We study the impact of a COVID-19 relief program on compliance with confinement measures in Italy, the early epicenter of the pandemic. We match information on the allocation of funds across Italian municipalities with data tracking citizens' movements drawn from mobile devices and vehicles' navigation systems, anonymized and aggregated at the municipality level. To assess the role of the program, we exploit a sharp kink schedule in the allocation of funds as a function of past income differentials that generated random treatment assignment in a neighborhood of the threshold point. We find robust evidence that, after the introduction of the program, mobility decreased with the amount of transfers. The impact is economically sizeable and resists bandwidth changes, with stronger effects holding in the proximity of the cut-off and the coefficient stabilizing with distance from the threshold. A battery of placebo tests supports the interpretation of results. Our evidence suggests that authorities could leverage targeted relief programs to nudge compliance with emergency measures at a relatively modest cost.