Virtually all policy reform occurs in the presence of major, continuing distortions. Unfortunately, we know from theory that recommendations for policy change must take into account the presence of these distortions, and from applied research (e.g., Loo and Tower, 1990) that continuing distortions may dramatically alter the measured welfare effects of policy changes. Fortunately, developments in economic theory and modeling have expanded our ability to evaluate the sign and magnitude of welfare effects of policy changes in the presence of distortions, so that the theory of the second best can be a practical guide to policy, rather than simply a caveat on policy recommendations.

Where a formal model of the economy is available, three distinct approaches to the evaluation of welfare change in distorted open economies are now in use. The first, most common in theoretical analyses, is the Balance of Trade Function approach (e.g., Anderson and Neary, 1992; Lloyd and Schweinberger, 1988, p. 293). The second involves evaluating the welfare effects of a shock using an expenditure function or utility function embedded in a general equilibrium model (e.g., Dixit, 1975; Clarete and Whalley, 1988). The third approach uses consumer and producer surplus measures and changes in tax and tariff revenue collections (Loo and Tower, 1990).

Each of the three approaches considered has the advantage of allowing the evaluation of discrete policy changes, rather than being restricted to marginal effects. Each approach has particular advantages and disadvantages.