Soybean yield prediction in Argentina using climate data

This paper was produced as part of the BIS Consultative Council for the Americas (CCA) research network and conference on "Macro-financial implications of climate change and environmental degradation", held in Bogotá on 2-3 December 2024.

Summary

Focus

This paper studies the impact of climate conditions on soybean yields in Argentina, a country where soybeans are a key export and economic driver. The authors use a detailed data set that combines daily weather data with crop-specific information across regions from 1980 to 2023. They examine how temperature, rainfall and major climate events – El Niño and La Niña – affect soybean production. They also consider how technological advances and global price trends influence yields. To measure these effects, they apply spatial models that take into account both local factors and regional interconnections.

Contribution

Soybeans make up a large share of Argentina's exports. When production drops due to droughts or extreme weather, the country's fiscal revenue, foreign reserves and exchange rate stability weaken. By better understanding how climate affects soybean yields, the authors aim to improve early warnings and guide economic policy responses. Their approach also allows for building climate risk scenarios, which can help prepare for future shocks.

Findings

The authors find that extreme heat reduces soybean yields, while moderate rainfall helps – up to a certain point. El Niño events, which usually bring more rain, tend to boost production. La Niña events, which bring drought, have the opposite effect. The authors also show that new technologies, such as genetically modified seeds, help increase yields. Finally, they compare their forecasts with those from the US Department of Agriculture and find that their models perform better when using early-season data. These results can help improve risk planning in agriculture and broader economic policy.

Abstract

Agriculture, and especially soybean production, has a critical role in Argentina's economy, as a major contributor to GDP and export revenue. This paper studies the impact of climate variability on soybean yields in Argentina using a novel department-level dataset spanning 1980–2023. We estimate a fixed effects spatial error model (SEM) to quantify the long-run effects of weather shocks - measured by extreme heat, precipitation, and ENSO phases - while controlling for economic and technological factors such as seed technology and relative prices. Our results show that extreme heat significantly reduces yields, while moderate rainfall boosts them up to a nonlinear threshold. El Niño phases increase yields, whereas La Niña events are detrimental. Technological adoption and favorable price signals also enhance productivity. These findings highlight the importance of accounting for both climatic and spatial dynamics when analyzing agricultural outcomes. The model provides a strong empirical basis for forecasting soybean yields and informing policy decisions under increasing climate uncertainty. These models can be employed as effective tools for anticipating yield outcomes under different climate scenarios and utilized in climate-related stress exercises. This work provides valuable insights for policymaking decisions, contributing to prepare for potential economic impacts stemming from climate risks on Argentina's agricultural sector.

JEL classification: Q10, Q12, C13, C32, C33

Keywords: Soybean Yields, Argentina, Forecasting, Model Selection