This blog post examines, based on the theory of suboptimality, the structural distortions that occur when Pareto optimality conditions are only partially satisfied, and why a greater number of conditions does not always guarantee a more efficient outcome.
To reach the most efficient resource allocation state—the Pareto optimal state—all optimal conditions must be satisfied simultaneously. Achieving Pareto optimality requires n conditions. If, for some reason, one condition remains unsatisfied while the remaining n-1 conditions are met, it’s easy to assume this state is better than one where only n-2 conditions are satisfied. However, Lipsey and Lancaster demonstrated that this common belief is not necessarily correct. That is, once one or more efficiency conditions have already been violated, it cannot be assumed that increasing the number of satisfied efficiency conditions will necessarily raise the overall efficiency of the economy. In real economies, it is common for some optimal conditions to be satisfied while others remain unsatisfied. The core question that arises at the aggregate economic level is whether it is truly desirable to continue maintaining the partially optimal conditions currently being satisfied. Since it is a frequently observed phenomenon in reality that correcting one distortion creates new ones, the task arises: rather than unconditionally striving to satisfy all optimal conditions, we must seek new conditions to find the most desirable resource allocation, premised on the reality that some conditions will always remain unsatisfied. In economics, this problem is called the ‘problem of second best’.
The problem of second best recurs across various fields of economics. Discussions on customs unions are considered a key example illustrating this problem. A customs union is an agreement to abolish all tariffs among member countries and impose tariffs only on goods from non-member countries. Proponents of free trade have long viewed a global state of free trade, where tariffs are completely eliminated among all nations, as the optimal situation. They argued that forming a customs union among some nations brings the world closer to this ideal state of free trade, and thus customs unions always enhance the efficiency of the global economy. However, Viner pointed out that customs unions can actually reduce the efficiency of the global economy. He explains the effects of a customs union by dividing them into trade creation and trade diversion. Trade creation refers to the phenomenon of new trade emerging between member countries. This increases efficiency by shifting the source of goods from countries with higher production costs to those with lower costs. Conversely, trade diversion occurs when trade with non-member countries shifts to trade with member countries. This results in a change of supply sources from countries with low production costs to countries with high production costs, thereby reducing efficiency. Therefore, whether a customs union increases the efficiency of the global economy depends on which effect—trade creation or trade diversion—is greater. If the trade diversion effect is larger, a customs union between some countries can reduce global economic efficiency. This observation applies with the same logic today and remains a key criterion in current discussions about regional trade agreements (RTAs) and the expansion of economic unions.
The second-best problem also arises in the long-standing debate over the relative merits of direct taxes levied on income versus indirect taxes levied on goods consumption. In economics, it is widely accepted that taxes distort markets and reduce the efficiency of resource allocation. While a state with no taxes imposed represents a Pareto optimum, the reality is that taxation is unavoidable. Therefore, the core challenge has been to find ways to impose taxes while minimizing market distortions. In this context, scholars like Henderson argued that indirect taxes imposed on a single good distort its relative price compared to other goods, making direct taxes—which do not affect relative prices—superior. However, Little criticized this argument, stating it holds only under the unrealistic assumption that direct taxes do not affect labor time and leisure. Indirect taxes on a single good make it harder to achieve Pareto optimality between that good and others, but direct taxes also hinder achieving Pareto optimality between leisure and other goods. Therefore, it is difficult to definitively state whether direct or indirect taxes are more efficient. Furthermore, Little suggested that imposing differential tax rates on multiple goods could potentially achieve higher efficiency than imposing only direct taxes or imposing indirect taxes only on specific goods, but he did not propose concrete, empirical methods. Subsequently, Colette and Hague discovered conditions under which direct taxes could be replaced with indirect taxes of the same revenue scale while increasing individuals’ labor hours and income. This approach involves imposing relatively higher tax rates on goods with a strong complementary relationship to leisure, and lower rates on goods in competitive relationships. For example, applying a high tax rate to goods like leisure goods, which have a strong complementary relationship with leisure, reduces their consumption. This, in turn, can curb leisure consumption itself, leading to an expansion in the labor supply.
Thus, the suboptimal problem demonstrates once again that a distinct standard of efficiency exists, one that can only be reached by considering suboptimal conditions within the structural constraints of the real economy. Because it requires exploring the most desirable outcomes within conditions that cannot be fully satisfied, rather than assuming an ideal world where all conditions are met, the suboptimal problem continues to serve as a crucial theoretical foundation in contemporary policy design and international economic analysis.
