We develop an equilibrium model where workers with different job loss rates use their social network to obtain job offers and decide to join or leave a network by considering the trade off between the cost of maintaining links (social contact) and potential benefit of regaining employment through their network. Our model is an extension of search and matching model by Pissarides that captures social networks in the matching process and allows for heterogeneity in job loss rates. We analyze survey data from PSID with information on job finding through social networks. Consistent with the data, our model predicts that workers with higher job loss rates are more likely to use their social contacts to find jobs; they have shorter unemployment durations and lower wages. Our endogenous framework allows us to analyze how networks are affected by labor market conditions. We show that a more generous unemployment benefit scheme will reduce the use of networks and there is a negative correlation between unemployment rate and the fraction of workers using social contacts to get a job. We also show that social networks exacerbate regional inequalities in the labor market.