Estimating Migration Flows from Population Stocks

University Of Colorado At Boulder, Boulder CO

Investigators

Abstract

This project will develop methods for inferring migration flows from two successive counts of population stocks. Such data on stocks in the past have been used to infer mortality and, indeed, of net migration. But a workable method for using population stock data to infer directional (not net) migration flows has not been developed. This project will draw on statistical representations of regularities in the age and spatial structures found in relatively complete data on migration streams to infer, in settings characterized by incomplete data, the directional flows that underlie and shape the population distributions enumerated by censuses and sample surveys. This new method will benefit a number of different user groups. Historical demographers will be able to identify changing mobility patterns hidden in recently released historical population censuses that lack a question on migration. Migration analysts will be able to study mobility patterns in data-poor less-developed countries. Furthermore, population researchers will be able to adapt to the prospective loss of migration data from the soon-to-be-eliminated "long form" questionnaire of the U.S. decennial censuses and its replacement by the American Community Survey, which is a significantly smaller continuous monthly sampling survey.