Reexamining a classical question: Does development push migration?

The migration curve hypothesis, a well-established concept in migration studies, posits that emigration rates initially rise with economic development but subsequently decline upon reaching a certain level of development. This hypothesis has been widely examined in international contexts, with many studies affirming its validity. This study aims to investigate whether the migration curve hypothesis is valid in internal migration contexts. It also aims to elucidate the effect of development on migration. Using data obtained from household surveys collected in Colombia between 2012 and 2019 and auxiliary variables from multiple sources, we extend the gravity models to control for spatial and nonspatial endogeneity to test the migration curve. The evidence from Colombia reveals a relationship between development and migration contrary to the international evidence. Some evidence of an inverted U-shaped curve is found when the sample when focusing on migration from rural areas. These results suggest that questioning the empirical studies used to test the migration curve hypothesis is still relevant. The results also suggest that disentangling the bidirectional effect between the relevant variables is crucial for deriving conclusions for policy analysis from this curve. It is also pivotal to test whether such a relationship changes over time and between areas. The findings also indicate a significant shift in migration patterns within Colombia, revealing a decreasing prominence of Bogotá as a primary migration destination over time.