Purpose – The purpose of this paper is to characterize that the marginal social cost of public funds and to estimate the response of labor supply to these publicly provided goods, and simulate the marginal social cost of cash-cum-in-kind transfers (MSCKT) for Brazil. Design/methodology/approach – The paper provides a theoretical model based on Wildasin to characterize the marginal social cost of public funds. Next it estimates using instrumental variables approach the variables necessary to calibrate our theoretical model. Findings – The marginal social cost of public funds depends on the relation between labor supply and the cash-cum-in-kind transfers. Last, the simulations suggest that MSCKT can increase up to 12.4 percent if compared with cases in which is assumed ordinary independence between labor and the bundle of goods provided by the public sector. Research limitations/implications – Further panel data experiments based on municipal public finance data should be conducted in order to circumvent the agents' heterogeneity problem inherent in cross section analysis – and individuals' labor supply response could be more sensitive at this data level. Finally, such cost-benefit analysis makes more sense when a specific project is considered and therefore its effects on the taxed good can be clearly estimated leading to a more reliable estimative of the marginal social cost of funding that project. Social implications – Governments should take the actual social cost of public policies into consideration before undertaking any new project. Originality/value – The paper is useful to characterize the marginal social cost of public funds, estimate the necessary parameters and, last, to calibrate its correspondent using Brazilian data.