Peer Reviewed Publications (in reverse chronological order)
Taking to the Streets: Protest as an Expression of Political Preference in Africa (with Erin Hern) forthcoming Comparative Political Studies
In this paper, we argue that protest behavior in Africa’s developing democracies is in fact a way for citizens to express their political preferences when more traditional behaviors like voting are insufficient. In the context of Africa’s democracies, there are a number of political systems in which formal channels for political expression are likely to be inadequate. In particular, expressing political voice is a problem if the electoral arena does not provide citizens with the opportunity to vote around salient issues. In African politics, three regularly documented political circumstances inhibit citizens’ ability to express their political preferences through conventional channels: dominant party regimes with low levels of political competition, volatile party systems where parties lack programmatic platforms, and countries in which ethnicity (or other identity variables) rather than programmatic preferences determine vote choice. In each of these systems, voters cannot rely on the electoral arena to demonstrate their preferences; as such, protest may be a more feasible way of expressing political ideas. In this conceptualization, protest accomplishes two things: first, it gives citizens a way to amplify their voices in expressing their preferences to the government. Second, it allows citizens to designate which issues are salient enough to warrant a willingness to challenge entrenched parties or elites.
Most research in developed countries on prejudice toward foreign-born minorities suggests that cultural rather than economic threat motivates xenophobia. Prior studies leave unanswered questions about the origins of anti-immigrant prejudice in developing countries, where one-third of worldwide immigration occurs. Alternatively, developing-country research simply assumes that economic threat drives prejudice in the Global South but has not presented credible empirical evidence. In this study, we seek to reliably measure anti-immigrant prejudice and examine possible determinants of prejudice and prejudice-based voting behavior. Through a list experiment conducted on a random sample of South Africans (N=1,088), we investigate the predictive power of economic threat theory in explaining prejudice toward immigrants in South Africa. The results show that significant prejudice towards immigrants exists among South Africans and that such prejudice is higher among the unemployed, but these sentiments do not seem to influence vote choice. The evidence suggests that the determinants of anti-immigrant sentiments due to South-South migration are distinct from South-North migration.
Elite and Mass Support for Foreign Aid Versus Government Programs: Experimental Evidence from Uganda (with Michael Findley, Helen Milner, and Daniel Nielson; Forthcoming at International Organization 2017)
Does foreign aid enable or constrain elite capture of public revenues? Building on prominent debates in the foreign aid literature, we examine whether recipient preferences are consistent with a view – called here donor control theory – that foreign donors wield substantial control over the flow of aid dollars, making elite capture more difficult and mass benefits more likely. We compare elite and mass support for foreign aid versus government spending on development projects through a survey experiment with behavioral outcomes on members of the Ugandan national parliament and a representative sample of Ugandan citizens. For two actual aid projects, we randomly assigned different funders to the projects. Significant treatment effects reveal that members of parliament support government programs over foreign aid, whereas citizens prefer aid over government. Donor control theory also implies that citizens should favor foreign aid more and elites less as their perceptions of government clientelism and corruption increase. We explore this and report on other alternative mechanisms. Effects for citizens and elites are most apparent for those perceiving significant government corruption, supporting donor control theory.
Data Replication Code
Ethnicity is frequently posited as an important factor in civil violence and other political contexts. Despite the attention that ethnicity receives, its effects depend on an important, but mostly ignored, assumption that ethnicity is identifiable within and across groups. There is likely considerable variation in peoples’ abilities to identify each other. Certain individuals within groups might be better at identifying others’ ethnicities; further, different types of information might aid identification better. We contend that the strength of an individual’s ethnic identity influences her ability to identify others correctly. We test this argument using an experiment in the Eastern Cape of South Africa in which individuals attempted to identify members of the major black ethnic groups. We find that the average individual struggles to identify ethnicity correctly in many conditions. Individuals with a stronger identity, however, are often better at correctly identifying the ethnicity of others relative to the average individual. When receiving contradictory information, individuals with stronger identities were sometimes deceived more easily than others. These results have implications for a diverse set of studies relying on the identifiability assumption.
