ashira-mo/mig_edu

Replication details for paper "The educational gradient of internal migration in Africa and Asia" # mig_edu Replication material for Demography 2026 journal article: "The educational gradient of internal migration in Africa and Asia", Ashira Menashe-Oren & Phillipe Bocquier This file explains the different steps taken, and needed to replicate the results of the paper, including: where to get the data and what variables are needed, how to generate in and out migration rates, and modelling approach. Data Census samples from IPUMS International are readily available at: international.ipums.org The downloaded data for each census should include variables of: individual identifier, person weights, age, sex, education, place of residence (rural/urban), region of current residence variables related to migration - either previous place of residence and years in current location, place of residence 1 year ago or 5 years ago Generating migration rates To compute in and out migration rates between the rural, urban and capital cities, from the census data, it is first required to clean the raw data, and then extract the migration rates. See ashira-mo/mig_urb2 repository for further information and STATA do files. The only difference in code is in adding the category of “capital cities” which is defined as the region in which the capital is located. Modelling internal migration The analysis of migration is based on Poisson models for Asia and Africa separately, using the computed migration estimates in the census, as well as the proportion urban per country-year taken from the World Urbanisation Prospects (WUP) 2018. The WUP data is available as excel file ("WUP2018-F02-Proportion_Urban.xls"), at: population.un.org The STATA do file "models" documents the modelling strategy.