DT 09-26 Enforcement Spillovers and the Efficiency Costs of Means-Tested Transfers: Evidence from Workplace Networks

Standard welfare evaluations of means-tested transfers typically treat eligibility enforcement as affecting only households directly exposed to the rule. We show that this direct-only approach is incomplete when enforcement events transmit information through workplace networks, inducing other beneficiaries to adjust formal employment and reported earnings. We study Uruguay's AFAM-PE conditional cash transfer, in which household income eligibility is verified monthly using administrative records of formal earnings and households exceeding the threshold may lose benefits, while most beneficiaries are largely unaware of the income rule. Linking program records, social security employment histories, and firm identifiers, we build workplace networks defined six months before each peer's exposure to enforcement and identify spillovers using a fuzzy regression discontinuity design that exploits quasi-random variation in whether a peer crosses the income threshold. Coworkers of peers just above the cutoff reduce registered labor earnings by 20\%, and the corresponding fuzzy-RD estimate implies a decline of about one-third of baseline earnings in response to a peer's suspension. These effects emerge within two months and persist for at least 12 months. The response operates almost entirely on the extensive margin of formal employment, and most of the decline reflects transitions to informal work rather than exit from the labor force. Spillovers are concentrated in small firms and among similar coworkers, are nearly three times larger when the suspended peer actively contacted the program's support service, and are absent among non-beneficiary coworkers. This pattern is consistent with information transmission as the main channel. Incorporating these indirect responses lowers the estimated MVPF of the program as implemented with enforcement from 0.53 to 0.36. Because the same information channel operates wherever means-tested programs are enforced through records of formal earnings, evaluations that count only directly exposed households systematically understate the cost of enforcement.

social assistance, income-testing, enforcement spillovers, peer effects, workplace networks, informality, MVPF.