Conference Agenda

I study how seasonal water forecasts influence farmers’ externally-costly, irreversible investments in groundwater well drilling in California. I use variation in 70 years of basin-level snow runoff forecasts and forecasted spread assigned to local and distant water districts based on pre-determined water rights. I find that farmers significantly respond to erroneously low surface water forecasts by drilling wells about four years earlier than they would have, resulting in a permanent level shift in the aquifer by about one centimeter, or 58\% of a district’s average annual decline. On the other hand, lowering erroneously high forecasts results in about half as many new wells as those avoided, and high private benefits for farmers. Further, I show empirically and using a conceptual model under tighter forecast spreads, farmers hold off drilling until lower snow runoff thresholds. Thus, improving a forecasting system by reducing spread reduces expected errors which both avoids depletion and creates private benefit, and also shifts the well demand curve inward. Where water management is politically infeasible, targeting forecast improvements can therefore substantially lower long-term external costs.

Rainfall uncertainty and groundwater depletion exacerbates production risks in agriculture. It endangers food security issues in developing countries, such as India, which heavily rely on groundwater resources and the timing of monsoon for production decisions. Micro-irrigation technologies, such as drip and sprinkler systems, are being subsidized and promoted in India for over a decade, can partially offset groundwater scarcity and shape a climate-resilient agriculture pathway. However, the adoption of these technologies are very low. Existing research on MI technology have often mostly relied on static models, failing to explain the dynamic nature of MI technology uptake, its sustained use, and disadoption. Using a high-frequency, plot-level panel data, covering 1,161 agricultural households in the Eastern and Semi-arid Tropics of India, we uncover patterns of adoption over half-a-decade period. Our duration analysis model identifies covariates that accelerates (and decelerates) the speed of MI adoption. We provide novel evidence on water-saving technology adoption response to negative rainfall shocks as well as pinpoint the role of community level adoption and past exposure to adoption as significant determinants of adoption speed. We also find that the households from marginalized castes (SC/ST) face a substantially lower probability of speedy adoption, suggesting the role of social structures in mediating access to technology. We also find evidence of disadoption of MI technologies, which can undermine the welfare gains anticipated under full adoption, warranting further research to understand factors that can deter disadoption.

We investigate the effects of climate-induced migration on criminal dynamics in destination countries. Leveraging restricted-access monthly and provincial-level data on human trafficking-related crimes, our difference-in-differences estimates show that the severe drought shock occurring in Southern Nigeria in 2016 increased human trafficking-related crimes committed by Nigerian nationals in Italy. Specifically, we estimate an increase of 18 additional crimes attributable to this climate shock, representing a 150% rise relative to pre-shock levels. We then explore the role of existing organized crime structures and community networks in facilitating the exploitation of migrants using provincial-level data. This analysis reveals that a stronger presence of criminal organizations and larger Nigerian communities in certain provinces explains a higher incidence of crimes related to human trafficking.

This paper examines the impact of recurrent droughts on farm size and land concentra-

tion in France over the period 1970–2022. While droughts in developing countries are often

associated with cropland expansion, we show that in France, they instead accelerate land

concentration. Using panel data covering French 716 small agricultural regions and more

than 344,000 farms, we analyze changes in farmers’ land size following drought episodes.

Drought exposure is measured using the Soil Wetness Index, capturing both absolute soil

moisture deficits and relative deviations from region-specific historical normals. Our results

indicate that severe droughts significantly increase farm size, particularly when they occur

during summer and autumn. These effects operate through two main channels: droughts

depress agricultural land prices and force smaller farms to exit the sector. Larger farms,

by contrast, are better able to expand along the extensive margin, facilitated by the with-

drawal of smaller producers. Moreover, we find that climate risk insurance and CAP Pillar

I subsidies amplify these concentration dynamics by supporting farm expansion rather than

mitigating inequality. Overall, our findings highlight the structural consequences of climate

shocks in developed agricultural systems and underscore the need for policy interventions

that promote both equity and climate resilience in agriculture.