When Household Heterogeneity Matters Optimal Fiscal Policy in a Medium-Scale TANK Model

We investigate the role of household heterogeneity in terms of marginal propensity to consume and of labor income for the design of optimal fiscal policy over the business cycle. We estimate a two agent New-Keynesian (TANK) medium scale model introducing aggregate shocks as in Smets and Wouters (2007) and allowing idiosyncratic shocks to impact household behavior. We further ensure that the government can set lump sum transfers and distortionary taxes to redistribute across households and finance deficit fluctuations across the business cycle. Estimating the model with US data on household earnings shows limited influence on the estimated parameters of the model, however it identifies heterogeneity across household types as a key driving force of the business cycle. Using the estimated model we solve an optimal fiscal policy problem assuming that a benevolent government sets taxes and transfers under commitment. Under optimal policy, fiscal variables display considerable volatility and respond considerably to shocks to labor income at the low end of the distribution. These shocks are also important for the optimal policy model to match the properties of fiscal variables seen in the US data