Predicting Migrant Workers’ Intention to Acquire Local Citizenship (Hukou) in Emerging Cities Using Machine Learning Models: A Case Study of Five Dongguan Factories

Abstract

This paper examines the data from a period when hukou migration was severely restricted, and aims to provide a reproducible analytical methodology with machine-learning methods to model migrants’ willingness to change their hukou. The data in this work is based on questionnaires and interviews collected in five factories in Dongguan from 2005 to 2007. After performing stepwise variable selection, our best-performing model is based on three main predictors: having children, education level, and household income. Analysis of interviews suggests that citizenship is regarded as an individual’s legal status and is used in exchange for other types of capital.

Keywords

• logistic regression
• hukou migration
• migrant workers
• dongguan
• hukou points system