Working papers

Source: Ilya REPIN. Arrest of a Propagandist. 1878

Governments often introduce limited political reforms to stabilize autocratic regimes, yet institutional openings frequently coincide with political violence. This paper examines when partial democratization pacifies dissent and when it instead reorganizes opposition toward clandestine activity. We study the introduction of local self-government (zemstvo) in late Imperial Russia, exploiting variation in the reform’s adoption across European Russia and its staggered implementation over time. Within treated districts, statutory voting rules generated discontinuous differences in elite control of assemblies. To measure underground mobilization, we construct and geocode a novel dataset of approximately 21,000 anti-Tsarist activists with individual-level information on social background and tactical choices. We find that the average effect of institutional opening on opposition is close to zero. However, in districts where assemblies were dominated by landed elites, reform reduced visible, low-risk dissent but increased clandestine, high-intensity violence. Where political power was less concentrated, institutional opening reduced both forms of opposition. These findings show that partial democratization can reorganize rather than reduce dissent when political inclusion is not credible.

JEL Classification:  D74, E65, I25, N34, O11, P16, R12.

Presented at:  PEDD Workshop University of Muenster 2026, Clio/EHA session ASSA 2026, IOS Regensburg 2025 Annual Meeting

Source: David Rumsey Map Collection

When states forcibly displace populations and then selectively restore their rights, does the displacement or the legal regime that follows determine their long-run outcomes? I exploit a setting that cleanly separates the two: Stalin's ethnic deportations subjected nine groups to the same coercive shock, and the 1956 Khrushchev Rehabilitation Decree then assigned them---by political fiat, not individual merit---to sharply different legal regimes. Drawing on Soviet census data (1926--1989) and a Spatial Durbin Model with instrumental variables, I show that the institutional regime, not the forced displacement, is decisive. Where the state restored rights, white-collar employment among non-Slavic locals rose 7.5% (p < 0.01) and collective farm employment fell 30% (p < 0.01), consistent with the portable-capital hypothesis. Where it sustained exclusion, neither settlers nor the local populations they lived among upgraded over three decades. A severity gradient sharpens the result: in host regions, complementarities emerged where exclusion was partial but collapsed where it was total and permanent---as for ethnic Germans, Crimean Tatars, and Koreans. Tertiary education gains that translate into employment only by 1989 expose a segmented labor market in which legal barriers, not human capital deficits, blocked upgrading. Restoring rights produced convergence; sustaining exclusion locked in divergence.

JEL Classification:  J15, J24, J61, N34, O15

Presented at: the Economic History Colloquium at Humboldt University 2021, the X. ICCEES World Congress 2021 in Montreal, the SES 2022, the Clio 2022, the 6th GCEG in Dublin, CEPR Economic History Symposium 2022, the 8th IAAE 2022, 17th EACES 2022, the 8th Annual Meeting of the Danish and Scandinavian Economic Society, the CEECON 2022 in Berlin, NEUDC 2022, ASEEES 2022, ASSA 2023, RES and SES Joint Annual Meeting 2023.

Working Paper: Enemies within the Gates (March 2026)

This paper studies intergenerational cultural transmission when dynasties embedded in a social network strategically misrepresent their traits. A central planner operates two instruments---centralized schooling and mobility incentives---on a DeGroot network with strategic parental display. Three sets of results emerge. First, a unique Nash equilibrium in displayed traits exists, and traits converge to consensuses determined by eigenvector centrality. Second, long-run traits are necessarily convex combinations of initial traits, so the planner cannot engineer convergence to an arbitrary target; moreover, the planner's instruments decouple, with schooling steering the consensus target and mobility incentives governing only the speed of convergence. Third, under a budget constraint, optimal targeting concentrates incentives on dynasties whose cultural positions load most heavily on the convergence bottleneck. The framework is applied to three phases of East German cultural dynamics: selective pre-Wall emigration, closed-network indoctrination, and post-reunification convergence.

JEL Classification:  C72, D83, D85, Z13

Presented at: 2022 ASREC conference (virtual), 2024 Workshop on Cultural Persistence and Transmission @QU Belfast

 Working Paper: Engineering of Consent (March 2026) 

We present “Can You Run Ireland,” a gamified workshop that uses economic history to introduce secondary-school pupils aged 14–17 to economic reasoning. Working in teams representing different government departments, pupils make budgetary and policy choices at five turning points in modern Irish history. Each decision contrasts the historical path taken with a counterfactual alternative, asking pupils how a different choice might have changed economic outcomes. We explain the pedagogical rationale, classroom design, and practical requirements, drawing on workshops delivered to over 800 pupils across 24 Irish schools. We identify lessons for implementing and adapting the workshop to different contexts.

Keywords: economics education, economic history, gamification, secondary education, economic literacy.

JEL Classification: A21, H50, N00.

Political violence does not always strengthen the forces of order. When anti-state violence makes insecurity salient, voters may demand authority, hierarchy, and repression, but attacks can also weaken right-wing forces when they reveal that the state cannot provide the security those forces claim to defend. Using newly digitized records from the Paris Okhrana archive and electoral returns from the 1917 Russian Constituent Assembly election, we study this mechanism in the Pale of Settlement. A shift-share exposure design shows that political assassinations shifted the urban-rural electoral gradient away from the ideological right. The effect is strongest for right-wing vote shares and a rightmost-party ideological measure and survives finite-sample spatial inference and controls for urbanization, Jewish population shares, pogroms, geography, wartime front distance, literacy, industrialization, serfdom, and military/Cossack proxies. Political violence can therefore erode the electoral foundations of order politics when it undermines the enforcement capacity of the state.

JEL Classification:  N33, N43, P20, P26, P37, P48, P51

Presented at:  2023 Summer Workshop in the Economic History and Historical Political Economy, Max-Planck PolEcon Summer Workshop, EHES 2023, EHA 2023, ASSA 2024, EHS 2024, EEA 2024, GSWG/VfS congress 2025, EPSA 2025

Working Paper: Upon Request