The paper surveys six WTO agreements selected on the basis of their potential relevance for Fund-supported programs, namely Customs Valuation, Subsidies, Safeguards, Antidumping, Trade-Related Investment Measures, and Trade in Services. It offers a critical reading of the rules, and highlights potential issues of concern for the Fund in its policy dialogue with member countries with selected country examples. As some rules have very different implications for various groupings of countries, the paper calls attention to policy consistency in the trade area of these countries, and points out some areas where implementation of the rules might pose problems.


The paper surveys six WTO agreements selected on the basis of their potential relevance for Fund-supported programs, namely Customs Valuation, Subsidies, Safeguards, Antidumping, Trade-Related Investment Measures, and Trade in Services. It offers a critical reading of the rules, and highlights potential issues of concern for the Fund in its policy dialogue with member countries with selected country examples. As some rules have very different implications for various groupings of countries, the paper calls attention to policy consistency in the trade area of these countries, and points out some areas where implementation of the rules might pose problems.

I. Introduction

The completion of the Uruguay Round in April 1994 and the creation of the World Trade Organization (WTO) in January 1995 paved the way to a new system of rules in the field of international trade. The system is composed of several agreements that are based on the GATT framework. However, there are also new areas not covered by the GATT, such as trade-related intellectual property rights and services. With the WTO framework, some key macroeconomic tools, such as tax or investment policies may now have stricter trade-related implications.

Although WTO rules apply to all member countries, 1/ the purpose of this paper is to present an overview of selected agreements which may affect Fund advice and program design in developing and transition economies. The paper is organized in a systematic way. Each Section covers one agreement with an identical structure composed of a short description of the rules and exceptions, an appraisal of the rules, and an analysis of potential implications for the Fund with country examples. It is intended to help country economists to understand the rules of these agreements.

Six agreements were selected on the basis of their potential relevance for Fund activities: Customs Valuation, Subsidies, Safeguards, Antidumping, Trade-Related Investment Measures (TRIMs), and Trade in Services. This selection would complement the existing work of the Fund on market access and the work of the World Bank on sectoral issues, such as agriculture and textiles.

Since the implications for the Fund crucially depend on the specifics of each agreement, highlighting general conclusions is difficult. However, it is important to emphasize that the establishment of the WTO expands the scope of the rules to all developing countries: under the GATT, countries could be selective in choosing the agreements they wished to ratify; by contrast, countries that ratify the WTO Agreement are automatically parties to all agreements. This calls for more attention to policy consistency in the area of trade in developing countries. As noted in the IMF document “Guidelines/Framework for Fund Staff Collaboration with the WTO,” 2/ Fund staff need to ensure that, in the context of surveillance and use of Fund resources, recommended policy measures and program conditionality are consistent with the members’ agreements under the auspices of the WTO. The main results of the paper are that:

  • the rules covered have very different implications for certain groupings of countries owing to different rights and requirements;

  • TRIMs, Subsidies, and Customs Valuation Agreements are likely to be most constraining, whereas that on Services is likely to be least constraining;

  • compliance with WTO notification requirements can increase transparency and, in some cases, affect rights to transition periods (in the agreements covered, TRIMs and some subsidies);

  • deadlines for phasing out prohibited measures can set maximum limits to policy advice; 1/ and

  • particular attention should be paid to transition economies in the process of applying to the WTO, especially in the area of customs valuation and TRIMs.

Country examples are included as often as possible, to illustrate potential or existing problems between WTO requirements and Fund policy advice. After consultation with different area departments, the following countries are referred to in some of the sections: Algeria, the Baltic States, Bangladesh, Brazil, Cameroon, Chile, Egypt, Ghana, Kenya, Nigeria, Pakistan, the Philippines, Poland, Rwanda, Senegal, Tunisia, Uganda, Vietnam, and Zaire.

Table 1.

Customs Valuation: Summary of Issues and Implications

article image
Table 2.

Subsidies and Countervailing Measures: Summary of Issues and Implications

article image
Table 3.

Safeguards: Summary of Issues and Implications

article image
Table 4.

Antidumping: Summary of Issues and Implications

article image
Table 5.

Trade-Related Investment Measures: Summary of Issues and Implications

article image

Prohibited TRIMs include local-content requirements, trade balancing requirements, exchange restrictions, and domestic sales requirements (illustrative list, see Chapter VI).

Table 6.

Trade in Services: Summary of Issues and Implications

article image
Table 7.

Special and Differential Treatment for Developing and Transition Countries

article image

II. Customs Valuation

1. Description of the rules framework

Any tariff agreement would make no sense without rules addressing the valuation of imported goods for customs purposes. Indeed, depending on the method of valuation, the protective effect of an ad valorem tariff can vary substantially. The same applies to revenue tax. Hence the need for some rules to establish uniform, transparent, and fair standards of imports valuation. This is what the WTO Agreement on Customs Valuation 1/ (“the Agreement”) attempts to achieve.

There are currently two international valuation systems: the Agreement based on the Tokyo Round Agreement (1979), and the so-called Brussels Definition of Value (BDV), established in 1953. Among the developing countries, only Botswana, Hong-Kong, and Lesotho were parties without reservations to the Tokyo Round Agreement; whereas Brazil, India, Malawi, Mexico, Turkey, and Zimbabwe were parties with reservations. When the Tokyo Round Agreement entered into force, there were 33 Contracting Parties to the BDV. This figure was reduced to 11 by 1994 (Algeria, Haiti, Israel, Kenya, Nigeria, Pakistan, Rwanda, Senegal, Tanzania, Uganda, and Zaire), 2/ as countries started to shift to the GATT valuation system:

Under the Agreement, the primary basis of valuation is the transaction value. 3/ The transaction value is the price actually paid or payable in the sale for export (Article 1) adjusted for certain specific costs (Article 8), including commission and brokerage fees, packing and container costs, certain goods and services provided to the seller by the buyer free of charge, or at a reduced cost, royalties and license fees and proceeds, or resale, use, or disposal. Insurance, transport, and handling costs may be added, depending on whether the country concerned chooses to use an f.o.b, c.i.f., or another form of valuation.

Certain limitations apply to the use of transaction value. One particular limitation is in the area of related party transactions. Article 15 defines “related persons” as officers/directors of one’s business, legally recognized partners in business, employer and employee, members of the same family, controlling together a third person, controlled together by a third person, or if a person owns 5 percent or more of the shares of both of them, and if one of them controls the other. The transaction value in a sale between related parties will only be acceptable as the basis for the valuation when it is established that the relationship did not influence the price. In this respect, the burden of proof rests on the importer. He must show that the transaction value closely approximates either the transaction value in sales to unrelated buyers of identical, or similar goods for export to the same country, or to the customs value of identical or similar goods under the deductive method, or to the customs value of similar goods under the computed value method.

If the transaction value cannot be used as the basis for the determination of customs value, either because the requirements of the Agreement are not satisfied, or because the importer has failed to produce sufficient evidence to satisfy the customs as to the truth or accuracy of the declared value, the following alternative methods of valuation must be attempted sequentially: 1/

  • 1. the transaction value of identical goods;

  • 2. the transaction value of similar goods;

  • 3. the deductive method consisting of the unit price at which the imported goods (or identical/similar imported goods) are sold at the greatest aggregate quantity to unrelated persons, subject to deductions of usual commissions and transport charges, c.i.f, and tax and duties charges (if applicable);

  • 4. the computed method consisting of the sum of the costs of material used in producing the imported goods, plus an amount for profit and general expenses equal to that usually reflected in sales of similar goods made by producers in the exporting country in sales in the country of importation; and

  • 5. if unable to determine the value by the above methods, the value is to be determined by reasonable means and available data. However, certain bases of value are specifically prohibited, including minimum values, arbitrary values, the price of locally-produced goods, and the price of goods on the domestic market of the exporting country.

Lastly, the exchange rate used for conversion purposes should be published by the importing country and should reflect the current value of the currency in commercial transactions in terms of the currency of the country of importation (Article IX.1).

2. Preferential treatment and exceptions

Developing countries that were parties to the Tokyo Round Agreement are not entitled to invoke provisions related to delayed application of the 1994 Agreement. However, those countries that were parties to the Tokyo Round Agreement with reservations related to delayed application of the provisions are entitled to maintain and extend them, following a decision of the Committee in January 1995. 1/

Other developing countries may delay the implementation of the Agreement—including elimination of minimum prices—for five years from the entry into force of the WTO Agreement for the country concerned. This grace period can be extended if the five years prove to be insufficient… Countries requesting extension must present valid arguments, and other member countries have to agree to it. Even if developing countries implement the Agreement immediately, or after five years, they can delay further the application of minimum prices, if other members so agree. There is no specific time limit set for this extension. A reservation can also be made to reverse the sequential order of options 3 and 4 listed above at the request of the importer, and delay the application of the computed value method for three years following their application of other provisions of the Agreement.

In August 1995, 30 out of the 74 developing countries that had ratified the WTO Agreement, had made reservations on the implementation of the Agreement on Customs Valuation, as indicated in Table 8. Although the remaining developing countries have not notified the WTO, they have five years to implement the Agreement. The Czech Republic, Hungary, Poland, Slovakia, Slovenia, and South Africa, as industrial countries, are requested to implement the Agreement immediately.

Table 8.

Reservations to the Agreement on Customs Valuation

(as of August 1995)1

article image
Source: WTO

Other reservations are possible depending on remaining ratifications. See Annex I, Table 20.

A five-year delay to accept the provisions of the Agreement.

A three-year delay to implement the computed value method from the date all other provisions have been accepted.

Zimbabwe was party to the Tokyo-Round Agreement with the same reservation. As covered by the January 1995 Decision of the Committee to allow developing countries to maintain same reservations, it did not need to notify the WTO. (The other countries concerned, i.e., Brazil, India, Malawi, Mexico, and Turkey, did not notify). It is reasonable to assume that Zimbabwe did it for transparency purposes.

3. Normative analysis

The adoption of the Agreement by the world’s major trading nations paved the way for the use of a single valuation system worldwide. This, together with the benefits of transparency and predictability, led the World Customs Organization (WCO) 1/ to favor the Agreement as the most suitable valuation system. It also suggests that the other valuation system, the BDV, is being abandoned by the majority of countries, a trend that is bound to continue as countries implement the provisions of the Agreement.

