In Progress

Existing scholarship on the legacies of state violence treats families as the primary vehicles through which repression shapes political behavior. Yet individuals are embedded in multiple social groups; a robust literature shows that both kin and non-kin relationships are consequential for political socialization. This paper examines the comparative effects of direct repression and the repression of family members, friends, and community members on subsequent political behavior. We draw on an original nationally representative survey of 1,200 adult Tunisians, supplemented by a targeted sample of 175 sibling pairs. We find that individuals who reported themselves or family members as repressed under the Ben Ali and Bourguiba dictatorships (1956–2010) are significantly more likely to report protest participation in the past year. Those who reported repressed friends and community members are more likely to report both recent protest attendance and electoral turnout. Analyses of the sibling pairs show that the relationship between repression of friends and communities and protest participation remains significant when controlling for unobservable family characteristics. Additional analyses indicate that, similar to family socialization, political discussion with friends is a key mechanism through which repression's behavioral effects are transmitted. Taken together, these findings suggest that scholars who focus narrowly on familial networks risk underestimating the broader reach of repression's political legacies.

• Working paper

Until recently, the empirical study of ideology in authoritarian regimes has received little attention. In these contexts, can ideology be used to describe the ways in which ordinary citizens organize their political attitudes and beliefs? Do these attitudes and beliefs exhibit structure and constraint? Answers to these questions in advanced democracies rely heavily on latent variables and their connection to ideological and partisan self-identification. But in many authoritarian regimes, the absence of clear political alternatives (along with weak and fluid partisan attachments) suggests a need for greater theorizing and a new set of tools. Using original data from a large-scale nationally representative survey, I study ideology in the Islamic Republic of Iran. I draw on recent advances in psychology and sociology to introduce an understanding of ideology as a network. I find that citizen preferences are organized along four dimensions: democratic/authoritarian institutions, secularism/Islamism, internationalism/nationalism, and gender inclusivity/exclusivity. Those who prefer authoritarian institutions, Islamism, nationalism and gender exclusivity are more likely to self-report support for the regime and exhibit greater ideological constraint. These findings suggest that political preferences are organized and constrained, but heterogeneous — and speak to the regime's informational advantage in sustaining popular support for authoritarian rule.

Recent waves of protest have given rise to movement parties, or parties with strong ties to social movements that combine anti-system activism with opposition to governing parties and elites. We study the effects of Lebanon's 2019 October Revolution, a series of nationwide protests that took place over several months and targeted an extractive elite that has governed for two decades. Difference-in-differences estimates show that localities where protests took place were more likely to support movement parties in the election that followed the uprising. We attribute this effect to the exclusionary nature of the electoral system and show that support for movement parties was greatest in areas neglected by existing parties. Our findings contribute new insight into the relationship between protest movements and electoral behavior. Movement parties persistently struggle to channel anti-government, revolutionary opposition into support for electoral alternatives — but not where they are able to exploit the weaknesses of existing parties.

• Working paper

Do incumbent autocrats benefit from early elections and premature dissolutions of legislative bodies? Building on insights from the literatures on opportunistic election timing, cabinet duration and durability, and legislative bargaining, we argue that assembly dissolutions in authoritarian regimes can generate substantial political backlash. Our empirical approach focuses on the consequences of assembly dissolution on electoral and legislative behavior in Kuwait, where the Emir enjoys broad dissolution powers — despite a constitutional requirement that obligates him to submit his policy agenda for legislative approval. Our analysis draws on 60 years of electoral and legislative data from Kuwait. We find that assembly dissolution leads to greater electoral support for more active legislators and prompts legislators to oppose the government more frequently. Our article challenges the conventional wisdom that autocrats benefit from manipulating electoral and legislative institutions. Closer attention to legislative and electoral behavior in electoral autocracies — rather than macro-level regime outcomes — indicates that executive attacks on legislatures can often backfire in ways that the existing literature has not previously explored. As a greater number of authoritarian regimes crack down on democratic institutions, our theoretical framework provides new insight into the consequences of executive aggrandizement and democratic backsliding in non-democracies.

