CHAPTER 5
TRADE IN GOODS
Governments pursue trade policies for a variety of reasons, including as a means to raise revenue, to protect specific industries (whether infant, strategic, senile or other), to shift the terms of trade, to attain certain foreign policy or security goals, or simply to restrict the consumption of specific goods. Whatever the underlying objective, an active trade policy redistributes income by transferring resources to specific industries and the factors of production employed there, generally at the expense of domestic consumers and taxpayers. It usually does so in an inefficient and nontransparent manner, and for precisely that reason tends to be supported by interest groups that lobby for import restrictions.
The GATT regulates the use of trade policies by WTO members. It does not address the basic question of whether governments should use domestic or trade policies to achieve particular objectives. That is, the issue of efficiency is not addressed directly. The premise is that inefficient instruments such as trade policy must be accepted, and that the best that can be achieved is to discipline the use of different types of trade policies and the level of the associated trade restrictions. Thus, although countries are free to use trade policies, the GATT generally encourages them to use fewer trade-distorting measures. General Agreement on Tariffs and Trade rules are mostly consistent with what economic theory would recommend in many circumstances, but only in the sense of moving governments to use second- rather than third-best instruments. The objective is to avoid the worst by accepting some bad in government intervention in trade.
Three broad categories of trade policy instruments can be distinguished: measures that affect quantities, restricting the volume or value of transactions; those that directly affect prices, involving the imposition of a monetary charge (tax) on foreign suppliers; and those that indirectly affect quantities or prices. Virtually any policy or action by a government may have an effect on trade (see Deardorff and
trade in goods 185
Stern, 1998, for a discussion of the many types of policies that have an effect on trade). As noted in Chapter 3, this is explicitly recognized in the GATT, in that dispute settlement procedures allow for so-called nonviolation complaints to be brought. Any policy—whether or not it is prohibited under GATT—can be contested if it acts to deny a benefit under the WTO.
This chapter summarizes the main GATT disciplines relating to specific instruments of trade control. For convenience, Table 5.1 lists the major articles of the GATT Articles dealing with contingent protection (Articles VI, XII, XVIII and XIX) and general exceptions (Articles XX and XXI) are discussed in Chapter 9.
5.1. Tariffs: Articles I and II
Article I GATT requires that a product made in one member country be treated no less favourably than a 'like' (very similar) good that originates in any other country. Because the initial set of contracting parties to the GATT was quite small (only 23 countries), the benchmark for MFN is the best treatment offered to any country, including countries that may not be a member of the GATT. Similar wording now applies under the WTO.
Most favoured nation applies to all measures applied or enforced at the border as well as to domestic policies covered by the national treatment rule (Article III). To determine if discrimination between suppliers occurs it is necessary to compare apples to apples. The question of how to determine whether products are 'like' each other has been a major area of case law on national treatment. In the case of MFN matters are simpler in that recourse can be and is made to the tariff classification that is used by WTO members: products that are classified under the same heading should be treated the same. We discuss the methods used by countries to classify goods later in this subsection.
The primary focus of both Article I and Article II is tariffs. The customs tariff is in principle the only instrument of protection allowed under the GATT for nonagricultural products, given that quantitative restrictions (QRs) are prohibited by Article XI and domestic measures are subject to the national treatment rule (Article III). The preference for tariffs is consistent with economic theory. Reasons why tariffs are preferable to QRs include the following (see also Annex 2):
• Tariffs maintain an automatic link between domestic and foreign prices, ensuring that the most efficient supplier continues to be able to serve the market. This link is cut with quotas.
186
the political economy of the world trading system
Table 5.1. Major GATT articles
Article Summary
I General MFN requirement. .
II Tariff commitments (schedules of bindings).
III National treatment.
V Freedom of transit of goods.
VI Allows antidumping and countervailing duties. Superseded by the GATT 1994
Agreement on Antidumping, and the Agreement on Subsidies and Countervailing Measures.
VII Requires that valuation of goods for customs purposes be based on actual value.
Superseded by the GATT 1994 Agreement on the Implementation of Article VII.
VIII Requires that fees connected with import/export formalities reflect actual costs
of services rendered.
IX Reaffirms MFN for labelling requirements and calls for cooperation to prevent
abuse of trade names.
X Obligation to publish trade laws and regulations; complemented by the WTO's
Trade Policy Review Mechanism and numerous notification requirements in specific WTO agreements.
XI Requires elimination of quantitative restrictions.
XII Permits trade restrictions if necessary to safeguard the balance of payments.
XIII Requires that quotas, if used, be administered in a nondiscriminatory manner.
XV Contracting parties to refrain from using foreign exchange arrangements to
frustrate the intent of the provisions of the GATT. Calls for the CONTRACTING PARTIES and the IMF to cooperate and consult with regard to exchange . arrangements and restrictions that are in the jurisdiction of the IMF.
XVI Notification and consultation requirements for subsidies. Export subsidies in
principle not to be used after 1958 if they result in under-cutting of domestic producers; may not result in a 'more than equitable share' of world markets. Superseded by the WTO Agreement on Subsidies and Countervailing Measures.
XVII Requires that state trading enterprises comply with MFN.
XVIII Allows developing countries to restrict trade to promote infant industries and to
protect the balance-of-payments (imposing weaker conditionally than Article XII).
XIX Allows for emergency action to restrict imports of particular products if these
cause serious injury to the domestic industry. Complemented by the WTO Agreement on Safeguards.
XX General exceptions provision-allows trade restrictions if necessary to attain
noneconomic objectives (health, safety).
XXI Allows trade to be restricted if necessary for national security reasons.
XXII Requires consultations between parties involved in trade disputes.
XXIII GATT's main dispute settlement provision, providing for violation and
nonviolation complaints. Complemented by the WTO Understanding on Rules and Procedures Governing the Settlement of Disputes (DSU).
XXIV Sets out the conditions under which the formation of free trade areas or customs
unions is permitted. XXVIII Allows for renegotiation of tariff concessions.
XXVIII bis     Calls for periodic MTNs to reduce tariffs. XXXIII Allows for accession. Superseded by Art. XII WTO.
Part IV        Calls for more favourable and differential treatment of developing countries,
trade in goods
187
• It is easy to ensure nondiscrimination between foreign sources of supply using tariffs; under a quota this is much more difficult. Quota allocation is often based on arbitrary decisions of officials.
• Tariffs are transparent. Once established, every trader knows the price of market access for specific products. This is not the case under a quota, where the conditions of market access may depend on timing (for example, under a first-come, first-served allocation scheme), past performance (if quotas are allocated based on historical utilization rates) or corruption (bribery of the officials responsible for licensing).
• Tariffs are also more transparent in that the level of nominal protection under tariff is easily calculated, whereas its estimation is more complex under a quota.
• Tariffs generate revenue for the government, whereas under quotas the tariff equivalent may go to the exporters or to intermediaries, depending on how the quotas are allocated. In most cases, governments do not receive the created rents—the extra revenue per unit sold that is due the price-increasing effect of restricting supply is transferred to those who have the quota rights. This is a major incentive for lobbying and rent seeking.
• Tariffs are also more efficient because they reduce lobbying incentives. Tariffs benefit the whole industry producing the protected good, reducing the returns for individual firms to lobby for protection. If quotas are an option, traders may seek individual quota allocations that are as large as possible, inducing socially wasteful lobbying.
Governments may levy tariffs on imports and exports, but import tariffs are by far the more important in practice. Customs tariffs may be:
(1) ad valorem (a percentage of the value of imported products);
(2) specific (a given amount of money per physical unit, say US$1.5 Per litre of wine);
(3) a combination of the two (e.g. 5 per cent ad valorem plus US$i per litre of wine).
The GATT does not favour one type of tariff over another.1 In practice, most tariffs are ad valorem. Each may have advantages in specific situations. Ad valorem rates are more transparent, and are indexed. If the nominal value of a product increases (because of inflation for example), then tariff revenue will keep pace with price increases. Specific tariffs have the advantage of not requiring customs authorities to determine the value of imports when entering the country, and are by definition not sensitive to changes in the value of goods.
' WTO members may switch from ad valorem to specific duties if they desire, as long as this does not result in a violation of a tariff binding.
188
the political economy of the world trading system
Product classification
For the GATT machinery to work it is necessary that traders know how their products will be classified for customs purposes, and that members have a common understanding regarding what goes where. Tariff commitments are made at the so-called tariff line level, which are often country-specific, even though most countries use internationally developed systems to classify imports. The main coding systems used for classification purposes during the first 40 years of the GATT era were the Brussels Tariff Nomenclature (BTN) and its successor the Customs Cooperation Council Nomenclature (CCCN). In the late 1980s, countries switched to the Harmonized Commodity Description and Coding System (HS). All these systems were developed by the Customs Cooperation Council in Brussels—which became the World Customs Organization in 1994.2 The HS allows for a greater range of product classification than its predecessors and is also used for reporting of trade statistics (Box 5.1). More than 200 countries and economies, representing 98 per cent of world trade used the HS in 2008.
