Requiem for the Fixed‐Transition‐Probability Migrant

In a recent paper titled “Requiem for the Net Migrant” Andrei Rogers persuasively argues for the use of a multiregional perspective rather than a uniregional one in measuring and projecting the dynamics of internal migration. In particular, he highlights the deficiencies of using net migration rates in population projections, giving illustrations of the very large differences that can occur if constant net migration rates are assumed versus fixed interregional transition probabilities. Net migration rates are conceptually unsound because their denominators are not true “at‐risk” populations. Fixed interregional transition probabilities, however, are inconsistent with a sound behavioral representation of migration system dynamics. Whereas such stationary Markov models posit a role for shifting origin region populations, they do not embed the assumptions intrinsic to gravity or opportunity model concepts about the role of changing destination region populations. This paper explores alternative, more behaviorally pleasing interregional models that posit a role for shifting destination populations in altering the attractiveness of migration alternatives. Density‐dampened, destination‐population‐weighted transition probability structures are explored. The importance of modeling intraregional migration separately from nonmovement is stressed.