Referral networks and inequality

I develop a theoretical model to study the welfare effects of using referrals in the labor market. In the model, firms use referrals to hire, workers are heterogeneous and the social network endogenous. Consistent with empirical evidence, referred workers are more likely to be hired, to receive a higher wage and to be more productive. The use of referrals exacerbates inequality among workers. Higher inequality is efficient if heterogeneity is driven by productivity differentials but is detrimental to efficiency if the probability of forming a match is weakly correlated with productivity, which is likely in the presence of discrimination.