Climate change and international migration: Exploring the macroeconomic channel

emigration function in this case redefines the position of the peak, mapping a portion of the countries on its left branch, where a loss in GDPc turns into a decrease in emigration. While few countries may find themselves on the left branch of the new shifted migration-hump, many countries get closer to the peak, where rates of change in emigration rates are lower and therefore the same impact on the GDPc would produce smaller changes under the TS assumption. Immigration, under the CR case, shows a similar pattern of impact as observed for emigra- tion but with a larger number of countries showing a reduced immigration (Fig 5a). Since the only varying factor in the CR assumption is the destination country relative GDPc factor, these results largely resemble the figure of impacts of climate change on the GDPc level (Fig 1). Escaping from this pattern, in countries like Mexico, despite a strong negative impact on GDPc, immigration is projected to increase. Importantly, in our model we account separately for the return migration flows (Eq 6b). These flows, not depending directly on economic fac- tors, may play a fundamental role in bilateral migration channels characterized by a large return migration flow (e.g. USA to MEX), and drive results that diverge from common global patterns. Under the T0 assumption we find that immigration increases in almost all the coun- tries and reaches higher levels of impact (Fig 5b). Importantly, we find that many countries see, under the T0 assumption, an opposite impact on immigration compared to the CR case.

This highlights the fact that the contribution from the GDPc at the origin dominates over the destination country GDPc factor, not only in defining the patterns of change in emigration but also for those regarding immigration. Under the TS assumption, results are similar as for T0, even though the levels of impact on immigration are smaller (Fig 5c). Altogether, for both the assumptions that include a migration transition mechanism, i.e. T0 and TS, we find a global pattern of higher mobility compared to the CS case. Many countries see an increase in immigration despite the projection of large negative and asymmetric economic impacts, which would reduce their relative capacity to attract migrants. This means that migrants from countries that are impacted negatively may still choose to migrate to another country which also faces negative impacts. This illustrates a more complex mechanism than would be expected by a simplistic implication where “environmental migrants” would move from coun- tries that see damages from climate change to countries that benefit. Finally, complementing the results from the country level we also look at the bilateral flows between ten major world regions (Fig 6. We construct these figures by selecting the subset of bilateral flows between countries included in each region (S2 Table in S1 File). We sum up the country-level bilateral flows for each region for both the baseline case and climate change sce- nario. We then calculate the difference between them, for each bilateral flow between regions.

Under the CR assumption, we find that climate change increases migration to North America, Europe, East Asia, and the Former Soviet Union, from most or all other world regions (Fig 6a). All other between-region flows see a decrease due to climate change. In line with the global- level results discussed above, the climate change effect on migration is overall relatively small under this assumption. Under the T0 assumption, we find much larger effects, mostly increases, in both migration between and within regions and within them (Fig 6b). These increases generally amplify exist- ing migration routes, i.e. we find increases in migration from poorer to richer regions of the world. However, we also find increases e.g. in migration from different parts of Asia to Africa. Note that these effects are expressed as percentage of the baseline flow size, so that large rela- tive changes also in small flows become prominently visible. Thus, these figures are indicative of the relative effect of climate change on between-country flows, but do not show which regions are more affected than others in absolute terms. When the climate change effect is instead expressed in absolute terms, it becomes visible that changes in flows to poor regions PLOS ONE Climate change and international migration: Exploring the macroeconomic channel PLOS ONE | https://doi.org/10.1371/journal.pone.0276764 November 16, 2022 18 / 25