We examine the effects of mass schooling on intergroup relations by studying an education reform in India that significantly expanded access to primary education in the 1990s. Using a regression discontinuity design, we find that the reform reduced reports of caste- and religious-based conflict — effects that persist up to 18 years after the program's launch, and are concentrated in religiously mixed localities with a history of Hindu-Muslim violence. We develop a framework in which schooling curbs conflict through bias reduction (via intergroup contact and inclusive curricula) and higher opportunity costs (via improved wages and employment); follow-up evidence favors the bias-reduction channel. Mass education can not only empower individuals but also serve as a scalable strategy for improving social cohesion.

I study how social media shapes political attitudes in Indonesia. In a field experiment among a nearly nationally representative sample (N = 1,502), I deactivated five social media platforms for thirty days, finding the intervention reduced perceived corruption — possibly because treated participants had worse access to salient information about incumbent performance. Deactivation also unexpectedly increased religious chauvinism. Follow-up evidence suggests a time-use substitution mechanism: participants blocked from social media reported greater involvement in religious organizations and more exposure to politicized sermons. I corroborate these findings with independent survey data, showing that respondents in villages with plausibly exogenous mobile-internet coverage report lower satisfaction with government anti-corruption efforts, increased religious tolerance, and reduced participation in religious organizations. The findings suggest that social media's effects on political attitudes are highly context-dependent and may chiefly reflect that which it crowds out.

We introduce the concept of political voyeurism: individuals' tendency to fixate on distant local political events that do not directly affect them. Surveying voters in the United States, we show that approximately one-sixth prefer other people's local news over their own, and we corroborate this with private web-browsing data. We argue that an uptick in voyeurism is partially driving the nationalization of local politics and threatening accountability. Compared to peers who consume similar quantities of news, voyeurs are more likely to be polarized and politically active; analyses of city-council transcripts show voters increasingly invoking distant events when calling for local action. Our findings suggest an overlooked translocal mechanism through which politics may have nationalized in recent years.

We study the mental-health effects of social media deactivation in a pre-registered randomized experiment among a nearly nationally representative sample of Indonesian adults (N = 1,502), comparing a status-quo control group to individual and household deactivation of five social media platforms. Deactivation improved subjective well-being and mental health on average, with effects concentrated among those assigned to household deactivation. Effects depend on users' offline household networks: deactivation benefits users with infrequent within-household interaction but reverses for those with tighter bonds, helping explain disagreement over social media's harms.

We study how economic development shapes bureaucratic selection. We present three stylized facts: bureaucrats are more educated than the average citizen but become less positively selected as countries develop; bureaucrats tend to be more pro-social than other citizens, and this gap widens with development; and bureaucracies become more representative as countries get richer. We propose a simple theory: individuals value the stability of salaried work, which is scarce in poor countries, so the most educated hoard government jobs there; as private-sector salaried work expands with development, credentialed citizens leave for wage premia, and only intrinsically motivated citizens sort into the bureaucracy. Because intrinsic motivation — unlike credentials — is distributed uniformly across groups, representation improves.

Public opinion surveys are an indispensable tool for studying politics in Southeast Asia, but publicly available data are often in short supply in the region. To this end, we introduce SIKAP, a harmonized and open-access dataset of 58 weekly surveys (N = 95,923) conducted in advance of, during, and in the aftermath of the 2024 Indonesian general election. We describe the data-collection procedures and assess the quality of the sample, and we demonstrate its utility by analyzing the effects of two political events on Indonesian voters' attitudes in almost real time.

I study whether and when electoral outcomes generate anti-democratic grievances on the part of candidates. I evaluate this question with a three-wave panel survey of candidates for local legislative office conducted during the 2024 Indonesian election. I find modest declines in both trust in electoral institutions and support for democracy among candidates following the election. The decline in trust is larger for losers and is concentrated among older and male candidates, as well as those with high expectations of victory. The findings improve our understanding of the extent to which elections effectively metabolize grievances, while also unearthing the limited conditions under which politicians develop beliefs antagonistic to democratic norms after defeat.

We surveyed voters and politicians in advance of the 2024 Indonesian election to measure preferences for environmental policy. We find that politicians underestimate voter concerns. We conducted an informational experiment with politicians to correct these misperceptions, and we document evidence of learning but no greater support for policy action. We explore three explanations for why voter preferences do not translate into policy: politicians only consider acting when their initial misperceptions are particularly large; elite capture prevents politicians from implementing environmental protection; and voters prioritize progress in other domains. Our results underscore the multiplicity of challenges facing climate action.

An essay on “political voyeurism” — the pull of digital media toward distant, emotionally charged local political conflicts — and what it means for pluralism and polarization in Indonesia and beyond.

The meritocratic recruitment of bureaucrats is thought to promote “Weberian” values through two distinct properties. The first concerns the use of examinations to sort aspirants on the basis of traits considered desirable for bureaucrats. The second relates to nondiscretionary selection, whereby recruits do not owe their position to a politician's patronage, making them less likely to be captured. Using a lab-in-the-field experiment among aspirants to the Indonesian bureaucracy, I disentangle the relative importance of these mechanisms by manipulating how “public officials” are chosen from groups to perform tasks on behalf of others. I find no differences on tasks gauging commitment to Weberian values across public officials selected via the merit system or lotteries — although both perform better than those selected under patronage. Because both the merit system and lotteries are nondiscretionary, the results question examinations as the ideal tool for selecting bureaucrats.

