Conference Agenda

Higher temperatures can increase substance abuse and physiologically exacerbate its effects on human body, increasing risks of impaired cognitive functions and violent behavior. Using administrative crime data and daily temperatures in the U.S. between 1991 and 2023, we show that two public policies regulating substance abuse, the opening of substance abuse treatment facilities and the reformulation of the prescription opioid OxyContin, substantially moderate the impact of temperature on interpersonal violent crimes. We monetize the ancillary policy benefits for intimate partner violence, the most widespread crime in the U.S., and show that regulations targeting substance abuse can be a cost-effective tool for climate adaptation.

In 2004, Italy introduced a national program to address heat-related health risks, combining public awareness campaigns, heat-wave warning systems, and hospital protocols. Leveraging administrative mortality data and high-frequency temperature variation, we show that the program reduced heat-related mortality by more than 57% on days at or above 30°C. To identify the mechanisms, we exploit the staggered introduction of heat-wave warning systems across provinces and show that treated areas experienced substantially larger reductions in heat-related mortality. We further document that information disclosure plays a key role in driving these reductions. Overall, our findings underscore the importance of public adaptation policies that rely on information provision to cost-effectively mitigate the health impacts of extreme temperatures.

Extreme heat is a growing public health risk, yet evidence on heat-related morbidity remains fragmented across countries and difficult to compare because studies rely on different data sources, exposure metrics, and empirical designs. We compile harmonized daily hospitalization microdata from eight Latin American countries covering more than 200 million people during 2015--2019 and merge these records with high-resolution ERA5-Land reanalysis weather data. Using a flexible temperature-bin dose--response framework with high-dimensional fixed effects that exploits within-municipality deviations from typical seasonal temperature patterns, we estimate pooled and country-specific impacts of extreme heat on hospital admissions. We document a striking convergence in aggregate morbidity responses across countries: days in the hottest temperature bins generate large increases in hospitalizations of similar magnitude despite wide differences in climate, income, and health systems. Cause-specific analyses show that extreme heat disproportionately increases admissions for communicable diseases and injuries, while non-communicable and maternal/neonatal admissions do not rise on average. In contrast to cross-country convergence, vulnerability within countries is sharply divergent. Heat-related morbidity increases steeply in low-development municipalities and is close to zero in the highest-HDI areas, and effects follow an inverted-U relationship with urbanization, peaking in intermediate-density municipalities. Together, the results indicate that while heat--health relationships are broadly similar across countries, climate--health vulnerability is shaped primarily by subnational inequalities in adaptive capacity and the built environment, with implications for targeting heat-health preparedness and infrastructure investments.

Natural disasters are becoming more frequent and more intense, presenting new

challenges for climate insurance. To better understand insurance take-up, the literature

focuses on experimental evidence and the short-run impact of actual disasters.

By contrast, we explore the medium-run consequences of a natural disaster. We

use an original survey of Swedish forest owners interviewed three years after Gudrun,

the ‘storm of the century’, which caused extensive damage to southern Swedish

forests in January 2005. We first document the timing of insurance take-up in response

to past hurricanes, then focus on the response to Gudrun. We exploit the

quasi-random spatial nature of the storm while conditioning on owners’ and forests’

characteristics observed through satellite imagery. Our empirical analysis provides

evidence of a persistent effect of actual damages on insurance demand. This is consistent with a long-term impact on subjective well-being, possibly due to capital destruction,

non-pecuniary losses, and fear. This welfare impact is partly mitigated by

owners’ insurance coverage.