Between Citizens and the State: The Bureaucratic Transaction Costs That Sustain Clientelism
My book develops a new theory about the conditions under which clientelism flourishes—as well as those under which it is weakened. In contrast to the extant literature, which has long held that clientelism is a consequence of low economic development, I argue that clientelism does not necessarily disappear as countries get richer. Instead, clientelist networks are sustained by the barriers and difficulties citizens face when interacting with the state, or what I call bureaucratic transaction costs. A particularly important avenue of interacting with the state is through claiming welfare. Perhaps even more than voting, claiming welfare has become the most important form of political participation in low- and middle-income countries, especially following the rapid expansion of social welfare that took place during the 1990s and 2000s.
I argue that the high bureaucratic transaction costs involved in claiming basic entitlements coupled with a high demand for these benefits create valuable arbitrage opportunities. These opportunities help create a market for intermediaries who exchange mediated access to state benefits and services for political loyalty or another form of political payment. These individuals are also known as clientelist intermediaries. Importantly, the intermediaries in this market do not often share how to access benefits, as doing so would shrink their market share. Thus, this clientelist market promotes and depends on the idea that welfare benefits are gifts that need to be repaid with political loyalty rather than rights and entitlements that the state guarantees.
Therefore, although potentially an efficient way of accessing welfare, I argue that mediated access to the state can have detrimental effects on building citizens’ civic skills, autonomous political participation, and ability to make policy demands on the state. These sets of incentives create greater dependency on clientelist networks as intermediaries between citizens and the state – further entrenching a clientelist equilibrium.
The dissertation on which this book is based can be found here.
Winner of APSA Best Experimental Research Dissertation Award, Honorable Mention, 2020.
Winner of 2017 Innovations in Transparency Prize given by the Mexican Government’s Instituto Nacional de Acceso a la Información y Protección de Datos Personal, Mexico.
Breaking the Clientelist Feedback Loop: Evidence from a Field Experiment in the Yucatan Peninsula (Conditionally Accepted, American Journal of Political Science.)
Abstract: Across the electoral cycle, individuals rely on clientelist intermediaries because they face high bureaucratic transaction costs when claiming welfare benefits. However, relying on intermediaries inhibits individuals from gaining useful bureaucratic experience and know-how, further entrenching their dependency on this clientelist system. A large-scale field experiment in rural Mexico that reduces the costliness of claiming social programs by providing a facilitator trained to assist citizens in the application process tests this argument. I find that reducing bureaucratic transaction costs nearly doubled the number of claims made through non-clientelist avenues. Treatment also weakened the belief that entitlements must be reciprocated with political support and diminished approval of quid-pro-quo exchanges, two key norms that sustain clientelism. Finally, the intervention reduced partisan identification, especially with traditionally clientelist parties. An important implication of this paper is that reducing the costs that citizens face in interacting with the state can weaken clientelism and bolster individuals' political autonomy.
Does the Built Environment Shape Voter Participation? Learning from Polling Place Imagery in Mexico (with Aditya Dasgupta, Conditionally Accepted, American Political Science Review.)
Abstract: Does the built environment shape political behavior, particularly voter participation in the democratic process? We use computer vision to learn high-dimensional relationships between the built environment and voter turnout in Mexico, where citizens cast ballots in variable physical environments. With a convolutional neural network applied to street-level and overhead imagery of polling places country-wide (𝑁 ≈ 80,000), we show that building imagery contains considerable information useful for predicting voter turnout. Using explainable artificial intelligence tools, we extract human-interpretable features encoded in imagery related to higher turnout, identifying polling place attributes that plausibly mitigated context-specific barriers to civic participation including heat, pandemic, and violence/insecurity. We leverage the quasi-random assignment of prospective voters to polling places within precincts to test for the existence of a causal effect of the built environment on turnout. The findings suggest that the built environment shapes inequalities in civic participation and demonstrates a promising approach to inductive inference from image data.
