2.3 The theory of the second best One issue that arises when we think about optimal policy, is that standard results in economic theory concern the characterization of an efficient outcome (for example, our social planner's 9 problems, or the first and second welfare theorems, which you might have seen in Econ 391). These results are useful in characterizing states in which welfare, defined using the Pareto efficiency criterion is maximized. While this is all well and good, it's applicability to optimal policy problems is unclear. For example, we know that competitive economies are efficient as long as there are no externalities, public goods, asymmetric information, etc... . It may be the case that optimal policy can eliminate all distortions in the economy, in which case efficient outcomes are attainable. However, suppose that this is not the case. That is, suppose that there are some distortions that cannot be fully eliminated. In such cases, a Pareto efficient outcome cannot be attained. Policy problems where an efficient allocation cannot be achieved are known as second best problems. This is because an efficient allocation is one of the best (or first best) possible outcomes. When an efficient allocation cannot be attained, the best implementable outcome is only second best. It is tempting to imagine that the optimal policy in second best involves eliminating whatever distortions can be eliminated. This turns out not to be the case. In fact, eliminating one distortion when others remain present might make things worse rather than better. The theorem of the second best states that if one distortion cannot be completely eliminated, the (second best) optimal outcome typically requires allowing some (or all) of the other distortions to remain. While we are not going to prove the theorem in Econ 306, the intuition behind it is easy to illustrate in an example. Suppose that output in some industry created a negative externality, such as pollution. At the same time, the market structure in that industry was that of a monopoly. In this case there are two deviations from perfect competition-the externality, and the market power of the monopolist. Suppose that the government could not directly address the externality, but did have the ability to break up the monopoly. In this case, it would not be optimal for the government to do so. The reason is that the two distortions partially offset each other. In isolation, the market power of the monopolist implies that output would be too low relative to what a competitive market would produce. However, the externality implies that the quantity produced by a competitive industry would be too high, because of the unpriced side effects of pollution. If the monopolist's market power were to be eliminated,that the quantity produced by a competitive industry would be too high, because of the unpriced side effects of pollution. If the monopolist's market power were to be eliminated, and the industry were to become more competitive, the externality would become worse. This could lead to an overall decrease in welfare. Hence, it would be better to maintain some degree of monopoly power in this industry. The extent to which it would be welfare 9Lipsey and Lancaster (1956) were rst to state the problem of the second best in a general form. 10 improving to reduce market power would depend on the severity of the externality relative to the deadweight losses created by monopoly pricing. In general equilibrium, second best problems are even more complicated because distor tions in one market affect the outcomes of other markets. This means, for example, that it isn't necessarily optimal to eliminate market imperfections in one market, if it is not possible to eliminate imperfections in the other markets. In fact, it is difcult to characterize sufcient conditions for a welfare improvement, as these depend on all of the imperfections in all of the sectors of the economy. This means that economic policy is difcult, and people are likely to disagree on what constitutes welfare improving policy. 10