Who publicly expresses support for authoritarian regimes while privately holding different views, and why? Existing work documents sizeable gaps between direct and indirect measures of regime support across authoritarian contexts, but has yet to produce a systematic individual-level account of overreporting that places competing mechanisms on a common theoretical footing. We organize the predictors of overreporting into three dimensions: normative and expressive identification (national pride, religiosity), which generates felt obligations to affirm the regime; behavioral compliance and conformity (fear of repression, self-censorship, social deference), which imposes external costs on candor; and cognitive and material resources (education, income, political awareness, efficacy), which enable citizens to form and voice independent judgments. We test our framework using a pre-registered double list experiment embedded in a nationally representative telephone survey of 1,999 adults in Iran, conducted in February–March 2025. Fifty-seven percent of respondents express trust in the authorities of the Islamic Republic when asked directly, but only 47 percent do so under the anonymity of the list experiment—implying that roughly one in ten respondents, or 18 percent of apparent supporters, overreport. Consistent with our framework, religiosity and national pride are the strongest predictors of overreporting.

• Pre-analysis plan

Do religious holidays shape legislative cooperation across sectarian lines? This paper examines whether Ramadan and Muharram — two Islamic holidays that prime religiosity in distinct ways — increase voting agreement between Sunni and Shia members of the Kuwait National Assembly (KNA) relative to non-holiday periods. Using roll-call vote data spanning 16 parliamentary terms from 1963 to 2013, we construct dyadic agreement measures for all MP pairs within each term and exploit the quasi-random overlap between Kuwait's legislative calendar and the Islamic lunar calendar as a source of exogenous variation in holiday exposure. A difference-in-differences design with dyad and term fixed effects isolates within-pair, within-term changes in cooperation attributable to holiday periods. We find that cross-sect legislative cooperation increases during both Ramadan and Muharram, and that this effect is concentrated among dyads where both legislators are Islamists, for whom religious priming is most directly activated. The effect is further concentrated in social policy votes, where religious identity is most directly implicated. These findings provide the first individual-level behavioral evidence that sacred times alter cross-sectarian legislative cooperation in an authoritarian parliament, contributing to debates on religious identity, political behavior, and institutional politics under autocracy.

This paper demonstrates how open-source large language models (LLMs) can extract structured career data from Arabic CVs, enabling systematic analysis of political elite networks across time. Studies of elite networks have traditionally relied on labor-intensive manual data collection, challenges that are particularly acute in authoritarian states where institutional opacity obscures the informal networks through which power operates. Recent scholarship shows the promise of gathering elite data computationally, but most existing work examines English-language sources using proprietary models. Critical methodological questions remain: Can open-source models match proprietary performance? How do prompting strategies affect extraction quality? We address these questions using 1,087 Kuwaiti political elite CVs from biographical directories spanning 1997–2012. Using the open-source Qwen3-4B-Instruct model, we compare information extraction quality across different prompting approaches, from minimal baselines to highly engineered prompts that incorporate domain-specific rules for institution name standardization and inference logic. Enhanced prompts achieve a 97.5% success rate and reduce missing institutional affiliations by 50% — critical improvements for generating network data. Post-processing recovers all remaining failures, yielding 100% coverage and 6,401 structured career entries. While our contribution centers on the extraction pipeline, the resulting data enable several analytical pathways central to elite network research. The structured career histories support the construction of institutional co-membership networks spanning ministries, the military, business, academia, parliament, and the diplomatic corps — relationships especially relevant in authoritarian settings where informal ties often cross organizational boundaries. The data also support career sequence analysis to identify pathways to influence and allow researchers to link institutional experience to legislative behavior. By producing high-quality structured data at scale, our approach provides an empirical foundation for rigorous, replicable elite research in the Middle East and beyond.