International Tax and Public Finance

The exchange of taxpayer-specific information between national tax authorities has recently emerged as a key and controversial topic in international tax policy discussions, most notably with the OECD's harmful tax practices project and the EU's savings tax initiative. This paper analyzes the effects of information exchange and withholding taxes, recognizing that countries which agree to exchange information do not forfeit the ability to levy withholding taxes, and also focusing in particular on the effects of innovative revenue-sharing arrangements. Amongst the findings are that: (i) the transfer of withholding tax receipts to the residence country, as planned in the European Union, has no effect on equilibrium tax rates, but acts purely as a lump-sum transfer; (ii) in contrast, allocating some of the revenue from information exchange to the source country—counter to usual practice (though no less so than the EU agreement)—would have adverse strategic effects on total revenue; (iii) nevertheless, any withholding tax regime is Pareto dominated by information exchange combined with appropriate revenue sharing; and, in particular, (iv) sharing of the additional revenues raised from information provided, while efficiency-reducing, could be in the interests of large countries as a means of persuading small countries to provide that information voluntarily.

This paper studies the design of education policies in a setting in which optimal redistributive labor taxation is available. It is usually argued that the crucial parameter to guide education policy is the complementarity of education and ability. This is true only when labor supply is fixed or when income taxes are not allowed. I show that, if labor supply is endogenous and if the government can tax income in a nonlinear way, the crucial parameter is how the education elasticity of wage changes with ability. Taking the elasticity criterion into account, education subsidies are optimal in cases in which, under the complementarity criterion, education taxes would be optimal. To do this, I use an asymmetric information setting that motivates nonlinear taxation of income and education.