Wisconsin - Madison P Institute for \ ' . Research on Poverty Discussion

Public and private sector activities often bring about undesired income-distributional effects, and these give rise to considerations of ways to compensate the losers. When "lump-sum" compensation is not possible, "second-best" compensation arrangements need to be identified and evaluated. The term "compensation" is often used to mean explicit payments from those who gain from a policy change, to those who lose. We find it useful to include in "compensation" any arrangement that reduces, from what it would otherwise be, an unintended net loss suffered by some person. We have in mind, in particular, those losses resulting from public sector activities, but our analysis should apply to any situation in which someone is harmed as an unintended side-effect of some otherwise desirable action. The objectives of this paper are to (1) identify the various alternative ways of providing compensation; (2) determine how the alternatives differ from each other in terms of their economic impacts; and (3) determine how to choose among the alternatives so as to maximize social welfare. We identify two broad categories of compensatory mechanisms: "Implicit" arrangements, which rely on constraints imposed ex ante on public actions in order to minimize the. harm imposed on certain groups; and "explicit" compensation arrangements, which avoid imposing such constraints and instead attempt ex post to offset some or all of the harm .done. Three criteria are identified as determining which compensation mechanism is I~est." In terms of equity, ·compensation arrangements may be .~~-~---_.._---~ --------------------i .J ---------------------compared on the basis of (1) target efficiency of the compensatory benefits, and (2) the manner in which the costs of compensation are distributed. In terms of efficiency, compensation mechanisms may be compared on the basis of (3) the real costs required to provide a given amount of compensation--costs which, in the real world of nonlump-sum taxes and transfers, may be sizeable. We show that there are identifiable situations in whiGh one form of compensation is preferable to others in terms of these criteria.