Brothers in Arms: Spillovers from a Draft Lottery

While family members tend to have similar labor market outcomes, measuring the contribution of behavioral spillovers is difficult. To identify spillovers between brothers, we exploit Denmark's largest random assignment of young men—to eight months of military service, where service status of brothers is correlated, but draft lottery numbers are not. We find average spillovers of elder brother service on younger brother service of 7 percent, and as high as 55 percent for closely spaced brothers without sisters. Elder brother's military service affects his own occupational choice and his younger brother's service by discouraging any refusal to serve.