After previous failed attempts, Prabowo Subianto won the 2024 Indonesian presidential election. We outline the events that unfolded in the run-up to the election before analyzing the drivers of Indonesian voting behaviour based on original survey data. Our analyses underscore the importance of transformations in the generational composition of the electorate: younger voters, especially those born after the return of democracy, are fuelling its erosion with their support of a former general accused of human-rights abuses. Beyond the backing of incumbent president Joko Widodo, we argue that Prabowo's victory was underpinned by a sophisticated social-media campaign, the waning memory of Indonesia's authoritarian past, and a young populace anxious about their economic prospects.

Most global population growth over the next two decades is projected to occur in small- and medium-sized cities in low- and middle-income countries. Expectations derived from the literature on fiscal federalism suggest this is a cause for concern, as larger cities are thought to deliver public goods more effectively than smaller ones owing to economies of scale. Drawing on detailed cross-sectional data from Brazil and Indonesia, we show that smaller municipalities tend to possess more basic public health clinics and schools per capita, scattered throughout their territory, than larger cities. We theorize that the greater prevalence of such facilities in small cities reflects a relative lack of non-state alternatives, fewer “urban” problems, and politicians' greater ability to secure and claim credit for such facilities. Our analysis underscores how jurisdiction size and non-state service provision can affect government services.

I surveyed the universe of recent applicants to the Indonesian civil service to study the effects of high-stakes examinations on political attitudes. Leveraging applicants' scores on the civil service examination, I employ a regression discontinuity design to compare the attitudes of applicants who narrowly failed with those who narrowly passed. I show that the simple fact of failure decreased applicants' belief in the legitimacy of the process and levels of national identification, while increasing support for in-group preferentialism. Applicants who were offered — and accepted — employment in the civil service reported higher satisfaction with the process, greater amity toward out-groups, and higher national identification. Because more applicants fail than pass, these results suggest that civil service examination outcomes may have unintended consequences for social cohesion — particularly where successful applicants disproportionately hail from specific ethnic, racial, or religious groups.

A prominent argument holds that the chief purpose of municipal civil service reform in the United States was to dislodge the overrepresentation of recent immigrants in city government. Using new data on all municipal employees from 1850 to 1940 and employing three research designs, we detect no evidence that the share of local government jobs held by foreign-born whites decreased following the introduction of reforms. Instead, we show that foreign-born whites — Irish immigrants in particular — experienced substantial gains in local government employment, concentrated in blue-collar occupations in small- and medium-sized municipalities. Our results call for a revisionist interpretation of Progressive Era reforms.

This article examines the consequences of the opium concession system in the Dutch East Indies — a nineteenth-century institution through which the Dutch auctioned the monopolistic right to sell opium in a given locality. The winners of these auctions were invariably ethnic Chinese. The poverty of Java's indigenous population, combined with opium's addictive properties, meant that many individuals fell into destitution. I argue that this institution set in motion a self-reinforcing arrangement that enriched one group and embittered the other, with consequences that persist to the present day. Consistent with this theory, individuals living today in villages where the opium concession system once operated report higher levels of out-group intolerance than individuals in nearby unexposed villages. These findings improve our understanding of the historical conditions that structure antagonisms between competing groups.

Despite the introduction of social restrictions designed to stem the spread of COVID-19, many Indonesians continued to attend places of worship, posing a major public-health threat. Using a nationally representative survey, we evaluated the efficacy of messages delivered from different authorities in encouraging Indonesians to worship at home. We find no consistent evidence that public-health messages changed Indonesians' attitudes toward communal prayer or their willingness to forgo it during the pandemic. Among well-defined subpopulations — non-Muslims and supporters of the president — we find suggestive evidence that messages increased the likelihood of praying from home.

Does the election of extremist political parties increase the incidence of violence and intolerance in the constituencies they represent? We test this using the case of Indonesia — whose party system pits secular nationalist parties against Islamist ones — with a natural experiment estimating the effect of Islamist (as opposed to secular nationalist) incumbency on religious violence and intolerance toward minorities.

We present findings from eight nationally representative surveys conducted during the 2019 Indonesian presidential campaign, in which we measured voters' reported belief in prominent pieces of misinformation. Younger, better-educated, and wealthier voters were more likely to believe the misinformation, and these results hold for stories about both the incumbent (Joko Widodo) and the challenger (Prabowo Subianto). These findings represent a significant departure from results in Western Europe and North America.

What can be done to encourage people to condemn intimate partner violence? Looking at Indonesia, I combine electoral data with a large-scale health survey and find that the narrow victory of a female candidate — as opposed to a male candidate — in local council elections leads to a decrease in the share of female constituents who agree that a husband is justified in assaulting his wife. I observe similar results for male constituents, although some estimates are not statistically significant. These results improve our understanding of descriptive representation as a cause, rather than simply a consequence, of changing attitudes.