On the marginal cost of public funds and the optimal provision of public goods

This paper argues that, in models with heterogeneous agents, the concept of the marginal cost of public funds (MCPF) will only be useful if it is compared with an analogous concept for the benefit side. The MCPF does not assume a unique value and is not particularly illuminating in and out of itself. Also gone is the benchmark status of MCPF = 1. Turning to the provision of public goods, using a mechanism design approach, the paper constructs a two-stage proof for Kaplow's [Kaplow, L., 1996. The optimal supply of public goods and the distortionary cost of taxation. National Tax Journal 49, 513-533.] proposition concerning the "irrelevance" of labor supply and distributional concerns in public good provision. This highlights the two fundamental ingredients for his result. First, the provision of public goods per se, when it satisfies the Samuelson's rule, is only potentially Pareto-improving. Second, the actual Pareto improvement will materialize when, or if, one reforms the income tax structure. If the reform is not forthcoming, the decision on public goods provision must rely on redistributional concerns. Finally, the paper generalizes Broadway and Keen's [Boadway, R., Keen, M., 1993. Public goods, self-selection and optimal income taxation. International Economic Review 34, 463-478.] result to a model with many types of agents, many private goods and without making any assumptions regarding which self-selection constraints are or are not binding.