Bilateral trade agreements: good or bad?

Infometrics economist Gareth Kiernan argues that bilateral trade agreements are worthwhile for New Zealand. There is much in favour of such an argument. But there are also some counterarguments. Jagdish Bhagwati has written a book explaining why such preferential trade agreements can have a down side. See Termites in the Trading System: How Preferential Agreements Undermine Free Trade.

Bhagwati is University Professor at Columbia University and Senior Fellow in International Economics at the Council on Foreign Relations. He is one of the world's leading trade economists and a long time fighter for free trade - see, for example, his books Free Trade Today and In Defense of Globalization. So when I saw he had written a book arguing that Preferential Trade Agreements (PTAs) are bad for free trade I had to read it. This is relevant for New Zealand since most PTAs are in the form of Free Trade Agreements (FTAs), a number of which New Zealand has signed in recent years - including the one signed this week with Malaysia.

The standard objection to PTAs, due to Jacob Viner, is simply that they could divert trade from the cost-efficient nonmember countries to the relatively inefficient member countries. The reason, of course, is that the nonmembers continue to pay the pre-PTA tariffs, whereas the higher cost member countries no longer have to. It is obvious that shifting production away from a low cost country towards a high cost country must sabotage the efficient allocation among countries and thus reduce total welfare. This process is known as trade diversion. Viner was the first economist to note the possibility of trade diversion arising with discriminatory reductions in trade barriers via PTAs. But the negative effects of trade diversion can go further. The liberalising country itself may also be hurt. How so? Because when a country (call it the "home" country) shifts to a higher cost within-the-PTA supplier it is buying its imports more expensively, incurring what economists call a "terms of trade" loss. The terms of trade is the ratio of export prices to import prices. As import prices increase the terms of trade decrease which implies that the volume of imports that can be bought with one unit of exports decreases.

Bhagwati notes, however, that

Trade diversion is not a slam-dunk argument against PTAs, for offsetting the loss from trade diversion can be a gain if trade creation takes place. Trade may grow because consumers in the home country now pay lower prices in their own markets; the higher cost supply from the member country is still cheaper than what the domestic consumers had to pay before the PTA was formed. Again, the import competing producers in the home country will reduce their own inefficient production as the domestic price of imports falls after the PTA comes into operation; this also leads to welfare-enhancing trade creation. Therefore, whether a specific trade-diverting PTA brings loss or gain to a country depends on the relative strengths of the trade diversion and trade creation effects. (p. 50)

The important point about trade diversion is that we can no longer assume that it does not matter how we liberalise trade. The use of PTAs is a two-edged sword in which we could end up impaled. It can matter whether we liberalise via bilateral or multilateral agreements.

Bhagwati goes on to argue that proponents of PTAs are too complacent about trade diversion. He considers seven arguments (p.52-7):

