Abstract: Does increased exposure to forced migration affect violent crime rates in developing host countries? To answer this question we exploit the unprecedented inflow of Venezuelans to Brazil. Contrary to fears propagated by the anti-immigration rhetoric, our two-stage least squares estimates reveal that the sudden influx of refugees did not affect violent crime in which natives were victimized. In fact, our results suggest that forced migration only increased violent crime involving Venezuelan victims. Victimization of migrants seems to have increased at a slower pace than their presence in the host country. Yet, it was concentrated among young males between the ages of 15 and 39 living in the border region of Brazil with Venezuela. Evaluating the causal impacts of forced migration in a developing context is crucial to providing governments and international agencies with rigorous evidence to support policy decisions. In absence of the latter, public perception can play a key role as host populations may pressure authorities for anti-immigration policies based solely on perceptions. Moreover, violence hinders migration’s documented long-term benefits by imposing high economic and social costs.

Media coverage: Blog Impacto (FGV)

Abstract: Climate change risk perceptions are subjective constructs that individuals use to interpret the potential harms of climate change and influence their engagement in mitigation and adaptation efforts. While research in high-income Western countries has identified cognitive processes, socio-cultural factors, and political ideology as key predictors of climate risk perceptions, their applicability to low- and middle-income regions remains uncertain. This study uses a cross-national survey conducted in Argentina, Brazil, Colombia, Chile, Ecuador, Peru, and Mexico (N = 5338) to assess climate risk perceptions in Latin America. We show that emotional responses, especially worry and perceived vulnerability to extreme weather, are the strongest predictors. In contrast, political ideology and socio-demographic factors exhibit weak and inconsistent associations, diverging from patterns observed in high-income countries. These findings highlight that climate change is not perceived as a politically divisive issue in the region, suggesting opportunities for cross-party collaboration on climate initiatives. Understanding these unique drivers in regions with emerging economies is crucial for developing effective, tailored risk communication strategies.

Abstract: Despite growing recognition of its importance for human capital formation, access to mental healthcare remains limited. This paper examines whether expanding access to mental health services improves educational outcomes. We study a nationwide reform in Brazil that replaced hospital-based psychiatric care with community-based Psychosocial Care Centers (CAPS), using administrative data covering the universe of Brazilian schools over two decades. Exploiting staggered municipal rollout in a difference-in-differences framework, we find that CAPS introduction reduces dropout by 4–9 percent and age–grade distortion by approximately 2.5 percent, with effects concentrating on enrollment margins rather than conditional pass rates. Educational gains operate through two channels: direct effects on youth, including rising outpatient utilization and declining self-harm; and broader stabilization of the municipal mental health environment, as adult psychiatric hospitalizations fall by 22 percent and interpersonal violence by 21.5 percent. These spillovers reduce the household disruption, violence exposure, and economic stress that children face. Our findings suggest that community-based mental health interventions generate broad social returns precisely because they treat the environments in which children develop, not solely the children themselves.

Presentations: Graduate Research Workshop (Duke University)
Upcoming Presentations: CESifo Junior Workshop on the Economics of Education (Munich), PAA Annual Meeting (St. Louis), and Duke Population Research Institute 

Abstract: We study how visa restrictions imposed by a transit country affect irregular migration to a final destination and legal tourism to itself. In 2021, Mexico reintroduced a tourist visa requirement for Brazilians, a nationality that had been a major source of both irregular migration to the United States and international tourists to Mexico. Using a difference-in-differences design, we compare Brazilians to unaffected nationalities in administrative data on U.S. border encounters and Mexican airport arrivals. The policy reduced Brazilian encounters at the U.S.–Mexico border by more than 75 percent, with larger declines among women and older individuals and a shift toward riskier crossing sectors. At the same time, Brazilian tourist arrivals to Mexico fell by nearly 50 percent, or about 170,000 visitors annually, with losses concentrated among younger travelers and women. Even under the extreme assumption that every irregular migrant first entered Mexico as a tourist, the contraction in genuine tourism remains of similar magnitude. The results show that transit-country visa policies can substantially reduce irregular migration, but only at the cost of large losses in legal mobility and tourism revenues.

Funding: Institute for Humane Studies (grant no. IHS019318)
Presentations: IHS Migration & Citizen Workshop (Rice University), APPAM's 47th Annual Fall Research Conference
Upcoming presentations: RFBerlin Migration Forum, Humans of LACEA 

Abstract: Large-scale forced migration can put pressure on local public services, especially in developing countries with limited fiscal and institutional capacity. This paper examines how the inflow of Venezuelan refugees affects local healthcare outcomes and municipal public finances within Brazil's decentralized universal healthcare system. We exploit cross-municipality variation in refugee exposure  and use distance to the country's only official border crossing with Venezuela as an instrument. Our findings show that municipalities with stronger exposure to the inflow of refugees experience a deterioration in local health outcomes. Furthermore, we find that healthcare spending does not keep pace with the rise in demand in affected municipalities. This reflects the relatively poor health of incoming refugees and congestion in the healthcare system. Municipalities finance the increased healthcare costs using their own revenues, diverting resources from other sectors, particularly education. Our findings emphasize the limitations of decentralized health systems in absorbing large, concentrated migration shocks and highlight the need for stronger national support.

Presentations: Graduate Research Workshop (Duke University), PPGE Seminar (UCB), APPAM's 47th Annual Fall Research Conference
Upcoming Presentations: PAA Annual Meeting (St. Louis)

Abstract: The Paris Agreement relies on reputational enforcement rather than sanctions to induce compliance. Existing research—based largely on advanced democracies—finds that foreign climate shaming generates domestic audience costs for noncompliant leaders and that elite counter-rhetoric rarely attenuates these effects. Using two survey experiments (N ≈ 6,100) and post-experimental focus groups in Brazil, we identify conditions under which this enforcement logic breaks down. We replicate a core finding: foreign shaming reduces incumbent support under partial, but not total, noncompliance. We then show that incumbent counter-rhetoric invoking historical responsibility, reciprocity, and sovereignty fully eliminates—and often reverses—the political costs of shaming. Qualitative evidence suggests that these messages activate distributive and sovereignty-based considerations that lead citizens to reinterpret foreign criticism as unfair or intrusive. Under these conditions, reputational pressure fails to generate audience costs, rendering defection politically costless.