Push-pull mechanisms in China's intercity population migration: Nonlinearity and asymmetry

Abstract

Previous research has often assumed a predefined, typically linear, relationship between migration and city characteristics. However, few studies have explored how these factors asymmetrically influence migration as push factors at the origin and pull factors at the destination. This study utilizes mobility data from the return-to-home phase of Chunyun in 2020 and employs the Random Forest (RF) model to construct a nonlinear explanatory framework for the push-pull effects and their asymmetry in intercity population migration. The findings reveal three key insights: Firstly, there is a notable asymmetry between push and pull effects in the predictive importance of feature variables. Pull effects at the destination are more prominent for economic and employment factors, whereas push effects at the origin are more influential for factors related to the living environment and gender differences. Secondly, we identify the phase transition points and threshold intervals of coupling effects based on the variations in the shapes, gradients, and thresholds of the nonlinear push-pull effects. In addition, we examine the disparities in the context of push and pull dynamics for the interaction effects of multiple variables. Thirdly, heterogeneity in demographic subsamples and different urban scenarios is observed, with within-group differences exhibiting various patterns of change. These findings provide valuable insights for policymakers to develop targeted population policies that account for the complex interplay of push and pull factors.