We investigate the impact of social networks on earnings using a dataset of over 20,000 senior executives of European and US firms. The size of an individual's network of influential former colleagues has a large positive association with current remuneration. An individual at the 75th percentile in the distribution of connections could expect to have a salary nearly 20 per cent higher than an otherwise identical individual at the median. We use a placebo technique to show that our estimates reflect the causal impact of connections and not merely unobserved individual characteristics. Networks are more weakly associated with women's remuneration than with men's. This mainly reflects an interaction between unobserved individual characteristics and firm recruitment policies. The kinds of firm that best identify and advance talented women are less likely to give them access to influential networks than are firms that do the same for the most talented men.
SAFE Working Paper No. 123
