The authors suggest that higher temperatures reduce agricultural productivity, and hence reduce incomes in poor and middle-income countries. These lower incomes encourage more people to migrate to higher income (in this paper, OECD) countries, since the net gains (higher incomes, minus the monetary and non-monetary costs of moving) from migrating are now greater. However, there is a key difference between poor and middle-income countries, and that is the level of income (duh!) and hence savings, which people can use to pay the costs of migration. People in poor countries are less likely to have the necessary financial resources to migrate than are those in middle-income countries. When high temperatures lower incomes (and savings) in poor countries further, even fewer people from these countries would be able to migrate. So, the authors expect to see higher temperatures associated with more migration to rich countries from middle-income countries, but less migration to rich countries from poor countries. And indeed, they find that:
...in very poor countries increasing temperature leads to lower emigration and urbanization rates, while in middle-income countries it leads to larger rates. We also show that long-run temperature increase speeds the transition away from agriculture in middle-income countries. Conversely, it slows this transition in poor countries – worsening the poverty trap – as poor rural workers become less likely to move to cities or abroad. We also find, for middle-income countries, emigration induced by higher temperature is local and is associated with growth in average GDP per person, while the decline in emigration and urbanization in poor countries is associated with lower average GDP per person.The idea for urbanisation is much the same. Rural folk in poor countries are less able to respond to increased temperatures (and reduced agricultural incomes) by moving to the city in search of work than are rural folk in middle-income countries.
These arguments do not extend to rich countries though, since agriculture is a much smaller share of national production in rich countries, and a much less important source of income for people, even in relatively rural areas. So, Cattaneo and Peri's results do not contradict mine for example, where the effects of climate change on internal migration in New Zealand are relatively small (see this post where I discuss my recent working paper on this).
Coming back to developing countries though, it is clear from Cattaneo and Peri's paper that climate change presents a particularly troubling potential poverty trap for the poorest countries.