(paper)
Berinsky, Adam, Tesalia Rizzo, Leah Rosenzweig, and Elisha Heaps. (2018). “Attribute Affinity: U.S. Natives' Attitudes Towards Immigrants.” Political Behavior.
Abstract: We examine the extent to which relevant social identity traits shared between two individuals—what we term “attribute affinity”—can moderate out-group hostility. We argue that in-group affinity is a powerful force in shaping preferences over potential immigrants. We focus on two closely related, yet distinct, dimensions of identity: religion and religiosity. Using evidence from three surveys that included two embedded experiments, we show that sharing strength in religious practice can diminish strong aversion to immigrants of different religious affiliations. We find that, among highly religious U.S. natives, anti-Muslim bias is lower toward very religious Muslims, compared to non-religious Muslims. This attenuating effect of attribute affinity with respect to religiosity on anti-Muslim bias presents the strongest evidence supporting our argument.
(paper)
Distributive Politics
Does Increasing Competition Increase Pro-Poor Preferences? Distributive Preferences of Partisan Brokers in the Yucatan Peninsula (with Mark Schneider)
Abstract: Do local partisan brokers change their distributive preferences when faced with competition? This paper measures local partisan intermediaries’ distributive preferences in 112 villages embedded in a larger field experiment conducted in Mexico that introduced non-partisan facilitators trained to inform and assist citizens in formally requesting government resources, a task that usually falls to a partisan broker. Using a conjoint experiment that measures capture preferred traits of potential recipients of social welfare benefit, we find that brokers in villages with a facilitator are 22% more likely to prefer to target the poorest category of voters than those in the control group. We interpret this change as shift in stated preferences as a result of increased competition for clients specifically in a context where the alternative provider is non-partisan.
(paper)
Political Brokerage as a Profession (with Pablo Balán)
Political brokers are central actors in developing-world democracies, yet little is known about their long-run careers. We fielded an original life-history survey of former precinct chiefs of a major Mexican party, combining two official rosters with a demographically matched sample of party sympathizers and a representative state-level survey. We document that political brokerage can be a stable profession, with average tenures exceeding two decades. Selection into brokerage is driven by early political interest, ambition, and family activism—not by socioeconomic advantage. Contrary to expectations that political engagement yields material returns, brokers are significantly poorer than both sympathizers and the general population and exhibit limited upward mobility alongside substantial downward mobility, especially among women. Politically, brokers develop strong partisan identification but do not transition into electoral office. By recasting brokerage as an occupation, the study broadens theories of political careers and class representation beyond formal officeholders and provides a blueprint for studying hard-to-reach political populations.
(paper available upon request)
Red Tape, Corruption, and Distributive Politics. (with Aditya Dasgupta)
Abstract: This paper investigates the distributive politics of red tape – the time-consuming bureaucratic hurdles attached to the application for government benefits and services – and its consequences for the distribution of access to government welfare programs. A simple model suggests that red tape has both progressive and regressive effects, tending to exclude not only the wealthiest individuals but also, in the presence of bureaucratic corruption, the very poorest individuals. This is because poorer individuals may be willing to pay but are constrained in their ability to pay the bribes required to clear red tape. This provides arbitrage opportunities for clientelist intermediaries that specialize in cutting red tape in exchange for the political loyalty of poor voters. Evidence is provided using household survey data on access to Below Poverty Line (BPL) cards in India.
(paper)
Methodology
AI-assisted Social Science Discovery: An End-to-End Framework for Learning About Human Behavior from Unstructured Data. (with Aditya Dasgupta and Yuhang Zhu)
Abstract: Humans make decisions in complex, high-dimensional environments. We propose usage of artificial intelligence (AI) algorithms for computer vision and natural language processing to assist in exploring this vast search space to discover potential determinants of human behavior that manual variable selection might miss. Our framework involves training a deep learning algorithm end-to-end to predict a target behavior of interest directly from unstructured data such as imagery and text that provides a naturalistic representation of the agent's decision-making environment. We then use explainable artificial intelligence (XAI) tools to discover underlying relationships learned by the algorithm. We provide information-theoretic grounding for the framework and develop new tools for translating machine-learned into human-interpretable relationships. We illustrate how the approach surfaces novel relationships in two cases: understanding how Chinese social media monitors choose what content to censor and how the built environment affects the turnout decisions of Mexican voters.