The use of the HS by WTO members is important to traders as it reduces uncertainty regarding the treatment of their goods by customs authorities. Knowing how products will be classified allows producers to finetune their production process and tailor and package their products in a way that minimizes the expected duty burden (Box 5.2.). The tariff structure of many countries is quite differentiated—close substitutes may be subject to widely varying tariffs, providing potential opportunities for producers to arbitrage across product (sub)headings.
WTO disciplines for tariffs
There are two basic rules for tariffs. First, tariffs and other policies may not discriminate between foreign products (Article I). Second, they must be bound (Article II). The main exceptions to the MFN rule are if countries are members of preferential trade agreements (Chapter io)> provide tariff preferences in favour of developing countries (Chapter 12) or confront imports from a nonmember country.
2 The genesis of the WCO began with a 1947 Study Group created by 13 European governments with the mandate to examine the possibility of establishing one or more inter-European Customs Unions based on the newly negotiated GATT. In 1948, the Study Group set up two committees—an Economic Committee and a Customs Committee. The Economic Committee was the predecessor of the Organization for Economic Cooperation and Development (OECD); the Customs Committee became the Customs Cooperation Council (CCC). The WCO is now a global organization with 173 members.
trade in goods 189
Box 5.1. The Harmonized Commodity Description and Coding System (HS)
The Harmonized System provides a legal structure and product typology for the purpose of tariff classification. It comprises 1,241 headings grouped into 96 chapters (in turn subdivided into 21 sections). The system's 5,000 subheadings are identified by a six-digit code, each of which is carefully defined and described, with rules to ensure uniform application. The first four digits indicate a product group or family. The fifth and sixth digits designate specific product lines. For instance, 8470.10 is the code for 'electronic calculators capable of operating without an external source of electronic power and pocket size data recording, reproducing and displaying machines with calculating functions', whereas 8470 is the code for electronic calculating machines as a group. The HS is implemented through an international convention that requires signatories to apply it in. a uniform fashion at the six-digit level. Beyond the six-digit level, countries may introduce national idiosyncrasies in their coding systems, as long as these are subclassifications of the applicable six-digit category (Article 3.3 HS). Given that such subclassifications are left to the discretion of countries, they can, in principle, be challenged before the WTO, although this has not occurred so far.
The HS nomenclature builds on an earlier standard, the CCCN, and incorporates certain aspects of the tariff schedules of the United States and Canada. The HS came into effect in 1988, and is used not only for customs purposes, but also for collection of trade statistics and associated transactions (such as transport and insurance). Before the adoption of the HS, countries tended to use the Standard International Trade Classification (SITC) for statistical purposes—developed under UN auspices. This differed substantially from the CCCN. The use of one classification system for both customs and statistical purposes greatly facilitates analysis and monitoring of tariff revenue collection and trade flows.
Source: WCO, The 2007 Harmonized System, Explanatory Notes (at http://www.wcoomd.org).
Most favoured nation implies that WTO members must extend any advantage immediately and unconditionally to all WTO members. Unconditional means no strings attached' For example, a country cannot condition a tariff on exporting countries satisfying a specific labour standard (that is, impose a higher rate on countries that do not satisfy whatever criterion is imposed) even if domestic producers are subject to it. However, if a country offers preferential access—that is, better than MFN—as is permitted for developing countries, this may be conditioned on the satisfaction of certain criteria (see Chapter 12). Disputes can be brought to the WTO alleging not just de jure violation of MFN, but also de facto violation.3 Moreover, a complainant need not show actual trade effects. It suffices
3 In Canada—Autos, a WTO dispute settlement panel was asked to consider whether a measure that limits the benefits of an import duty exemption to a certain class of domestic importers without imposing any restrictions regarding the origin of the imported goods constitutes a de facto violation of MFN. The panel and the AB found against Canada, arguing that the extensive intrafirm trade in automotive products between Canada and the US implied that the Canadian duty exemption scheme was likely to benefit imports from the US.
190
the political economy of the world trading system
Box 5.2. Product strategy and customs classification
Average customs tariffs on agricultural products tend to be substantially higher than on industrial goods. Product classification by customs consequently may have important implications for an exporter's competitive position in a given market. In such cases managers will analyse the differences between product duty rates in order to adjust their product strategy. Three examples:
• The Canadian Dehydrators Association exports alfalfa feed products all over the world. It monitors any modifications in customs rates for particular feed product lines in order to modify its export product mix to minimize import charges. One option that is used is to mix dehydrated alfalfa products with other types of feed (such as grains) in order to benefit from lower customs rates.
• An East European company exporting food ingredients to Japan used to mix its powder milk with fat and then separate it again in Japan. It did this to pay lower customs duties on 'frozen cheese' rather than higher duties levied on powder milk and on animal fats when exported separately.
• A US sugar refinery imports 'stuffed molasses', a syrup with a high sugar content, to produce a liquid sugar product for sale to ice-cream and confectionery companies, allowing it to avoid high duties on sugar.
Source: Kostecki (2001).
that a WTO member creates more favourable competitive opportunities for some countries that are not extended to others.
World Trade Organization members may not raise tariffs above the levels they have bound in their schedules. The tariff concessions made by members upon accession or in periodic MTNs are expressed in the form of tariff bindings inscribed in each member's tariff schedule (Article II). By binding its tariff, a member undertakes not to impose a duty on a specific product that is higher than the bound rate. A binding may be identical to the currently applied rate; it may comprise a so-called ceiling rate that is higher than the applied rate; or it may consist of a pre-commitment, a negotiated rate that is lower than the currently applied rate. The last possibility often arises after a MTN has been completed, with the negotiated rate entering into force at a specified future date. Tariff binding does not preclude levying a tax at the border that is equivalent to an internal tax as long as national treatment is satisfied, i.e. governments can collect applicable excise or other indirect taxes at the border (Article III:ia). For example, in India—Additional Duties on Imports, a 2008 panel report found that additional duties imposed on US imports were equivalent to internal taxes imposed by Indian States and thus legal. (Note that the burden of proof in such cases is put on the complainant.)
A tariff binding establishes a benchmark for the conditions of market access that a country commits itself to. Any measure taken or supported by a government that has the effect of nullifying or impairing the concession implied by its tariff bindings justifies complaint by trading partners. There is no need to show an impact on
trade in goods
191
trade. Thus, the binding not only restricts the possibility of raising tariffs, but also limits the possibility of using measures that have an equivalent effect, such as indirect taxation (however, as discussed in Chapter 9, there are various ways around this constraint).
World Trade Organization members have agreed to apply the HS at the six-digit level of disaggregation. Thus, up to that level tariff classifications are identical. Beyond the six-digit level—that is, for finer degrees of differentiation—countries are free to define subcategories of products. Although the HS helps to make the customs environment more predictable, it is important to note that WTO members sometimes bind their tariffs at the 10 or 12 digit level of disaggregation. At this level, classification is country-specific. However, members may not impose tariffs at a finer level of disaggregation that exceed the binding that has been made at a more aggregated level.
To a significant extent the tariff schedules determine the relevance of GATT rules. Formally, the product coverage of the GATT for each member is determined by a positive list approach. Each member includes in its schedule the products (tariff lines) on which it is willing to make tariff commitments. These schedules form an integral part of the GATT. The comprehensiveness of tariff bindings for members has traditionally varied considerably. For most industrialized market economies, the share of bound tariffs in the total number of tariff lines has always been high for manufactured goods. The coverage of bindings for most developing countries has historically been very low or nonexistent, and usually limited to ceiling bindings that are significantly higher than applied rates.
Many contracting parties joined GATT 1947 after becoming independent in the late 1950s or in the 1960s. Such former colonies were allowed to accede to GATT without tariff negotiations. Developing countries were also granted special and differential treatment, allowing them not to offer concessions in MTNs (see Chapter 12). This implied that only nondiscrimination disciplines applied, as countries are free to raise tariffs if there are no ceiling bindings. During the Uruguay Round an attempt was made to expand the coverage of bindings by requiring all WTO members to submit tariff schedules. In contrast to GATT 1947, WTO membership requires a schedule of commitments. Moreover, all WTO members are obliged to bind 100 per cent of their agricultural tariff lines—a major change in comparison with GATT 1947, under which agriculture had largely become exempt from disciplines (Chapter 6). Although there are no rules concerning the product coverage of tariff schedules for nonagricultural goods, developing country participation as measured by the scope of tariff bindings increased substantially during the Uruguay Round. This reflected a realization on their part that greater participation in the multilateral trading system was beneficial, as well as significant unilateral liberalization undertaken by many countries since the early 1980s. The share of industrial tariff lines bound by developing countries increased from 22 to 72 per cent (Table 5.2).
192
the political economy of the world trading system
Table 5.2. Tariff bindings for industrial products, pre- and post-Uruguay Round
Country Group	Number of Lines	Percentage of Tariff Lines Bound Pre-Uruguay Post-Uruguay	
Developed countries	86,968	78	99
Developing countries*	157,805	22	72
Transition economies	18,962	73	98
Note: "Data span 26 countries accounting for 80 percent öf the total trade of countries participating.
