Conference Agenda

This paper addresses the challenge of accurately modeling and estimating climate change damages. Time series approaches rely on weather shocks, while cross-sectional analyses capture climatic differences but suffer from omitted variable bias. Climate is defined as the statistical pattern of weather that persists over time and allows for adaptation, unlike unpredictable weather realizations. To assess econometric approaches, I (i) integrate forward-looking adaptation into a full-fledged integrated assessment model of climate change permitting an analytic solution and (ii) generalize the insights based on a dynamic stochastic envelope argument. I show how a carefully designed time series (or panel) estimation strategy can comprehensively identify the costs of climate change, including the indirect identification of unobserved adaptation costs. The paper also presents the first explicit formula for the social cost of carbon under forward-looking adaptation. This result is not only insightful in its own right but also valuable for clarifying and refining prevailing envelope-theorem arguments in the literature and for emphasizing that adaptation costs are part of the social cost of carbon.

Does land development adapt to changing climate risks? Extreme precipitation increases flood risk, yet land-use decisions may overlook rare events. Drawing on nearly half a century of climate and land-use data for Europe (1975–2020), this paper finds that extreme precipitation slows the expansion of built-up areas, though only slightly. The effect is stronger over longer horizons, consistent with a clearer climate signal. Adaptation is concentrated in fast-growing urban areas. An accounting framework that combines these effects suggests that relocation in response to shifting precipitation patterns has reduced damages by roughly 5%.

We study the evolution of majority preferences for climate policy in response to rising economic inequality and climate risk. Households differ in age, income and beliefs, forming blocks to vote on preventive policy. Beliefs on climate risk converge over time, while beliefs on public effectiveness persist and lead to persisting disagreement. The initial majority favors a low intervention level, slowly rising as climate risk rises. The political majority shifts at a tipping point driven by young skeptical voters exposed to rising climate damage, unless beliefs are very far apart. Loss of habitat leads to a rising ratio of house prices to income. A second tipping point may arise in response to rising inequality, though its effect on policy is ambiguous. The political conflict combined with the "tragedy of the horizon" effect produces underinvestment under any coalition.

Climate-related disasters such as floods are increasingly severe due to climate change. This is particularly the case in cities witnessing rapid urbanization and widespread poverty. In these settings, mobilizing local communities for flood prevention may be key to economic development. This paper follows a community-level intervention in a Mozambican coastal city with increasing population pressures and recurrent floods. It nudged communities and leaders to select and implement their own climate adaptation projects. This was achieved through the organization of a sequence of community meetings and the provision of technical information, validated by local and international water experts and made specific to the circumstances of each block of the city. The technical information included details of suitable small-scale infrastructural improvements, e.g., drainage canals, as well as of appropriate waste management. The intervention also offered seed grants to help local resource mobilization. We employ a clustered randomized controlled trial design with multiple survey rounds, supplemented by data from structured community activities, enumerator observations, photographs, and focus group discussions. Our results during/after the first rainy season following the intervention show that the treatment shifted attitudes toward personal responsibility and collective efficacy of local block leaders, who are seen as central to enabling impact, and improved observed localized cleanliness. We also register less flooding in flood-prone points and increased walkability. We conclude that a community-driven approach to disaster risk prevention can be effective. These are preliminary results based on short-term outcomes; a further round of endline data collection is pending.