“Not recognizing climate change related mobility will not obliterate it, but likely lead to more clandestine mobility”

Professor Filiz Garip, Professor of Sociology and Public Affairs at the renowned Princeton University, gives a Special Lecture at the IAB on 1 December 2023. She will discuss the complex interplay between climate change, migration, and inequality. In this accompanying interview, she describes the connection between weather extremes and migration, shows a window to respond for policy makers, and draws conclusions for future migration and return migration scenarios in Germany.

Research on the impact of weather events on international mobility in the Mexico – U.S. setting between 1980 and 2018 revealed that weather indicators offer at best a modest improvement in migration prediction. Can you please describe your approach to optimise the prediction performance?

Machine learning methods involve fitting complex models in a flexible way. With a sufficiently complex model, however, one can perfectly fit every data point and capture the idiosyncracies in the data. To prevent that, researchers split their data. Part of the data is used to train different models and to select one among them; and the other part is reserved to test predictive performance of the selected model. That way, one can balance optimizing predictive performance and selecting a model that generalizes the data that the model has not seen.

In our case, we used this approach to try several modeling strategies in the exploration data, and then evaluated the best-performing model in the test data. We found that our best-performing model (which includes hundreds of weather measures) only marginally out-performs a simple model with a hand-selected set of weather indicators in predicting migration behaviors. One reason for that is individual-level indicators (such as age, gender, household income) included in the simple model already account for most of the variation in migration choices.

“We also relied on some of the lessons we learned from our fieldwork in Mexico, where we asked farmers about the weather conditions that matter most to them.“

When you opened the “black box,” what did you find?

The complex models used in predictive modeling do not allow us to directly interpret the effect of different inputs (say, weather indicators) on migration. These models are used for prediction tasks, and not for estimating the effect of a given factor on the outcome of interest. This is what we mean by the ‘black box.’ In the follow-up to our predictive modeling, we went in the other direction, and used a well-specified simple model to relate weather indicators to migration choices. We used some of the insights from predictive modeling (for example, the weather measures that turned out to be most predictive of migration) in this effort. We also relied on some of the lessons we learned from our fieldwork in Mexico, where we asked farmers about the weather conditions that matter most to them. This analysis allowed us to discover the importance of sequence weather events (where one adverse event, like a drought, follows another one), as well as their combination (where extreme heat occurs at the same time as extremely dry weather).

Do migrants’ networks play a role when they decide to migrate for climate reasons?

This is an area we are currently working on. This is a difficult problem because past migration patterns (which set up migrant networks between origin and destination places) might be a result of past weather conditions; and past weather conditions might be correlated with current weather. These interlinkages make it difficult to identify the effect of networks on weather-related migration.

“One finding from our work is that consecutive weather shocks have a stronger association to migration than shocks that follow ‘normal weather’ years.”

Does gender play a role? Do female and male migrants make decisions differently?

We have not explored this question yet. We do find that households with better resources are able to respond more quickly to adverse weather conditions and send migrants to the United States. It is quite costly to cross the border for undocumented migrants. This means that poor households cannot immediately fund migration. If weather-related adversities further reduce their income, these households may be further delayed in sending a migrant as a coping strategy.

“Governments can intervene in regions after a single year of adversity and give people the option to stay in their home regions.”

Given that migration as a climate-adaptation strategy is only available to selected households in the developing world, how should policy makers react?

One finding from our work is that consecutive weather shocks have a stronger association to migration than shocks that follow ‘normal weather’ years. This pattern means that we might have a window to respond; governments can intervene in regions after a single year of adversity and give people the option to stay in their home regions if the following year also happens to present extreme conditions.

We also find that farming communities with irrigated land are less vulnerable to weather shocks and less likely to send migrants in their aftermath relative to communities without irrigation. That, again, presents a potential intervention point. We can invest in irrigation to support smallholder farmers in the developing world.

“Undocumented migrants in the United States delay their return to Mexico if their origin communities experience adverse weather shocks.”

In Germany, we find that migration is often temporary. Many migrants return to their country of origin, others continue their migration on to other countries. Do you see any connections between climate change and onward or return migration?

This is a very important question. We have a working paper on that. We show that undocumented migrants in the United States delay their return to Mexico if their origin communities experience adverse weather shocks. In other words, continuing weather extremes make migrants more likely to settle in destination countries.

“Extreme weather conditions, which are likely to increase with anthropogenic climate change, can disrupt return migration and lead to a more settled migrant population.”

What conclusions can be drawn from research on Mexico-U.S. migration, climate change and inequality when considering migration to Germany?

Extreme weather conditions, which are likely to increase with anthropogenic climate change, can disrupt return migration and lead to a more settled migrant population, especially if migrants’ origin countries are vulnerable to weather shocks.

“The global community is currently debating whether ‘climate change related mobility’ constitutes a distinct category and whether countries have an obligation to protect such migrants.”

What kind of future scenario could be plausible with regard to climate change and migration to Europe or to Germany?”

One thing we have learned from this research is that it is difficult to make predictions. But we can still prepare for different climate (and migration) futures. The global community is currently debating whether ‘climate change related mobility’ constitutes a distinct category and whether countries have an obligation to protect such migrants. Of course, it is challenging to connect weather events to mobility choices; it is also difficult to attribute such weather events to climate change. That said, the connection we identify in our work between weather extremes and undocumented Mexican migration to (and prolonged stay in) the United States clearly shows what is at stake. Lacking access to a legal category, Mexican migrants are resorting to dangerous and costly clandestine channels, putting their lives at risk to cross the border, and becoming vulnerable to exploitation in the United States. Research shows that restricting migration does not bring it to a stop, but rather reorients migrants to seek alternative channels of entry. Not recognizing climate change related mobility will not obliterate it, but likely lead to more clandestine mobility.

More information on the IAB Special Lectures Series

For more information on the IAB Special Lecture with Filiz Garip on 1 December 2023 and how to register for attending it on site or online, please go to:

https://iab.de/en/iab-veranstaltungen/climate-change-migration-and-inequality/

For more information on the IAB Special Lectures in 2024, please go to:

https://iab.de/veranstaltungen/special-lecture-series/

DOI: 10.48720/IAB.FOO.20231129.01