INTRODUCTION

1. This paper reviews the current market access situation in agricultural markets at the global level for the post-Uruguay Round (UR) period. It identifies those commodity groups in the food and agricultural sector for which tariff peaks exist, and examines how various tariff cutting formulae would affect those tariffs. The related issues of tariff escalation, tariff complexity, and the administration of Tariff Rate Quotas (TRQs) are also briefly reviewed. The main objective of this paper is to investigate the implications of alternative tariff-reduction options for the tariff structure of developed and developing countries for selected commodity groups. No attempt has been made, at this stage, to provide a detailed study of the significance of these tariff cuts for agricultural markets or trade flows.

2. The study uses data available from the Agricultural Market Access Database (AMAD)¹, and covers the 1995-98 period. For some countries, where available, 1999 data are also taken into account. All data in the database are drawn from publicly available sources. These sources are presented in Annex Table 1.

3. The countries included in the AMAD database are members of the World Trade Organization (WTO) that scheduled TRQs and those additional members that listed tariff commitments on a tariff line basis in their Uruguay Round schedules. Currently, there are 46 WTO members included in the AMAD database.² New entrants to the WTO with schedules meeting the above criteria will be included as the database is updated. The present study is based on data for the 46 countries in the AMAD database, which together account for around 81 percent of the value of world agricultural trade.

ANALYSIS OF TARIFF PEAKS IN AGRICULTURAL MARKETS

4. Tariff peaks are defined here as rates above 20 percent ad valorem³. In order to identify the incidence of peak tariffs, the AMAD data on Most Favoured Nation (MFN) applied and bound tariffs were reviewed. Only those tariff lines with rates above 20 percent ad valorem or in ad valorem equivalent were taken into account to compute the average applied and bound tariff peaks for the following commodity groupings: bovine products, ovine products, pork products, poultry products, other meat products, dairy products, coarse grains, rice, wheat, oilmeals, oilseeds, vegetable oils, sugar, fruits and vegetables, cocoa, coffee, tea, tobacco, cotton, hard fibres and hides and skins. Each commodity grouping includes both primary and processed products (e.g. cocoa includes cocoa beans, cocoa butter, cocoa powder and chocolate).

5. A major problem in the process of tariff peaks identification stems from the complexity of the tariffs. A substantial proportion of peak tariffs are specific or combined tariffs, particularly in the developed countries⁴, and, accordingly, ad valorem equivalents had to be estimated. The ad valorem equivalents of specific and combined tariffs were calculated using national import unit values derived from the AMAD database at the level of 6 to 11 digits in the Harmonised System (HS) classification. For missing values (for example, because of a prohibitive tariff) national import unit values by basic items at the 6 digit level or world import unit values were used as proxies. A detailed specification of the formulae used to compute ad valorem equivalents is presented in Annex Table 2.

6. Table 1 shows that high bound and applied tariffs are still common in global agricultural markets, even after the implementation of the UR Agreement on Agriculture commitments. Moreover, the analysis shows that the frequency of heavily protected tariff lines, bound and applied, is still significant for many commodity groupings. In the developed economies the highest tariff peaks are in the oilseeds, dairy, meat and wheat sectors. In the case of developing countries, the most prominent peak levels are in the meat, oilseeds and coarse grains sectors. It should be recalled, however, that this analysis on the basis of MFN tariffs does not account for the preferential market access that is frequently extended to developing or least developed countries.

7. For the countries in the AMAD database, the developed countries have more tariff peaks in each commodity grouping and higher average peak tariffs than the developing economies. Applied peak levels are much higher in the developed than in the developing countries for most commodity groups. Exceptions are found for oilmeals, coffee, tea, cotton, and hides and skins where the developing countries have higher applied peaks. Bound peak levels also tend to be higher in the developed countries except for oilmeals, tobacco, cotton, hard fibres and hides and skins.

8. In general, applied peak tariffs are significantly lower than bound peak tariffs for developed and developing countries. It should also be noted that, for some commodities, the majority of trade takes place under relatively low applied tariff levels. A detailed examination of trade conditions for each commodity grouping is beyond the scope of this paper, however, since the study tries to highlight tariff peak levels in the different commodity groupings that could be subject to reductions in further trade negotiations.

9. The tariffication process under the UR, which disciplined the use of non-tariff measures, often resulted in higher bound tariffs than prevailed before the Round. This fact has allowed for the persistence and even the increase of peak applied tariffs. In the developed economies, the application of the special agricultural safeguard clause has also led to additional duties in the beef, sugar, poultry and dairy sectors⁵.