Source: GATT (1994c).
Binding coverage was above 95 per cent of tariff lines in developed and transition economies in 2007, but developing-country tariffs are either not bound or bound at relatively high levels. Low binding coverage is particularly prevalent for Asian countries (averaging 40 per cent). There is also considerable variation in the share of bound duty-free tariff lines. For developed countries this ranges from 17 per cent for Switzerland to close to 50 per cent for Japan. The share of lines with bound zero duties in developing countries is very low (Bacchetta and Bora, 2004).
Most developing country bindings relate to 'ceilings' (maximum rates), not applied tariffs. Ceiling rates are less valuable than if applied rates were inscribed, but do have value, as they establish an upper bound on the downside risk confronted by traders and investors (Francois and Martin, 2004). The difference between the average applied and bound rate is a measure of what has been called the binding overhang. The greater this difference, the deeper the average tariff cut that must be realized in a MTN for the outcome to imply actual liberalization. It is often not realized that the focus of MTNs is on tariff bindings, not on applied rates. An implication is that a claim that a MTN reduced average tariffs by, say, 30 per cent does not necessarily imply that applied tariffs have fallen by that extent.
Of the 153 members of the WTO, most have scheduled tariff bindings that are far above applied tariff rates. Only eight members have mostly bound their tariffs at applied levels: the EU, the US, Japan and Canada (the Quad), plus China, Chinese Taipei (Taiwan), Hong Kong and Macao. These eight 'full' members account for 67 per cent of world trade and 78 per cent of global GDP (Table 5.3). Although most trade is therefore fully subject to WTO rules, these statistics also imply that one-third of global trade is not 'secure' in that the governments concerned can sign-ificandy increase tariffs if they desire without violating tariff bindings. On average, the majority of WTO members can raise applied tariffs threefold if they wish
Table 5.3. Average applied and bound tariff levels, 2008
WTO Members Global Import     Global GDP Simple Average Tariff Import Weighted Tariff
Share Share------
Industry Agriculture Industry Agriculture
			Bound	Applied	Bound	Applied	Bound	Applied	Bound	Applied
Eight large 'full' members	67.1	78.1	4.1	3.9	13.1	13.4	3.6	3.4	11.4	11.8
26 largest other members	26.2	17.5	27.6	7.9	65.8	19	21.5	7.1	60.2	22.8
All other members	2.0	4.4	34.3	9.9	62.3	17	26.5	9.6	49.7	19.8
Source: Messerlin (2008a).
194
the political economy of the world trading system
without incurring any penalties. Least developed countries have even greater 'headroom' to raise tariffs. A post-Cancun Decision of the General Council called for LDCs with a binding coverage less than 35 per cent, while exempt from making tariff reductions, to bind 100 per cent of their tariff lines at the average applying for all developing countries.
Why don't countries bind tariffs at applied rates? In many cases it simply reflects the mercantilism underlying the GATT (that is, a perception that bindings are negotiating chips). In addition, finance ministries may be opposed to losing a revenue raising tool. The latter consideration is particularly important for low-income countries that are dependent on taxation of foreign trade for revenue. Finally, countries with overvalued exchange rates and resulting foreign exchange shortages and rationing often have no wish to be subjected to GATT surveillance in instances where measures are required to safeguard the balance of payments (see Chapter 9).
In a MTN context, a country must determine what the cost-benefit ratio is of binding or not binding at applied rates. One source of benefit is the quid pro quo that may be realized in terms of improved access to foreign markets. This is the benefit that most often appears to be uppermost in the minds of policymakers. However, this is arguably not the key dimension of the tradeoff, especially for small countries with little if any negotiating power. More important is to take into account that binding at or below applied rates can be a powerful tool with which to signal to investors that the government is serious about reducing uncertainty regarding its future trade policy stance. This can have substantial payoffs in terms of reducing risk premia demanded by investors and stimulating capital inflows.
Tariff-related disputes
Numerous disputes have been brought over the years relating to allegations of violations of tariff bindings. These disputes have also clarified the rules of the game. An example is a 1997 case brought by the US against the EU, which had reclassified for tariff purposes certain local area network (LAN) adapter equipment and personal computers with multimedia capability. The US alleged that this violated EU tariff bindings (Article II:i GATT). The panel found against the EU. On appeal, the Appellate Body (AB) reversed the panel's conclusion, ruling that the US had not satisfied the burden of proof and that the panel should have examined the HS in interpreting the EU tariff commitments. But the AB did not come to a clear conclusion regarding the matter because it may not rule on factual dimensions of a case, but only on the legal reasoning of the panel. The EU eventually abolished tariffs on the products concerned as a result of the Information Technology Agreement (see Chapter 11).
trade in goods
195
A 2003 case, also against the EU (EC—Chicken Cuts), illustrates the type of classification disputes that may arise. The EU had bound its tariffs on meat, 'salted, in brine, dried or smoked' at an ad valorem 15.4 per cent. Tariffs on fresh, chilled or frozen poultry were specific, at €102.4 per 100 kilograms, which implied an ad valorem equivalent between 40 and 60 per cent. Thai and Brazilian firms had been exporting frozen chicken cuts treated with salt under the salted meat tariff heading until 2002 when the EU instructed customs to classify chicken into the lower tariff item only if salt was added for the purpose of long-term preservation. The complainants argued that it was irrelevant what the purpose of the salt was. The AB found against the EU—on the basis of arguments that have been contested in the literature (Howse and Horn, 2008).
Other fees and charges on imports
World Trade Organization members are constrained regarding the use of fees and specific import taxes that have an effect equivalent to tariffs. Examples include taxes on foreign exchange transactions, service fees affecting importers and special import surcharges. Such para-tariffs as they are sometimes called used to be particularly important in developing countries. Data for a sample of 41 developing countries in the early 1980s indicated that at least one-third of revenue from import taxation was generated by para-tariffs. Such measures were frequently subject to arbitrary implementation and were nontransparent (Kostecki and Tymowski, 1985). They were often driven by specific interest groups that had successfully lobbied for earmarked taxes to finance their activities. Para-tariffs continue to be used today. For example, in 2004 an industrial development surcharge and supplementary duties accounted for some 38 per cent of the overall average level of protection in Bangladesh (World Bank, 2004). In such cases the average statutory customs duty will give a misleading picture of the actual structure of protection, especially if the para-tariffs vary by good or sector.
In contrast to GATT 1947, the GATT 1994 requires that the nature and level of other duties or charges be listed by tariff line in each WTO member's schedule. Allowance is made for the imposition of fees or other charges, as long as these are commensurate with the cost of services rendered (Article II:2c). Article VIII (on fees and formalities related to trade) requires that all such service fees must 'be limited in amount to the approximate cost of services rendered and shall not represent an indirect protection to domestic products or a taxation of imports or exports for fiscal purposes' (Article VIII:i). Examples of such fees include consular transactions, licensing, statistical services, documentation, certification, inspection, quarantine, sanitation and fumigation. Article VIII applies irrespective of whether a country has bound its tariffs. The requirement that service fees be cost-based aims to prevent circumvention of tariff bindings.
196       the political economy of the world trading system
In a 1988 GATT dispute concerning the imposition of a uniform ad valorem customs user fee by the US, the panel concluded that such fees must be service-specific (GATT, 1994b: 251). Imposing an average fee equal to the total cost of customs administration divided by the total value of imports was not acceptable. Although the US altered its customs user fee to conform to these findings, other countries continued to maintain ad valorem fees that were inconsistent with the GATT. In part this was because service fees in existence on the date of a country's accession to GATT were grandfathered and thus immune from scrutiny. Developing countries also had greater leeway than industrialized countries, as their tariffs often were not bound. As of the mid-1990s, some developing countries continued to maintain customs user fees of 5 per cent ad valorem or higher (Messerlin and Zarrouk, 2000). After 2001, terrorist threats contributed to increases in customs user fees in a number of countries. For example, the US Customs and Border Protection increased certain charges by 10 per cent in 2007 to cover the costs of stricter border controls.
With the establishment of the WTO, in principle all grandfathered policies that are inconsistent with WTO rules had to be abolished. It is likely that over time WTO members will be less tolerant of fees on imports that are not cost-based. The issue arose several times in the WTO. For example, in 1998 the EU contested a US Harbor Maintenance Tax (a 0.125 ad valorem duty on imports) as a violation of among others, Articles II and VIII, as well as the Uruguay Round Understanding on the Interpretation of Article ll:ib (WT/DS118). This Understanding is quite important from a transparency perspective, as it requires members to notify all the duties and charges that apply in connection with importation of goods. The case led to a MAS, based on a US commitment to alter its legislation.4 In Argentina—Textiles and Apparel (WT/DS56) the US contested a system of minimum specific import duties and a statistical services tax in Argentina. In its 1998 ruling, the AB found both measures to be inconsistent with Article II:i£> GATT as they resulted in customs levies in excess of those provided in Argentina's schedule.
