Introduction

Public finance teaches that in a multijurisdiction setting, such as an economic union of nations, there is a strong case for assigning the responsibility for income redistribution to the central, supranational government. The rationale is twofold. First, this makes it possible to avoid the depressive effects of tax competition; in a decentralized setting, national jurisdictions may tend to lower their tax rates in order to attract, say, capital or high-income earners. Second, this is the only way to redistribute income across nations.

Even though this argument sounds strong and rather general, it is not completely clear-cut. It has been argued that decentralized jurisdictions may not be engaging in less income redistribution than central authorities, but rather may be engaging in a different kind of redistribution. This is because they can exploit informational or organizational advantages relative to the central authorities.

If indeed national governments do possess better information or more motivation to effect income redistribution, it does not necessarily follow that the central government ought to abandon all responsibility for income redistribution. Rather, the central government should take advantage of all such information and motivation in designing its optimal policy.

That is the viewpoint adopted in this study. We assume that both supranational and national governments attempt to redistribute income within their respective boundaries. Both rich and poor nations join the confederation, and the rich ones know that they are going to lose in the process of redistribution across countries. However, the central government has only imperfect information on individual countries' ability to pay: It observes only the aggregate redistribution effort of each country.