10. As noted above, developed countries tend to have a higher number of applied tariff peaks per commodity group than developing countries. For developed countries, the number of applied tariff lines above the 20 percent level is highest in the following commodity groups: fruits and vegetables, dairy, meat, coffee, cocoa and tea. In the developing countries the frequency of applied peak tariffs is highest for the fruits and vegetables, meat, vegetables oils, coffee, cocoa and tea sectors. If compared with the developed countries, the frequency of applied peak tariffs in the developing economies is higher for coarse grains, oilmeals, vegetable oils, fruits and vegetables, coffee, tea and tobacco.

ANALYSIS OF FURTHER TRADE EXPANSION OPTIONS

11. This section examines the potential effects of applying different tariff cutting formulae to the identified bound peak levels. These tariff cutting formulae were selected on the basis of tariff reduction proposals that were presented in the previous rounds of GATT negotiations, and that have already been compiled and analysed in a systematic way⁶. Although not all of these formulae have been put into practice, some of them, such as the Swiss formula, have been subject to intensive debate in academic and professional circles during the post UR period. The formulae, which are defined in Table 2, include the following: the average linear tariff cuts agreed in the UR Agreement on Agriculture (36 percent for developed countries and 24 percent for developing countries); the 50 percent linear cut applied to industrial goods in the Kennedy Round; four formulae proposed during the Tokyo Round; and the compromise Swiss formula finally adopted in that Round.⁷

12. Since some formulae would yield different results depending on whether they were applied to average commodity group tariffs or to individual tariff lines, it was decided to apply each tariff cutting formula to every distinct bound tariff line. The resulting rates were then aggregated according to the different commodity groupings.

13. Figures are given in Table 3 for the tariff rates that would result after the application of each of the formulae. Table 4 shows the effect of the different formulae on the dispersion of tariffs within each commodity category. Perhaps the most notable finding is that a repetition of the cuts implemented under the UR would leave significant tariff peaks remaining in all commodity groups, both for the developed and for the developing countries.

14. The application of more aggressive linear cuts, such as those proposed in the Kennedy and the Tokyo Rounds, would lead to lower protection levels, although significant bound tariff peaks could still persist in most agricultural sectors. The only formulae that would eliminate all tariff peaks are the iterative cuts formula and the Swiss formula. This suggests that linear cuts, even at higher levels than those implemented in the UR, would not eliminate tariff peaks. Only the application of harmonizing formulae, such as the Swiss formula, would effectively reduce tariff peaks, both for the developed and the developing countries. However, the results of the application of the Swiss formula depend on the formula coefficient, which would need to be defined. In this study the coefficient was set at 0.16, which means that all tariffs above 16 percent are driven below this level. In any case, this example illustrates that such a formula is well suited for eliminating tariff peaks and compresses tariffs within a narrow range without necessarily imposing deep cuts in tariffs below the formula coefficient.

THE PROBLEM OF TARIFF ESCALATION

15. Some tariff peaks may represent the related problem of tariff escalation. Tariff escalation occurs when higher tariffs are levied on products resulting from higher stages of processing. Tariff wedges between processed commodities and their corresponding primary commodities protect domestic processing industries, allowing them to compete - perhaps unfairly - with exporters. Escalating tariffs are regarded as one of the obstacles facing developing countries in their efforts to establish processing industries for their primary exports.

16. In a previous study⁸ FAO assessed the changes in tariff escalation resulting from the UR tariff concessions, examining the agricultural import markets of the EU, Japan and the United States, which together account for 45 percent of world imports of processed agricultural products. After a detailed comparison of base tariffs with the bound tariffs resulting from the Agreement on Agriculture commitments for an extensive range of commodity pairs, it was concluded that more than 80 percent of the scheduled tariff wedges between processed and primary products had decreased as a result of the UR, creating thus additional opportunities for developing countries to diversify their exports into higher value-added processed products. However, after full implementation of the UR commitments more than 50 percent of the commodity pairs examined would still have bound escalating tariffs, with an average nominal tariff wedge of 17 percent. Therefore, tariff escalation still represents a significant problem for those countries trying to diversify their exports. In the selected markets, the highest post-UR bound tariff escalation was present in the dairy, sugar, fruits and vegetables, tobacco and hides and skins sectors.

