This study investigates the impact of air pollution on hours worked in Chile using a nationally representative sample. We construct downwind smoke plumes originating from wildfires and com bine them with the presence of thermal inversions to causally identify the effect of air pollution. Our findings reveal a 2.2% reduction in hours worked due to increased fine particulate matter from an extra smoky day. Our study suggests that focusing solely on labor productivity underes timates the economic cost of air pollution by 18-43%. The effect is more pronounced for female workers as well as for workers engaged in outdoor tasks. We show a high degree of heterogeneity in the effect across rural and urban areas. In urban areas, high-income workers experience a larger impact of air pollution, but this is not the case in rural areas.
I study the impact of locust swarms on secondary education in Ethiopia. Combining high-resolution spatial data on locust swarm activity with administrative records from secondary schools, I use a difference-in-differences event-study design to identify the causal effects of locust shocks. I find that locust shocks reduce secondary school enrollment and completion by approximately 9.1% and 9.5%, respectively, while increasing the secondary school dropout rate by about 2.9 percentage points. These effects are particularly pronounced in remote and highly agriculture-dependent woredas. Further analysis provides suggestive evidence that these educational disruptions are associated with agricultural income losses, increases in local food prices, and declines in junior secondary school enrollment and completion. These findings highlight how biological disasters can disrupt educational progress in developing countries.
This paper examines optimal environmental regulation based on emission or performance standards in the presence of consumer search and firm innovation. With an exogenous market structure, we find that the emission and performance standards lead to the same equilibrium level of innovation, which decreases in the degree of competition in a Schumpeterian way, and these two policy tools are equivalent in terms of total welfare but not in distributional terms because they give rise to different levels of consumer and producer surplus. Here, low search costs can lead to price-increasing competition in concentrated industries and spillovers determine the time consistency of the standards. With an endogenous market structure, the two regulatory instruments lead to different levels of total welfare, they both yield excessive firm entry, and the magnitude of the inefficiency from excessive entry is relatively greater under the emission standard.
I study whether reforming the delivery of agricultural input subsidies can generate environmental benefits. I exploit Zambia's staggered transition from a traditional physical fertilizer distribu tion program to a flexible electronic voucher system and combine this policy variation with a 5 km & times; 5 km panel of satellite-derived water quality measures from 2010 to 2023. Using a staggered difference-in-differences design, I find that the e-voucher reform significantly improves downstream water quality, reducing chlorophyll-a concentration by 13.0%, phycocyanin concen tration by 11.0%, and total suspended matter by 7.5%. The effects are larger in districts with denser private agro-dealer networks and in areas with highly leachable soils and intensive maize monocropping. Household-level evidence indicates that the reform reduces synthetic nitrogen use and increases crop diversification toward nitrogen-fixing legumes. These responses are consistent with reduced nutrient losses and improved soil retention. Overall, the findings suggest that increas ing flexibility in agricultural subsidies can improve downstream water quality while encouraging more sustainable agricultural practices.
Environmental land registries are increasingly used to support forest governance, yet their po tential contribution to reducing rural land conflict remains poorly understood. Brazil's Cadastro Ambiental Rural (CAR) requires landholders to self-declare property boundaries and land use for environmental compliance but neither confers legal title nor resolves competing land claims. Before becoming mandatory under Brazil's 2012 Forest Code, the CAR was mandated in the Amazonian state of Par & aacute; beginning in 2008. We exploit this early implementation to examine whether an environmental land registry can reduce rural violence in frontier regions charac terized by weak property rights and limited state capacity. Using municipality-level panel data (2001-2012) and a synthetic difference-in-differences design, we estimate that early CAR imple mentation reduced land conflict events by 11.5%, families affected by conflict by 33.4%, and land-related death threats by 10.0%. Murders and attempted murders are too rare for the de sign to detect effects of plausible magnitude. The estimated reductions are not uniform across outcomes or institutional environments. Death threats decline consistently across all estimation samples, whereas reductions in conflict events and families affected are concentrated in municipal ities with complementary governance programs and stronger pre-treatment enforcement capacity. These findings are consistent with environmental registration operating through complemen tary informational and institutional channels. Documenting land occupation appears sufficient to discourage lower-intensity forms of conflict, while larger reductions in realized conflict require institutions capable of acting on the information the registry provides. More broadly, the results suggest that information-based land administration can improve land governance before formal property rights are resolved.