Although a balanced comparison of advantages and disadvantages of the Agreement and the BDV is certainly of little relevance at this stage, it may be useful to understand the rationale behind the preference given to the former, as well as to the practical consequences involved.

Technically speaking, both systems lead to virtually the same values despite methodological distinctions. However, as the value is positive in the Agreement (based on the price actually paid for the goods) and notional in the BDV (based on the price of goods that could be reclaimed conceptually in the course of trade), differences may arise. 2/ In addition, rules are more explicit in the Agreement, as compared to the BDV. 3/ However, one of the most important practical distinctions is that under the BDV the customs administration is in a stronger position, and has more discretion than under the Agreement, to reject the invoice price as a basis for valuation when it appears unacceptable, and to choose alternative methods of valuation.

More problematic is that the BDV has been interpreted and is still applied as a minimum, or reference value system by many developing countries. As a result, some countries claim to apply the BDV, but have practices which differ from their principles, such as official lists, widespread reference prices, or automatic uplifts, which are typical instruments of protection. In practice, such abuses help create price distortions, which is clearly second-best, and do not ensure transparency. They may also be used to combat under-invoicing to increase the tax base for revenue collection purposes.

The rules of the Agreement for establishing customs values in a correct way are certainly more explicit than in the BDV, and offer a solid and detailed valuation framework. However, three practical issues may reduce its strength and/or pose implementation problems:

  • a. Many countries lack the necessary administrative framework to implement the provisions of the Agreement. For instance, without a proper information value system, it is hard to imagine how countries could follow the set of alternative methods to be used (in a hierarchical order) to assess the value when the transaction price is not available. As indicated in the previous Section, the Agreement takes up this problem in allowing developing countries to make reservations on the implementation of some specific provisions, with a possible extension in case of technical difficulties. LDCs are likely to get an extension easier, but it may be less certain for other developing countries. This means that the (correct) application of the Agreement, like any other valuation system, requires a serious and bold reform of customs administrations, structures, and practices in most developing countries, which raises the issue of technical assistance. Indeed, valuation depends less on the legal base than on the manner in which rules are implemented.

  • b. The issue of minimum prices, which is closely linked to the previous point, is very sensitive. Countries have an obligation to change their valuation system to conform to the principles and rules provided in the Agreement, but minimum or reference prices can be maintained for a limited and transitional period. This can be done either by delaying the implementation of the whole Agreement, or making a reservation on Annex III, paragraph 2, and having it accepted by trading partners. In economic terms, their use creates open price distortions yielding welfare costs. However, due to crucial problems of customs valuation in most developing countries (high incidence of under-invoicing, invoice falsification, smuggling, poor bookkeeping standards of importers, low administration performance), reference prices are frequently used to guarantee minimum customs revenue. Their removal is therefore linked to effective customs reform to avoid negative revenue effects.

  • c. It seems that voluntary (non-binding) minimum prices, 1/ to be used as reference by customs officials, for instance, remain permissible. This may create a bias for the use of voluntary rules in developing or transition economies bypassing the Agreement, for the same reasons presented before.

4. Potential or existing concerns for Fund advice

For the reasons presented in the previous Section, compliance with the WTO framework on customs valuation may require technical assistance and appropriate policy advice for developing countries that have requested a grace period for the implementation of the Agreement. In countries that rely heavily on tariffs for fiscal revenue, insuring a smooth transition can be valuable in constructing Fund-supported policies. Since the Agreement can hardly be applied in stages, 1/ countries should use their transition periods (if applicable) to begin preparing the organizational and procedural requirements for implementing the Agreement. This could include ensuring that manuals are developed, staff are trained, and developing an information system that would provide customs valuators the means to make comparisons between prices of goods identical to the imported goods, and to obviate the need for reference prices.

Interesting developments in the field of future technical assistance are the valuation pilot projects under discussion between the WCO, the Fund, and the World Bank. It has been decided that the WCO and the IMF would consult, in order to present the project to the World Bank, and that initially certain countries would be chosen to test the action plan. So far, the focus seems to be on African countries. If the pilot programs prove to be satisfactory, the funding organizations would be inclined to implement this project in a larger number of countries. The usual practice would require the recipient country to submit a technical assistance request directly to the IMF and the World Bank. For countries which are WCO members, the requests for assistance could be formulated in consultation with the WCO which could, in the first instance, help the country to identify the type, duration, and cost of assistance required and, secondly, provide the beneficiary with services necessary to meet the assistance request. For the time being, the WCO provides technical assistance in case of problems for countries implementing the Agreement. 2/

It could also be added that preshipment inspection agreements with international private firms can help in establishing values. However, it offers no long-term solution to the problem of effective customs valuation, because inspection of values is to be conducted on the basis of national valuation regulations. 3/ The preshipment inspection and customs valuation issues are in fact closely related. 4/

An illustration of problems with customs valuation is provided by past World Bank/IMF advice, given to the Philippines in 1992-93. The Philippines was advised to change its home-consumption valuation system 1/ with transition to the BDV as a first step towards the GATT rules on valuation. This gave rise to disagreement between the GATT and the Bretton Woods institutions, but owing to political domestic difficulties, the customs valuation reform is still pending. In the near future, as countries are left with no choice but to implement the Agreement, such problems are unlikely to happen again. However, the Philippines should request assistance from the WTO or the WCO to prepare the transition towards the Agreement.

Another area of crucial concern for Fund advice is the practice of minimum prices linked to customs valuation. On the one hand, as any systematic reference or minimum price system acts as a price distortion mechanism, countries should be encouraged to abandon such practices as soon as possible. Under WTO rules, temporary use of minimum prices remains permissible during the transition period, and even thereafter, if prolongation is approved, which acts as maximum time constraints only. On the other hand, it is certainly fair to admit that many developing countries have resorted, to a smaller or larger extent, to minimum pricing practices as the only secure valuation method because of the significant revenue losses owing to serious customs problems. 2/ Therefore, it seems difficult for countries to remove minimum values immediately without being offered any other revenue alternatives. In other words, all other things equal, the benefits of removing the price distortions might be outweighed by fiscal costs. In this context, it would seem appropriate to encourage countries to develop and use an appropriate value information system, so that the removal of minimum prices becomes feasible in fiscal terms.

A direct illustration of this problem is Kenya. In December 1994, Kenya notified the WTO that “due to revenue considerations, Kenya is still using the BDV procedure when valuing imports for taxation purposes. The country would like to continue with this mode of valuation, while looking into ways of smoothly adopting the WTO Agreement on customs valuation.” Kenya’s reservation certainly reflects the authorities’ fear that a change in valuation system might yield a revenue loss because minimum prices yield higher revenues. A second example is Ghana. Due to severe problems of import under-invoicing, Ghana has started to apply a system of “minimum presumptive duties” in 1994 to a specific list of imports. At the same time, Ghana notified the WTO the wish to delay the implementation of the Agreement for five years.

In both countries, technical advice and assistance appear to be necessary to develop appropriate customs administrations and structures within the five-year period or less, and to eliminate minimum pricing. Again, although this deadline may be extended, it is important to note that it has to be reviewed by the WTO. These remarks also apply for other countries still using the BDV, i.e., Algeria, Haiti, Kenya, Pakistan, Rwanda, Senegal, and Zaire. 1/

For non-members wishing to apply for WTO membership and using minimum prices, faster adjustment might be needed, depending on their country status. Examples include Latvia, Lithuania, and Vietnam.

In Lithuania, to increase tax revenue, the present threshold price system applied to agricultural imports is to be replaced by reference prices. Although the related draft amendment to the Agricultural Law has not yet been passed, it would seem to imply that reference prices, rather than the transaction value, would be the basis for customs valuation. This would go against WTO rules. One solution recommended by the WTO Secretariat is to make the reference price system non-binding. Thereby, a reference price system would be one of many indicators assisting customs inspectors in detecting under-valuation of imports. However, it is crucial how this system is actually implemented. One may suspect that the present system might complicate accession to the WTO, because Lithuania is likely to join as an industrial country with no right to transitional arrangements.

This also applies to Latvia, where a system of reference prices based, in theory, on European Community prices is “optional” for the valuation of imported agricultural products. As EU (producer) prices are known to be high, and depending on its implementation, this system could yield sizable price distortions. In any case, this system could pose problems for WTO accession as Latvia is also likely to join as an industrial country.

The case of Vietnam is different. It has recently applied to the WTO as a developing country. The Fund provided technical assistance in developing the necessary administrative structure for implementing the Agreement. In this context, the system of minimum prices is to be maintained in the short run.

Lastly, regardless of the distortion effects and fiscal implications of minimum prices, and although they may be used during the transition towards the implementation of the Agreement, problems may arise in some cases (e.g., adverse effects on a trading partner). During accession, countries may be requested to give up their right to the transition period. Also, if customs valuation practices affect the interests of other member countries (i.e., nullification or impairments of benefits), a challenge in WTO would be also possible. This means that even if, conceptually, developing countries may benefit from transition periods, the possibility of a challenge under the WTO is not totally ruled out.

To conclude, although the notification requirements in the Agreement do not affect the rights and obligations of member countries as in other agreements (e.g., Subsidies and Countervailing Measures), it would be in the interest of the Fund to be informed of the situation of each country, as this might affect Fund-supported programs. As the previous discussion indicates, deadlines for phasing-out proscribed measures (e.g., minimum prices) only set out maximum time constraints for the countries concerned, and Fund-supported programs require faster adjustments on economic grounds. However, when countries fail to respect WTO deadlines or notification requirements, they may encounter problems. This suggests that these issues could be usefully raised in Fund’s policy dialogue, when necessary.

A simple practical example illustrates this point. Uganda ratified the WTO Agreement on Customs Valuation, without reservations, in January 1995. However, minimum prices are still in place and the authorities have stated explicitly that they want to implement the Agreement by the year 2000. 1/ As the issue of minimum prices is still sensitive, and to follow its WTO obligations and avoid problems, Uganda should notify the WTO on the delay.

