Immigrant Benefit Receipt: Sensitivity to the Choice of Survey Years and Model Specification

Abstract

Receipt of unemployment insurance by immigrant men and social assistance by immigrant families are analysed using thirteen surveys from Canada. Estimates from a cohort fixed effects model are found to be sensitive to the choice of survey years. This is due to the mis-specification of the fixed effects model which is rejected when tested against a model allowing for separate year-since-migration effects by arrival cohort. The estimates from the more general model provide little evidence of higher receipt of these benefits, ceteris paribus, for more recent cohorts or that immigrants assimilate toward greater receipt of these benefits.