Conference Agenda

As climate change exacerbates weather shocks, there is growing interest in understanding whether social assistance programs can support coping among poor rural households and whether program effects vary by gender. We assess whether a social assistance program – the Transfer Modality Research Initiative (TMRI) –

influenced the effects of prior monsoon flooding on household consumption and adult diets in southern Bangladesh. TMRI provided cash or food transfers, with or without nutrition behavior change communication, from 2012-2014. Within the study sites and years, flooding was substantial but moderate. Our findings suggest that, without TMRI, a one-standard-deviation increase in lagged flooding led to households smoothing consumption by drawing down savings and reducing diet quality among both men and women. In contrast, among TMRI treatment households, lagged flooding did not reduce savings, and both men’s and women’s diet quality improved. Effects on diet quality appeared largely driven by legumes and by fruits and vegetables, and improvements appeared strongest among households receiving both transfers and behavior change communication. Results indicate that social assistance can help households cope with effects of moderate flooding in southern Bangladesh, protecting household savings and improving both men’s and women’s diets.

Climate-induced inland floods represent a growing threat to U.S. communities, accounting for nearly one-quarter of all billion-dollar weather events since 1980. While coastal flooding has received substantial attention, the impact of inland floods on community composition and housing tenure remains understudied. This paper examines how severe inland flooding affects homeownership rates and migration patterns, linking tenure decisions with relocation choices to understand how floods restructure local demographics. We combine a detailed case study of the catastrophic 2008 Cedar Rapids flood with a nationwide county-level analysis spanning 2006-2022 to assess both immediate and longer-term effects. Using a spatial regression discontinuity design that exploits sharp inundation boundaries, we find that homeownership rates in flood-affected areas within the 100-year floodplain declined by 12.2 percentage points in the two years following the Cedar Rapids disaster. This decline was driven by two mechanisms: heightened outmigration among owners and selective in-migration of renters. Incoming residents from outside the affected county were 80% less likely to purchase homes immediately after the flood, though this effect dissipates within one year. The nationwide analysis confirms these localized patterns at scale, revealing a 2-percentage-point decline in county-level homeownership rates between years 2-5 following extreme precipitation events, with effects gradually reverting to baseline thereafter. Our findings suggest that flood-induced tenure shifts are substantial but temporary, as risk perceptions fade during recovery. However, the concentration of rental housing in hazard-prone neighborhoods raises concerns about wealth accumulation and inequality. These results underscore the need for recovery policies that address vulnerable populations, strengthen flood-risk communication beyond mapped floodplains, and provide equitable pathways for climate adaptation.

Efficient adaptation to climate change requires the spatial sorting of the population away from high-exposure zones. We investigate whether political polarization introduces frictions into this sorting process, leading to maladaptation. Using a novel panel of 13 million registered voters in Florida (2022–2024), we document a widening partisan gap in exposure to flood and hurricane risk. We find that this divergence is fueled by asymmetric adaptation: voters who switch from Republican to Democrat move to census tracts with significantly lower expected annual losses, whereas those switching from Democrat to Republican exhibit much smaller elasticity with respect to risk. These patterns hold even for inland and riverine flood risks, suggesting that the sorting is not driven solely by coastal amenities. The results point to a "separating equilibrium" where high-risk areas become increasingly politically homogeneous, complicating the political economy of disaster relief and insurance reform.

Natural disasters create environments in which information and misinformation travel at extraordinary speed, potentially increasing hazards and general insecurity. We analyse more than 160,000 tweets posted during and after the 2024 DANA in Valencia (Spain), finding that more than 10% are related to mis/disinformation issues. Tweets classified as mis/disinformation show consistently higher levels of user interaction than those from other categories. On average, misinformation tweets receive more retweets (22.16 vs. 14.6), likes (52.09 vs. 37.2), replies (2.62 vs. 1.9), and views (4013 vs. 3426). These differences indicate that, despite representing a smaller share of total tweets, mis/disinformation-related content achieves greater visibility and generates stronger audience engagement. This pattern suggests that mis/disinformation is disproportionately amplified within the broader conversation, reinforcing its potential influence beyond its numerical prevalence. Our results support the need for crisis-sensitive algorithmic policies to prevent further impacts related to climate extreme events. These might include temporary mechanisms to dampen the diffusion of content that combines high emotional load with indicators of low veracity; priority amplification of verified coordination messages from high-trust sources; and greater transparency about how engagement signals are used to rank content during emergencies.

What effects natural disasters have on individual preferences is still an open question. Although some studies show that large natural disasters increase individual impatience and risk aversion (Tanaka et al. 2016; Cassar et al. 2017), other studies show that frequent exposure to smaller disasters, such as floods, can lead people to become more tolerant of losses (Cheong . We study the effects of living in a flood area frequently exposed to shocks and having a native mother tongue on attitudes toward risk, losses, and ambiguity. To do so, several traditional tasks in experimental economics were used to capture the individual attitudes of 676 residents of a riverside area of a large metropolitan area of Bolivia. Bolivia is considered one of the most vulnerable countries to natural disasters (The World Bank 2017). Our findings indicate that those with a native mother tongue living in flood-prone areas are more tolerant of risk, regardless of the decision framing (gain, loss, or ambiguity). Thus, the results of this study highlight the multidimensional and complex relationships of those communities exposed to natural disasters.