On March 13th, Juan Pablo Rud from Royal Holloway at the University of London, who is also a fellow at IFS, presented his paper titled: “Extreme Weather, Climate Change, and Agricultural Productivity: Effects and Responses in Rural Peru.” This topic is of special importance for the livelihood of poor farming households in developing regions since estimates predict a temperature increase of 2ºC in tropical regions, which could harm wheat, maize and rice crop yields. The paper sets out to examine the effects of extreme weather in poor agricultural areas in Peru, using data from a monthly survey to agricultural households. One contribution of their research is to geo-match household information with high-frequency satellite data on daylight temperature and total accumulation.
Using a production model approach, Rud pins down how extreme weather events decrease total factor productivity, and most importantly, relates this to a decrease in yields, which is defined in the paper as output per unit of land used. This is achieved by considering that conditions not only have an impact in a botanical setting, but also, in an economic setting, such that adjustments are made by producers who face imperfect markets and various constraints that hinder complete mitigation.
Rud considers nonlinearities in the relationship between temperatures and yields which are found in the literature. That is, only if temperatures are too high- or too low- will they negatively impact yields. Considering that Peru is not prone to frosting or hail, Rud only considers high temperatures. Thus, three sets of measurements, which count degree days (DD), high degree days (HDD) and very high degree days (HDD), are constructed. The threshold temperature- namely the highest degree after which a day is considered a HDD- with the highest explanatory power was chosen. Geographical variation was considered as well. In Peru, the coast is warmer and wealthier, on average, than the highlands.
Results suggest that HDD are considerably harmful to total factor productivity across both regions. However, mitigation methods differ. In the coast, high temperatures have a negative impact on output and producers adjust by hiring less paid labor and relying more on labor supplied by family members. Conversely, the highlands, which are closer to subsistence levels, increase their use of land such that output does not fall.
Ultimately, Rud simulates several climate change scenarios by assuming different temperature increases, uniformly, across the coast and highlands. It seems that, on average, the coast would incur great losses while the highlands would benefit from higher temperatures. Therefore, their simulations predict a redistribution effect because of climate change.