Research Highlights Article
July 15, 2026
Climate change and migration
Extreme temperatures in El Salvador destroy farm jobs and push workers abroad.
Source: zenpix
During the Great Recession, migration from Mexico to the United States began to fall, a pattern economists attributed to weaker job prospects and tightening immigration enforcement. Migration from El Salvador, Guatemala, and Honduras, however, did not follow suit. Even as the pull of the American economy faded, the flow from the Northern Triangle kept rising. For economist Andrea Velásquez that divergence presented a puzzle.
"We see that the pull factors are tightening, and we see a response in Mexico, but we don't see that response in the Northern Triangle," she said in an interview with the AEA. "What are the push factors?"
In a paper in the American Economic Journal: Economic Policy, Velásquez and coauthors Ana Maria Ibáñez, Juliana Quigua, and Maria Jimena Romero show that extreme heat damaged El Salvador's corn harvest, eliminating farm jobs, and prompting workers to migrate abroad with support from large migrant networks in the United States, where more than a quarter of Salvadorans now live.
The authors combined three data sources to study migration from El Salvador: an annual agricultural survey of yields, land, and inputs to farming, covering 19,325 subsistence corn producers from 2013 to 2018; a nationally representative household survey that tracks employment and whether a family member migrated abroad, covering 207,249 households from 2009 to 2018; and satellite temperature data from NASA and the ERA5-Land, which allowed the researchers to measure heat at a fine geographic scale.
Temperatures in El Salvador
The chart below shows the ERA5-Land averages by year in El Salvador, weighted by area. The green line shows the smoothed trend in average temperatures.
The effects of extreme temperatures can be especially significant in regions like El Salvador where agriculture is overwhelmingly rain-fed and small-scale, with only 2 percent of farmland being irrigated. Farmers in hot places adapt to chronic heat but are not equipped to respond instantaneously to extreme spikes. To measure extreme temperature shocks, the researchers counted the number of weeks during the main corn-growing season in which temperatures ran two standard deviations above the historical average. This definition helped to isolate unanticipated weather events.
Comparing municipalities over time, with controls for rainfall, local trends, and homicide rates, the authors traced weather shocks through the rural economy and found that one additional week of extreme heat during the growing season cut corn production by 3 percent and yields per hectare by nearly 4 percent. Because planting and fertilizing decisions were made before the heat arrived, producers had little choice but to cut back on hired farm help, decreasing demand by a little over 3 percent. One extra week of extreme heat raised the probability that a corn-farming household sent a member abroad the following year by 0.15 percentage points, a 23 percent increase relative to the average. Nonagricultural households showed no such response.
It's a story of extreme weather events and how that affects agricultural households. When they are not able to relocate to other agricultural sectors, having this network allows them to migrate as a way of coping with this extreme weather event.
Jimena Romero
Migrant networks were a critical link between home and host country. Households in municipalities with the largest shares of migrants abroad responded most strongly to heat shocks while those in the least-connected municipalities showed no significant migration response at all.
"It's a story of extreme weather events and how that affects agricultural households," said Romero. "When they are not able to relocate to other agricultural sectors, having this network allows them to migrate as a way of coping with this extreme weather event."
The researchers also distinguished between two types of moves to better understand the underlying mechanisms. First they note that a bad year triggered what they call a distress migration.
"You have a contemporaneous weather shock, low agricultural production, bad income, and no risk-coping mechanisms,” Velásquez said. “You then migrate, but this is a distress migration. This was not planned."
By contrast, runs of good harvests allow households to finance a planned, investment-style migration. The authors found that in El Salvador, the distress effect dominated.
For policymakers, there are several lessons to take away from this work. Where farming remains viable, insurance, credit, irrigation, and early-warning systems could blunt the shocks that force families out. Where migration is an essential coping strategy, policymakers may want to lower the costs of migrating, creating more opportunities for safer, investment-style migration.
“A planned migration is a pathway out of poverty. Let's not break those pathways," Velásquez said.
♦
“Responses to Extreme Temperatures: Migrant Networks and International Migration from El Salvador” appears in the May 2026 issue of the American Economic Journal: Economic Policy.