Political Communication and Accountability in Uganda


Macartan Humphreys
Professor of Political Science


In many developing economies, poor social and economic outcomes are attributed in part to political decision-making that is unresponsive to the needs and preferences of citizens. Weaknesses of governance structures and social welfare in turn are associated with fragile security outcomes and economic failures that may in turn have implications for US economic and security interests. In light of these adverse outcomes, large amounts of development aid has been provided by the United States to bolster democracy and governance interventions in developing areas. We focus on one such intervention implemented by the Parliament of Uganda and the National Democratic Institute (NDI) with support for the US National Endowment for Democracy (NED). The intervention seeks to improve accountability by opening a new channel of communication between Members of Parliament (MPs) and their constituents, allowing citizens to send messages to MPs via an SMS platform in parliament. Uniquely the intervention is structured with a random assignment of MPs into treatment and with a random assignment of the cost of communication facing voters. The experimental structure provides an unprecedented opportunity to assess the political effects of citizen communication. 

Intellectual Merit: The intellectual merit lies in contributions to our understanding of the effects of the structure of channels of political communication on political behavior. The study contributes to debates on the role of technology in political participation, assessing how the channels and price of access affects the types of bias in the information politicians receive; the role of cost in the ability of politicians to extract signals from constituent communication; and the impediments to political communication. Taking advantage of a unique collaboration with the National Democratic Institute and the Ugandan Parliament, this study combines a strong identification strategy with extraordinary access to data, including face-to-face surveys with Members of Parliament and a sample of their constituents, and all direct communication between politicians and constituents. The study is one of the few field experimental studies to operate at the level of national political elites, pushing the frontiers on the use of experimental methods to address theoretically motivated questions of elite level politics. In addition, the study pushes the frontiers of research using mixed methods by piloting the use of an approach to formally integrate qualitative tests (from intensive analysis) with quantitative tests (from an extensive analysis) in a Bayesian framework.

Broader Impact: The results of this research are likely to affect institutional choice in Uganda and possibly elsewhere. Critically, the results are relevant for the Ugandan Parliamentary Commission which needs to decide how to develop the system. More broadly, while there is much enthusiasm for interventions of this form, the Uganda experience to date suggests that actual take up can be low. This research will provide evidence on effectiveness as well as an analysis of the impediments to success that can help guide US and other democracy supporting interventions. The study is specifically designed to examine whether an unmediated and impersonal IT-based communication channel can increase the access of marginalized populations. While many stakeholders hope changes will be for the good, a key motivation of this study is precisely to be able to assess biases and impediments to communication from marginalized populations. The study strengthens capacity among US and international researchers as well as among practitioners. This is achieved through continuous interactions with practitioners, through collaboration with developing country researchers, and through the involvement of graduate students at Columbia University in experimental design and implementation.