Policy effectiveness

Does the Pareto criterion discriminate among policy choices when the policymaker does not know the correct model of the economy? If the policymaker can specify ex ante preferences for each agent, there will typically be some policy change that improves the welfare of each agent relative to a status quo that suffers from a preexisting distortion. If in addition there are at least as many commodities as states, the second welfare theorem will typically apply: for almost every Pareto optimum, there is a policy that attains this allocation. Moreover, agents must trade under these policies; optimal allocations cannot be instituted by government fiat as they can be in the standard formulation of the second welfare theorem. The downside is that ex ante preferences impose interpersonal welfare comparisons. If we instead require that policy changes increase all possible social welfare functions, and we are allowed to perturb a base model with additional states, then all policies including the distorted status quo are optimal. The methodology of perturbations is problematic, however, and robust cases exist where at least some policies are suboptimal. Finally, the set of policies that maximize some welfare function is open; consequently, small changes in the environment usually do not call for any policy response