Modeling the Network of Inter-County Migration Flows in the U.S. Using Dynamic Valued ERGMs

Peng Huang, Carter Butts


Despite the notion that the United States is “a land of mobility,” its migration rates have been declining for decades, especially during this century. This paper studies how various network mechanisms, together with demographic, economic, political, and geographical factors, influence the directions and magnitudes of inter-county migration flows in the United States. We first introduce an approach to model large, dynamic, valued networks under the framework of exponential-family random graph model using Maximum Pseudo-Likelihood Estimation. We model the population flows between the 3,140 U.S. counties from 2011 to 2015 using data from the American Community Survey. The results offer implications about how the network approach might contribute to better understandings of internal migration.

← Schedule