Dangerous Liaisons: A Social Network Model for the Gender Wage Gap

Abstract

We combine stylized facts from social network literature with findings from the literature on the gender wage gap in a formal model. This model is based on employers' use of social networks in the hiring process in order to assess employee productivity. As a result, there is a persistent gender wage gap, with women being underpaid relative to men after controlling for productivity characteristics. Networks exhibit inbreeding biases by productivity and by gender, which in combination with women's lower network density cause women to be hired less often through referral, as well as receive a lower average referral wage premium. Finally, we use 2001-2006 UK Labour Force Survey data to test the hypotheses implied by our model. We find that networks do indeed account for a significant part of the gender wage gap for newly hired workers.

Keywords: social networks, gender wage gap, imperfect information

JEL Classification: J16, J31