert technocratic implementation of policy programs targeting the poor in order distribute these state resources clientelistically. It specifically focuses on when politicians will manipulate distribution of agricultural subsidies as a highly salient policy issue in societies that remain largely agrarian. Given the need to win rural votes in the multiparty electoral context, it should be quite tempting for politicians to manipulate state resources targeting the rural poor. However, these programs are typically implemented by bureaucratic agencies, meaning that politicians do not have direct control over their distribution – a fact overlooked by existing studies of distributive politics. The relationship between politicians and bureaucrats is a principal-agent relationship characterized by considerable uncertainty. Thus the question becomes when politicians will interfere with the bureaucratic implementation of such programs given the inherent uncertainty.

I argue that politicians’ incentives to clientelistically manipulate the bureaucratic implementation of policy programs given the uncertainty are jointly determined by two features of a country’s political party system – party system institutionalization (PSI) and opposition competitiveness. In contexts of low PSI, it is difficult for politicians to establish credibility with voters as the options facing voters change from one election to the next, raising incentives for politicians to build credibility through direct clientelistic exchanges with voters. If the party system is also highly competitive, the incumbent will face particular incentives to clientelistically distribute state resources rather than party-controlled resources to gain an advantage over the opposition. This leads to the expectation that clientelistic distribution of agricultural subsidies and political inference with bureaucratic implementation is most likely in contexts of low PSI and high opposition competitiveness.

I test this argument in the context of agricultural subsidy programs in Malawi and Burkina Faso, two countries with similar socioeconomic contexts and agricultural sectors but different party systems. To test the claim the subsidy distribution will be clientelistic in Malawi but not Burkina Faso, I pair household agricultural survey data with district election data in both countries. I find that agricultural subsidies were distributed clientelistically in Malawi but not in Burkina Faso. Politicians could manipulate subsidy programs either by politicizing the Ministry of Agriculture or by seeking to circumvent the bureaucracy. Through an original survey of Ministry of Agriculture bureaucrats in both countries, I find low levels of ministry politicization in both countries. Instead, interviews with subnational bureaucrats demonstrate that bureaucrats in Malawi were able to circumvent the bureaucracy and distribute agricultural subsidies directly.

This dissertation makes several contributions to the study of distributive politics and African politics. It introduces party system institutionalization as a factor contributing to incentives for clientelism. It recognizes that whether distribution of goods is controlled by the party or the state has important implications for the logic of clientelism. And it demonstrates that bureaucrats must be taken seriously by scholars of African politics as independent actors able to affect policy outcomes.