Tariff peaks and escalation
The structure of tariffs maintained by WTO members varies considerably in OECD countries. High tariffs are particularly frequent for agricultural imports. In the EU, about one-quarter of post-Uruguay Round imports of manufactured goods is
4 The tax initially applied to exports as well but this was declared to be unconstitutional by the US Supreme Court in 1998 because it was a tax and not a user fee (the US constitution prohibits the use of export taxes). The result of exempting exports was that the tax not only was inconsistent with WTO rules—not being a cost-based user fee—but also violated the national treatment principle.
trade in goods
197
duty-free, and some 40 per cent are subjected to tariffs below 5 per cent. Tariffs above 25 per cent are imposed on a negligible share of imports of manufactures. In the US, duty-free imports account for some 80 per cent of the total; a tariff rate of more than 10 per cent is imposed on less than 3 per cent of imports. Developed countries' tariffs show important dispersion in bound tariff rates and significant peaks especially in products such as textiles, clothing and leather goods. These three product categories have the lowest share of bound duty-free lines across countries, the highest tariff averages and the most lines with peaks (Bacchetta and Bora, 2004).
World Trade Organization members are free to determine the structure of their tariff regimes. Many countries tend to impose a structure that taxes inputs and unprocessed products less than finished goods. Such tariff escalation has been a problem for developing countries seeking to process commodities before they are exported. The more escalated is the tariff structure maintained in export markets, the greater the difficulty for such countries to generate value-added at home, as the low tariffs on raw materials (usually duty-free) provide an incentive not to process commodities before they are exported. A group of products where tariff escalation has often been a source of particular concern to developing countries are natural resource-based products, defined in GATT to include nonferrous metals and minerals, forestry products and fish and fishery products.
Escalation is a feature of almost all tariff structures. The exceptions are the few economies that essentially have a free trade stance (e.g. Hong Kong) or have implemented a uniform tariff, that is, apply the same tariff to imports of all goods (examples include Bolivia, Chile, the Kyrgyz Republic and Estonia before it acceded to the EU). At the product level, tariff escalation may occur even if the overall structure exhibits little or no escalation. This is often the case for textiles and clothing and leather products, where inputs—cotton, yarn, raw hides—tend to have lower duties than does the finished product.
There are good economic arguments against escalation. A uniform tariff structure ensures that all activities are treated equally, leaving it to the market to determine what activities are profitable to undertake. Uniformity also has important benefits in terms of incentives for rent seeking, as it sends a signal to firms that there will be no exemptions and that it will not pay to lobby for either higher or lower tariffs. Thus, a uniform rate saves on lobbying costs, and induces less corruption and social waste. Lobbying is unattractive because efforts to increase the tariff will affect virtually the whole economy, and solicit opposition by users of imported inputs. Indeed, a uniform tariff, because it puts pressure on industries that are dependent on imports, creates incentives for lobbying in favour of reducing the average rate. In the case of Chile, which maintained an 11 per cent tariff for some years, a decision was made to gradually lower it to 6 per cent, starting in 1998.
Another advantage of a uniform tariff is that it lowers transaction costs—there is no scope to argue about how products should be classified for customs duty calculation purposes. Uniformity also tends to make it easier for the government
198      the political economy of the world trading system
to monitor the performance of customs agencies—given data on the value of imports it is easy to calculate how much revenue should have been collected. A uniform tariff is more difficult for countries to maintain in MTNs, if these take the form of bid-offer negotiations. All a government can do is to make offers of formula cuts. As a result of this 'constraint'—which is really an advantage from an economic perspective—statements have been made in the WTO Council opposing the use of uniform tariffs. This is an example of putting the cart before the horse: the desire to use a certain negotiating technique potentially impeding the use of an efficient trade policy by WTO members.
In addition to deliberate tariff escalation, dispersion in the structure of tariffs often results from duty exemption programmes. Exemptions may be granted to enterprises that engage in certain 'favoured' activities—a frequent example is foreign investors, who are often granted tax holidays and duty exemption for imported inputs and capital equipment. The welfare effect of such discrimination is ambiguous, but is likely to be negative as it will distort resource allocation incentives and may generate socially wasteful lobbying. This is not the case for duty exemptions granted to firms and industries that produce for export. Exporting firms must be able to import inputs, including machinery, at world prices. Otherwise they will be impeded in their ability to compete on international markets. To ensure that this is not the case many countries have implemented duty drawback or temporary duty-free admission schemes for inputs used in export production as part of their customs procedures. Although such mechanisms are an important tool of export development if they function efficiently, from a political economy viewpoint they suffer from the problem that they reduce the incentives for exporters to support general import liberalization. In general, non-export-related duty exemptions are a costly policy instrument as they can encourage corrution and rent seeking, reduce revenue collected by customs, and distort economic incentives.
The Doha ministerial declaration stipulated that special attention be paid to reducing tariff peaks and escalation. This motivated the use of a nonlinear formula approach, with flexibility for developing countries (see Chapter 4).
5.2. National Treatment: Article III
The national treatment rule is the second leg of the nondiscrimination principle. It requires that foreign goods—once they have paid tariffs and satisfied whatever border measures apply—be treated no less favourably than like or directly competitive goods produced domestically in terms of internal (indirect) taxation (Article 111:2 GATT). That is, goods of foreign origin circulating in the
trade in goods
199
country must be subject to the same taxes and charges that apply to identical goods of domestic origin. A similar obligation applies to nontax policies (regulations) (Article 111:4 GATT). In both cases, members are to provide treatment 'no less favourable' to foreign products as they accord to domestic goods. Although Article III distinguishes between fiscal and nonfiscal measures, the AB increasingly takes the view that Article III.i GATT discourages the protectionist use of domestic legislation no matter the type of policy involved (Mav-roidis, 2007).
National treatment is a general obligation in the GATT, although not in the GATS (see Chapter 7). Its role is to ensure that liberalization commitments are not offset through the imposition of domestic taxes or other regulatory measures. By requiring that foreign products be treated no less favourably than domestically produced competing products, foreign suppliers obtain greater certainty regarding the regulatory environment in which they must operate. The purpose of Article III GATT is to constrain circumvention of tariff bindings through the discriminatory application of domestic policies (Box 5.3). The importance of Article III has increased over time as tariff levels have fallen. Especially in high-income countries, tariffs are now quite low on average, and NTMs increasingly determine the conditions of competition on markets. Some 30 per cent of all tariff lines in OECD countries are subject to NTMs of some kind—not taking into account indirect taxes (Figure 5.1). Many of the NTMs in use today are subject to specific WTO disciplines that are the subject of the remainder of this chapter and subsequent ones. All are subject to the national treatment rule.
Article 111:8 GATT excludes subsidies and government procurement from the reach of national treatment. In the case of purchases by government entities, the carve-out from national treatment is only for goods that are not resold. Subsidies are regulated by the WTO Agreement on Subsidies and Countervailing Measures (discussed later in this chapter), whereas procurement is the subject of a plurilateral Agreement on Government Procurement (discussed in Chapter 11). The reason both of these policy areas were excluded from Article III was that the initial GATT contracting parties wanted to be able to use these instruments to favour domestic industries.
The national treatment principle has often been invoked in dispute settlement cases brought to the GATT. It was the basis for the first case brought against China. In its Accession Protocol China bound tariffs on cars at 25 per cent and on parts at 10 per cent. In 2005, China imposed a tax of 25 per cent on cars assembled in China from imported parts. The 2008 WTO panel ruling held that this was a discriminatory internal tax. Even if not regarded as an internal tax it clearly violated the tariff binding. National treatment is a very wide-ranging rule. The obligation applies whether or not a specific tariff commitment was made, and whether or not a tariff is bound. It covers taxes as well as other domestic regulatory measures: any and all policies must be applied in a
200
the political economy of the world trading system
Box 5.3. The rationale for the National Treatment Principle
The standard theoretical rationale offered for the national treatment rule in the literature is that it prevents 'concession erosion' through indirect taxation or other policies that discriminate against imported goods once they have entered a customs territory. The principle is very far-reaching, which can give rise to tensions. There may be good reasons for governments to adopt regulatory measures that discriminate between foreign and domestic products—for example, if foreign firms are much more efficient and the market is imperfectly competitive higher tax rates on the foreign firm(s) may induce them to lower prices charged to consumers. The national treatment rule precludes the use of domestic tax policy to offset such domestic market distortions. However, it will only rarely be the case that differential taxation of foreign and domestic goods is the optimal instrument to offset domestic distortions. In the example just mentioned, the problem can be addressed more directly through the imposition of tariffs on the foreign products. But as noted by Horn (2006), in practice governments may not be able to negotiate tariffs that internalize whatever distortions occur at the border. If so, the national treatment rule will prevent them from using domestic policies instead—thus potentially reducing welfare.
