The Use of the Compensation Principle in Evaluating the Efficiency Costs of Tax Distortions

The sum of Hicksian compensation measure is shown to be the incorrect criterion for determining potential Pareto improvements. The correct criterion is a weighted average of households' compensation measures and is not, in principle, independent of the distribution of costs and benefits across households. This correct criterion yields the conventional welfare cost expression under exactly those conditions for which aggregate consumer demand systems can be consistently estimated.