Estimating interregional population and migration operators from interregional population distributions

A common constraint in analyses of interregional population systems is the absence of reliable data for describing the behavior of the fundamental components of population change. As a result, demographers frequently have had to rely on crudely constructed measures of natural increase and net migration.Recent efforts to express interregional population growth in matrix form, however, suggest a method for estimating the regime of growth of a multiregional system solely on the basis of historical data on interregional population distributions. This paper describes and illustrates such a method. Since the results are largely unsatisfactory, a large portion of the paper is devoted to an analysis of the conditions leading to the poor performance of the method, by evaluating, in particular, alternative estimators and by considering the possibility of "data smoothing."