Saggi and Sara (2008) argue that quality differentials across countries may have important implications for the effects of national treatment. They note that while the notion of likeness of products is central in the application of the rule, a practical problem is that competing products from different countries are often differentiated from one another. Are a Toyota Camry and a Ford Taurus—two cars that contest the same segment of the US market—like products? Although very similar cars, they are not of identical quality and both are inferior to a higher end competitor such as a BMW. Even for relatively homogenous goods, national origin itself can serve as a signal of quality and therefore a source of product differentiation. Motivated by such considerations, Saggi and Sara analyse when and why heterogeneous countries producing differentiated goods find it in their mutual interest to abide by national treatment. In their model, discrimination against the foreign product is in each country's interest. By definition, a national treatment agreement between the two countries rules out such 'beggar-thy-neighbour' tax discrimination. They find that if market size is equal across countries, the high-quality country benefits from national treatment, but the low-quality country loses. On the other hand, when market size differs across countries, national treatment can make both countries better off relative to tax discrimination. Thus, asymmetry in market size helps counterbalance the quality differences across countries, but only when the quality gap between goods is not too large. This result corresponds quite well with the emphasis put on the 'likeness' criterion in the WTO.
nondiscriminatory fashion to like domestic and foreign products. It is irrelevant whether a policy actually hurts an exporter (has an impact on trade). What matters is discrimination per se.
For a violation of Article III GATT to occur, a complainant must establish that a WTO member has intervened through regulatory means so as to afford protection
trade in goods
201
4 6 8 10 12
log GOP per capita
Figure 5.1. Average tariffs and NTMs and per capita income, 2006
Source: Hoekman and Nicita (2008).
to domestic competing (like) products.5 General Agreement on Tariffs and Trade/ WTO case law has clarified that for likeness to be determined: (1) demand-side factors are relevant; (2) economic criteria may be used; and (3) all like products have to be directly competitive or substitutable. In the GATT years criteria to establish the likeness of goods or that a product was directly competitive or substitutable included whether end-uses in a given market were similar and the product's properties, nature and quality. In the WTO era greater use has begun to be made of economic criteria such as the cross-price elasticity of demand in determining likeness.
For products to be 'like' they have to share some properties beyond what two directly competitive or substitutable products share. So far, WTO case law has offered one such extra property: customs classification. Two products that fall under the same HS classification are considered to be like. However, this is not necessarily sufficient. In the EC—Asbestos dispute, for example, the AB found that asbestos-containing construction material and asbestos-free construction material are unlike products for the purposes of Article III.4 GATT, even though from an economic end-use perspective they fulfill exactly the same need. Thus, physical
5 What follows draws on Hoekman and Mavroidis (2007).
202
the political economy of the world trading system
characteristics may be an important criterion (Horn and Weiler, 2004; Howse and Tuerk, 2006).6
The WTO agreements on Technical Barriers to Trade (TBT) and Sanitary and Phytosanitary Measures (discussed later in this chapter) address specific types of domestic regulation. If a measure is deemed to violate national treatment but is covered by one of these agreements, the rules of the specific agreement apply. There is a hierarchy of these agreements: the SPS agreement, if applicable to a product, dominates the TBT agreement, while the TBT agreement dominates the more general GATT disciplines (see Mavroidis, 2007).
Regulation 'so as to afford protection'
With respect to fiscal policies, Article III.2 GATT requires that foreign products not be taxed in excess of like domestic products so as to afford protection to domestic directly competitive or substitutable products. World Trade Organization case law has defined the term 'excess' as a nonzero difference in tax rates. But for a measure to afford protection, the tax differential must have an effect. That is, it must be more than de minimis, which can only be determined on a case-by-case basis. When it comes to nonfiscal measures, Article III.4 GATT requires that imported products not be treated less favourably than domestic like products. Even if two products are 'like', for a measure to be inconsistent with Article 111:4 a complaining member must establish that 'like' imported products are accorded less favourable treatment than 'like' domestic products.
To establish whether a regulation operates in a protectionist manner, a successful complainant does not have to show either protective effects or protective intent. Trade effects and/or regulatory intent do not matter when determining whether a domestic policy measure has the effect of affording protection. Moreover, likeness or direct comparability/substitutability may be established through various means or criteria. The resulting broad reach and flexibility can result in panels and the AB outlawing legislation that may not have protectionist effects or intent.
An example of the type of ambiguity that may arise was a case involving a Chilean law that distinguished between three categories of alcoholic beverages: drinks with less than 35 per cent alcoholic content; drinks between 35 per cent and
6 In this dispute the AB reversed the panel on its determination of likeness by introducing the construct of a 'reasonable consumer', which allowed them to apply available econometric data to a counterfactual situation where this 'consumer' had not been confronted with a choice between asbestos- and nonasbestos-containing products. A result of this approach was that it provided grounds for certain groups to argue that the result of this case is that the WTO applies to production and processing methods (PPMs)—see below.
trade in goods
203
39 per cent; and drinks with alcoholic content above 39 per cent. The products in the first category were taxed at 27 per cent ad valorem and those in the third one at 47 per cent ad valorem. The complaining parties argued that some imported products of slightly more than 39 per cent were directly comparable to Chilean products of less than 35 per cent and that the tax differential operated so as to afford protection. Chile responded that the majority of the affected products were domestic and that no protection could therefore result. The AB dismissed the relevance of this fact, arguing equality of competitive conditions for all directly competitive or substitutable imported products was required in relation to domestic products, and not just the treatment of imported products within a particular fiscal category. What mattered in particular was the cumulative consequence of the policy: that approximately 75 per cent of all domestic production of the distilled alcoholic beverages at issue fell under the lowest tax rate, whereas approximately 95 per cent of the directly competitive or substitutable imported products were subject to the highest tax rate.
Many countries maintain tax systems under which tariffs paid on imported goods that are used in the production of exports are rebated to the exporter. Many countries also maintain value added tax (VAT) systems under which tax paid on imports are rebated on export. Such drawback systems are permitted by GATT as they do not violate national treatment: what matters for national treatment is whether foreign and domestic like goods are treated the same when sold on the domestic market. However, the VAT and other tax rebates may constitute actionable subsidies if they do more than return actual taxes paid.
5.3. Quantitative Restrictions, Exchange Rates and Import Licensing
General Agreement on Tariffs and Trade rules on QRs were written when the use of such measures was widespread. During the GATT years QRs were particularly prevalent in trade in agricultural products, textiles and clothing, and steel. Over time, the relative importance of QRs declined substantially. Notwithstanding the reduction in the use of QRs, the GATT provisions banning their use continue to be a frequently invoked basis for dispute settlement. Many of these cases concern developing countries; they also often concern measures that are argued to have an effect that is equivalent to a quantitative restriction.
The economic case against the use of quotas was summarized earlier. A quota cuts the link between domestic and foreign prices, is generally discriminatory, may not allow the changing pattern of comparative advantage over time to be reflected
204
the political economy of the world trading system
in imports, is less transparent, and more open to administrative abuse and corruption. For all these reasons interest groups seeking protection tend to prefer QRs to tariffs. Another motivation for the use of QRs by governments is that if the associated licences are allocated to exporters they offer some automatic compensation (see Annex 2).
WTO disciplines
Articles XI-XIVof the GATT provide the legal framework addressing QRs. Article XI prohibits them in principle, except for agricultural commodities if concurrent measures are taken to restrict domestic production (Chapter 6). Article XII provides another exception, allowing QRs to be used for balance-of-payments (BOP) reasons. If this is done, Article XIII requires that such quotas in principle apply on a nondiscriminatory basis, whereas Article XIV allows for a request that the Council waives this requirement. These provisions are complemented by the Agreement of Safeguards, which bans the use of VERs, a specific type of QR that is applied by the exporting country rather than the importer (see Annex 2).
The basic obligation imposed on members in Article XI:i is to refrain from introducing or maintaining QRs, whether on imports or exports. Quantitative restrictions are banned not so much because of economic considerations— although the ban is certainly consistent with economic first principles—but to prevent governments from circumventing tariff bindings. Article XIII requires that if QRs are used they abide by the MFN principle. The economic rationale for this is that a global quota is more efficient than selective QRs. Under a global quota traders (importers) are left free to determine from where to source. The direction of trade (sourcing of imports) will then be responsive to changes in prices, quality and transportation costs. From the standard reciprocity perspective—maintaining a balance of concessions—the MFN rule disciplines the country-specific allocation of quota rights, which can easily imply de facto discrimination. Country-specific allocations are usually based on historical market shares, the idea being to reduce all exporters' market access rights proportionally.
World Trade Organization case law has defined QRs broadly to include any measure that is equivalent. For example, a GATT-era panel report, Japan— Semiconductors held that if a government provides incentives for private parties to act in a manner inconsistent with Article XI GATT, such behaviour is GATT-inconsistent. In the semiconductor case, Japanese firms raised prices (which led to reduced exports) as a result of incentives (including administrative guidance and monitoring of costs and prices) by the Japanese government. Thus, the term 'QR covers anything that might operate as a QR, irrespective of whether the subject of the challenged activity is the state or the private sector. However, for the behaviour by firms to be subject to Article XI GATT, they must be attributable to an action by
trade in goods
205
a government that creates an incentive to cut back imports or exports. Moreover, to attack such de facto QRs a successful legal challenge requires the complainant to demonstrate a causal link between the contested measure and a reduction in trade. In the case of dejure QRs, the need to document trade effects is less: what matters is that competitive opportunities of imported products have been affected.
