Following is from a series of blog posts by MPSA members about their research that has received funding by either the National Science Foundation (NSF) or the National Endowment for the Humanities (NEH). Here, Kirk A. Harris a Ph.D. Candidate at Indiana University summarizes his NSF-funded research onPolitical Science: Ethnicity, Civic Participation, and Public Goods in Kenya.”

It looks like ethnic diversity is harmful to public goods provision. Diverse countries tend to invest less in things like education, infrastructure, and public health that promote economic growth. While explanations for this abound, a common conclusion amongst those who study African countries is that spending on public goods suffers when politicians channel resources towards their cronies and co-ethnic kin at the expense of other groups in society.

This type of kinship-oriented service provision exists in both autocracies and in “patronage democracies” (Chandra 2004), where leaders allocate resources disproportionately towards voters with whom they share an ethnic identity – or at least voters believe that they do. But this ethnic targeting of development resources is certainly not ubiquitous, and observing this pattern isn’t the same as explaining why it happens. So, when do leaders decide to target development resources towards co-ethnics, and why?

Kenya’s Constituency Development Fund (CDF) provides an excellent context in which to examine the relationship between ethnicity and public goods provision. The CDF gives all 290 members of Kenya’s National Assembly who are elected from single-member districts significant financial resources to form a local committee that makes decisions about funding clinics, schools, roads, and other community infrastructure projects within their parliamentary constituency. These committees have substantial autonomy over how the CDF is managed and where projects should be located.

As part of my dissertation research, I carried out over 150 semi-structured interviews with individuals in six different Kenyan parliamentary constituencies as well as representative public opinion surveys in three of these constituencies in order to understand the politics of resource allocation, and to find out just who benefits from CDF projects in these regions.

The constituencies that I study contain groups who distinguish between themselves on an “ethnic” basis – group membership is determined by socially-defined, “visible”, descent-based characteristics. But the nature and importance of these ethnic cleavages differs across constituencies. In some regions, groups differentiate between themselves on the basis of their membership in locally-relevant ‘clan’ groups even as they acknowledge a common overarching ethnic identity that sets them apart from other Kenyans. In other, more “cosmopolitan,” regions members of different groups perceive no such overarching kinship tie connecting them to their neighbors.

In these latter constituencies, ethnicity often serves as a reliable heuristic for how citizens vote in local and national elections. The “image” of political parties is strongly associated with different ethnic groups and candidates from a given ethnic community feel compelled to stand as representatives of “their” group’s party if they are to have any chance of victory (cf. Ferree 2006 , Posner 2005).  Because voters’ choice of candidates in such contexts is determined more by their ethnic identity and party identification than by an incumbent’s performance in office, politicians have an electoral incentive to allocate resources towards their co-ethnics as a way of shoring up support amongst their kin.

By contrast, in diverse constituencies where ethnicity is not a politically salient feature politicians lack the incentives to engage in this type of ethnically-oriented patronage and must rely on other criteria to guide the distribution of CDF resources. In these settings, politicians are able to leverage the CDF to appeal to voters from a range of ethnic groups within the constituency. And challengers from other parties can likewise make credible appeals across ethnic boundaries about their ability to serve all constituents rather than simply those with whom they share an ethnic identity.

These findings suggest that a narrow focus on ethnic diversity as detrimental to public goods provision is misguided. It is not diversity itself, but the political salience of ethnic divisions that motivates the ethnically-biased distribution of local development resources in Africa’s new democracies. Elected politicians will use ethnic criteria in the distribution of development resources only when they stand to gain votes by doing so.

Kirk A. Harris is a Ph.D. Candidate in the Department of Political Science at Indiana University who specializes in the study of democracy and development in Africa. His dissertation research, on ethnicity and resource allocation in Kenya’s CDF, was funded in part via a Doctoral Dissertation Research Improvement Grant (DDRIG) from the National Science Foundation (SES-1423998). Harris can be contacted via email at