Paul Wonnacott Middlebury College and Ron Wonnacott University of Western Ontario
In a recent article in the American Economic Review, Kyle Bagwell and Robert W. Staiger (1999, hereafter B & S) have made a major contribution to answering the oft-posed puzzle (p. 215): What can governments gain from a trade agreement? Their answer is that:. . . a trade agreement is appealing provided that an inefficiency . . . exists under unilateral tariff setting.. . . Previous authors have established that the classic terms-of-trade externality creates an inefficiency in unilateral trade policies.
For example, suppose a country unilaterally raises its tariff to improve its terms of trade, and its partner -- faced with a resulting terms-of-trade loss -- responds by raising its tariff, with both countries losing. A trade agreement that prevents such unilateral actions will allow these countries to avoid this loss.
The difficulty is that B & S conclude that this terms-of-trade issue is the only reason for governments to enter a trade negotiation (p. 216, ital. added): "trade agreements are appealing to governments solely as a means to remedy the inefficient terms-of-trade-driven restrictions that arise when trade policies are set unilaterally." Specifically (p. 219, ital. added), "trade agreements are appealing to governments if and only if they serve as a means to remedy the inefficient terms-of-trade-driven restrictions in trade that arise under unilateral trade policies."
While this conclusion follows from the model that B & S use, surely it cannot be the whole story. In practice, protection may be motivated, not just by unilateral attempts to improve terms of trade, but also by a multitude of other objectives, including the desire to:
1. increase incomes (or prevent trade-driven losses) in import-competing industries
2. increase employment
3. improve the balance of trade.
It may or may not be sensible for countries to increase tariffs for any of these reasons, but they do.
In each case, trade agreements may be appealing to governments as a way to prevent an inefficiency that arises from unilateral trade policies. For example, suppose one country unilaterally raises its tariff to increase incomes in its import-competing industry. Faced with the resulting damage to its export industries, its partner may respond by raising its tariff, and both countries lose. A trade agreement that prevents -- or reverses -- such unilateral action allows both countries to avoid this loss. In short, the justification for the GATT is that it restrains unilateral protection, regardless of its political or economic motivation, whether it be terms of trade, employment, the protection of rents, pure vote seeking, or whatever. Moreover, this conclusion holds even if each country has a different motive, or combination of motives, for protection.
How then can B & S conclude (p. 223, ital. added) that "the terms-of-trade externality is indeed the only inefficiency that a trade agreement can remedy"? The answer is that B & S include "political" motives for protection, such as 1-3 above (in addition to the traditional motive of increasing real national income), in each government's welfare function, where they are accepted as given.[1] But actual trade negotiations frequently do not take such objectives as given, but instead aim at modifying them. For example, the GATT reduces the freedom of political leaders to change trade policies in order to gain special-interest votes at public expense. Thus, many justifications of the GATT are missed if the B & S model is used.[2] If, in analyzing negotiations, economists include "political" objectives as given in the welfare functions, some of the main objectives of actual trade negotiations will be excluded.[3] In addition, some of our most cherished trade principles may disappear, such as the theorem that a small country can improve its welfare by unilaterally eliminating its tariff.
The appropriate conclusion is really quite simple: countries engage in multilateral trade negotiations because they can get something of value -- namely a modification of foreign tariff barriers or other policies -- which is not achievable through unilateral action. There is a strong presumption that a country gains when its trading partners reduce their barriers -- regardless of the original motive for those barriers. B & S (e.g., pp. 215-216) recognize this benefit only when terms of trade provide the original motive for these barriers.
The "something of value" may go far beyond tariffs, to include (for example) modification of quotas or modification of restrictions on services. Such steps may result in a change in the terms of trade, but this does not have to be the case. For example, a relaxation of foreign quotas or other restraints on a country's manufacturing exports might result in a gain to that country, without there necessarily being a change in price; the price exceeds the marginal cost.[4]
Our main conclusion -- that any country has an interest in the reduction of foreign barriers -- would scarcely surprise a business executive or a trade negotiator. The puzzle is: Why have we been puzzled that countries engage in multilateral trade negotiations?