III. Subsidies

1. Description of the rules framework

The WTO Agreement on Subsidies and Countervailing Measures (“The Agreement”) disciplines the use of subsidies in the goods area (a separate agreement further elaborates disciplines on agricultural products). The GATT rules on subsidies always suffered from lack of clarity because no agreed definition of the term “subsidy” was given. By contrast, the Uruguay Round Agreement on Subsidies and Countervailing Measures provides a first definition of subsidies: a “subsidy” exists when “a financial contribution by a government or any public body” 2/ is involved, and “a benefit is thereby conferred.”

There are three categories of subsidies defined in the Agreement, namely “prohibited”, “actionable”, and “non-actionable.”

  • Prohibited (Article 3): Subsidies contingent upon export performance or upon the use of domestic goods over imported goods are prohibited (with the exception of some practices covered in the Agreement on Agriculture). By definition, such subsidies are specific.

  • Actionable: 1/ Other specific subsidies granted to domestic industries can be maintained, provided they do not cause adverse effects (defined in terms of injury, impairment of benefits, or serious prejudice) 2/ to other members.

  • Non-Actionable: Non-specific and specific subsidies for industrial research and pre-competitive development activities, for adapting to new environmental requirements or assistance to disadvantaged regions, are non-actionable, i.e., not subject to countervailing duties or any other remedial actions, provided they are notified to the WTO Committee on Subsidies and Countervailing Measures before implementation.

Actions can be taken against prohibited (linked to export performance or domestic content requirements) or actionable (i.e., causing adverse effects) subsidies, only if those subsidies are specific to an enterprise or an industry (or even to a group of enterprises or industries). 3/ Such actions can follow two tracks: multilateral challenge through the WTO system, and unilateral countervailing duties at the border, 4/ except in the case where the country granting the subsidy has a negligible share of the complainant’s imports, or when the level of subsidy is de minimis. 5/ Both tracks may be pursued simultaneously, but only one remedy may be applied.

Examples of specific and non-specific subsidies are provided in Table 9, below.

Table 9.

Type of Subsidies and WTO Action

article image

Except in the case of duty drawbacks on energy imported and consumed as production inputs by exporting firms, as defined in Annex II. Such duty drawback schemes are not deemed to be export subsidies, so long as such drawbacks do not exceed the actual tax/duty liability.

2. Differential treatment and exceptions

The Agreement distinguishes several categories of countries in giving them different deadlines to phase out measures inconsistent with the Agreement and different treatment in the context of multilateral subsidies disciplines and countervailing measures. All developing countries are exempt from the presumption of serious prejudice. 1/ In addition, action based on nullification or impairment of benefits can only be taken with respect to developing and transition countries in the event of displacement of other members’ imports or in the event of injury. This means that multilateral action could (conceptually) be taken against certain developing countries’ subsidies during the transition periods in these circumstances. Furthermore, nothing prevents the countries concerned from facing unilateral countervailing actions even within the transition period.

a. Phase-out rules:

  • Industrial countries must phase out all prohibited subsidies within three years.

  • LDCs are not subject to the prohibition on subsidies linked to exports. They are also exempt from the prohibition on subsidies contingent upon the use of domestic over imported goods for a period of eight years, starting in 1995. However, export subsidies for products, 2/ in which a share of at least 3.25 percent in world trade for two consecutive years has been reached, should be phased out within eight years.

  • Countries whose GNP per capita is less than US$1,000 per annum are not subject to the prohibition on subsidies linked to exports. They will be exempt from the prohibition on subsidies contingent upon the use of domestic over imported goods for a period of five years, starting in 1995. These countries are Bolivia, Cameroon, Congo, Côte d’Ivoire, Dominican Republic, Egypt, Ghana, Guatemala, Guyana, India, Indonesia, Kenya, Morocco, Nicaragua, Nigeria, Pakistan, the Philippines, Senegal, Sri Lanka and Zimbabwe. However, export subsidies for products, 2/ in which a share of at least 3.25 percent in world trade for two consecutive years has been reached, should be phased out within eight years.

  • Other developing countries (self-defined) will be exempt from the prohibition on subsidies linked to export for an eight-year period (with a possible extension), provided that subsidies are progressively phased out during this period. They will also be exempt from the prohibition on subsidies contingent upon the use of domestic over imported goods for a period of five years, starting in 1995, However, export subsidies for products, 2/ in which a share of at least 3.25 percent in world trade for two consecutive years has been reached, should be phased out within two years. In addition, the level of existing export subsidies should not be increased (from 1986 levels).

  • Transition economies: there is a seven-year period for bringing prohibited subsidies into conformity with the Agreement, provided proper notification is made at the earliest practicable date (and not later than end-1996). This transitional period may be extended in “exceptional circumstances.”

b. Privatization: In all developing countries, subsidies given to private purchasers of state-owned enterprises as incentives are safe from multilateral challenge, provided that they are notified to the Committee. 1/ One example of such subsidies is the case where the value of past subsidies granted to a state-owned enterprise is not included in the selling price of the firm to private purchasers. The same applies to direct debt forgiveness and subsidies covering social costs within the framework of a privatization scheme, provided that privatization actually takes place. 2/ By contrast and rather surprisingly for countries in transition, there is no mention of privatization programs: only direct debt forgiveness is safe from multilateral challenge for seven years.

The different transitional arrangements and preferential regimes are summarized in Table 10.

c. Notification of subsidies: There are various obligations linked to the notification of subsidies. In the case of developing countries, whether or not they notify their existing prohibited subsidies does not affect the different phase-out periods to which they are entitled. For example, a developing country member that uses prohibited export subsidies, whether or not it notifies these subsidies as prohibited ones, will have eight years to eliminate them. 3/ Most constraints apply to countries in transition. Failure to notify the existence of prohibited subsidies to the WTO at the earliest practicable date (and in any case not later than the end of 1996) will result in the denial of any extended transitional period to remove such subsidies (the normal three-year period will apply).

Table 10:

Subsidies, Phasing-Out Periods, and Country Status

article image

Unless displacement of other member’s imports occurs, or in the event of injury (nullification or impairment of benefits).

3. Normative analysis

The Agreement on Subsidies and Countervailing Measures aims at controlling the use of subsidies with, among other things, a more precise distinction between different categories of subsidies. In this respect, the achievements of the Agreement are certainly positive because the rules are more forceful than those replaced.

As a general comment, the Agreement disciplines more the use of specific subsidies in contrast to general ones. This emphasis on the “fairness” of subsidies may reflect the mercantilist underpinnings of GATT/WTO rules in general, with less consideration for broader economic efficiency. In economic theory, it is not evident why general subsidies would be less costly in terms of welfare than specific ones. It is true that subsidizing certain sectors may seriously distort incentives and result in inappropriate allocation of resources, with substantial costs for the economy. However, although general subsidies may not distort incentives toward a specific sector, resource allocation would be affected by such artificial conditions, 1/ and the fiscal implications might be enormous. In addition, the theory of domestic divergence makes it clear that in the presence of certain distortions or market failures, some subsidies may be first-best instruments. 2/ These can be either specific subsidies (product X in a partial-equilibrium analysis), or general subsidies (knowledge, education), depending on the divergence. These few conceptual elements indicate that there is no clear welfare analysis behind the GATT/WTO’s separate treatment of specific and general subsidies, which, in turn, suggests that subsidies allowed under the WTO may have very different welfare implications, depending on their nature.

This being said, in economic terms, any move in the direction of less subsidies certainly makes sense. First, because subsidies have well-known adverse effects on the domestic economy in terms of both resource allocation distortions and fiscal imbalance. The latter deserves close attention because it may have serious implications for macroeconomic adjustment. Second, because the widespread use of subsidies has fostered the mushrooming of countervailing duties and related regulations.

In the Agreement, however, differential treatment is still granted to developing countries, which is often described as a major drawback (OECD, 1994). This raises two different problems. First, although the Agreement seems to give some support to the usefulness of subsidies in developing countries (which is questionable), the criterion chosen (e.g., self-definition as “developing”, GNP per capita below US$1,000) can be arbitrary. As GNP figures are often distorted, this creates an artificial limit to subsidies (regardless of judgment on the usefulness of subsidies). Second, with a cost-benefit approach in mind, it is somehow ironic that the right to use subsidies is conferred on those countries which can the least afford them. These two problems, stemming from opposite approaches, lead to a certain paradox: the poorest countries can use more subsidies than others, although they can the least afford them. Furthermore, despite transitional arrangements and exemptions, no developing country should believe that it is protected against either a multilateral challenge, or unilateral countervailing action, although preferential rules exist for developing countries in the context of negligible import levels and de minimis subsidy levels. This may have very important implications because, in some cases, the positive effect of a subsidy could be offset by the countervailing duty levied by another country, or even, the antidumping action launched by another country.

From a political, economic perspective, one could argue that, in case of (proven) injury caused by the export subsidies of a least-developed country, some industrial countries would prefer to use countervailing duties (or antidumping) rather than bringing the case before the WTO. Several reasons could justify this choice:

  • From a practical perspective, the investigation and procedure in any countervailing duty case take place within the domestic political framework which is more convenient than a multilateral challenge under the auspices of the WTO.

  • In international trade practice, countries often act unilaterally before opting for multilateral options.

  • From a mercantilistic point of view and regardless of the total welfare effects on the economy, if the objective is to protect a domestic industry that is being harmed by the foreign subsidy, a countervailing duty is a more suitable tool than a multilateral challenge; WTO challenge will result in either the removal of the subsidy in the foreign country in the medium term, or in the ability to retaliate against the foreign country. In the first case, the removal of the subsidy will not solve the injury problem in the short run, and in the second case, the aim of retaliation is not to protect the domestic industry, but to damage the foreign industry.

  • Reputation considerations would certainly play a role in the decision to challenge an LDC before the WTO.

  • Past evidence in the area of subsidies indicates that even amongst industrial countries, the countervailing duty option (or antidumping) was the rule whereas the multilateral challenge option was the exception.

