Social Networks, Employee Selection, and Labor Market Outcomes

We provide a direct empirical test of Montgomery's 1991 notion that firms hire workers through social ties of productive employees as these workers know others with high unobserved productivity. We focus on coworker networks and show that firms recruit workers with better military draft test scores but shorter schooling when hiring previous colleagues of current employees, suggesting that firms use these networks to attract workers with better qualities in hard-to-observe dimensions. Incumbent workers' abilities predict the incidence, abilities, and wages of linked entrants. These results suggest that firms rely on the ability density of the studied networks when setting entry wages.