17. Although a detailed analysis of tariff escalation by commodity pairs is beyond the scope of the present study, the problem of tariff escalation was further examined with a focus on those aggregated commodity groupings that presented applied tariff peaks. Table 5 summarizes the applied tariff peak levels presented in Table 1 classified according to the level of processing. The figures indicate that applied tariff peaks levels are in general higher for processed than for primary commodities, both in the developed and in the developing countries.

18. These findings seem to confirm that tariff peaks are more prominent for processed products than for primary products, and suggest that tariff escalation may also be present when applied tariffs are taken into account. Tariffs peaks for processed commodities are particularly high in the developed economies, although it should be noted that preferential tariffs for certain developing and least developed countries reduce the degree of tariff escalation they face. There are indeed other obstacles to food exports from developing countries, which can be particularly relevant when it comes to processed products. Nevertheless, tariff escalation can have a significant impact on developing countries, by limiting opportunities for vertical diversification of their food industry exports.

OTHER ISSUES

19. Two other important issues concerning the post-UR market access situation are briefly examined here: tariff complexity and the administration of TRQs. Ad valorem tariffs are more transparent than complex tariffs (comprised of specific and ad valorem components) or TRQs. The prevalence of non ad valorem tariffs makes it more difficult to aggregate tariffs along commodity groupings and to arrive at inter-country comparisons. In addition, the presence of specific or complex tariffs makes it more difficult to monitor the real degree of protection over time. Since specific duties are imposed on each unit of import, the ad valorem equivalent varies inversely with the import price. A decline in the price yields an increase in the level of protection. As a result, the degree of protection associated with specific and complex tariffs is not easy to gauge and can change independent of the tariff. Table 6 provides information on the share of non ad valorem tariff peaks in developed and developing countries.

20. TRQs were intended either to preserve current access (often under preferential arrangements) or to open up minimum access opportunities after tariffication. Minimum access TRQs were to be opened on a Most Favoured Nation basis. Thirty-six WTO members have tariff quota commitments in their Schedules, with a total of 1 370 individual quotas for agricultural products. Of these, 59 percent were established in the developed countries. As Table 7 indicates, the largest number of TRQs occurs in the fruits and vegetables, meat, cereals and dairy sectors. The total volume of the scheduled TRQs in 1995 as a percentage of world trade in the products concerned ranged typically from 3 percent to 7 percent. For product groups such as dairy, meat and sugar, it exceeded 10 percent.

21. The issue of TRQ administration is of particular relevance for these products. TRQs could lead to new trading opportunities for all WTO members, particularly if they are administered in a transparent way, such as a "first-come, first-served" basis, an auction basis or an "applied tariffs" basis (imports are allowed in unlimited quantities at the in-quota tariff rate or below). According to a recent WTO study⁹, 62 percent of the existing TRQs are administered following the above mentioned methods. However, a significant proportion of TRQs are administered following less transparent options, such as allocation to historical importers, state trading entities, producer groups or associations, or mixed allocation methods. In any case, Table 8 shows that fill rates tend to be higher for the latter administration methods as compared with the former. Further research is, therefore, needed to clarify whether other constraining trade restrictions (e.g. higher in-quota tariffs) are responsible for the lower fill rates.

CONCLUSIONS

22. After examining agricultural tariff peaks in the developed and the developing countries, this study has arrived to the following conclusions: First, despite the implementation of the UR commitments, high average tariff peaks persist in global agricultural markets, and the incidence of peak tariffs is still significant for many commodity groupings. The incidence of peak tariffs is also more prominent in the developed than in the developing economies both in terms of frequency and average levels. Some of these tariff peaks could have substantial effects on trade flows, although the analysis of the actual significance of their impact is left as a subject for further research.

23. Second, if the level of tariff cuts implemented under the UR were to be repeated, tariff peaks would remain in most commodity groups. The application of more aggressive linear cuts would lead to lower protection levels, but would not eliminate tariff peaks. Only a harmonizing tariff cutting formula would eliminate tariff peaks and compress tariffs within a narrow range, without imposing deep cuts into tariffs that are already at low levels.

24. Third, although tariff escalation has decreased as a result of the Uruguay Round cuts, many commodity groups still have significant escalating tariffs. Tariffs peaks for processed commodities represent a substantial problem for those developing countries trying to diversify their exports.

