This paper argues for an efficiency approach to the evaluation of policy changes.  Rather than comparing the utility allocations that arise before and after a policy change is introduced, this  approach evaluates a policy change by comparing it with other possible changes which might be made from the status quo.  The main merit of the approach is that it is founded on the Pareto criterion rather than on a distributional value judgement.  The paper provides a precise statement of the approach and applies it to a number of examples.  Some objections to the aproach are also anticipated and discussed.