Hence, the preferential regime granted to various groups of countries is not clear owing to numerous ambiguities in the wording of the Agreement. For example, it is not clear from the text (Article 27.4) whether the ban on the increase in export subsidies refers to a developing country (including LDCs), or to developing countries, except the least developed ones. The second interpretation seems to be predominant, which actually makes sense with respect to the purpose of the Article. Therefore, from a legal point of view and although the wording is very ambiguous, one can assume that LDCs (so defined in Annex VII of the Agreement) are exempt from this rule. Also the term “increase in export subsidies” remains unclear and could be subject to various interpretations. The same comment applies to Article 27.13, which states that subsidies granted in the context of privatization schemes may be permitted, provided that they are given for a “limited period” and notified to the Committee. The precise meaning of “limited period” is vague.

Another important problem is the difference of treatment between developing and transition economies with respect to privatization subsidies. As Section 2.b, above, indicates, such subsidies are permissible in developing countries, but not in transition countries. This situation certainly reflects, in part, the relative negotiating strength of the category of countries concerned during the last round of negotiations.

It can be argued that the export-competitiveness provision could affect some dynamic developing countries whose specialization is in a few products, provided such products are subsidized. 1/

Lastly, the countervailing part of the Agreement, which is central in the text, it is worth mentioning. Although the procedural requirements in the determination of injury, provisional measures, undertakings, and imposition of countervailing duties are identical to the rules in the Antidumping Agreement (see Chapter V), there are differences as well as similarities between countervailing duties and antidumping duties: the main difference is that countervailing duties may be used only against prohibited or actionable subsidies, whereas antidumping duties are used against “dumping” problems with no proper definition of dumping (see Chapter V.3). This means that either a countervailing duty or an antidumping duty, depending on the case, could be levied against a prohibited/actionable subsidy. The main similarity is that the economic rationale behind both instruments may be questioned because of the absence of genuine cost/benefit analysis on the effects of such measures on the domestic economy. The economics of countervailing is therefore questionable, as it only considers the producer’s point of view. It is clearly second best, with adverse effects in terms of resource allocation and welfare.

4. Potential or existing concerns for Fund advice

The new set of rules on subsidies can affect tax policy advice in developing countries, and may have some implications for exchange rate policies.

Broadly speaking, the tendency toward less subsidies, as well as the freezing and phasing out of export subsidies, are in line with Fund-supported programs in most developing countries. In general, this means that in future, as the scope for trade-related subsidies is reduced, the need for exchange rate adjustments might increase. 1/

For some developing countries (i.e., those not listed in Annex VII of the Agreement), there may be some maximum constraints on available policy options because they are subject to shorter deadlines than LDCs, to bring their export subsidy programs into line with the Agreement. In any case, shorter transition periods than those available under the WTO may be desirable, given the adverse welfare effects of subsidies. A typical example is Tunisia, where export subsidy programs exist. 2/ As Tunisia’s GNP per capita is well over US$1,000, the measures, which are inconsistent with the Agreement, will have to be phased out in eight years at the latest. This requires a careful examination of the existing measures: some fiscal incentives, such as certain income tax or real property tax rebates, may be prohibited by the Agreement. 3/

Another example is Colombia, which, in June 1995, notified a subsidy program that contained an export promotion scheme. This scheme consists of tax credits equivalent to a percentage of the f.o.b. value of the exports, and represents all, or part, of the indirect taxes and other levies prepaid by the exporter. The implication of the Agreement for Colombia could be that such a program should be phased out within eight years or less. Fund recommendation could therefore focus on the analysis of the program to verify its nature, the economic costs involved, and the length of any applicable (or desirable) phase-out period.

The case of Turkey deserves close attention. Turkey’s notification contains two different subsidies that are given different treatment under the Agreement:

“Within the aim of compensation for the disadvantage due to the high energy costs, vis-à-vis the average world energy costs levels, the firms manufacturing final products to be exported can purchase electricity, natural gas, and liquified petroleum (LPG), consumed in the production of the exported final products, at discounted prices that are determined by the Money and Credit Board, taking into account the energy prices in the EU countries. Also, for fuel oil, an exemption from customs duty and all levies collected on imports are being implemented.” 1/

Any drawback of customs duties, sales taxes, or value-added tax on imports that are used as inputs in the production of exports are not deemed to be export subsidies under the Agreement (Annex II). This certainly applies to the second measure notified (fuel oil), provided that imported fuel oil is actually used as an input in the production of exportables. In this case, the measure is allowed and does not impose any constraint on Fund advice (regardless of economic considerations). By contrast, the first measure notified (applying to electricity, gas, and LPG) does not seem to encompass any customs duty drawback. Under the Agreement (Annex I), any government subsidy on a product granted to exporters on terms or conditions more favorable than those commercially available on world markets for competitive products to exporters is a prohibited export subsidy. This would suggest that, unless Turkish exporters get electricity, gas, and LPG on world markets at the same price as under the government scheme, 2/ the measure notified may be understood as an export subsidy. If so, Turkey may have to phase out its first notified measure within eight years (maximum).

An example of the problems arising from the interpretation of the Agreement is a tax incentive scheme proposed by the Fund staff in the Philippines for export industries and firms with a good export potential. The potential problem is that the incentives may be considered as new export subsidies and fall into the category of prohibited subsidies. For the time being, the Philippines is part of the group of countries listed in Annex VII and therefore is exempt from the ban on new export subsidies. However, it is really on the borderline as its GNP per capita was close to US$989 in 1994, 3/ and therefore is likely to outstrip the US$1,000 threshold soon. If so, export subsidies would have to be eliminated within a maximum of eight years starting from the year the GNP threshold is reached. 4/

Another potentially important implication for Fund-supported programs is the obligation of transition economies to notify the WTO their prohibited subsidy programs before the end of 1996. Failure to notify will prevent them from benefitting from the transition period to phase out such subsidies, apart from the normal three-year period to which they are entitled. For instance, Poland implemented an export promotion program in 1994, including tax and financial incentives (income tax relief for investors in export promoting investment, property tax rebate for restructuring investment purposes in the area of exports, accelerated depreciation rates for modernization investments in exports, and attractive refinance credits). 1/ As some of these measures may fall into the category of export subsidies defined in Article 3 of the Agreement, formal notification is needed as soon as possible. Indeed, Poland may face two problems: first, a trade partner may challenge Poland soon if such measures are discovered; second, if Poland notifies with substantial delay, a trade partner may challenge Poland on the grounds that the notification was made too long after the “earliest practical date” (as required by the Agreement).

Currency retention schemes involving a bonus on exports are prohibited under the Agreement (Annex I). Although several exchange rates for exports with various retention schemes were used in 31 countries in 1994, 2/ most are either countries listed in Annex VII of the Agreement (and therefore exempt from the ban on export subsidies) or former members of the Soviet Union which are currently negotiating their WTO accession. Only four countries would be affected, i.e., Brazil, Colombia, Suriname, and Thailand. If existing retention schemes were found to fall into the category of prohibited subsidies in those countries, the latter would have eight years (maximum) to phase them out. Non-member countries joining the WTO as developing countries would also have an eight-year period to phase out prohibited measures from the date of ratification of the WTO Agreement.

Lastly, an important general comment should be made about any Fund’s recommendation related to subsidies. Although the Agreement offers some transition periods to some countries to bring their subsidies into line with the Agreement, such countries remain exposed to other members’ countervailing actions, and, in particular, to countervailing duties. This is an additional potential cost to be taken into account in any policy dialogue with the countries concerned. To conclude, the various phase-out periods provided by the Agreement should be understood as maximum deadlines, but the Fund can promote faster policy adjustments based on economic efficiency.

IV. Safeguards

1. Description of the rules framework

Trade liberalization is certainly welfare-enhancing; nevertheless, there may be exceptional situations (e.g., severe damage to a specific sector) in which parties to a trade agreement may wish to take temporary protection measures, without invalidating their commitments vis-à-vis other countries. This was the idea behind providing the GATT with a so-called “escape clause” in Article XIX (among other contingent protection measures) allowing for temporary safeguard measures. Such an escape clause has been used only infrequently in recent years. Measures falling outside GATT discipline, such as voluntary export restraints (VERs), have circumvented safeguard measures and undermined the credibility of the Agreement on Safeguards (the “Agreement”). This motivated efforts to clarify and reinforce the discipline in the new Agreement on Safeguards and eliminate measures that escape such control.

Under the Agreement, a country may apply a safeguard measure to a product only if two conditions are satisfied:

  • imports of a product must be increasing, relatively or absolutely, or

  • such imports, combined with the member’s WTO obligations must seriously injure, or threaten to injure a domestic industry producing like or directly competitive products. 1/

Investigations on injury to a domestic industry must be based on “all relevant factors.” 2/ Then, safeguard actions can be taken in the form of tariffs or quantitative restrictions in a non-discriminatory way, 3/ but the use of any VERs or orderly marketing agreements (or similar measures) is prohibited. In addition, existing VERs and orderly marketing agreements (so-called “grey areas”) have to be phased out by January 1, 1999, 1/ which required notification to the Safeguards Committee before March 1995.

In critical situations (where delays can cause damage), provisional safeguard actions can be taken, provided that there is clear evidence that increased imports caused—or were threatening to cause—serious injury. However, only tariffs can be used, the duration of which shall not exceed 200 days. Safeguard actions, including the applications of provisional measures, shall not exceed four years, with a possible extension to eight years if the serious injury continues.

Lastly, a member facing a safeguard measure is entitled to receive equivalent compensation in the form of a trade concession from the country applying the measure. This should be the result of consultations between the countries concerned. If no agreement is reached, and when the safeguard measure is taken after an absolute increase of imports, the country facing the safeguard can retaliate (i.e., suspend the benefit of an equivalent measure) after three years following the implementation of the safeguard action. However, if no agreement is reached, in case of a safeguard measure taken in the context of a relative increase of imports, the country facing the safeguard can retaliate anytime.

2. Preferential treatment and exceptions

There are basically two implications of the new Safeguard Code for developing countries. One is the use of safeguards against developing countries, and the second, the use of safeguards by such countries. The first is that safeguard measures shall not be applied against a developing country’s product as long as the latter’s share is less than 3 percent of total imports of that product in the importing country. 2/ The second is that developing countries may extend the application period of a safeguard measure for ten years instead of eight. They will also have the right to reintroduce a measure on a product after a shorter period than other countries.

As there are no provisions for transition economies, their implications depend on how they classify themselves: Romania, for example, considers itself a developing country, which means that the above provisions apply. By contrast, Poland, as an industrial country, will have to comply to the four-year period.

