Pre-print, Working paper: This article analyzes whether the large scale provision of non-contributory health services encourages workers to move away from jobs that pay contributions to social security (formal employment). Using a difference-in-differences design, that exploits the variation generated by the municipal level roll-out of an intervention of this kind in Mexico, this paper finds that contemporaneous program exposure has no impact on the ratio of formal to total employed and that lagged exposure leads only to a small (0.78 percentage points) decrease. Two proxies of spillover effects further reveal that this estimate is robust and that the upper-bound of program effect is only moderately larger (1.5 percentage points).