Essays on the Theory and Empirical Evidence Regarding Spatial Tax Competition.

Abstract

The United States commodity tax system is fiscally decentralized: state, county, and municipal governments have taxing authority. Geographic borders create “lines,”' which are demarcations of goods into different tax rate categories; tax rates change discontinuously at lines. A discontinuous change in tax liabilities is a “notch”' and results in deadweight loss from cross-border shopping and tax driven product innovation. This dissertation studies the consequences of lines and notches on tax competition and utilizes variation in local taxes to understand tax policy. The dissertation focuses on sales taxation, but the theories and empirical methodologies are informative for tax policy more generally and are applicable to issues of preferential taxation, optimal line drawing, and time and characteristic notches. Chapter II extends the standard model of sales tax competition to allow for tax rates to vary not only across states, but within states. I demonstrate that municipal sales taxes are lower in level terms on the high-tax side of a state border relative to municipal sales taxes on the low-tax side of a border. Using a new methodology for studying municipal responses to discontinuities, I show that municipal taxes are a function of distance to the nearest state border. After accounting for differences in tax rates within a state, discontinuous tax changes at borders become smaller. Chapter III answers whether geographic differentiation of tax rates is optimal from a state government's perspective. I demonstrate that under relatively weak assumptions, differentiation of tax rates is almost always optimal. Replacing a large distortion with many smaller distortions everywhere is optimal from a social welfare perspective. Chapter IV studies how municipal governments react to county tax rates. Using an identification strategy grounded in a theoretical model, I identify strategic reaction functions. Jurisdictions mimic the tax rates of their neighbors. I also demonstrate -- unlike state-federal reaction functions estimated in the literature -- that a higher county tax rate implies strategically lower municipal tax rates. The results in this dissertation imply that fiscal federalism and both spatial location and spatial interaction must be considered in order to understand tax competition and optimal commodity taxation.