An agent-based computational approach to explaining persistent spatial unemployment disparities

Abstract

This paper explores possible reasons for persistent spatial unemployment disparities using agent-based computational methods. The method relies on observing the actions of thousands of individuals within an artificial society. The paper models the effect of unemployment insurance, wage disparities, region specific amenities and innate residential preferences on regional labour market interactions, accounting for both migration and commuting. An empirical example of Rogaland county in south-west Norway is given, where unemployment disparities have proved remarkably persistent for decades. The model provides non-trivial insight into the nature of spatial unemployment disparities as well as making a valuable contribution to the policy debate.