3. Normative analysis

Any assessment of the safeguard measures should clarify that the economics of safeguards are linked to the costs of ordinary protection, i.e., distortions in resource allocation and incentives, as well as consumption losses. The effects of protection are further accentuated in developing countries where production is highly dependent on imported intermediates. Therefore, safeguards are not more than second-best solutions. In addition, safeguards may yield some serious time-inconsistency problems when a government commits to a liberalization scheme, and enterprises do not adjust in expectation of obtaining protection.

However, nothing in the WTO rules separates trade restrictions that enhance the national economic interest from those that do not. The fact that the “escape clause” has not been frequently used recently 1/ does not mean that countries are aware of the economic costs of protection, but rather that other contingent protection measures have been used for convenience reasons under the GATT, such as VERs, countervailing duties, and antidumping. The fact that safeguards involve compensation (or retaliation) is certainly important in this respect (see Rom (1994)).

The rationale behind the revision of the Agreement was to strengthen GATT discipline, provide flexibility, but prevent abuses. However, many shaded areas remain. First, although the definition of the circumstances under which safeguards may be used has changed, there is considerable room for discretion in the interpretation of the terms “serious injury” and “threat of serious injury.” For instance, imports of a given product have to increase in absolute or relative terms. However, imports of a product may increase relative to domestic production simply because of exogenous domestic factors (e.g., fall in demand), thus complicating the interpretation problems. Second, the enforcement of the ban on VERs is a central problem because VERs may take many forms, even go underground and become bargaining gambits between governments (OECD (1995), page 55). The last point is particularly interesting because the Agreement encourages members to notify the Committee on Safeguards of any “non-governmental measure” appearing to be a VER. In some instances, however, a country would certainly hesitate before accusing a neighbor because of fears of retaliation. Third, the legitimation of quantitative restrictions is a drawback to the Agreement, as it may contradict the non-discriminatory spirit of safeguards, at least in relative terms. Article 5b indicates that an importing country can seek an agreement with exporters on how to distribute the quota among them, but cannot do so according to historical shares. The importing country may use “quota modulation” if it can prove that the imports from a specific country have increased to a much larger proportion than the total imports of the product. Therefore, there is room for selectivity in relative terms in giving different shares of a quota to specific suppliers. 1/ Fourth, although the interests of affected parties are mentioned in the listing of criteria and procedures for the determination of injury, there is no explicit public interest clause.

To conclude, the effectiveness of the Agreement will depend on the ability of the WTO to control the use of safeguards or prevent the use of VERs underground. While the Agreement may improve disciplines and restrict the use of VERs, the problem remains that antidumping measures may still be chosen as the preferred path to WTO-consistent protection.

4. Potential or existing concerns for Fund advice

The Agreement on safeguards may provide some useful guidelines for trade policy advice. Although far from perfect, the Agreement could prove to be the least damaging, (“next-best”) to national economy of the WTO-legal instruments for emergency protection (e.g., countervailing, antidumping), provided that transparency, and specific constraints on its use are followed. So far, 30 developing and transition countries have notified the WTO their existing safeguard framework as shown in Table 11.

If requested, recommendations should insist on the use of temporary price-based measures (tariffs) instead of quantitative restrictions, as the works of Baghwati (1989) suggest. Safeguards being second-best by definition, the use of quantitative restrictions would yield further distortion effects. In addition, useful provisions could supplement any WTO-legal safeguard framework. In particular, the inclusion of an explicit public-interest clause in the investigation procedure is crucial, as well as precise dispositions concerning the maximum duration of safeguard measures. A possible candidate—among others—for such advice is Cameroon that recently expressed interests in developing either a safeguard or antidumping framework to struggle against unfair competition. 2/

Table 11:

Notification of Safeguards Legislation to the WTO

(as of August 1995)

article image
Source: Notifications of countries to the WTO.

GATT 1994 Safeguard Agreement has simply become national law.

No current legislation, but interested in developing one.

It should also be clear for countries, whether or not they notified their existing VERs by March 1995, that they have four years to remove such measures. So far, only seven countries notified their existing VERs, i.e., Cyprus, the EU, Korea, Mauritius, Slovenia, South Africa, and Thailand.

V. Dumping and Antidumping

1. Description of the rules framework

The new Agreement on Antidumping 1/ (the “Agreement”) deals with procedures for determining dumping, assessing injury, and taking appropriate actions. According to the GATT framework, dumping occurs when the price of a product exported from one country to another is less than its “normal value”, that is:

  • less than the comparable price, in the ordinary course of trade, for a like product when destined for consumption in the exporting country (i.e., export price lower than domestic price), or

  • in the absence of such a domestic price, less than either the highest comparable sales price of the like product for export to any third country, or the cost to production of the product in the country of origin plus a reasonable addition for selling costs and profit.

If a preliminary investigation shows that injury, dumping and causality from one to the other are present (i.e., when the dumped imports have caused injury to the domestic industry producing a like product), preliminary measures can be taken. These are generally in the form of duties, the implementation of which can only enter into effect not before 60 days after the initiation of the preliminary investigation. The antidumping duty may not exceed the dumping margin (defined as the difference between the “normal value” of the product and its price when sold for export). Then, if the final determination of dumping, injury, and causality is positive, an antidumping duty can be imposed.

A country taking an antidumping action will have to publish information and explanation at each stage of the antidumping action (initiation, preliminary and final determination, imposition of a preliminary or final duty, etc.). In addition, all interested parties shall have full opportunity to defend their interests in the dumping investigation, which should ensure transparency. Antidumping duties are supposed to be corrective, not punitive, and refundable if dumping ceases, or is not found. In addition, the Agreement specifies that antidumping duties must be terminated at the latest after five years (“sunset clause”) unless a review determines that expiry of the duty would likely lead to the continuation of dumping and injury.

2. Preferential treatment and exceptions

There is a special reference to developing countries with respect to using antidumping actions against them: “it is recognized that special regard must be given by developed country members to the special situation of developing country members when considering the application of antidumping measures under this Agreement.” Accordingly, possibilities of constructive remedies should be explored before applying antidumping duties where they would affect the essential interests of developing country members.

3. Normative analysis

The traditional users of antidumping in the last decade were Australia, Canada, the EU, and the United States. In the past, targeted exporters have been evenly distributed between OECD and non-OECD countries, and concerned traditionally sensitive sectors, such as chemicals, metals, textiles, consumer electronics, and machinery. However, the alleged success of such antidumping actions (in terms of protection of domestic economy) and the new efforts in trade liberalization in several developing countries calling for new forms of protection explain the recent spread of antidumping policy to other users. 1/ Among the main economic justifications supporting antidumping actions, three deserve attention to give a normative assessment of such policies.

The first justification is that dumping may be predatory (e.g., pricing below marginal cost), and that antidumping is needed to reestablish fair competition to avoid domestic firms from being driven out of business. Although much debated, this is hardly convincing for several reasons. From a conceptual point of view, predatory pricing can yield monopolistic profits only if the firm using that strategy is able to sustain its monopolistic position in the target market. In other words, to cover the initial losses of predatory pricing, a firm should be able to drive out of business, not only domestic firms but also foreign competitors, which is unlikely. Empirical evidence suggests that predatory pricing is a remarkable exception rather than the rule. 2/

The second justification is that dumping may be temporary (for cyclical reasons for instance), and that antidumping is needed on a temporary basis to avoid short-term detrimental resource allocation effects. However, the operational implications of such a rationale, i.e., “temporary” antidumping duties, is not reflected by empirical evidence since most antidumping duties have a high propensity to last more than ten years. 3/

The third justification is that antidumping provisions act as a safety valve in the process of broad-based liberalization efforts. Accordingly, countries are less reluctant to agree on further trade liberalization given the possibility of protection provided through antidumping. There are two main objections to this justification, as noted in Leidy 1994. First, even though the idea of safety-valve is to be accepted, the question remains whether antidumping is the most economically efficient policy given its costs. Second, the “escape clause” provided in Article XIX of GATT was precisely designed to give breathing space to countries (the fact that antidumping substituted in practice for safeguards reflects the greater facility in using the former). 4/ In addition, the safety-valve argument has a positive spirit (the country requires a breathing space to adapt itself to a new economic situation, and compensates other member countries affected by the measure), whereas the essence of antidumping is to target unfair traders (the country puts the blame on another country), hence a genuine contradiction.

The above comments suggest that existing antidumping policies can hardly be justified in terms of either predatory pricing, cyclical effects, or safety valve. Antidumping, the enforcement of which is based on national laws and regulations, has become an entitlement for protection-seeking interests (Finger (1995)), with substantial distortion effects on the nature of competition and incentives, and substantial costs for the consumers. There is indeed no meaningful cost-benefit approach in antidumping. The typical supply-side approach taken in antidumping actions only considers the producer’s point of view. However, the consumers’ side is never really addressed, and no disposition makes consumers’ interests part of the decision. The recent findings of the U.S. International Trade Commission on the economic effects of antidumping and countervailing duty measures in the United States indicate that the net cost to the economy is higher than the gains to protected industries. 1/ This is apparently also the message of another recent draft report prepared by independent experts for the OECD, according to which antidumping measures often impairs prosperity, job creation and investment in importing countries by sheltering producers at the expense of consumers and competition. 2/

Antidumping may also act as a means of signaling to foreign competitors to restrain their sales, which has been described as the “harassment effect” (Leidy, (1994)). In such cases, competition is reduced and antidumping acts as a tax on trade. However, antidumping petitions can lead to de facto VERs to avoid the final determination of dumping. Under certain conditions, these “agreements” may be profitable to an exporter (i.e., the latter will get the quota rent). This would act as a perverse incentive for some exporters to produce more and to sell at dumped prices, in the expectation of negotiating a VER. This is a serious issue because GATT 1994 has not excluded resort to such disguised VERs. Indeed, although VERs are to be banned within four years of the WTO establishment in the framework of the Safeguard Agreement (Article XIX), countries may be tempted to use antidumping actions to obtain de facto VERs (OECD, (1994)).

