A Different Kind of Model to Estimate Migration due to Climate Changes

 

 A recent New York Times report summarizes recent approaches in estimating the impacts of climate change on migration flows. Alan B. Krueger, a labor economist specializing in statistics, and Michael Oppenheimer, a leading climate geoscientist at Princeton, developed an econometrics model for climate migration in Mexico. They examined the statistical relationships of census data, crop yield, historical weather pattern, and other factors to get a better understanding of how farmers respond to drought. The model predicts a measure of farmers’ sensitivity to environmental change. The study found that Mexican migration to the United States “pulsed upward during periods of drought and projected that by 2080, climate change there could drive 6.7 million more people toward the Southern U.S. border.”

Oppenheimer’s approach to predict climate migration has sparked controversies. The model is built upon assumptions and cannot include all factors that influence human decision-making. However, there are no better publications for predicting climate migration, and econometrics has been commonly used for climate-related modeling. 

In one study, researchers have found that crop losses due to climate change “led to unemployment that stoked Arab Spring uprisings in Egypt and Libya.” In North Africa’s Sahel, droughts and extraordinary population growth have killed 100,000+ people due to water shortages and poverty. The United Nations predicted that “some 65 percent of farmable lands have already degraded.” The World Bank also projects that 17 to 36 million people will be uprooted from South Asia to the Persian Gulf and India’s Ganges Valley. Despite drought and crop losses, climate scientists have estimated that some 150 million people globally will flee due to rising sea levels. 

A two-year study published in 2018 included a gravity model, which assesses the relative attractiveness of destinations, to predict where migrants will go. In 2019, more environmental data were added[1] to the model to make it more sensitive to climate changes. Existing data sets on political stability, agricultural productivity, food stress, water availability, social connections, and weather [were added] to approximate the kaleidoscopic complexity of human decision-making.” However, even with more layers of data added, individual decisions and consequences are difficult to predict since those data do not exist. The model instead uses decision-making patterns of entire populations and apply them on various scenarios (different levels of growth, trade, border control, etc). More than 10 billion data points were included. Tests were done with past cause and effect events to see if results match. The model is so large that it took a supercomputer four days to calculate its estimated migration from Central America and Mexico. However, the results were built upon assumptions about complex relationships. Although some relationships, such as how drought and political stability relate to each other, can change over time, the model assumes that the relationship is linear. With these data limitations, the model is far from definitive. 

Post by Isabel Wang, Colgate Class of 2021.

Source: Lustgarten, Abrahm. “The Great Climate Migration,” The New York Times, 23 July, 2020.

---

[1] This study was done by the Times Magazine, and ProPublica, with support from the Pulitzer Center, hired an author of the World Bank report, Bryan Jones.