Conference Agenda

The effects of environmental stressors are likely to vary across individuals, firms, and locations, in part due to adaptation to local conditions. Here, I examine how this relationship between the effects of stressors and the exposure regime can induce important contamination biases if heterogeneity goes unmodeled. For example, failure to model heterogeneity in the impacts of high temperatures can contaminate estimated effects of lower temperatures, or failure to model heterogeneous effects of temperatures could contaminate estimated effects of precipitation. I first show how the conditions for contamination bias can arise using a simple model of adaptation to environmental stress and illustrate the resulting bias with simulations. After detailing a range of solutions, I assess the practical importance of this issue by revisiting published work. I re-analyze links between temperature and asylum applications and crop yields, finding that contamination bias may either inflate or dampen estimated effects. Finally, I connect this issue to recent work on related biases.

Identifying the economic impacts of climate change requires assuming that producers pursue optimal resource allocation, conditional on their beliefs about local climate. This basic assumption places certain requirements on research design and data that are particularly challenging to fulfill for regions exposed to a large diversity of climates. We propose an empirical strategy that addresses this aspect by mapping latent climates considering multiple climate dimensions that shape sample units’ production possibility sets. By establishing plausible correspondence between regional climate conditions and production technology sets, this procedure ensures the validity of the Envelope Theorem, thus enabling consistent inferences about marginal climate effects. We implement this approach using a fixed-effects model that accounts for differences in within-variation patterns between groups of study regions exposed to different climates. The model is estimated with agricultural and weather data for 77 regions in Kazakhstan, Russia, and Ukraine over 2000–2020. Our results reveal substantial heterogeneity in climate impacts. Recent temperature trends reduced land productivity in three of six regions in Kazakhstan and in eight of 49 regions in Russia, while a handful of regions in Ukraine and Russia, along with one in Kazakhstan, should have benefited from higher temperatures. Our predictions for a 1°C increase in growing-season average temperatures indicate that, given the current extent of adaptation, land productivity can diminish in four regions in Kazakhstan (−5.8% to −18.4%), 13 in Russia (−4.8% to −11.2%), and two in Ukraine (−5.0% to −5.3%), whereas only three regions are predicted to benefit. These findings highlight the uneven distribution of climate impacts across the Eurasian grain belt and underscore the importance of region-specific adaptation strategies. While developed for KRU, our latent-climate framework is broadly applicable and may help assess climate impacts in other major agricultural regions characterized by considerable climatic diversity.

Global cereal production is increasingly constrained by rising temperatures, yet the timing and extent of changes in crop yield sensitivity to climate remain unclear. We apply Bai–Perron structural break detection to identify candidate instability points of yield-climate data for six major crops across 156 countries (1962–2022), followed by first-difference panel regressions to quantify shifts in climate-yield relationships. Structural breaks predominantly occurred between the late 1980s and early 2000s, coinciding with agronomic and policy transitions. Post-break, yield sensitivity to temperature intensified by approximately 50–100%, while precipitation effects remained weaker and more variable. C₄ crops like maize exhibited more frequent breaks, reflecting physiological vulnerability to heat and moisture stress. These findings indicate that climate-driven yield sensitivities have accelerated historically, underscoring the need to refine yield projections and target adaptation and risk management strategies in a warming world.

El Ni˜no–Southern Oscillation (ENSO) is a global climatic event with important consequences for agricultural production. Several studies have investigated its impact on

yields, considering different crops and varying spatial scales. However, less attention has

been paid to its effects on food prices. The present paper tries to fill this gap through

a global analysis of the effect of ENSO shocks, divided into its main events, namely

El Ni˜no and La Ni˜na, on the retail prices of several food commodities and crops. To

achieve this, a large panel dataset of monthly retail market prices combined with climatic

indices envisaged to capture ENSO shocks has been analysed by adopting the

local projection methodology. Impulse response functions spanning till one year after a

shock are adopted to interpret the results. The impact of ENSO shocks on food prices

varies considerably according to the event, the region under consideration, and the same

food item. However, it can be observed a general tendency of increasing prices for several

staples in Sub-Saharan Africa in consequence of El Ni˜no shocks. This deserves

particular attention in light of the vulnerability of this region.