In Argentina—Hides and Leather (WT/DS155), the EU held that the presence of representatives of the domestic leather (tanning) industry in Argentine customs sufficed to establish a QR, as it was in their interest to seek to restrict exports of hides. The panel rejected this claim, arguing that the presence of representatives of the domestic industry did not suffice for establishing a violation of Article XI GATT. The panel held that the burden of proof was on the EU to provide evidence that the result was to reduce or restrict exports of hides. Surveillance by competitors could not be shown to be a QR.
Despite the general prohibition on QRs, they were widely used during the GATT years by governments to protect domestic import-competing industries. In the Uruguay Round disciplines in this area were strengthened, with the outlawing of VERs, the tariffication of agricultural quotas and the agreement to phase out textile and clothing quotas administered under the auspices of the MFA (see Chapters 6 and 9). Formal QRs were used especially in the agricultural context by industrialized countries and for BOP purposes by developing countries. Avery popular form of QR—used increasingly in the 1970s and 1980s—was the VER, often negotiated as 'undertakings' agreed to by exporters to reduce supply or raise prices in the context of antidumping investigations. Following the ban on VERs and agricultural QRs in the Uruguay Round tariff quotas have become much more important in agriculture (Chapter 6).
Examples of disputes where policies were found to violate WTO rules on QRs include Canada—Periodicals (an import ban); EC—Poultry (tariff rate quotas); India—Quantitative Restrictions (QRs for balance-of-payments purposes); India— Autos (a trade balancing requirement); Dominican Republic—Cigarettes (a requirement for importers to post a bond operated as a QR); and Brazil—Retreaded Tires (import prohibition).
Foreign exchange arrangements and restrictions
Article XV calls on contracting parties not to use exchange rates to frustrate the intent of the provisions of the GATT (Article XV:4). It also requires that the CONTRACTING PARTIES consult fully with the IMF on monetary reserves, balances of payments or foreign exchange arrangements. In such consultations, the CONTRACTING PARTIES are to accept findings of statistical and other facts presented by the Fund and the determination of the Fund as to whether action by a contracting party in exchange matters is in accordance with the Articles
206       the political economy of the world trading system
of Agreement of the IMF (see Box 5.4). These disciplines were drafted at a time when exchange rates were fixed. The IMF's mission was to monitor exchange rates and assist countries dealing with balance-of-payments problems through the provision of loans to finance deficits and provide breathing room for countries to adjust to an external shock. As countries shifted to systems of (managed) floating exchange rates, the main role of the IMF continued to revolve around surveillance and provision of assistance to members to deal with macroeconomic imbalances.
A country can use the exchange rate to achieve trade policy objectives. Indeed, by undervaluing the exchange rate, the same effect can be achieved as a combination of import barriers and export subsidies. This was recognized by the drafters of the GATT and explains the inclusion of Article XV. The existence of the IMF helps understand why the GATT disciplines are not very specific: it was left to the IMF to address exchange rate misalignments. It is not surprising therefore that there have been no disputes under the GATT or the WTO regarding the level of exchange rates.
Substantial progress was made to eliminate exchange restrictions in the decades following the creation of the GATT. The IMF actively promoted current account convertibility through consultations, technical assistance, programme condition-ality and research showing the benefits of such liberalization. Domestic factors also
Box 5.4. What's an exchange restriction? Dominican Republic—Cigarettes
In a 2004/5 dispute involving two developing countries a WTO panel considered, among others, the definition of exchange restrictions. Responding to Honduras' complaint, the Dominican Republic argued that a foreign exchange fee charged on certain import transactions was an exchange measure within Article XV: 9(a) of the GATT and thus justified even if otherwise inconsistent with GATT Article II. It argued that the fee 'is prescribed by monetary authorities, not by trade or customs authorities; it applies to exchange actions, not to import transactions as such; and it is a charge on foreign exchange transactions imposed through the banking system, not a charge on import transactions levied by customs authorities' (WT/DS302). The panel rejected the above defence under Article XV:9 on the grounds that 'since Article XV:g of the GATT exempts exchange measures that are applied in accordance with [IMF] Articles, from obligations under other Articles of the GATT, the guiding principle that the IMF prescribed as the criterion for the determination of what constitutes an "exchange restriction" should be respected by this Panel'. Thus, the panel asked the IMF for a legal opinion as to whether the Dominican Republic measure was an 'exchange restriction', and the IMF replied that in its current form, as a fee charged on importation rather than all foreign currency transactions, the measure was not an 'exchange restriction. The panel followed the IMF's opinion and found that the fee in question was not an exchange measure justified by Article XV:9(a) of the GATT; the AB essentially upheld the panel's finding (DS202 and Howse, 2008).
trade in goods
207
played a role, for example the rise of the private sector in developing countries, which generally had an interest in dismantling foreign exchange restrictions and controls. Most important was the recognition of the inefficiencies associated with maintaining fixed exchange rates and the need for greater exchange rate flexibility to maintain balance-of-payments equilibrium.
Exchange rates have often been the source of economic tension between trading partners. Countries that are running large current account deficits—that import much more than they export—frequently claim that surplus countries are 'unfairly' manipulating the exchange rate so as to give their exporters a competitive advantage— increasing the domestic prices of imports and reducing the foreign currency denominated export prices. In the 1980s such claims were frequently made against Japan and the Asian 'tigers'; in the 2000s it was the turn of China to bear the brunt of such complaints. Given that the IMF cannot force governments to adjust exchange rates—it can only advise—there have been suggestions that the WTO members should agree to stronger rules in this area that can be enforced through the DSU.
Mattoo and Subramanian (2008a) argue that if there is a clear finding of undervaluation and this is clearly due to government action, this should be regarded as fully equivalent to a violation of import tariff bindings and the ban on export subsidies. They recognize that undervaluation can result from a number of factors, including fiscal and monetary policies, policies related to capital flows, taxes and subsidies, and intervention in foreign exchange markets, but argue there is a clear hierarchy of policy actions in terms of proximate causation. Prolonged one-way intervention in foreign exchange markets by the central bank or by government and quasi-government agencies, redenomination of domestic debt into foreign currency, and extensive forward market operations are policy actions that can clearly be identified as causes of undervaluation. In such cases, they propose that the IMF—as is already the case when members invoke the GATT balance-of-payments articles (see Chapter 9)—assess whether the member's exchange rate was misaligned and whether it was a consequence of clear government action if a dispute is brought to the WTO.
Import licensing
Quantitative restrictions are generally enforced by means of licences. A separate Agreement on Import Licensing Procedures, which applies to all WTO members, aims to strengthen general GATT obligations in this domain. The agreement resembles closely the code on licensing that was negotiated in the Tokyo Round. It aims to enhance transparency of licensing systems and calls for publication of licence requirements, the length of licence validity and the right of appeal of decisions.
208
the political economy of the world trading system
Licensing can become a source of disputes when the allocation of quota rights is perceived to be biased and to violate the GATT nondiscrimination principles. Traditionally, licences tend to be allocated on the basis of historical market shares. An alternative would be to auction off the licences to the highest bidder. The latter approach has the economic advantages of generating revenue for the government and reducing the resource allocation distortions generated by quota systems, but the political disadvantage of eliminating discretion (patronage possibilities) and the opportunity for powerful vested interests to obtain quota rents. The WTO rules on import licensing played a major role in the Bananas case discussed earlier, which revolved to a great extent around the procedures used by the EU to allocate import licences (see Chapter 3).
5.4. Customs Clearance—Related
Provisions
Customs clearance requires the valuation and classification of imports for purposes such as levying tariffs, determining origin, enforcing foreign exchange controls and collecting statistics. Customs procedures may become NTBs if officials incorrectly classify goods or assign goods a value greater than appropriate. An agreement to reduce and bind tariffs would be practically meaningless without a set of rules concerning valuation and classification of imported goods. Arbitrary customs procedures could then be used to ensure that a government collects as much revenue as desired, independent of the formally negotiated tariff schedule. Import-competing industries might also bribe officials to harass importers. To reduce the likelihood that a country's published tariff schedule is not representative of the nominal tariffs that are actually applied, GATT establishes certain rules and principles regarding customs valuation.
Valuation
The provisions on customs valuation contained in the GATT 1947 (Article VII) were not very precise, basically requiring that goods be valued on the basis of their actual value. Before the launch of the Tokyo Round (1973), a number of contracting parties, led by the EEC, felt that certain national valuation practices were restricting international trade. US methods were a major bone of contention, in particular the so-called American Selling Price criterion, which established the value of some imported goods on the basis of the selling price of similar domestically produced
trade in goods
209
goods. Although this clearly violated GATT rules, the US was able to employ this method because it had grandfathered the practice when acceding to GATT. Largely motivated by US practices, in the Tokyo Round a Customs Valuation Code was negotiated which supplemented GATT's valuation provisions and outlawed practices such as the American Selling Price. As was the case with all the Tokyo Round codes, participation was voluntary, and most developing countries did not sign the agreement. The US did sign the agreement and reformed its valuation practices to comply with it.