1. B & S make reference to the other,"political" objectives on p. 217 and especially their fn. 2, p. 216, where they write: "A political motivation for trade agreements might arise if governments seek such agreements to gain commitment relative to their private sectors." However, they then throw the importance of this point into question: ". . . whether this commitment theory of trade negotiations offers an interpretation of the basic principles of GATT is still an open question." In their concluding comments (pp. 242-43), it is clear that their framework excludes the possibility that governments seek a trade agreement as a commitment device.Our main point is that the objectives such as 1-3 above, which are not analyzed by B & S, are central -- not peripheral -- in understanding the motivation of actual trade negotiators. In contrast, the terms-of-trade motivation is not very relevant to the real world. As Harry Johnson noted (1953-54, p. 153, fn.): "A further implication. . ., and one which is too obvious to deserve elaboration, is that the optimum tariff theory is not a very useful approach to the explanation of tariff history."
As an aside, note that, even though policy makers may pay little attention to terms-of-trade issues, such issues cannot be ignored in any economic analysis comparing alternative trade policies. For example, those who have tried to prove that a policy of unilateral tariff reduction (UTR) is at least as beneficial as a reciprocal trade agreement have often assumed no change in the terms of trade. But this "small country" assumption -- that ensures the home country against a terms-of-trade loss under UTR -- also provides a terms-of-trade benefit from reciprocal trade agreements with a large partner. (By removing or reducing its tariff, the large country allows the small country to trade at, or closer to, the large country's more favorable domestic terms of trade. When the large country's tariff is reduced, its offer curve rotates, thus improving the terms of trade of the small country.) Because they have ignored this (and other) benefits available to the home country under reciprocity but not UTR, proofs that UTR is weakly superior to negotiated agreements are not valid.To Text.
2. As an aside, one can also ask whether the B and S model is missing something in arriving at what seems to be a broad theoretical rejection of preferential agreements (p. 218): "More generally, we offer as our third broad conclusion that preferential agreements pose a threat to the efficiency properties of the existing multilateral system."This conclusion holds only within their framework but not beyond, where there are many ways that a preferential agreement may increase the efficiency of the existing trading system. For example, increased competition within a CU or FTA may reduce protection-based rents there and shrink previously-protected industries and their political power to resist multilateral liberalization. This may well make member countries willing to accept deeper tariff cuts in the GATT. In the B & S model, any such weakening of protectionist interests moves member countries' welfare frontier out, closer to the Pareto optimum, thus allowing members to realize a higher level of real income while still satisfying their domestic political objectives (which, due to the preferential agreement, are now less restrictive).
Of course, by creating preference-based rents, a CU or FTA may on balance have the opposite effect of increasing the resistance to a GATT agreement that would reduce or eliminate these rents. For these and other reasons, no general, a priori theoretical judgment for or against preferential agreements is possible; each must be judged on its own. To Text.
3. On p. 216, B & S write:The assumption that governments maximize national income, however, stands in contrast to the manifest political constraints under which real governments operate. It is thus important to consider further the rationale for a trade agreement, within a richer model in which governments may have political concerns.This may leave the impression that the topic is being broadened to deal with political concerns, when in fact its being narrowed. Specifically, their framework does not allow the analysis of real-world "political" negotiating objectives such as agreements to limit increased protection for specific industries. To Text.
4. In addition, one of the GATT agreements -- namely, that limiting subsidies -- is aimed at preventing countries from taking steps that would cause a negative movement in their terms of trade. Within the B & S framework, this agreement -- to prevent countries from taking steps that would improve their partners' terms of trade -- is a complete mystery. Yet such an agreement has in fact been negotiated. To Text.
Bibliography
Bagwell, Kyle and Robert W. Staiger. "An Economic Theory of GATT." American Economic Review, March 1999, 89 (1), pp. 215-48.
Harry G. Johnson. "Optimum Tariffs and Retaliation." Review of Economic Studies, 1953-54, 21(2), pp. 142-53.
Beginning of Paper
Return to List of Papers
Return to Econ Faculty page