Research

Working Papers

Settlers and Seekers: Immigrant Proximity and Voter Polarisation (with C. Berti Ceroni, L. Piemontese, G. Prarolo) [under review]
CEPR DP19236

This study explores the crucial role of local immigration history in shaping the relationship between the arrival of refugees and support for anti-immigration parties. Using unique georeferenced individual-level survey data, we show that opening refugee hosting facilities close to voters polarises support for anti-immigration parties: in neighbourhoods with established immigrant communities such support reduces, while it increases in areas with fewer long-term immigrants. These findings are consistent when using official electoral outcomes. We extend the analysis beyond the direct effect of exposure to immigrants by examining the role of voters’ demographics. Our results are mainly driven by male, childless, and low-skilled working-age voters. Moreover, facilities that highlight the presence of short-term immigrants amplify the main effects. These findings emphasise that, beyond individual differences, the composition of the immigrant population itself is crucial to fully understand the complex dynamics through which contact affects voters, particularly the interaction between long- and short-term immigrant presence and the mediating influence of voters' demographics.

The Quiet Payoff: Mafia Electoral Support and Policy Inaction (with A. Carrozzo Magli and G. Righetto) [under review]
Awards: SIEP Prize 2025
NICEP WP

Organized crime groups are known to provide electoral support to politicians, but the rewards they obtain in return remain poorly understood. We develop a theoretical framework suggesting that modern mafia support hinges on parties’ willingness to weaken anti-mafia policies, specifically by neglecting the reallocation of confiscated mafia assets. Judicial records indicate that when these assets remain unassigned, crime families can quietly repossess them, turning policy inertia into a hidden payoff. Using data from Sicilian municipalities between 1992 and 2022, we first detect vote manipulation in tightly contested majoritarian races—particularly in smaller towns—indicating strategic vote buying by the mafia. A regression discontinuity design, restricted to comparable municipalities quasi-randomly sorted around the threshold, reveals that narrowly won Forza Italia victories trigger a sharp fall in asset reallocations only within mafia-controlled areas. To measure variation in vote‑buying capacity, we exploit the mafia’s abrupt 1987 withdrawal of support from the Christian Democrats. Municipalities suffering larger DC vote losses—our proxy for historical mafia influence—experience steeper post‑election cuts in asset reallocations, but only during Berlusconi’s governments. Instrumenting modern Forza Italia support with these historical shifts further supports a causal relation between mafia vote buying and national‑level policy concessions.

Can data disclosure improve local governments' performance?  Evidence from Italian municipalities   (with B. Lockwood, F. Porcelli, M. Redoano) [under review]
CESifo Working Paper No. 10155

We exploit the introduction of an open data online platform, OpenCivitas (OC), introduced by the Italian Government in 2014 for municipalities in ordinary regions, as a natural experiment to analyze the effect of data disclosure on mayors' expenditure and public good provision. We investigate the effect of mayors' attention to data disclosure within treated (ordinary) regions by tracking their daily access to the platform, which we instrument with the day of release of newspaper articles mentioning municipality expenditure in OC. We find that mayors react to this disclosure by decreasing spending, primarily expenditure on administration. Attention also leads to a reduction of service provision, and a small decrease in efficiency. We use data disaggregated by service to further investigate the channels. Finally, provide an "extensive margin" analysis by comparing municipalities on the border between treated and untreated (special) regions. We find a similar pattern of behavior to the "intensive margin" i.e., mayors in treated regions decrease expenditure on administration. These findings are consistent with the principal-agent literature on the unintended consequences of increased transparency.  

Networks and Tax Setting in the Digital Age: Evidence from Italian Municipalities (with B. Lockwood, F. Porcelli, M. Redoano)   [under review]

This paper examines the impact of OpenCivitas, a data disclosure program introduced by the Italian government that allows mayors to access detailed expenditure data of other municipalities through a dedicated website. Through the analysis of website views, we construct a directed network and investigate the characteristics and strategic behaviors of mayors in this network. Mayors in the network are on average younger, more educated, often from larger cities and are more likely to be affiliated to traditional parties, although populist parties are the ones usually relying more on the web for communication and political activities. Using directed dyadic models, we show that mayors tend to form links with those who share similar age, city size, and region. Interestingly, links are more likely to be formed when mayors differ in gender, education, and party affiliation. Mayors in this network do not engage in yardstick competition with neighbouring municipalities as the other mayors do, and rather compete with each other, despite the physical distance. We find that the network existed prior to the website opening, but after data disclosure, yardstick competition within the network becomes strongly driven by mayors who are up for re-election, which was not the case before. For the other municipalities, yardstick competition between neighbours remains uncorrelated with mayors' term limits.

The Democratic Costs of Lobbying: Policy Quality and Participation (with A. Carrozzo Magli and G. Righetto) [under review]

We develop a probabilistic-voting model to study how lobbying shapes both policy outcomes and democratic participation. In our model, a lobby provides blocs of votes in exchange for weaker policy commitments, but departing from existing frameworks we allow partially ideological voters to abstain when policies tilt too strongly toward lobby interests, thereby linking lobbying not only to policy concessions but also to voter turnout. The model yields a paradoxical prediction: stronger electoral competition can worsen policy quality and depress participation, as parties increasingly depend on lobby support to secure victory. We examine these predictions in the context of Mafia lobbying in Sicily between 1994 and 2018, exploiting variation from close national contests between Forza Italia and the Democratic Party of the Left. Consistent with the theory, tighter competition in Mafia-influenced municipalities led to fewer reallocations of confiscated Mafia assets—one of Italy’s central anti-mafia policies—and to lower electoral participation. Our findings highlight a perverse consequence of democratic competition: when votes can be mobilised by powerful lobbies, more competition may erode rather than enhance accountability and participation.

Selected work in progress

Red "Scare" or Red "Scar"? Early Voting and Habit Formation in Ideologically-Charged Elections (with C. Berti Ceroni, G. Bellettini, R. Di Leo)

CIVITAS: Councils In View: Italian Transcripts and Analysis System (with G. Romani, A. Saia)

From Rates to Ranks: How Tax Systems Alter Perceptions of Income Status (with D. Diez and J. Boccanfuso) 

Questioning the Role of Marginal Tax Rates for Compliance