Thiago da Costa

Concerns about negative impacts of student phone use have led to calls around the world for tighter restrictions on phones in schools. This paper evaluates the impact of a 2023 policy that banned non-pedagogical uses of phones within schools in Rio de Janeiro, Brazil. To isolate the causal effects of the policy, we contrast middle schools that already had strict rules on phone use prior to the policy (“control schools”) to similar schools that did not have strict rules (“treatment schools”), before and after the ban. While restrictions were imperfectly implemented both before and after the ban, we show that in-school phone use fell substantially in treatment schools relative to control. We then show that test scores, which were trending similarly in the two groups prior to the ban, improved by 0.06 s.d. in treatment schools relative to control.

Economic shocks reshape labor markets, yet their political consequences remain underexplored. This paper examines how economic disruptions influence political entry, candidate composition, and electoral outcomes in legislative elections. We extend Caselli and Morelli (2004) model to accommodate heterogeneous shock effects across the ability distribution, enabling analysis of differential impacts on political entry and candidate quality. Under a broad class of equilibria, negative wage shocks increase both candidate numbers and the high-ability share by eroding low-ability candidates' comparative advantage, reshaping entry incentives to favor higher-ability individuals. Empirically, we exploit Brazil's exposure to the China Shock to test these predictions. Using a shift-share design, we find municipalities exposed to import competition experienced sustained increases in political entry, electoral competition, and more educated candidates—driven by deteriorating labor market opportunities consistent with supply-side responses. In contrast, the export boom had minimal effects, highlighting asymmetric trade shock impacts. We further document shifts in candidate demographics and ideology, with import shocks reducing left-wing electoral success. Our findings show economic dislocation shapes political representation by altering the candidate pool, independent of voter demand responses.

Many studies compute correlations between variables aggregated at different levels. We show that ecological fallacies can arise not only from inferring individual-level correlations from group-level ones, but also from analyses that combine outcomes measured at fine levels of disaggregation with explanatory variables available only at higher levels — characterizing the conditions under which such mismatch can even reverse the sign of underlying population parameters. We illustrate the phenomenon with unique school-level data on school staff's implicit racial preferences, measured through implicit association tests (IATs) in a representative sample of Brazilian K–12 schools. In the data, the variance of commuting-zone-average IATs is about 70% lower than that of school-level IATs; aggregation induces non-classical measurement error in racial preferences; race gaps in disciplinary measures are not systematically associated with commuting-zone-average preferences; but a significant association between race gaps and racial preferences — with the opposite sign — emerges when using school-level IATs instead.

While an extensive literature studies the fluidity and inconsistency of gender and race classification, this paper quantifies the phenomena first-hand for disability classification, documenting its causes and consequences during childhood and adolescence. Drawing on longitudinal school census data, we document that disability classification fluidity lies within the same range as racial fluidity among Brazilian students, corroborating the claim that disability is also socially constructed. Using novel data from a representative survey of Brazilian K–12 schools that asked students directly about their sensory and motor challenges, we further document that official statistics hugely understate the share of students with functional disabilities, and that school-reported prevalence does not even help predict students' self reports. Interviews with school staff reveal a multiplicity of biases that render students with major difficulties to see, hear, move, or carry objects — but lacking formal medical reports — ‘invisible’ in official statistics. Among severely delayed middle-school students in São Paulo State, 37% experience vision, hearing, or motor challenges, a share more than four-fold its national prevalence and increasing in students' learning gaps. Leveraging a cluster-randomized controlled trial of a remedial tracking intervention, we show that the strictly pedagogical intervention helped students with functional disabilities significantly less.

This report analyzes federal healthcare expenditures on treating firearm injury victims in Brazil. In 2024, 15,800 hospitalizations for firearm injuries consumed R$42.3 million in federal health resources. Over the past decade (2015-2024), total federal spending on these hospitalizations reached approximately half a billion reais (R$556 million), averaging R$56.6 million per year. The report documents persistent patterns of victimization, with 89% of patients being men, 82% being Black individuals, and 52% being young people (15-29 years). Aggression remains the primary cause of firearm-related hospitalizations (77.3%), with significant regional disparities—the Northeast accounting for 42% of hospitalizations in 2024. The analysis reveals that treating firearm injuries is costly: the average hospitalization cost is 2.6 times higher than federal per capita health spending, and hospitalizations due to armed assault cost approximately 80% more than those for other types of assault. These findings underscore armed violence as a major public health issue requiring integrated policy responses from both health and security sectors to strengthen control over firearms circulation.