Duration Dependence in Migration Behaviour: Cumulative Inertia versus Stochastic Change

Abstract

In this paper a new theoretical framework and supporting empirical evidence on the relationship between movement probabilities and length of stay are presented. Individuals' evaluations of the relative value of alternative locations are assumed to evolve stochastically, with a possible tendency either to cumulative inertia or to cumulative stress. In general this yields a nonmonotonic duration function, with probabilities of movement starting at zero, rising and then falling—a pattern consistent with either cumulative tendency, or neither. A version of the model fitted to data on household movement intentions, from the UK General Household Survey, confirms the hypothesised form of this function and indicates a dominance of cumulative stress over cumulative inertia.