Distributional Consequences of Fiscal Policy and Computational Methods for Heterogeneous Agent Models

This dissertation consists of two independent chapters on the use of heterogenous agent rational expectations models for studying the distributional effects of macroeconomic policy, as well as new computational methods for solving such models. 

In the first chapter, I develop a multi-country dynamic incomplete-markets model with open capital markets and non-cooperative tax setting to study how within-country income and wealth inequality shapes governments’ incentives to tax mobile capital. I derive an analytical welfare decomposition that highlights efficiency, redistribution, and tax-base erosion trade-offs, and show that inequality intensifies short-run downward pressure on capital taxes while long-run savings responses moderate this effect. A quantitative application to the four largest EU economies shows that cross-country differences in marginal propensities to consume—rooted in inequality—help explain the lack of agreement on coordinated capital taxation.

In the second chapter, I develop new perturbation methods for solving heterogenous agent rational expectations models in which the state of the economy is an infinite-dimensional distribution. I derive an analytical linear state-space representation that accommodates kinked decision rules and mass points, and propose a first-order approximation method that captures nonlinearities without imposing certainty equivalence or suffering from the curse of dimensionality. The approach achieves accuracy comparable to global methods and is particularly suited for studying distributional effects of aggregate uncertainty and long-run structural change.