Given that public spending will have a positive impact on GDP if the benefits exceed the marginal cost of public funds, the present paper deals with measuring costs and benefits of public spending. The paper discusses one cost seldom considered in the literature and in policy debates, namely, the volatility derived from additional public spending. The paper identifies a relationship between public spending volatility and consumption volatility, which implies a direct welfare loss to society. This loss is substantial in developing countries, estimated at 8 percent of consumption. If welfare losses due to volatility are this sizeable, then measuring the benefits of public spending is critical. Gauging benefits based on macro aggregate data requires three caveats: a) considering of the impact of the funding (taxation) required for the additional public spending; b) differentiating between investment and capital formation; c) allowing for heterogeneous response of output to different types of capital and differences in network development. It is essential to go beyond country-specificity to project-level evaluation of the benefits and costs of public projects. From the micro viewpoint, the rate of return of a project must exceed the marginal cost of public funds, determined by tax levels and structure. Credible evaluations require microeconomic evidence and careful specification of counterfactuals. On this, the impact evaluation literature and methods play a critical role. From individual project evaluation, the analyst must contemplate the general equilibrium impacts. In general, the paper advocates for project evaluation as a central piece of any development platform. By increasing the efficiency of public spending, the government can permanently increase the rate of productivity growth and, hence, affect the growth rate of GDP.