Random social networks, unemployment and wage inequality

Empirical studies of labor markets show that social contacts are an important source of job-related information. At the same time, wage differences among workers may be explained only in part by differences in individual background characteristics. Such findings motivate our model in which differences in "social connectedness" among otherwise identical workers result in wage inequality and differences in unemployment rates. The model of this paper allows for heterogeneity in the number of connections among workers within the Pissarides model of labor market turnover. The paper derives conditions for which a unique labor market equilibrium exists. It also shows that such heterogeneity has important consequences. Workers with more connections both receive a higher wage and face a lower rate of unemployment at equilibrium. For the specific cases in which connections follow Poisson and negative binomial distributions our numerical results show that variability in connections can produce substantial variation in labor market outcomes. One lesson from the computational analysis is that (changes in) the social structure sometimes affect labor market outcomes in nontrivial ways. For example, when society becomes more connected, the average unemployment level falls but the unemployment rate of workers with few connections rises.