Key points are not available for this paper at this time.
Despite recent research that has demonstrated the clear superiority of a multiregional perspective in measuring and projecting the dynamics of internal migration flows, many scholars continue to adopt the uniregional perspective that is forced to focus on net migrants, a nonexistent category of individuals. Net migration models are misspecified because the rates that they use confound changing migration propensities with changing population stocks. Moreover, they obscure regularities in age profiles of migration and thereby further misspecify the spatial dynamics generating observed settlement patterns. Thus, the use of the net migration rate as the dependent variable in explanatory models of migration can produce a misspecification of the fundamental relationships that are the subject of inquiry. This paper considers deficiencies of the net migration concept and illustrates them with numerical examples.
Andrei Rogers (1990) studied this question.