With the creation of the WTO, the Tokyo Round code became the Agreement on Customs Valuation (formally the Agreement on Implementation of Article VII of GATT). It binds all WTO members. The main impact of the customs valuation agreement was on developing countries, as valuation practices often did not conform to the provisions of the agreement. Separate disciplines were also added on pre-shipment inspection—the practice of requiring the inspection of goods in the country of production before they are shipped—and on rules of origin.
The customs valuation agreement seeks to establish uniform, transparent and fair standards for the valuation of imported goods for customs purposes. The agreement outlaws the use of arbitrary or fictitious customs values. In principle, valuation should be based on the transaction or invoice value of the goods—the price actually paid or payable for the goods (subject to adjustments for freight and other charges). The invoice value method should be applied when:
(1) there are no special restrictions on the disposal or use of goods;
(2) the buyer and seller are not related;
(3) no proceeds of the subsequent sales accrue to the exporter; and
(4) the sale or price is not subject to special conditions that cannot be quantified.
The agreement does not prescribe a uniform system regarding shipping, insurance and handling charges. A country may opt for a cost, insurance and freight (c.i.f.), a cost and freight or a free-on-board (f.o.b.) valuation basis. If customs authorities have reasons to believe that the transaction value is inaccurate, they are required to proceed sequentially through five alternative valuation options:
(1) the value of identical goods;
(2) the value of similar goods;
(3) the so-called deductive method;
(4) the computed value method; and
(5) an unspecified 'if all else fails' method.7
7 The deductive value method consists of the unit price at which a significant quantity of imported goods is sold to unrelated persons, subject to deductions for commissions, profit margins, transport and insurance costs. The computed method consists of summing the cost or value of materials and other inputs employed in producing the imported goods, and adding an amount for profit and general expenses equal to that applied in sales of similar goods by other producers.
210       the political economy of the world trading system
It is only when the customs value cannot be determined under one of these options that the next option in the sequence can be used. However, an importer may request that the computed method be used in preference to the deductive method. In many instances, refusal to accept the invoice price will be connected to there being a relationship between buyer and seller. The mere fact of such a relationship is not sufficient grounds for the authorities to reject the invoice price. What matters is that the relationship influences the price. If the value is questioned by customs, the burden of proof is on the importer.
Reflecting fears voiced during the Tokyo Round regarding fraudulent invoicing, especially between related parties, a Protocol to the code gave developing country signatories somewhat greater regulatory flexibility in their customs procedures. Also, technical assistance in implementing code procedures was promised. Despite this, developing country participation in the code remained limited. Fears of reduced tariff revenue as a result of underinvoicing, a wish to maintain discretion in valuing imports, or the administrative burden of implementing code provisions were among the major concerns. In the Uruguay Round, a number of developing countries put forward the view that the need to accept declared values was the main factor prohibiting greater participation in the code. Consequently, they proposed that the agreement be amended to allow more scope for rejecting transactions (invoice) values.
In recognition of developing country implementation concerns, in the Uruguay Round developing countries that were not party to the code were given the right to delay implementing its provisions until 1 January 2000. Application of the computed value method could be delayed for an additional three years, and reservations could be registered in respect of any of the provisions of the agreement if other members consent. Developing countries that valued goods on the basis of officially established minimum values could request approval to retain such practices on a limited and transitional basis, subject to the terms and conditions required by the other members. Requests for such derogations require approval from the WTO Council. An annex to the agreement allows developing countries to request extension of transition periods.
During 1999, in the run-up to the Seattle ministerial meeting, over 40 developing countries requested and obtained extensions of Uruguay Round implementation deadlines. They argued that their customs systems did not yet allow a switch to invoice-based valuation. A problem with using a transactions-based valuation system that puts the burden on customs authorities to query invoice values is that mechanisms and tools must be available for the authorities to determine whether declared values are reasonable. Most OECD countries have sophisticated computer-based systems and risk-assessment techniques that allow them to identify suspect claims, and access to databases that allow them to roughly determine the market value of most commodities that pass the frontier. This is not the case for many developing countries, notwithstanding the technical assistance efforts in this
trade in goods
211
area by organizations such as the WCO, UNCTAD and the IMF. Implementation in developing countries involves much more than issuing a decree that valuation is to conform to WTO rules. Customs administrations need to be automated, infrastructure improved, staff trained, and so forth. This generally takes a significant amount of time and requires the investment of substantial resources. Even Canada, a developed country with ample resources, took five years to complete implementation of the valuation agreement (Staples, 2002).
There have been relatively few dispute settlement cases relating to valuation. One was a case brought by the EU in 1996 against Mexico, claiming that Mexico applied c.i.f. value as the basis of customs valuation of imports originating in non-NAFTA countries, and f.o.b. value for imports originating in NAFTA countries. This was argued to raise the valuation basis for imports, although in practice it should not result in discrimination if intra-NAFTA flows are duty-free.8 In 2000, the US consulted with Romania in respect of the use of arbitrary minimum and maximum import prices for such products as meat, eggs, fruits and vegetables, clothing, footwear and certain distilled spirits. The dispute was resolved rapidly as a result of Romania changing its regime, which it was obliged to do in any event in the context of its effort to converge to EU norms. In 2003, Guatemala requested consultations with Mexico concerning certain customs rules, procedures and administrative practices that impose officially established prices for customs valuation in Mexico. It also contested Mexico's practice of requiring a deposit or bond to guarantee the observance of these officially established prices. As a result of consultations a MAS concerning footwear and brushes, two areas that were severely affected by the measures, was agreed in 2005.
The Singapore ministerial conference gave the WTO the mandate to take a more comprehensive look at trade facilitation including issues related to customs valuation. This work was conducted in the Doha Round Negotiating Group on Trade Facilitation in cooperation with several other international organizations (TN/TF/i). These negotiations are discussed further in Chapter 13.
Rules of origin
A rule of origin is a criterion used by customs authorities to determine the nationality of a product or a producer. Rules of origin are necessary when there is a desire to discriminate between sources of supply. The WTO has no rules regarding rules of origin. The only multilateral convention dealing with rules of origin is the 1974 International Convention on the Simplification and Harmonization of Customs Procedures (known as the Kyoto Convention), negotiated
8 This case was brought under Article XXIV: 5 fc, that is, the GATT provision dealing with free trade areas, not the valuation agreement.
212
the political economy of the world trading system
under the auspices of, and administered by, the WCO in Brussels. This convention, which was revised after lengthy negotiations in 1998, provides a short list of products that should be considered to originate in a country because they are wholly produced or obtained there, that is contain no imported materials. These are largely natural resource-based products extracted or obtained from the territory of the country concerned. Where two or more countries are involved in the production of a product, the Convention states that the origin of the product is the one where the last substantial transformation took place, that is the country in which significant manufacturing or processing occurred most recently. Significant or substantial is defined as sufficient to give the product its essential character.
Various criteria can be used to determine if a substantial transformation occurred. These include a change in tariff heading (CTH), the use of specific processing operations which do (or do not) imply substantial transformation, a test based on the value of additional materials embodied in the transformed product, or the amount of value added in the last country where the good was transformed. Under a CTH the value added may be high or low for a given product, so that a value-added criterion may or may not lead to the same result as a CTH test. Different rules of origin may therefore vary widely in their economic effects. If written in ways that make it difficult to satisfy them, rules of origin can be effective protectionist devices. Thus, the setting of rules of origin may be accompanied by rent-seeking activities, as import-competing lobbies have an incentive to either try and make the rules as restrictive as possible, or to influence the way they are applied. Restrictive rules are particularly important in the context of the negotiation and implementation of preferential trade agreements (see Chapters 10 and 12).
In the application of MFN trade policy the problem is often vaguely defined criteria that generate uncertainty by giving discretion to the importing authorities to determine if a particular rule has been met. The more discretion officials have in this area, the greater the incentive to lobby. Such problems are especially prevalent under value-added criteria, as enforcement of such rules requires detailed investigations of the financial accounts of exporting firms. Box 5.5 gives an example of the operation of rules of origin in one of the major nonpreferential trade policy areas where they are important—antidumping.
