Working Papers

Drafts available on request.

The Discretionary Color Line in International Institutions

Abstract

PDF

A growing literature documents racial hierarchy in international institutions. I test whether formally equal rules eliminate that hierarchy in practice. The Schengen visa regime offers a hard case because member states process applications under identical law. I show that ancestral distance, a measure of perceived racial difference, predicts visa refusals. Contemporary ancestral distance drives the result. A pre-1500 measure does not, consistent with the social construction of race. A consulate-level analysis isolates the discretionary channel. Busier consulates discriminate more against ancestrally distant applicants, even though a nationality's risk profile does not change with consular workload. Routine operational pressures activate the categorical shortcuts that institutional design was supposed to prevent. A further test shows the effect is stronger where national populations hold more restrictive racial attitudes, not where institutional capacity is weakest. Harmonized rules do not eliminate international institutional racism. They push it into discretionary gaps that formal law cannot close.

Reliable Panel Regression: A Default Workflow for Slow-Moving, Mismeasured Variables

Abstract

Political scientists use two-way fixed effects (FE) as the default specification for panel regressions, on the reasoning that any coefficient surviving FE has been purged of time-invariant confounding. But many of the variables in those models—Polity, V-Dem indices, log GDP, trade share, oil rents—are slow-moving and measured with error. Under those conditions, FE does not purify the cross-sectional estimate; it strips out the signal (which lives between countries) and keeps the noise (which sits in the within-unit variation). Two-thirds of country-year variables in the Quality of Government dataset sit in this range. Larger panels make the problem worse because increased statistical power tightens confidence intervals around a biased point estimate. This paper develops a default workflow for the modal observational panel regression in political science. A sign test built on the empirical intraclass correlation of the regressor and the agreement between pooled OLS and fixed effects picks between random effects, fixed effects, and partial-identification bounds. The paper then applies this workflow to three recent panel-data papers and shows examples where it rescues and corroborates the findings.

Traces of the Past: Regime Histories and Anti-Foreigner Violence in Post-Unification Germany
with Hannah M. Alarian and Michael Bernhard.

Abstract

Anti-foreigner violence is a pressing problem in European democracies. Most interventions look to contemporaneous solutions to reduce or stigmatize such violence. Yet this focus fails to consider how historical legacies continue to shape the persistence and geographic contours of violence. We address this possibility, theorizing that divergent memory politics, reconciling with Nazism's violent legacies, influence current anti-foreigner violence. Using novel anti-foreigner violence data between 1992 and 2021, we reveal that past Nazi support remains associated with violence toward foreigners in post-unification Germany. Moreover, we demonstrate that memory politics can mitigate this violent legacy. The evolution of a culture of contrition in the former Federal Republic specifically diminished the frequency of attacks on foreigners, in contrast to the former German Democratic Republic, where the social roots of Nazi-era hatred remained unconfronted. State-level memorial and individual survey data corroborate these relationships, revealing that memory politics can blunt the impact of violent legacies.

Who Deserves to Belong? Public Opinion and U.S. Naturalization Decisions
with Juliana K. Mucci and Hannah M. Alarian.

Abstract

Citizenship is one of the most important and politically sensitive policy issues in democracies worldwide. Naturalization is a crucial aspect of democratic inclusion, as it requires states and their citizens to accept immigrants as fellow citizens. American attitudes toward immigration, in particular, are well-established, with Americans often favoring immigrants from culturally similar origins or those who hold high-status positions. Yet, little is known about Americans' preferences for naturalization. This article addresses this gap by examining Americans' evaluations of United States (US) citizenship applicants. Supplementing an existing study (osf.io/5ctkj), we test this theory in another original conjoint experiment to isolate attitudes toward naturalization in the US. We hypothesize that the "hidden consensus" that has been found to exist across party lines regarding immigration attitudes does not hold in the context of naturalization attitudes, especially in a polarized political climate. This study provides a further nuanced understanding of how Americans' own political identities shape their attitudes toward citizenship.

Races at the Bottom: Local Impacts of Official Discrimination on Election Outcomes
with Conor O’Dwyer and Matthew Stenberg.

Abstract

Why do illiberal parties keep deploying symbolic exclusion as electoral strategy when, by some accounts, voters punish them and, by others, do not? We argue the existing literature treats parties as the unit of evaluation and so misses where the cost actually lands. Using the Polish anti-LGBTQ "zone" resolutions of 2018–2020, we show that voters punished PiS sharply, by sixteen percentage points of council vote share by 2024, but that the punishment was structured to spare the elite agents who made the decision. The PiS mayors who personally signed and championed the resolutions retained office at the same rate as PiS mayors elsewhere. The PiS-affiliated mayoral candidates whose campaigns advertised the affiliation lost ground; those who held PiS membership privately did not. The lost vote did not flow to opposition parties of any ideology, did not produce abstention, did not transmit to four subsequent national elections. It flowed locally, to non-partisan committees, at ballot items where the PiS brand floated free of any personally-known officeholder. The literature has not described this kind of accountability failure. Voters do punish symbolic policy, but the punishment lands at unanchored brand-as-list ballot items in the arena where the policy was made, away from the elites who chose it. Symbolic exclusion remains attractive to illiberal parties not because it is cost-free but because the cost is paid by others within the party.

Voting for Extremists: How Extremist Protests Influence Far-Right Voting Behavior in Germany
with Hannah M. Alarian and Sebastian Elischer.

Abstract

The far-right Alternative für Deutschland (AfD) is a prominent force in German politics, risking social solidarity, increasing anti-democratic attitudes, and advancing exclusionary policies. The AfD's ascent in national polling underscores a striking paradox: the party's popularity persists alongside its growing recognition as an extremist organization. Is extremism attracting voters, or rendering the far-right anathema to the mainstream electorate? This article addresses this question, examining the proximity of the AfD to right-wing extremist civil society and its impact on the party's electoral performance. Using a novel dataset of AfD and extremist organization protests, we show that AfD participation in protests alongside extremist actors significantly increases its vote share between the 2013 and the 2017 elections. However, between 2017 and 2021, cooperation with right-wing extremist groups declined, and so did the contribution of that collaboration to the overall vote share. The findings contribute to the ongoing debate in political science about the impact of extremism on far-right parties.