Institute for Research on Labor and Employment

I examine the phenomenon of occupational agglomeration – the observation that workers with similar skills tend to co-locate geographically. Extant explanations point to the fact that industries also tend to agglomerate – thereby creating a need for a particular type of employee to locate there. However, labor markets can pool even when propinquity to employers is not beneficial. I argue that particular types of work become associated with specific geographical locations. This association becomes a categorical stereotype – which leads employers to prefer employees from particular geographic regions because they will seem more appropriate – a form of “spatial signaling.” I test this theory in an online, virtual marketplace for freelancing services. I find that the greater the association between a particular job category and a country – what I termjob specific geographic identity– the more likelyanyfreelancer from that country will win a job in that category. I also find this effect is stronger when a freelancer has no previous relevant experience but a bad experience by a buyer (at this job/country intersection) can eliminate this positive effect. This effect holds net of other explanations such as spatial mismatch, knowledge spillovers, and input cost advantages.