In contrast to GATT 1947, the WTO includes an agreement on rules of origin. The WTO Agreement on Rules of Origin aims to foster the harmonization of the rules used by members. The agreement calls for a work programme to be undertaken by a Technical Committee, in conjunction with the WCO, to develop a classification system regarding the changes in tariff subheadings based on the Harmonized System that constitute a substantial transformation. In cases where the HS nomenclature does not allow substantial transformation to be determined by a CTH test, the Technical Committee is to provide guidance regarding the use of
trade in goods
213
Box 5.5. Origin rules and antidumping
One area where the application of rules of origin is an issue is in the enforcement of antidumping (AD) actions. A European case illustrates how rules of origin can be used to achieve the objectives of a specific lobby. In the mid-1980s, the EU imposed a 20 per cent AD duty 01112 Japanese exporters of photocopiers. In 1988, three years after the AD duty was imposed, a so-called anticircumvention case was brought by the EU industry. It claimed that Japanese exporters had circumvented the AD duty by establishing assembly operations inside the EU that imported most of the parts of photocopiers from Japan, adding very little local value. What is relevant here is not the mechanics of AD— discussed in Chapter 9—but the role of origin rules. The aim of AD is to protect a domestic industry that is injured by dumping. But, in a world where companies establish alliances with—-and equity stakes in—rival enterprises, establishing which firms constitute the domestic industry is not always easy. In the photocopier AD investigation, Canon (Japan) subsidiaries located in the EU were regarded as foreign firms, while a Xerox (US) affiliate was treated as a European firm. Similarly, in the follow-up anticircumvention case, the Canon subsidiaries were investigated to determine how much local (EU) value was added in the production process. What is interesting about this case is not only that the composition of the domestic industry was determined arbitrarily, but also that a number of the firms who petitioned Brussels for protection had value-added performances that were lower than those of the Japanese firms. These EU firms were basically in the business of importing and distributing photocopiers. They did not produce them. Some even had formal connections with Japanese companies. The AD case was therefore not about dumping, or to protect a national industry, but simply part of a strategy used by individual multinational firms competing for market share. The lack of clearly defined rules of origin introduced one of the elements of discretion that made the strategy attractive to petitioning firms. Source; Messerlin and Noguchi (1991).
supplementary tests such as value-added criteria. The harmonization programme was to be completed in July 1998 but the Technical Committee was unable to meet this deadline due to the complexity of the task.
The results of the technical review undertaken by the WCO were submitted to the WTO by the revised deadline of November 1999, but as of mid-2008 the harmonization work programme had yet to be completed by the Committee on Rules of Origin. This reflected differences in views between members regarding the specifics of some of the rules, the costs associated with adopting a common set of nonpreferential rules, while at the same time countries remain free to adopt idiosyncratic preferential rules of origin, and the extent of discretion that should be left to importing countries regarding the criterion to be applied in instances where no agreement on a common rule exists.
The economic impact of a rule of origin depends on the specific criterion that is used and on the degree of uniformity with which the rule is applied. Rules of origin
214
the political economy of the world trading system
have been problematical mostly in the context of PTAs. The harmonized set of WCO rules of origin, assuming they are ever adopted by WTO members, will not apply to preferential trade. This is not an oversight. Many countries do not want constraints imposed on their policy freedom with regard to the implementation of PTAs or their unilateral trade preference programmes for developing countries. As discussed further in Chapters 10 and 12, rules of origin are a major instrument for 'fine tuning' the effective scope of preferential liberalization at the product level. First principles would suggest that preferential and nonpreferential rules should be the same and that rules of origin should be simple to administer and transparent. Currently, this remains far from being the case.
Pre-shipment inspection: outsourcing customs
Trade facilitation is a key instrument for countries seeking to reduce transaction costs for traders associated with customs clearance and related regulatory enforcement. Many countries have pursued alternative mechanisms-—on either a transitional or longer term basis—to facilitate trade while ensuring that revenue and other objectives are attained. One instrument that may be used to address classification, origin and valuation problems and more generally reduce customs-related compliance costs is pre-shipment inspection (PSI).
As the name suggests, PSI consists of inspection of goods by specialized firms before they are shipped to the country of importation. Governments of importing countries usually decide to engage the services of PSI firms in order to reduce the scope for exporters and importers to engage in either overinvoicing or under-invoicing of imports.9 Overinvoicing may occur in contexts where there are foreign exchange controls, this being a classic way to transfer capital outside the country. Underinvoicing is usually driven by tax evasion considerations: by under-reporting the value of an imported item, traders may seek to reduce their tax obligation (partially evade applicable tariffs).
Governments use PSI in large part because national customs administrations are not able to undertake the required activities. This may reflect a lack of institutional capacity, or problems related to rent-seeking and corruption. Government-mandated PSI should be distinguished from PSI services that are required as part of a contract between buyers and sellers of a product. Most firms that are internationally active in providing inspection services provide pre-shipment certification and inspection of goods because buyers require it. Such services focus on the specifications and quality of goods, not their value. Government-mandated PSI is predominantly concerned with the determination of the quantity and value of goods
9 PSI may have nonrevenue objectives as well. An example would be to ensure that imports meet national (or international) standards of safety or quality.
trade in goods
215
imported into their territories, and tends to be motivated by a desire to reduce revenue leakage for the government and control fraud.
Pre-shipment inspection became an issue for GATT in the 1980s because exporters objected to some of the methods used by inspection firms (Low, 1995). Under the WTO agreement on PSI, countries that use PSI agencies must ensure that such activities are carried out in an objective, transparent and nondiscriminatory manner. Verification of contract prices must be based on a comparison with the price(s) of identical or similar goods offered for export from the same country of exportation around the same time. In doing this, PSI entities are to allow for the terms of the sales contract and generally applicable adjusting factors pertaining to the transaction. The selling price of locally produced goods, the export price of other producers, the cost of production or arbitrary prices may not be used for price verification purposes.
The PSI agreement requires that WTO members who use PSI establish appeals procedures. Complaints may also be brought to an Independent Entity, established under Article 5 of the PSI agreement. Through 2008 only two cases have been brought to the entity. In 1996, a working party on PSI was established by the General Council, with a mandate to review the agreement (as provided for under Article 6 of the PSI agreement). The working party consulted with the International Federation of Inspection Agencies (IFIA), the ICC and a firm of PSI auditors. Overall, the experience with the PSI agreement appeared to have been positive. Problems that were identified included across-the-board, 100 per cent inspection requirements, delays in shipments caused by inspectors being unavailable or failing to arrive for scheduled inspections, or inspectors allegedly having little knowledge of the products concerned.
In 2008 more than 36 developing countries used PSI services, including Argentina, Bangladesh, India, Indonesia, Kenya, Liberia, Mexico, Mozambique, Nigeria, Paraguay, Peru, Philippines, Uzbekistan and Zimbabwe (www.Cotecna.com). Data submitted to the PSI committee in 2008 by the International Federation of Inspection Agencies showed a marked reduction in traditional PSI programmes. The trend resulted from a considerable shift towards modern forms of shipment inspection that are considerably less intrusive for exporters (G/VAL/W/63/ Rev. 10).
Pre-shipment inspection is not a panacea. As discussed by Low (1995), PSI companies have been found to have engaged in bribery to obtain contracts, and circumvention of PSI by traders through exploitation of loopholes in the system (for example, minimum value thresholds) have reduced the effectiveness of these systems in a number of countries. A lack of ex post checking of revenues collected against the reports issued by PSI agencies to Ministries of Finance also reduces the usefulness of PSI for countries that use it, especially taking into account that the costs of PSI for governments can be significant. Pre-shipment inspection firms frequently charge a fee of up to 1 per cent of the value of goods
2l6      the political economy of the world trading system
inspected.10 The prevailing consensus in the trade and development community is that PSI may be helpful in the short term but that in the longer run what is required is serious customs reform and institutional strengthening to allow a government to manage the process itself. The potential downsides of PSI have been considerably reduced over the last decade as a result of new approaches and tools, such as the security label and the electronic transmission of results. These innovations permit achieval of higher cost-effectiveness and speed up the flow of information to customs authorities, ensure documentary security and reduce delays in the importation process.
5.5. Subsidies
Subsidies, and measures to counter their impact on trade, frequently gave rise to disputes in the GATT era. This section discusses the WTO rules on subsidies; disciplines on countervailing subsidization are discussed in Chapter 9. Subsidization may benefit import-competing or export industries. If subsidization distorts trade (expands or reduces trade above or below the free trade level) it may work to offset market-opening commitments negotiated in a MTN. From a rule-making and public policy perspective, a problem that arises is that subsidies may be a desirable form of government intervention. Tax-subsidy schemes may be required to bring marginal private costs or benefits into alignment with marginal social costs or benefits. The need for this arises when externalities (market failures) cause social and private costs and benefits to diverge. Usually this implies that private agents are not given an incentive to take into account the costs or benefits of their actions on others in the economy (for an overview of the theory, see Bhagwati, 1971, and the summary in Annex 2).
Tax-subsidy schemes may be an appropriate means of offsetting externalities or distortions associated with overvalued exchange rates or labour market rigidities if these problems cannot be dealt with directly. They may also be used to redistribute income. Aside from their efficiency advantages—in contrast to measures imposed at the border, which distort both producer and consumer incentives, a subsidy can target either consumers or producers—subsidies are superior from a governance viewpoint. This is because they are more 'visible' to taxpayers—and the Ministry of Finance—than trade policies. As a result, subsidies can be expected to be subject to
IC This practice has given rise to discussions in the PSI Committee concerning the consistency of such fees with Articles II and VTII GATT (discussed above). Proposals have been made that flat fees would be more appropriate.