Optimal Taxation in a Federal System of Governments

An implemented tax system causes distortions which leads to a minor overall welfare level compared to a system without taxes. This deviation in social welfare is often denoted by excess burden or dead weight loss (DWL) of taxation. So the traditional optimal taxation approach comprises the implementation of a tax system which minimizes the excess burden and hence the distortions caused by the levied taxes. Therefore, the policy maker has to anticipate possible behavioral adjustments of the market participants when choosing its optimal tax policy. Assuming the policy maker will do so all effects (i.e. distortions) caused by the tax system will be internalized which means that no fiscal externalities would arise from implementing the (optimal) tax system. However, the traditional optimal taxation approach abstracts from any intergovernmental relations as the existence of only one government and accordingly only one level with fiscal jurisdiction is assumed. The question here is whether and to what extent federal structures (i.e. multileveled government structures) affect the optimal tax policy decision.
The first attempt to take into account the characteristics of a federal system
related to optimal tax policy goes back to Gordon (1983) who applied the methodology
of the traditional optimal taxation approach to fiscal federalism. Therein each
unit of government (i.e. the federal and usually several state governments) decides
independently how much of public goods to provide and in particular which tax
policy to use in funding the provided public goods. Hence, we now consider a decentralized
form of decision-making in which each unit of government chooses the
optimal tax policy in the best interest of its own residents. As a consequence of this
solely intrajurisdictional externalities are internalized analogous to the traditional
optimization approach. Though, it isn’t obvious whether this solution is also optimal
in the sense of an inter jurisdictional point of view. Sobel (1997), Wrede (1999)
and also Keen/Kotsogiannis (2002) stated that a common pool problem emerges
given that subordinated governments (i.e. state governments) are allowed to levy
taxes as well as the federal government. This means that taxation at multiple levels
lead to a shared tax base which is the fiscal analogue to the common property
resource. Due to this overlap in tax bases any separately considered optimal tax
policy at a certain level may affect the optimality character of the ...