Ethnically Proximate Swing Voters: The Limits to Identity and Group-Based Voting
The key contribution of my dissertation is to provide answers to the following two questions: who are swing voters in Africa’s “ethnic census” elections? Why do some voters not politically align with their ethnic group? The vast literature that seeks to explain the effect of ethnicity on voter behavior focuses on the role of co-ethnicity, and these studies repeatedly find that voters, especially in Africa, tend to support the party that represents (or the candidate from) their own ethnic group. While this finding is robust across contexts and research designs, data from twenty African countries shows that between 30-53% of voters do not support their ethnic group’s party. In fact, while in many multi-party contexts a plurality of members of an ethnic group do support their ethnic group’s party, often the majority of the group’s members support various alternative parties. Thus, past studies that focus on the average tendency of groups, due to their focus on co-ethnicity, overlook the vast within-group variation in political preferences and cannot explain why some voters do not support their co-ethnic candidate. When this within-group variation is acknowledged in the literature, it is attributed to strategic or policy considerations, which sets aside ethnicity as a possible mechanism, but I argue this is premature. My research moves beyond the convention of co-ethnicity in order to better understand the ethnic nature of within-group political preference variation.
Why are some voters less likely to align with their group when group-based voting norms are strong? Further, are these voters more likely to be swing? I argue that the answer to these questions can be found in the extent to which individuals are apparently consistent with the prototypical individual in their group. I develop a concept of racial distance, which provides a continuous conceptualization of race that can be measured exogenously. Empirically, I investigate this relationship in South Africa using an original panel survey, which brackets the 2014 national elections. The results show that those who are more racially distant from the prototypical member of their group are significantly less likely to vote with their group and more likely to change their vote due to an election campaign. Using data from the US and Brazil, I conclude that this relationship holds for racial majorities but only certain racial minorities.
Inter-marriage is rapidly altering the ethnic landscape across Africa. Estimates from the most recent Demographic and Health Surveys (DHS) show that the median country-level rate of inter-marriage is over 20% across a large sample of countries. This paper explores how the blurring of ethnic lines affects voter behavior. We focus on the electoral choices of mixed individuals who descend from parents with different ethnic backgrounds. Drawing on two large surveys from Malawi and Kenya, we find that individuals of mixed ethnic heritage are less likely to engage in ethnic voting than mono-ethnics. We explore the empirical and theoretical implications of this finding and investigate potential mechanisms.
Politicians and Policy Positions: Evidence of Positional Voting from Uganda’s 8th Parliament (with Macartan Humphreys)
Many accounts of political behavior in Africa focus on clientelistic behavior, with an implicit assumption that ideological position taking is either absent or irrelevant. Yet while there is evidence that clientelism matters, there is little evidence that ideological position taking does not. We explore the question here using the wordscores text analysis procedure developed by (Laver et al., 2003) to convert MP contributions in plenary sittings of Parliament into estimates of ideological ideal points. We try to validate these ideal point estimates with MP self-reported policy preferences and we use our text-based MP ideal points to test the canonical voter proximity theory in the Ugandan context. Our preliminary results nd support for the idea that political speech captures ideological position taking in Uganda and that Ugandan voters vote as a function of the ideological position of politicians (although the evidence for proximity voting is mixed). In addition, and consistent with the literature, we find no evidence of policy coherence within political parties.
Civil Service Management Practices for a More Motivated, Committed, and Ethical Public Service in Uganda (Co-authored with Immaculate Apio Ayado, Christian Schuster, Jan Meyer-Sahling and Kim Sass Mikkelsen)
This report is part of an international research project that seeks to help governments across the world – including Uganda’s – to make evidence-based decisions about civil service management practices, including recruitment, pay, promotion, performance evaluation, career management and leadership. It does so by assessing the effects of civil service management practices in government institutions on the attitudes and behavior of public servants – their work motivation, job satisfaction, commitment to public service, performance, integrity and ethical conduct. Based on this assessment, governments can learn which of their civil service management practices have positive effects and which do not; and which management practices from other countries might be worthwhile adopting. The report concludes that civil service reforms need to be tailored to the realities in each government institution, that while Uganda’s public servants are motivated to serve the public they are not motivated to work hard nor are they satisfied with their job conditions, and that civil service management practices are highly correlated with work motivation and job satisfaction (and any reforms would do well to focus on this relationship).