"COVID-19 Lockdowns and Domestic Violence: Evidence from a Domestic Violence Hotline in Argentina" (with Enrique Carreras), 2020 (Chapter 1 of IDB Technical Note Nº IDB-TN-1956, link)
The Government of Argentina responded to the declaration of the COVID-19 pandemic by decreeing a quarantine that came into effect on Friday, March 20, 2020. We study the impact of this lockdown on domestic violence reports using information from calls to a domestic violence hotline in the City of Buenos Aires. Using a difference-in-differences strategy to account for seasonal variations in the number of calls, we find an increase of 32% (almost 6 calls per day). Breaking down the calls by type of violence, we find that this increase is mainly explained by a spike in reports of psychological violence. We exploit that the hotline is used not only by victims but also by other institutions (which report incidents brought to them) to assess the change in reporting channels due to the lockdown. We find a large substitution between reporting channels: calls to the hotline received from the police and other institutions fall sharply (62%) while direct calls from the victims increase by 127%. The results underscore the need to complement the mobility restriction measures imposed to combat the pandemic with specific services to respond to the domestic violence that the lockdown generates.
Competition for public office is an essential feature of democracy but having many candidates competing for the same position might lead to voter confusion and be counterproductive. In current democracies, ballot access regulations limit citizens' right to become candidates, seeking to balance this trade-off by discouraging frivolous contenders. This paper examines the causal effect of signature requirements -- a widespread ballot access regulation -- and finds that their impact goes beyond this goal. I use data on Italian local elections and apply a regression discontinuity (RD) design to estimate the effects of these requirements on electoral competition, candidates' selection, and voter participation. I find that signature requirements reduce the number of candidates running for office, decrease electoral competition, lead to a more experienced pool of candidates, and reduce voter turnout. The positive effects of this policy are observed in municipalities with fragmented political systems, where signature requirements lead to fewer wasted votes and fewer spoiler candidates. The downside is observed in municipalities with concentrated councils: signature requirements increase the frequency of uncontested races and reduce voter participation. Findings reveal how this barrier to entry impacts key dimensions of democracy and indicate that designing efficient electoral institutions requires a clear understanding of local political contexts.
"Using Centralized Assignment to Evaluate Entrepreneurship and Life-Skills Training Programs in Argentina" (with Diego Ubfal), 2018.
We implemented a centralized assignment mechanism (random serial dictatorship) to fill seats for a set of training courses offered by the government of Buenos Aires in schools across the city. The mechanism generates a seat assignment that is random conditional on people's preferences over schools. We exploit the fact that we know these preferences and the assignment rule to analytically compute individuals' propensity scores (i.e., their probability of obtaining a seat). We then use propensity score stratification to evaluate the short-term impact of two of the courses (an entrepreneurship course and a life-skills course). This strategy leads to important sample size gains relative to full preference stratification, allowing us to fully exploit the random variation in treatment assignment. Using survey-based information collected three months after the courses, we find positive effects on course-related knowledge for both programs. Entrepreneurship training helps participants start a business from an initial idea, leading to increased business ownership and self-employment. Life-skills training leads to higher job-search rates. There are no significant effects on soft skills.
Work in Progress
"Do Women Compete like Men? Evidence from the Track" (with Samuel Berlinski).
We use information on middle-distance races to obtain insights on how men and women compete in gender-segregated, high-stakes settings. We document important differences in competitive behavior across genders and examine the empirical validity of different theories by estimating a model of behavior in tournaments with interim-feedback.
"Paying Politicians: A Semi-Structural Approach".
I examine how the wage paid to politicians affects citizens' decision to participate in politics and their performance in office. I estimate a political agency model to retrieve unobservable characteristics of the political environment, such as the quality of politicians and the proportion of office-motivated and intrinsic-motivated candidates. I fit the model using data on administrative efficiency in two samples of Italian municipalities on each side of an arbitrary population threshold established by law to determine a jump in mayors’ remuneration. Results show that higher wages improve candidates’ managerial ability but reduce their public service motivation. Relative to the existing empirical literature on the topic, the semi-structural approach in this paper serves to better quantify and disentangle the different mechanisms at play, providing a stronger link to the theoretical literature.
"Seeds and Roots of Prejudice: Lessons from Contemporary Africa" (with Alexsandros Cavgias).
We provide original descriptive evidence on the correlates of prejudice towards homosexuals in the African continent. We document a negative and significant relationship between material conditions and sexual prejudice. We also find a non-monotonic relationship between education and prejudice: relative to the population with no formal education, people with completed primary education display higher rates of prejudice. On the contrary, post-secondary education is associated to lower levels of prejudice. We finally test the hypothesis that exposure to Christian missions increased individual-level prejudice toward homosexuals in contemporary Africa. We find evidence against this hypothesis: exposure to missionary activity at the village level has no significant relationship with individual level prejudice toward homosexuals.