25. Fourth, a significant percentage of tariffs rates are specific or combined tariffs, particularly in the developed countries. Converting all tariffs into ad valorem rates would improve transparency, as would the shift to administration of TRQs on an applied tariff, auctioning or a "first-come, first-served" basis.

26. The Committee may wish to indicate its views with regard to these conclusions and whether, drawing on the experience of Members, it considers that further analysis should be undertaken to obtain a better indication of the tariff structures affecting major trade flows, on a commodity basis and for selected groups of countries.

Table 1. Tariff Peaks in selected Developed and Developing Countries, 1995-98 (percentages)

Table 2: Tariff Cutting Formulae

Table 3. Resulting tariffs after the Application of Tariff Cutting Formulae to Selected Commodity Groups with Peak Tariffs in the Developed and Developing Countries (average percentage)

Table 4. Effects of Alternative Tariff Cutting Formulae on the Dispersion ^1/ of Tariff Peaks in the Developed Countries Group

Table 5. Applied Tariff Peaks for Primary and Processed Commodities, 1995-98

Table 6. Percentage of Non-Ad Valorem Tariff Peaks in Relation to the Total Number of Tariff Peak Lines in the Developed and Developing Countries for Selected Commodity Groups, 1995-98

Table 7. Percentage of Tariff Quotas by Product Category for all WTO members

Table 8. Tariff Quotas by Principal Administration Methods for All WTO members,1995-99

1 AMAD is an inter-agency cooperative effort among Agriculture and Agri-Food Canada; EU Commission, DG Agriculture; OECD Directorate for Food, Agriculture and Fisheries; UNCTAD, TRAINS Database unit; USDA, Economic Research Service; and FAO, Commodities and Trade Division. The AMAD database is publicly accessible through the Internet at http://www.amad.net, and includes information on tariff protection and market access conditions for agricultural products.

2 Argentina, Australia, Barbados, Brazil, Canada, Chile, Colombia, Costa Rica, Czech Republic, EC-15, Ecuador, Egypt, El Salvador, Guatemala, Hungary, Iceland, India, Indonesia, Israel, Japan, Malaysia, Mexico, Morocco, Namibia, New Zealand, Nicaragua, Norway, Pakistan, Panama, Paraguay, Philippines, Poland, Republic of Korea, Romania, Singapore, Slovakia, Slovenia, South Africa, Swaziland, Switzerland, Thailand, Tunisia, Turkey, Uruguay, United States of America and Venezuela.

3 There is no standard definition of tariff peaks. Given the large price variations in agricultural products, 20 percent could be regarded as a low threshold, however an earlier WTO/UNCTAD study used 12 percent as a threshold level. No account is taken of the possibility that a low bound tariff plus an associated special safeguard duty could also result in a higher tariff than the peak as defined here. See The Post-Uruguay Round Tariff Environment for Developing Country Exports: Tariff Peaks and Tariff Escalation, (TD/B/COM.1/14/Rev.1), January 2000, UNCTAD, Geneva.

4 The definition of developed countries used here is that of the United Nations.

5 The Post-Uruguay Round Tariff Environment for Developing Country Exports: Tariff Peaks and Tariff Escalation, (TD/B/COM.1/14/Rev.1), January 2000, UNCTAD, Geneva.

6 See for example Evaluating Alternative Formulas for Reducing Agricultural Tariffs, J. Wainio, P. Gibson, and D. Whitley, Economic Research Service, USDA, Washington DC, 2000; and Multilateral Approaches to Market Access Negotiations, S. Laird, in M. Rodriguez, P. Low and B. Kotschwar, eds. Trade Rules in the Making Challenges in Regional and Multilateral Negotiation, Organization of American States, Brookings Institution Press, Washington DC, 1999.

7 Although the AoA gives some discretionality to member countries to distribute the tariff cuts among different tariff lines (provided that a 15 percent minimum cut is applied for each line), the analysis assumes that a 36 percent or a 24 percent linear cut would take place for each tariff line.

8 The Impact of the Uruguay Round on Tariff Escalation in Agricultural Products, 1997. J. Lindland, Commodities and Trade Division. FAO, Rome.

9 Tariff Quota Administration Methods and Tariff Quota Fill", Committee on Agriculture Special Session (G/AG/NG/S/8), May 2000, WTO, Geneva.

Annex Table 1. Principal Sources of Data used in AMAD

Annex Table 2. Formulae for converting specific tariffs to their ad valorem equivalents