Friends in High Places: Minimum Wage Shocks and Social Network Propagation

Abstract

Do social network connections to high-wage labor markets improve local economic outcomes? We construct a novel population-weighted measure of network minimum wage exposure using Facebook's Social Connectedness Index that captures the breadth of a county's connections to high-minimum-wage areas. Using an instrumental variable strategy exploiting out-of-state network connections, we find that a \$1 increase in the population-weighted network average minimum wage raises county-level earnings by 3.4% and employment by 9% (a local average treatment effect concentrated among high-connectivity counties; USD-denominated semi-elasticity specification). The first stage is strong ($F > 500$). A critical specification test reveals that probability-weighted exposure—capturing only the share of connections in high-wage areas, ignoring scale—shows no significant effects despite a robust first stage ($F = 290$). This divergence indicates that the breadth of network connections, not merely the network share, drives labor market responses. Distance-restricted instruments show effects strengthening as connections are limited to more distant origins, placebo shocks produce null effects, and Anderson-Rubin confidence sets exclude zero. Job flow analysis reveals increased hiring and separations consistent with heightened labor market dynamism, while IRS migration data show that migration responses are small relative to employment effects and unlikely to mediate the main results.