Workplace Segregation between College and Noncollege Workers

How segregated are American workplaces in terms of education and skills? For 60 years, social scientists have researched residential segregation (Taeuber and Taeuber 1965; Murray 2012; Athey et al. 2021), but the typical employed American spends about eight hours at work on a normal workday. This time at work is far more than the less than one hour per day that adults usually spend “socializing and communicating” and in “organization, civic and religious activities.”1 Moreover, events at work seem far more likely to shape long-term skill acquisition by workers and advance their economic outcomes than neighborhood events. Indeed, the level of human capital in the workplace may beget more human capital and productivity in the future (e.g., Becker 1964; Lucas 1988; Duranton and Puga 2003; Moretti 2004; Mas and Moretti 2009). We measure the level of educational segregation in US workplaces experienced by noncollege workers using the Longitudinal Employer-Household Dynamics (LEHD) data. During the 2000–2020 period, educational dissimilarity of the workplace was comparable in magnitude to neighborhood racial segregation for housing and rose somewhat. Workplace isolation was particularly high for young and male workers without college degrees. A companion paper, Dillon, Glaeser, and Kerr (2025), finds that workplace isolation negatively impacts the future careers of noncollege workers who are less exposed to college-educated workers.2 The growing educational segregation of our workplaces may therefore limit the ability of younger, less-skilled workers to learn on the job from their more educated peers, diminishing acquisition of skills that promote higher-paid future employment (e.g., Blair et al. 2020).