1. There is evidence of fierce competition in many products and sectors today, with few managing to escape with "thick" margins of competitive advantage that provide comforting buffers against loss of comparative advantage.Thus, even small tariffs are compatible with trade diversion as tariffs are removed from members of a PTA while they remain in place on nonmembers.
2. The thinness of comparative advantage also implies that today we have what I have called kaleidoscopic comparative advantage, or what in jargon we economists call "knife-edge" comparative advantage. Countries can easily lose comparative advantage to some "close" rivals, who may be from any number of foreign suppliers. So even if preferences today do not lead to trade diversion, the menu of products where you develop comparative advantage in a world of volatility and rapidly shifting comparative advantage will be forever changing, and any given preferences may lead to trade diversion in the near future, if not today.
3. While Article 24 requires that the external tariffs not be raised when the PTA is formed so as not to harm nonmembers, the fact is that they can be raised when the external (MFN) tariffs are bound at higher levels than the actual tariffs. In these cases, a member of the PTA is free to raise the external MFN tariffs up to the bound levels, whereas typically the scheduled tariff reductions in the PTA, when a hegemonic power is involved, will be hard to suspend. This is in fact what happened during the Mexican peso crisis of 1994, when external tariffs were raised on 502 items from 20 percent or less to as much as 35 percent, while the NAFTA defined reductions in Mexican tariffs on U.S. and Canadian goods continued. So the prospect of trade diversion actually increased, despite the intent of those who drafted Article 24.
4. Article 24 freezes only external tariffs when the PTA is formed, with no increase in the external tariff allowed. But it does not address the modern reality that "administered protection" (i.e., antidumping and other actions by the executive) is both elastic and can be used and abused more or less freely in practice. Once you take into account the fact that trade barriers can take the form of antidumping measures, which are arbitrary in their design and protectionist in their practice, there is a real danger that initially welfare-enhancing trade creation can be transformed into harmful trade diversion through antidumping actions taken against nonmembers. Thus, if a member country is gaining a market in the member "home" country, creating trade by replacing inefficient home country production with less inefficient production and imports from another member country, that pressure could be accommodated, not by allowing domestic industry to yield to these imports from a member country, but by discouraging imports from the nonmember countries by using antidumping actions against them. Thus trade-creating imports from member countries could be replaced by trade-diverting restrictions on imports from nonmember countries.
   Such an "endogenous" response of the external trade barriers, typically in the shape of antidumping actions, violates the spirit of Article 24, which explicitly prohibits trade barriers on non-members from being raised but is confined to tariffs and does not extend to "administered protection."
5. There is plenty of evidence that trade diversion can occur through content requirements placed on member countries to establish "origin" so as to qualify for the preferential duties. Thus, typically, to qualify for the preferential tariffs in PTAs that include the United States, one must satisfy requirements such as that the imports of raw materials and components must come from the United States. For example, if apparel exports to the United States are accorded preferential tariffs, they must be made with U.S. textiles. This naturally diverts trade in textiles from efficient nonmember suppliers to inefficient U.S. textile producers.
6. Many analysts do not understand the distinction between trade diversion and trade creation and simply take all trade increase as welfare-enhancing. However, some recent analysts who are familiar with the phenomenon of trade diversion have tried to estimate it using what is called the "gravity model." Dating back some decades, this equation simply explains trade between two countries as a function of income and distance. Adapting this simple equation to their use, the economists Jeffrey Frankel and Shang-Jin Wei, who pioneered the use of gravity analysis to estimate trade creation and trade diversion, estimated total bilateral trade between any pair of countries as a function of their income and per capita incomes, with bilateral distance accounted for by statistical procedures. If the countries belonged to, say, the Western hemisphere and they traded more with each other than with a random pair of countries located outside the region, that would mean that the PTA between countries in the Western hemisphere had led to trade creation. But it is clear that even if one disregards other objections, the real problem with the analysis is that more trade between partners in a PTA can take place with both trade creation and trade diversion, so that one simply cannot infer trade creation alone from this procedure. Hence, the recent estimates based on gravity equation, which are improved variations on the original Frankel-Wei approach and which sometimes (but not always) suggest that PTAs in practice have led to more trade creation than diversion, cannot be treated as reliable guides to the problem of determining whether or not a PTA has led to trade diversion.
7. Several economists have suggested that we need not worry about trade diversion and that beneficial effects will prevail if PTAs are undertaken with "natural trading partners." The initial proponents of this idea, Paul Wonnacott and Mark Lutz, declared, "Trade creation is likely to be great and trade diversion small if the prospective members of an FTA are natural trading partners." One criterion proposed for saying that PTA partners are natural trading partners is the volume of trade already between them; the other is geographic proximity. Neither really works.
   At the outset, note that though some writers, including Paul Krugman and Larry Summers, both heavy hitters, have occasionally argued as if the two criteria go together, they do not. There is no evidence that pairs of contiguous countries or countries with common borders have larger volumes of trade with each other than do pairs that are not so situated, or that trade volumes of pairs of countries arranged by distance: between the countries in the pair will also show distance to be inversely related to trade volumes. This is evident from Table 3.1 [on p.58], which contains destination-related trade volume for major regions in 1980, 1985, and 1990. There are some compelling examples. Chile shares a common border with Argentina, but in 1993 it shipped only 6.2 percent of exports to and received only 5 percent of imports from Argentina. By contrast, the United States does not share a common border with Chile , nor are the two countries close geographically. Yet in 1993, the United States accounted for 16.2 percent of Chilean exports and 24.9 percent of its imports. The volume-of-trade criterion would thus make the United States, not Argentina, Chile's natural trading partner, clearly contradicting the claim that the volume-of-trade criterion translates into the regional criterion, even in a broad-brush sense. The two criteria, and their inappropriateness in ensuring that trade diversion will be minimized and beneficial effects of the PTA guaranteed, must therefore be assessed separately, as immediately below. [This is done on p.57-60.]

