Enforcement Spillovers and the Efficiency Costs of Means-Tested Transfers: Evidence from Workplace Networks

Standard welfare evaluations of means-tested transfers typically focus on households directly exposed to eligibility enforcement. We show that this approach misses important spillovers when enforcement events transmit information through workplace networks. We study Uruguay's AFAM-PE conditional cash transfer, where income eligibility is monitored monthly using administrative records of formal earnings, but many beneficiaries have limited knowledge of the income rule. Linking program records, social security data, and firm identifiers, we construct workplace networks and estimate spillovers using a fuzzy RD design around the income threshold. Coworkers of peers just above the cutoff reduce registered labor earnings by 20%; the fuzzy RD estimate implies a decline of about one-third of baseline earnings following a peer's suspension. The response is persistent, operates mainly through formal employment, and largely reflects transitions to informal work. Spillovers are stronger in small firms, among similar coworkers, and when peers contact the support service. Incorporating these indirect responses lowers the estimated MVPF from 0.53 to 0.36.