Measuring the Restrictiveness of International Trade Policy

Abstract

James E. Anderson and J. Peter Neary Measuring the Restrictiveness of International Trade Policy The MIT Press. December 2005. ISBN: 0-262-01220-0 6 x 9, 352 pp., 34 illus. Price: $37.50/£24.95 (cloth). When you open the first page of the book, you realise that this is economics at its finest. Good economics combines theory and measurement, and this is what the authors do in developing their concept of measuring restrictiveness of trade policies through a theoretically founded index number. The basic idea is to find the uniform tariff that would yield the same welfare as the actual tariff structure. The welfare-equivalent uniform tariff is the preferred trade restrictiveness index (TRI), but the concept remains valid if alternative reference levels are used. For example, if the level of imports is kept constant instead of welfare, one obtains the mercantilist TRI (MTRI) which could have some appeal for trade negotiators. The main thrust of the index number approach is to take into account behavioural responses to changes in policies, which contrasts this approach to conventional aggregate measures such as the tariff averages. A trade-weighted average implicitly uses observed past behaviour in the aggregation, while the correct index number incorporates changes in trade shares due to policies. It will consequently not be haunted by the well known substitution bias that leads to diminishing weights for demand-elastic goods on which a tariff is imposed. This theoretically correct index number approach is indeed very appealing, but it comes at the cost of additional information needs in the form of relevant elasticities, or even complete economy-wide models, needed to perform the calculations. For example, Kee et al. (2004) estimate more than 4,600 import demand elasticities that are consistent with GDP maximisation to subsequently derive TRIs.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 161 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.