The General Agreement on Tariffs and Trade (GATT) was designed to reduce trade barriers via multilateral trade negotiations. Exceptions to the multilateral nature of negotiations had to be explicitly provided for, Article 24 - referred to above - is such an exception for free trade areas and customs unions:

Article 24 --Territorial Application; Frontier Traffic; Customs Unions and Free Trade Areas

Customs unions and free trade ease (FTAs) are exempted from the MFN clause, but such an arrangement must not increase existing levels of trade restrictions affecting nonmember countries. If existing trade barriers are raised to outsiders, compensation may be required. The arrangement must lead to significant liberalization --in particular, it must cover "substantially all" trade between participating countries --and interim arrangements should lead to formation of Ff As or customs unions within a reasonable period of time. Article 24 also provides that, regardless of political status, any area that maintains its own tariffs and commercial regulations may be treated as a contracting party.

Another problem with the ever increasing number of PTAs is the "Spaghetti Bowl" that they give rise to. There are two basic problems here. The first is that when a country enters into a number of FTAs, a given commodity will be subject to different tariff rates if the trajectories of tariff reductions vary across FTAs. This is normally the case. The second issue is the fact that tariffs on specific goods must depend on where a product is supposed to originate which gives rise to inherently arbitrary "rules of origin". Bhagwati writes

With PTAs proliferating, the trading system can then be expected to become chaotic. Crisscrossing PTAs, where a nation had multiple PTAs with other nations, each of which then had its own PTAs with yet other nations, was inevitable. Indeed, if one only mapped the phenomenon, it would remind one of a child scrawling a number of chaotic lines on a sketch pad. (p.61)

Rules of origin are there to determine which product is made by whom. But in this globalised world where multinational firms source components from all around the world trying to determine the origin of a given good is in Bhagwati's description "a mug's game". It is virtually impossible to say which product is whose. This gives rise to endless problems. As Bhagwati explains

There are in fact numerous cases where such questions have led to disputes that come for resolution before arbitration and bilateral dispute settlement panels. In a classic case, the U.S. Customs Service refused to certify Hondas produced in Ontario, Canada, as "North American," and hence eligible for duty-free exports from Canada to the United States, on the grounds that, in its own estimation, Canadian Hondas did not meet the local content requirement of more than 50 percent imposed by the Canada-U.S. Free Trade Agreement (CUFTA). Honda countered that its estimates showed that they did. There is no surefire, analytically respectable way to determine the truth in such a case: it all boils down to who has greater stamina and whether Honda is willing to put moneys into legal costs. (p.68)

Such problems so not arise if there is a multilateral agreement which imposes the same tariff on goods from all countries. Bhagwati quotes Hong Kong businessman Victor Fung, from the Financial Times, on the distortions and costs imposed on business by the spaghetti bowls,

Bilateralism distorts the flow of goods, throws up barriers, creates friction, reduces flexibility and raises prices. In structuring the supply chain, every country of origin rule and every bilateral deal has to be tacked on as an additional consideration, thus constraining companies in optimizing production globally. In each new bilateral agreement, considerations relating to "rules of origin" multiply and become more complex. This phenomenon is what trade experts call the "spaghetti bowl effect." While larger companies have a hard time keeping track, for small groups it is impossible. Bilateral agreements cause the business community to work below its potential. In economic terms, bilateral agreements destroy value. If left unchecked, their continued growth has the potential to hinder the development of the global production system. (p.70)

Additional problems enter the picture when "trade-unrelated" demands are placed on an FTA. Such issues are easier to put in PTAs than multilateral agreements where the possible number of parties who will oppose the move is much greater. Issues such as intellectual property protection, which has more to do with collecting royalties than with trade, is an obvious example. Other examples would be "values-based" demands on things such as labour standards and environmental standards. In many cases demands to harmonise such standards are just a form of protectionism for oneself against foreign rivals.

So the issue of New Zealand's FTAs just got a whole lot more complicated. How we deal with these issues will determine just how beneficial our FTAs turnout to be. If there is anything to the arguments above it would suggest we may need to rethink a position with regard to bilateral v's multilateral trade agreements.