We analyze the procyclicality of fiscal policy in an economy where government expenditures provide a public good complementary to private consumption and the existence of an informal labor market makes the tax base more sensitive to economic conditions. We solve for the optimal policy in a small open economy with these characteristics subject to technology and foreign interest rate shocks. The model rationalizes the observation that government expenditures are procyclical and labor taxes move countercyclically in emerging economies. The results are robust to the addition of foreign government debt subject to default risk.

We study how the threat of tax revolts constrains the optimal design of tax policy. In our model, low-skilled workers disagree about whether tax cuts on the rich will ultimately benefit them. Workers are Bayesian and form beliefs about the elasticity of substitution between skills in production—the key determinant of the strength of trickle down—based on observations of past tax changes and their own, idiosyncratic history of wages. Following a tax cut on top earners, workers with low wage realizations will believe the elasticity of substitution to be high, and will therefore favor high taxes on top earners. Workers accumulate a stock of anger that grows when the tax schedule is likely to be much more regressive than the optimal level given their beliefs. We characterize optimal tax policy for a government that faces a trade-off between maximizing social welfare and minimizing the occurrence of a revolt.

The design of a fiscal rule involves a fundamental trade-off between committing governments to a fiscally responsible budget and giving governments the discretion to respond to shocks. This paper quantifies the trade-off for members of the European Union. Novel evidence of a Pareto tail of the distribution of shocks to the need for government spending implies that a well-designed fiscal rule for EU members should grant more discretion than the discretion granted by the Stability and Growth Pact.

Using detailed household expenditure and embodied emissions data, we document that low income household expenditures have higher embodied emissions per dollar than that of high income households. This implies that a carbon tax may be regressive. Using a simple model with clean and dirty goods with low and high income households, we show that the optimal carbon tax is decreasing in income inequality. We then build a standard incomplete markets model with climate change to quantify the optimal carbon tax and how it depends on income and wealth inequality.