Lastly, antidumping actions may be challenged as effective instruments to prevent dumping from happening because they seem to tackle the problem from the wrong angle. On a conceptual basis, dumping is possible because of the market segmentation. 1/ Antidumping actions are targeting the consequence of (likely) markets segmentation, but, as noted by Hoekman (1995), “no account is taken of whether price discrimination or selling below cost is the result of market access restrictions.” More generally, antidumping actions do not seem to address the source of the problem, i.e., government policies artificially segmenting markets through various direct or indirect means.

At first sight, the antidumping dispositions contained in the Agreement seem to define an explicit framework of action against dumping problems. However, a closer look at the Agreement reveals problems that may limit its effectiveness.

As such, the Agreement does not provide a genuine definition of dumping practices, and certainly does not define dumping as predatory behavior of foreign firms (such as selling large amounts below marginal cost) aiming at driving local firms out of business. In other words, it does not explicitly condemn dumping per se. In addition, the assessment of the prospect for predation plays no role in antidumping practice under the WTO. In fact, the controversy surrounding antidumping regulation comes to a large extent from the nebulous terminology used in the Agreement.

Indeed, exact determination of the “normal value” and the “export price of the product” under “the ordinary course of trade” presents a real challenge as the definition of each of these terms is subject to various interpretations. The same comment applies to measurement of “costs of production” of the product, and to the “reasonable addition for selling cost and profit.” Such ambiguity allows for a degree of government discretion.

This discretion is also likely to affect the method of calculation of the dumping margin, i.e., the weighted average of the normal value (comparable price for the like product) minus the weighted average of prices of all comparable export transactions, or the normal value (comparable price for the like product) minus the prices of all comparable export transactions. 2/ As averaging methods sometimes inflate normal values, it can be argued that this calculation procedure is one of the most serious flaws of the Agreement, owing to its bias towards finding dumping (Hoekman (1995b); Finger (1995)). 1/

The burden of proof, that is the evidence of injury and causal link, is also a very controversial issue in the Agreement. The determination of material injury must be based on positive evidence and involve an objective examination of both: (i) the volume of the dumped imports and the effect of the dumped imports on prices in the domestic market of the like products, and (ii) the consecutive impact of these imports on domestic producers of such products. A significant increase in dumped products, either in absolute or relative terms according to production/consumption in the importing country, is a necessary condition for finding injury, and dumped imports must be found to cause injury to the domestic industry because of dumping. This means in principle that the necessary causality must be established on the basis of all “relevant evidence before the authorities.” 2/ However, the term “domestic industry” is ambiguous because it refers to the domestic producers of the like product as a whole, or to those whose collective output of the products constitutes a major proportion (more than 50 percent) of the total domestic production of those products.

Furthermore, the required causal link is often called “weak causality” because the Agreement does not require injury to take place. A threat of injury (“clearly imminent” injury) is sufficient, which means that the burden of proof is very soft. 3/

In addition, although the Agreement states that all interested parties shall have full opportunity to bring evidence in the dumping investigation, Article 6.11 defines a narrow set of potentially interested parties (exporters, exporting government, and import competing producers). Consumers, or other groups, may be treated as “interested parties”, but need not be. In other words, consumer groups may provide information relevant to the investigation of injury to an import-competing industry, but are not in a position to defend their interests.

The above comments regarding the determination of injury and the burden of proof also apply to countervailing duties (see Chapter III. 3). However, by contrast with antidumping duties, countervailing duties can only be used against certain subsidies that are rigorously defined in the Agreement on Subsidies and Countervailing Measures.

4. Potential or existing concerns for Fund advice

The main implication of the Agreement is that antidumping is becoming an increasingly common tool in developing countries, as many of them recently passed or are in the process of passing antidumping laws. As Table 12 shows, about 20 developing and transition countries have notified their national antidumping laws to the WTO, and 15 other countries are in the process of drafting or passing such legislation. However, the list below is far from exhaustive because it is based on countries’ notifications. For example, among African countries, only Zambia has notified antidumping legislation, although some African countries have already passed such legislation (e.g., Zimbabwe in 1984, and Senegal in 1994), or are interested in developing one (e.g., Cameroon). Some countries also have protection regimes very close to antidumping laws such as Tunisia. 1/ In Asia, Pakistan is in the process of drafting antidumping legislation, although no notification has occurred. 2/ Among potential future WTO members, available information suggests that the three Baltic States have also developed antidumping frameworks.

In general, the use of antidumping should be opposed on the grounds of economic costs, on the one hand, and on the futility of the entire exercise for countries whose tariff lines are weakly bound 3/ to the GATT/WTO.

First, antidumping procedures are long, complicated and very costly because such actions are “lawyer-intensive.” They are often described as the “rich man’s trade remedy.” Indeed, the training and expertise required to develop expertise in the detailed antidumping procedure is certainly immense, and the WTO does not provide training to countries interested in developing an antidumping framework (apart from basic training to understand the rules). Therefore, it is realistic to assume that a developing country, acting as a total beginner in this field and initiating antidumping procedures against any experienced industrial country, and bringing the case to the WTO’s Dispute Settlement Body, would be likely to lose the case.

Second, in countries where most tariff lines are not bound in the WTO, antidumping legislation can be unnecessary because the same results (restricting imports) can be obtained through tariff increases on specific products. 4/ However, with the current move towards increased trade liberalization, antidumping laws are de facto perceived as alternative and practical protection means.

Table 12:

Antidumping Legislation Notified1 to the WTO by Developing and Transition Countries

(as of August 1995)

article image
Source: Notifications of countries to the WTO.

Note that under Article 18.5 of the Agreement, members shall inform the WTO Committee on Antidumping Practices of any change in their laws and regulations relevant to the Agreement and in the administration of such laws and regulations.

Legislation exists but is being complemented with new provisions GATT 1994.

Antidumping code has simply become national law.

The Indonesian authorities notified that no legislation existed, but according to Trade Policy Review (Indonesia), GATT, 1995, the Ministry of Trade had already prepared a draft legislation.

Third, there is a potential inconsistency in WTO rules between allowing export subsidies (for certain groups of countries at least in the short run) and antidumping regulations and procedures.

In the light of the above comments, the first-best option for countries could be to place antidumping within a competition policy framework, or simply to replace it by competition laws addressing predatory pricing. 1/ The next-best solution could be to develop a safeguard framework (see Chapter IV).

In countries where national antidumping laws are already in place, the challenge for Fund advice is, first, to ensure that the country verifies that such legislation conforms to WTO disciplines as a legal floor, and, second, to suggest that the country supplements such legislation with additional provisions aimed at minimizing the negative effects of antidumping. Implementing an economically sensible antidumping regime requires rising above the WTO-consistency floor, and advice could focus on the adoption of rules that achieve greater discipline in the use of antidumping. Accordingly, the next-best policy recommendation in the area of antidumping could include the following measures:

  • Competition authorities could have a role in the investigation process, as first suggested by Applebaum (1988) and, more recently, by Messerlin (1995). 1/

  • Tighter sunset provisions than the five years provided in the WTO framework.

  • The application of the “lesser duty rule” should be required (it is only made “desirable” in the WTO framework).

  • A public-interest clause should be added, which would mean that antidumping duties could be levied only if a cost/benefit analysis shows that the advantages granted by the antidumping duties for the economy as a whole significantly outweigh their costs. 2/

VI. Trade-Related Investment Measures

1. Description of the rules framework

Foreign direct investment is an important element in fostering export-led growth, both directly and indirectly. Therefore, the measures adopted by governments to attract and regulate FDI include a wide variety of incentives (fiscal incentives, loans, tax rebates, and provision of services on preferential terms) and requirements on the use of domestic resources to promote the national economy (e.g., manufacturing or export performance requirements, technology transfers, etc.). Whenever such investments are related to trade in goods, they are labeled TRIMs. An illustrative list of TRIMs is provided in Annex I, Table 14.

The TRIMs Agreement of the Uruguay Round (the “Agreement”) covers some trade-related investment measures, with reference to trade in goods only. It prohibits those measures that already are inconsistent with GATT 1994 Article III (national treatment) and Article XI (Elimination of Quantitative Restrictions). TRIMs that violate these articles need to be mandatory, operate as conditions for achieving certain advantages, are enforceable under domestic law, and which:

  • require the purchase or use by an enterprise of products of domestic origin or from any domestic source (local-content requirement);

  • require that an enterprise’s purchases, or use of imported products be limited to an amount; related to the volume or value of local products exported (trade balancing requirement);

  • limit the import of products by restricting an enterprise’s access to foreign exchange to the amount of foreign exchange inflows attributable to the enterprise (exchange restrictions);

  • limit the export of products specified in terms of volume or value of local production (domestic sales requirement); and

  • limit the import of products to an amount related to the quantity or value of local products exported (trade balancing).

The above list of TRIMs, which are inconsistent with the WTO framework, is only illustrative.

2. Preferential treatment and exceptions

The Agreement provides for the immediate notification and subsequent phasing out of non-conforming TRIMs after two years for industrial countries, five years for developing countries, and seven years for LDCs. 1/ An extension is possible if the country concerned demonstrates particular difficulties in implementing the Agreement.

Eighteen developing and transition countries submitted notifications of prohibited TRIMs by the April 1995 deadline: Argentina, Barbados, Columbia, Costa Rica, Dominican Republic, India, Malaysia, Mauritius, Mexico, Pakistan, Peru, Philippines, Romania, Slovenia, Thailand, Uruguay, Venezuela, and Zambia. 2/ Of those countries, Mauritius, Slovenia, and Zambia notified that they had no TRIMs. This deadline does not apply to countries which, for exceptional or constitutional reasons, have not yet ratified the Uruguay Round Agreement. These “late comers” have 90 days to notify their, TRIMs following ratifications. 1/ Five of these notified by end-1995 (Cuba, Cyprus, Egypt, Honduras, and Poland). However, the start of the transition periods is not affected by such late notifications. Some of the notifications (e.g., Mexico) are not always clear, as indicated in Table 8a.

Developing countries are allowed to deviate temporarily from the provisions concerning prohibited TRIMs, for instance, in case of balance of payment problems (provided it could be justified under the balance of payment provisions under Article XVIII(b), or for the purpose of infant industry development (provided consistency with Article XVIII(a),(b),(c),(d)).

