Democratic Accountability and the Politics of Service Provision

How do citizens get politicians to respond to their wants and needs… to actually experience “rule by the people?” What are some of the opportunities and challenges of democratic politics for improving the quality of life, particularly through government services? (Here is a review article I wrote on the politics of service delivery; and you can learn about my research on the effects of ethnic politics on service delivery here.)

An informal settlement in the West Rand of South Africa, where residents live without basic infrastructural services.


Formal institutions of democratic rule are thought to be fundamental for generating the policies and practices needed to improve levels of human development. In short: we expect citizens to “throw the bums out” when they don’t deliver the goods and services we need, and to keep them in office when they do! But as it turns out, it’s not so straightforward.

An important place to look is the relationship between voting patterns and service delivery. Do citizens, as voters, respond in the manner that we might expect from democratic theory? Specifically, do they vote for (against) incumbents once services are (not) provided?


Daniel de Kadt (Asst Prof at UC Merced; PhD from MIT) and I investigated the politics of voting and service delivery in Southern Africa, and we published this study in the British Journal of Political Science, but you can get the gist of what we learn in this post on the Monkey Cage.

We find, to our surprise, that when citizens in Southern Africa gain access to basic infrastructural services they are less likely to vote for the dominant incumbent party. We discuss why we think this might be going on, and highlight that there is good reason to believe that citizens may be taking into account several other important considerations, including perceptions of corruption, that help to explain the results.

Beyond citizens, it’s also critical to consider the role of politicians, bureaucrats, and other service providers in the accountability “matrix.” In particular, what drives politician attitudes and behaviors in relatively new democracies? To what extent do they work to represent the needs and preferences of their constituents? To what extent do they accept or resist efforts to hold them accountable? Do certain circumstances or particular initiatives produce better politicians?


In the Eastern Cape province of South Africa, along with several Princeton undergraduate and graduate students, I carried out research on the various challenges to democratic accountability in the context of decentralized and polycentric institutional environments, and in this work, I paid particular attention to local politicians.  Despite the great optimism for cooperative or polycentric governance, I show that diffuse accountability mechanisms can lead to substantial under-performance. A related paper from that research described the gender- and race-specific policy preferences among those politicians, particularly with respect to health- and AIDS-related policies.


During the past few years, I have been working to develop and to extend this research in some exciting ways with various terrific collaborators. Along with Philip Martin (Asst. Prof. George Mason University) and Nina McMurray (PhD student, MIT), and more recently with Gabriel Nahmias (PhD student, MIT), we are working on projects investigating accountability and responsiveness by directly studying African politicians in South Africa, focusing on the thousands of democratically-elected local councillors. Specifically, we have been studying how these individuals relate to both their constituents and to other political principals (i.e. parties, and other levels of government), and the extent to which their individual life histories shape their attitudes and performance in elected office. This research includes analyses of an original survey, electoral and administrative records of over 100,000 local government candidates, and field experiments. You can learn more here: South African Councillor Panel Study (SACOPS) and check out a preliminary research report published by a funder, Making All Voices Count.

The Dullah Omar Institute and Konrad Adenauer Foundation hosted a seminar at the Cape Town office of the South African Local Government Association, where I presented our findings.

The research on South African councillors was designed as a longitudinal study, and I hope to have some interesting results following some planned follow-up research around the 2021 local elections.

As part of the first stage of work, we engaged in some collaborations with students and faculty from the University of Cape Town and University of the Witwatersrand, and thanks to a MISTI grant, we were able to bring over several of the students and a faculty member (Vino Naidoo) for a fund and productive week of discussions.

Information and Active Citizenship

It’s one thing for a country to have elections, and for citizens to vote every few years. It’s another to have a polity where citizens are actively engaged in monitoring and sometimes co-producing public goods and critical services. For example, in the context of Free Primary Education (FPE) in East Africa (and elsewhere), many parents are not even aware that they could or should be monitoring school quality or making sure that their children are studying.


A few years ago, as this concern was increasingly recognized by scholars and development professionals, NGO’s, funders, and starting working to close this gap, and to increase “active citizenship.” Twaweza is one such organization that became very involved in this “space,” and between 2010-4, I worked with collaborators, Dan Posner (UCLA) and Lily Tsai (MIT) on an impact evaluation of their Uwezo project. Their aim was to provide various forms of information to citizens to promote better government accountability around education in East Africa. I continue to be interested in questions about how different types of informational interventions affect citizen attitudes and behavior, particularly in terms of their relationships with politicians and service providers.

Currently, I am collaborating with Yang-Yang Zhou (Asst Professor, University of British Columbia) on a study of efficacy and active citizenship in Tanzania, in partnership with Twaweza. We completed two field experiments in Tanzania to study the effects of a novel intervention, Validated Participation, on efficacy and active citizenship. We find that validated participation does cause citizens to feel greater self-efficacy; and in a smaller two-year study, we find that validated participation results in greater parent involvement with their children as learners and in their children’s schools. We are excited about these findings and plan to release more information about how to implement the VP intervention in different settings, with the hope that it may have a positive impact, and that its effects may be rigorously studied in other contexts.