(paper available upon request)
Militarization
Support for Military Involvement in Elections: The Impact of Normalizing Military Presence in Civil Affairs on Democratic Processes in Mexico (with Rodrigo Castro Cornejo, Abby Córdova and Víctor Hernández Huerta)
Abstract: This paper examines how the growing military presence in Mexican civilian life impacts democratic values and processes. Over the past two decades, the Mexican government has expanded the military’s role to include distributing social welfare, safeguarding elections, and patrolling streets. We argue that increased exposure to the military normalizes its presence and boosts support for its involvement in civilian domains. Using a survey experiment embedded in a nationally representative sample of Mexicans, we measure how exposure to images of military personnel in civilian roles shapes public attitudes. Our findings reveal that priming respondents with pictures of the military safeguarding ballot transportation, patrolling streets, and providing social welfare significantly increases support for their presence inside and outside polling stations during elections and increases respondents’ overall trust in the military. Viewing multiple images of these activities, compared to a single image, further amplifies support for military oversight of elections. Finally, we also find suggestive evidence of a spillover effect where exposure to the military conducting civilian activities raises approval for their involvement in other traditionally civilian domains, such as public works, operating ports, and education. These findings highlight an overlooked mechanism in the militarization of public life: normalization through repeated exposure, underscoring the potential risks of expanding military roles in civilian governance.
(PAP, paper available upon request)
Electoral Politics
How Election Officials’ Political Preferences Affect Electoral Outcomes (with F. Daniel Hidalgo).
Abstract: In Mexico, ordinary citizens are invited to participate as poll workers if the first letter of their last name and their month of birth is randomly drawn. Although the random draw of invitations is meant to form impartial polling committees, accepting to participate can be politically motivated. We match registered poll workers for the 2012 and 2009 elections in a northern state in Mexico to a unique data set of party sympathizers to test whether poll workers’ party sympathies affect election outcomes. Using the proportion of eligible party sympathizers to serve as poll workers, we instrument for citizen poll workers’ political preferences. We find that having a party sympathizer as a poll worker increases the vote share for that party by about 0.4%. Our findings have important implications for poll worker recruitment through lottery systems where compliance can be a problem.
(available upon request)
Motivated Brokers: An Ethnography of Political Brokers during the Electoral Off-Season
Abstract: I use data from in-depth interviews with party leaders and political brokers during the electoral off-season to show that brokers are not only motivated by electoral payoffs. Brokers also procure their clients’ esteem across electoral cycles to secure their positions. I argue that motivation to maximize electoral return during elections differs from the one to maximize their tenure during the electoral off-season. While the first requires strategic distribution according to electoral preferences, the second requires brokers to procure esteem and trust within their communities.
(paper)
Navigating the State as an Undocumented Immigrant: Evidence from a Field Experiment in California's Central Valley (with Nihan Karagul)
This study investigates the impact of providing information on social benefit program eligibility and intermediary assistance on immigrants’ engagement with government services and their trust in government redistribution policies. Given the complex eligibility requirements tied to immigration status, immigrants often find accessing social benefits challenging, leading to a higher reliance on intermediary organizations. We aim to assess whether providing information about eligibility and intermediary organizations who can assist can motivate higher interest in claiming benefits. To do this, we implement a within-subjects factorial survey experiment that randomizes information on eligibility and intermediary organization and estimates its effects on interest in applying, claim-making behaviors, and trust in government. Additionally, we include a simple priming experiment where we randomized a hypothetical person’s citizenship status and estimated respondents’ stated expectations of government
(PAP)