As there is no special regime for transition economies, those countries self-defined as developing countries (i.e., Romania) are subject to the above mentioned transition period, whereas the others have to phase out inconsistent TRIMs within two years.

3. Normative analysis

Although TRIMs are apparently praised by many developing countries as part of their development strategy, assessing the costs and benefits of such policies is far from easy. In fact, it seems that there are two main reasons for using TRIMs in developing countries: to provide local linkage, and to assist the balance of payments. 2/ Table 13, below, presents some popular TRIMs in relation with these two objectives as well as the status of such measures in the context of GATT 1994.

One could argue that TRIMs, in the form of local-content requirements, may lead to welfare losses in perfect competition (both for the country of origin and for the host country) because of the price distortions involved, as shown in Grossman (1981). On the incentive side, TRIMs work like indirect subsidies, and on the requirement side, they work as indirect taxes on potential foreign inputs. TRIMs in the form of trade balancing and foreign exchange requirements result in restraining imports and, therefore, act like quantitative restrictions. Some authors have shown that export requirements may be either welfare reducing or welfare increasing, depending on the model chosen. 3/ However, it is reasonable to assume that TRIMs induce distortions in terms of economic efficiency.

It has also been argued that the rationale behind TRIMs was often the presence of trade distortions (Moran & Pearson (1988)), or domestic production distortions (Greenaway (1992), Low and Subramanian (1995)), or even distortions at the level of the foreign firm. 1/ Furthermore, the removal of TRIMs could result in other policy adjustments whose effects on trade are difficult to assess.

As to the Agreement itself, many shaded areas remain. Unlike the Agreement on Subsidies, there is no definition of a TRIM, which opens a window for interpretation and differences in opinion. Many areas are not explicitly covered, such as technology transfer requirements, and the Agreement falls short of being a complete set of WTO-consistent investment rules (OECD (1994), pp. 189-190). For instance, export performance requirements remain permissible, as well as some other measures (i.e., manufacturing requirements) simply because there are no explicit legal prohibitions against them (Low and Subramanian (1995), UNCTAD (1994)). However, the border between export performance requirements and export subsidies (which are banned for certain countries) is sometimes thin, and the separate treatment of both issues can be questioned. In any case, it is intuitive that large developing countries with a tradition of protection (e.g., India) have more interest in using export performance requirements (in exchange of access to their large domestic markets) to attract foreign direct investment (FDI), whereas small countries are less able to do so. Furthermore, some overlap remains between TRIMs and subsidies, as local-content subsidies linked to imports are prohibited in the Agreement on Subsidies and Countervailing Measures (see Chapter III). Some financial incentive-related TRIMs (e.g., tax exemptions, excise duty rebates, corporate tax rebates) are very similar to subsidies. Conceptually, export performance requirements may be caught by GATT 1994 rules regarding subsidies, because export performance requirements and investment incentives when used together operate as indirect subsidies (Morrissey and Rai, 1995). However, it is crucial to note that developing and least-developed countries will not be subject to this kind of “subsidy argument” owing to the special provisions of the Subsidy Agreement.

Lastly, although local-content requirements are pervasive worldwide in the area of government procurement, the Agreement does not address them. Many developing countries are not signatories of the WTO Agreement on Government Procurement, 2/ and, hence, face no constraint on the use of local-content requirements in the context of government procurement (Low and Subramanian (1995)).

Table 13:

Trade-Related Investment Measures and Their WTO Status

article image
Source: Morrissey and Rai (1995), page 709.

TRIMS that are prohibited under the Agreement.

Although export performance requirements are not explicitly prohibited, they could become actionable under the Subsidies Agreement subject to exceptions for some developing countries (with the possibility of facing countervailing duties, however).

4. Potential or existing concerns for Fund advice

The Agreement is in a sense retrograde as it recognizes that countries were in violation of their GATT obligations. This calls for the elimination of all the existing prohibited TRIMs, with a grace delay only for countries notifying on time (April 1995, or 90 days after ratification). However, it should be clear that the forbidden measures exclude those with no trade effects and those affecting services (the latter are covered in the General Agreement on Trade in Services). In addition, exchange restrictions approved or maintained by the Fund are excluded from the discussion on prohibited TRIMs. Such restrictions cannot constitute the basis of action within the WTO.

The Agreement calls for the gradual elimination of prohibited TRIMs within the corresponding time periods (seven years for LDCs, five years for developing countries, and two years for transition economies that consider themselves industrial countries), or preferably, within shorter periods depending on economic efficiency assessments.

An example of such phasing out is provided by the motor vehicle industry in the Philippines. TRIMs notified to the WTO in 1995 include local-content requirements in the manufacture of automotive parts and foreign-exchange requirements for imports (see Table 17a for details). Recent estimates provided by a World Bank study indicate substantial benefits to the assembly and components industries TRIMs, 1/ but large losses to vehicle purchasers. The study concluded that measures to eliminate the domestic-content and compensatory export requirements should be accompanied by simultaneous reductions in tariffs on assembled vehicles. In any case, this should be done within five years (maximum) since the Philippines is not an LDC.

The Agreement may also have important consequences for countries that fail to notify within 90 days following ratification of the WTO Agreement because they will not be granted a transition period to phase out their existing prohibited TRIMs. This holds unless the measures concerned can be justified under other agreements (e.g., the Agreement on Subsidies, Article XVIII). The possibility of challenge in the WTO may call for immediate policy revision in the countries concerned. 2/ GATT/WTO Trade Policy Review reports indicate that potentially prohibited TRIMs exist in the following countries, but were not notified within the 90 days required (see Tables in Annex I): Bangladesh, Chile, Ghana, Kenya, Korea, Nigeria, Senegal, and Uganda. 3/ Other countries with existing potentially prohibited TRIMs include Cameroon, but the implications are slightly different. Cameroon ratified the WTO Agreement in December 1995, and is required to notify its measures within 90 days hereof. In general, the notification record appears to be low in some developing countries, especially in Africa. This might reflect the lack of incentives for timely notification, and the legacy of GATT practices such as the ability to be selective in “picking” Agreements and the lack of disciplinary action.

An interesting and yet unresolved case is provided by Brazil. In July 1995, Brazil notified a new set of (apparently temporary) measures planned or already implemented on trade policies in the automobile sector. Such measures include imports conditional upon export and local-content requirements. Most of these measures are likely to go against the WTO Agreement. The risk here is a challenge under the WTO by one of Brazil’s partners. In January 1996, Brazil stated that it is planning to apply for a waiver for the measures.

The Agreement may have implications for recommendation in tax matters involving present and future incentives for foreign direct investment. While export performance requirements remain permissible, local-content requirements are banned. It may shift incentives from one type to another with unclear impact on economic efficiency. In addition, although the desire to attract foreign direct investments is understandable, whatever constraint the Agreement implies, the countries concerned should not forget that the most important condition for foreign investors is macroeconomic (and political) stability. The most appealing tax incentives will not do any miracle, and the key component of a successful FDI regime is sound macroeconomic policies.

According to Article 5.4, TRIMs introduced less than 180 days before the entry into force of the WTO Agreement will not benefit from transitional arrangements. Together with Article 2.1, this means that measures introduced after July 1994 should be phased out without delay. This will constrain future investment incentives in members. 1/

Table 14.

Examples of Trade-Related Investment Measures

article image
Sources: Low and Subramanian (1995), OECD (1995).

Means that the measure is covered by the TRIMs Agreement.

Note: Above examples relate only to those investment measures that have an impact on trade in goods.
Table 15a.

Trade-Related Investment Measures in Africa Notified to the WTO (End-1995)

article image
Source: WTONote: The date in parenthesis is the deadline for phasing out the notified measures. However, extensions are still possible.
Table 15b.

Existing Trade-Related Investment Measures in Africa not notified to the WTO

article image
Source: GATT various Trade Policy Review reports, 1995.
Table 16.

Trade-Related Investment Measures in Latin America notified to the WTO

article image
Source: WTO Committee on TRIMs, 1995Note: The date in parenthesis is the maximum deadline for phasing out the notified measures. However, extensions are still possible.
Table 17a.

Trade-Related Investment Measures in Asia notified to the WTO

article image
Source: WTO Committee on TRIMs, 1995Note: Date in parenthesis is the deadline for phasing out the notified measures. However, extensions are still possible.
Table 17b.

Existing Trade-Related Investment Measures in Asia not notified to the WTO

article image
Source: GATT various Trade Policy Review reports, 1995.
Table 18.

Trade-Related Investment Measures in Europe notified to the WTO

article image
Source: WTO Committee on TRIMs, 1995.Note: Date in parenthesis is the deadline for phasing out the notified measures. However, extensions are still possible.

VII. Trade in Services

1. Description of the rules framework

The framework of rules and principles for trade in services provided in the General Agreement on Trade in Services (GATS) aims at facilitating trade in services under conditions of transparency and progressive liberalization, with a particular emphasis on the expansion of the service sector in developing countries. The first main achievement of the GATS is to provide a definition of trade in services: it is defined as the supply of a service through four modes of supply, namely, cross-border supply, consumption abroad, commercial presence, and presence of natural persons suppliers of services. The definition includes the movement of factors, as well as of consumers.

Some precise examples of transactions for the four modes of supply are:

  • Cross-border supply: a French bank purchases securities from a Swedish bank.

  • Consumption abroad: a German resident crosses the border to deposit money in a Swiss bank.

  • Commercial presence: Hungary allows the establishment of foreign banks to offer banking services in its territory.

  • Personnel movement related to financial transactions: e.e., Thailand allows a British bank branch located in Bangkok to hire one or two (non-local) top financial managers for a given period of time. 1/

Unconditional most-favored-nation (MFN) treatment is the core general obligation of the Agreement. Accordingly, any foreign service or service supplier must be treated no less favorably than any other foreign supplier, for the four above-mentioned modes of supply. Exceptions to the MFN requirements are possible and defined in an Annex to the Agreement, but they are limited in time (not more than ten years) with a review requirement after five years. Market access and national treatment obligations are scheduled by mode of supply and apply only to listed sectors and subsectors. They can be subject to conditions and limitations (either across all modes of supply or for a specific mode).

