How do dispersal policies affect labour market integration of refugee immigrants subjected to such policy? To investigate this, we estimate the effects of location characteristics and the average effect of geographical mobility on the hazard rate into first job of refugee immigrants subjected to the Danish Dispersal Policy 1986-1998. We correct for selection into relocation to another municipality by joint estimation of the duration of the first non-employment spell and time until relocation. The main estimation results are as follows: First, the hazard rate into first job is increasing in the concentration of fellow countrymen and decreasing in the regional unemployment rate, the size of the local population and the percentage of immigrants in the local population. The two latter findings support dispersal policies. The two former findings emphasize that refugees should be dispersed in big clusters of refugees of the same ethnic origin across regions with low unemployment. Second, on average, geographical mobility had large, positive effects on the job finding rate, suggesting that either relocations were carried out to improve employment prospects, or they were carried out to improve place utility and thereby lower the reservation wage. Hence, restrictions on placed refugees? subsequent migration (or on their initial choice of location) would hamper labour market integration.