Despite the extensive literature devoted to measuring the `cost of protection', or the welfare loss caused by trade restrictions relative to free trade, traditional indices may be misleading. For example, the trade-weighted average tariff typically falls with the elimination of trade in a commodity as its trade weight falls to zero. In Discussion Paper No. 683, James E Anderson and Research Fellow J Peter Neary instead propose a `Trade Restrictiveness Index'; this has a firmer basis in welfare economics and measures the degree of restrictiveness of a system of trade protection.

When only tariffs restrict trade, this Index is equal to the uniform tariff that has the same effect on aggregate welfare as the existing tariff system. Changes in the Index therefore affect economic welfare normalized by the total welfare effect of the initial tariff system, which should allow consistent comparisons of changes in the Trade Restrictiveness Index across countries and time. Where only quotas restrict trade, the Trade Restrictiveness Index equals the equiproportionate reduction in permitted import volumes that is welfare equivalent to the initial structure of quotas. When both quotas and tariffs restrict trade, the index equals the equiproportionate increase in tariffs and reduction in quota levels yielding the welfare of the initial system of trade restrictions; such goods should be viewed as quota constrained, with the tariff merely ensuring that some rents accrue to the importing country.

Anderson and Neary then illustrate some empirical short-cuts that may be used to estimate changes in their Trade Restrictiveness Index. If the protected goods are separable in a general equilibrium sense, this provides a rigorous foundation for partial equilibrium analysis; the Trade Restrictiveness Index can also be adapted to allow for different forms of rent-sharing and countries' abilities to influence their terms of trade. Measuring their index for exports of textiles and clothing from Hong Kong to the US, which face binding quotas under the Multi-Fibre Arrangement, they find that restrictiveness increased for both sides during 1982-8. Increased restrictiveness need not imply quotas were tightened, however: if there is overall economic growth, constant or even rising import quotas may imply tightening of the protective system. Finally, Anderson and Neary compare their results with those based on the trade-weighted average of the `tariff equivalents' and find that the two measures have opposite implications for trade restrictiveness in two-thirds of the observations in their case-study. They conclude that their own measure is empirically operational.

A New Approach to Evaluating Trade Policy
James E Anderson and J Peter Neary

Discussion Paper No. 683, June 1992 (IT)