In market access, an (in principle) exhaustive list of (possible) limitations is defined in Article XVI and includes the following six measures:

  • the limitation of the number of service suppliers,

  • the limitation of the total value of service transactions or assets,

  • the limitation on the total number of service operations or quantity of service output,

  • the limitation of the quantity of foreign services in total production or consumption,

  • the limitation of the type of legal entity through which a service supplier is permitted to supply a service (i.e., branch versus subsidiary), and

  • the limitation of participation of foreign capital in terms of maximum percentage limit for foreign shareholding or the absolute value of foreign investment.

As far as national treatment is concerned, foreign service suppliers must receive treatment de facto “no less favorable” than national suppliers. This is consistent with “de jure” differences in cases where identical treatment may actually worsen the competition conditions for foreign-based firms (OECD (1995), page 121). Exemptions to this can take many forms in the schedules.

Apart from the quantitative or other limitations listed above under market access and national treatment, there are also both general and specific exemptions to the rules. These include measures deemed to be necessary to protect public morals, human, animal or plan life, or health, to maintain public order, or to prevent either fraud, or the effect of a default on service contracts (Article XIV). Some qualitative restrictions are also possible on qualifications requirements, technical standards, and licensing requirements, but this should not constitute unnecessary or discriminatory barriers to trade (Article VI). Members can also restrict trade for balance of payment purposes. In such cases, restrictions are possible on trade in services listed in specific commitments, including payments and transfers (Articles XI and XII). Both capital and current account restrictions are possible, subject to qualifications, as will be discussed in Section 2.

In general terms, all MFN exemptions, market access, and national treatment commitments apply only to services that are included in the schedules of members. Such specific commitments are scheduled by mode of supply and by sectors according to a mix of “positive list” and “negative list” approaches; the “positive schedule” simply lists the services which are actually open to competition and the measures taken to comply with the broad principles of liberalization, whereas the “negative list” describes the sector-specific qualifications, conditions, and limitations that continue to be maintained. All this means that commitments are very complex.

2. Preferential treatment and exceptions

The large flexibility of the system makes it in principle less dependent on exceptions. Few limitations on national policy are imposed under the GATS, given the ability of countries to schedule commitments à la carte.

Apart from special rules applying to preferential (regional) liberalization that are similar to the GATT, there are few articles explicitly related to developing countries. Basically, with the aim of promoting liberalization, as well as increased participation of developing countries in trade in services, those countries benefit from “appropriate” flexibility, certain facilities, and “technical co-operation.” In any case, many developing countries acceded to the GATS with very minimal commitments, which means that the GATS has minimal implications for the majority of developing countries (Hoekman (1995a), page 27).

Countries may adopt or maintain restrictions on trade in services on which they have undertaken certain commitments in case of serious balance of payments or external financial difficulties. This applies in particular to developing and transition economies to ensure, inter alia, the maintenance of a level of financial reserves. However, the restrictions should be consistent with the Articles of Agreement of the IMF. In dealing with payments for cross-border transactions, the Agreement separates current account and capital account issues.

Current account transactions: In principle, countries must guarantee convertibility on current account transactions for commitments made. However, restrictions approved (or maintained) under the Fund Articles VIII and XIV are exempt or specifically allowed. Restrictions linked to balance of payment problems should be notified and discussed with the WTO’s Balance of Payments Restriction Committee (BOP Committee), even though some of them have to be approved separately by the Fund. Also, countries can invoke restrictions on prudential grounds (i.e., “to ensure the integrity and stability of the financial system”).

Capital account transactions: Although free capital mobility is required in the GATS, it refers specifically to the commitments of mode 1 (cross-border transactions, both inflows and outflows) and mode 3 (commercial presence, inflows only). Capital controls can be introduced either at the request of the Fund, 1/ or if consultations are held with the BOP Committee. Lastly, countries can also invoke restrictions on prudential grounds.

Restrictions on both current and capital account transactions can be justified on various grounds, which seems to give freedom to deviate from the commitments. How this will work in practice, and whether or not it will give birth to potential challenges, remains to be seen.

3. Normative analysis

Gradual liberalization of trade in services is certainly a positive development, and developing countries should reap substantial gains from the overall process in the long run. These gains are linked to improved capital access and increased efficiency in allocating financial or other resources, and, in particular, to better quality or lower price of inputs to export industries.

The GATS should be perceived as a first attempt to cover services, but, like all “first drafts”, it is not entirely satisfactory for a number of reasons:

First, in many areas, the Agreement is still temporary or incomplete depending on sectors. In financial services, countries can start withdrawing their commitments in November 1997, which may render the life of the commitments short. Countries that modified their schedules in financial services 1/ have until July 1996 to ratify the second protocol of the Agreement. In the telecommunication and maritime transport sectors, for instance, negotiations will continue until mid-1996. Rules on safeguards, subsidies and public procurement remain to be negotiated.

Second, the hybrid structure of the GATS makes it complex, and thus reduces the importance of bindings. Countries are allowed to schedule commitments à la carte and resort to many general or specific exceptions. An example of such discretion is the commitments made by Egypt. In all sectors, the supply of mode 3 (commercial presence) is subject to an assessment of the “economic needs” of the country, which is vague. In addition, in many countries the provisions of services is subject to licensing by non-governmental bodies (e.g., medical associations), or in the area of financial services, to approval by the central bank. In such cases, any restriction based on prudential measures is subject to wide interpretation.

Third, the positive list approach is restrictive and lacks transparency. For instance, there is no information on sectors/sub-sectors in which no commitments are scheduled. It automatically leaves out new products or sectors, and therefore provides few incentive to future effective liberalization (Hoekman (1995a)). This contrasts with a negative schedule approach with which the country lists only those services that continue to be protected from competition, which provides an actual basis for negotiation. Furthermore, the positive approach “will make it very difficult for anyone other than experts on the service in question to evaluate the schedules” (OECD (1995), page 123).

Fourth, many market access barriers are absent from liberalization schedules, which may reduce the effectiveness of liberalization in some sectors. These are labor laws, tax regimes, restrictions on land availability, ownership or use, competition policies, regulation of monopolies, and enforceability of contracts.

The first three elements mentioned make it difficult to give any economic assessment of the impact of the Agreement at this stage. A first reading of the different schedules of commitments calls for a few comments. In all sectors, it is fair to say that, in most countries, commitments reflect the existing status of liberalization.

For example, in numerous developing countries, mode 1 (cross border) and mode 2 (consumption abroad) are unbound, whereas mode 3 (commercial presence) is open, but restricted. Extreme examples of protection include India and Pakistan. 1/ In general, transition economies seem to be fairly open in non-financial services, restrictions being essentially put on mode 3 (commercial presence). This may stem from the fact that as services were underdeveloped under central planning, there is less need for protection. In the financial services sector, most developing countries’ commitments are non-existent in mode 1 (cross-border) and restrictive in mode 3 (commercial presence), owing to numerous conditions and limitations. For instance, in the Philippines, cross-border borrowing or lending is subject to commercial presence, which in turn is subject to its own limitations (e.g., reciprocity requirements). It should be noted that restrictions on payments, transfers and capital flows are maintained by a number of developing and transition countries, which could explain the modest commitments made especially in cross-border trade in financial services. For example, Brazil, Colombia, Egypt, and the Slovak Republic (among many others), all maintain current account restrictions, and all have very limited commitments in cross-border transactions. Without current account convertibility, commitments in mode 2 (consumption abroad) can be difficult to make. In mode 3 (commercial presence), market access is often restricted by national monopolies (e.g., telecommunications) or restrictions on foreign (or private) ownership.

In more general terms, even if liberalization proceeds further, in the absence of facilitating domestic conditions with respect to competition and regulation, few tangible results are to be expected apart from a simple rent redistribution process. In addition, specific commitments on commercial presence may impose substantial restrictions limiting market access and national treatment. According to Messerlin (1993), consumption quotas are the most frequently used measures for regulating services in numerous sectors. These can be limitations on production or business scope (equivalent to domestic sales requirements) for a limited range of services. Input quotas are also sometimes used, which include measures imposing domestic content requirements, technology transfer requirements or limits on foreign participation in capital. A straightforward example is provided by Cameroon’s limitation on commercial presence: 25 percent of the value of input must be supplied by services provided in Cameroon, which is a typical local-content requirement and against national treatment. In addition, fear of migration and security reasons may be used in the future to restrict movements of persons (mode 4), especially in industrial countries.

4. Potential or existing concerns for Fund policy advice

As mentioned before, the GATS is so flexible that it is likely to have limited implications for many developing countries. Therefore, it is unlikely to set many policy constraints for Fund advice. In some cases, there might be inconsistencies. For instance, in financial services, Romania committed itself to open mode 1 (cross-border) in lending and deposit taking without restrictions, which requires free capital transfers in these operations. At the same time, it maintains both current and capital account restrictions. Similarly, Sierra Leone provides unrestricted cross-border trade in many banking services, but maintains capital account restrictions. As Romania’s capital account restrictions are not maintained on prudential grounds, Romania may have to consult with the WTO’s BOP Committee to maintain them if they affect its commitments. This could justify a check on countries’ restrictions in payments and transfers and their consistency with GATS commitments.

Fund approved current account restrictions are automatically exempt: for instance, free cross-border trade in repair services is subject to any current account restrictions approved by the Fund. The Fund’s role is less clear with capital account restrictions. Capital controls can be introduced either at the request of the Fund, 1/ or if consultations are held with the WTO’s BOP Committee. This could mean that the Committee could decide alone on some capital account restrictions in case of BOP or external financial problems in a given country; it would do so presumably on the basis of the Fund’s assessment on the country’s BOP and external financial situation. A country could also introduce capital controls on its own under prudential regulation.

Fund advice could also help in increasing gains from trade in services by focusing attention to accompanying policies, such as competition policies, and to related areas not incorporated in the present GATS, such as monopoly regulation, tax regimes, and labor codes. Along these lines, there is also room for useful complementary policies in the area of privatization. In many countries with large state-owned sectors, liberalization commitments may simply not affect state enterprises. In the Czech Republic for instance, the state monopoly over motor vehicle insurance is exempt from liberalization. Privatization may be a first step towards liberalization. Foreign partnership can be an important complement to foster competition.