Table of Contents

I. INTRODUCTION

II. THE NEW BANANA IMPORT REGIME IN THE EC (AGREEMENT OF APRIL 2001)

A. MAIN PROVISIONS OF STEP I

B. STEP II

III. A QUANTITATIVE ASSESSMENT OF THE TRANSITIONAL TARIFF-RATE QUOTA REGIME

A. OVERVIEW OF RESULTS

B. EC MARKET SUPPLYING

C. ACP AND DOLLAR ZONE COUNTRY EXPORTS

D. QUOTA A/B RENT AND TARIFF EQUIVALENT

IV. GENERAL CONCLUSIONS

ANNEX I - BANANA SUPPLY IN THE EUROPEAN COMMUNITY (THOUSAND TONNES)

ANNEX II

I. INTRODUCTION

1. Along with the United States of America and Japan, the European Community (EC) is among the three most important import markets for export bananas, accounting for nearly 30 percent of global imports. Therefore, since the original Common Market Organization for Bananas (CMOB), laid down in EC Council Resolution 404/93 of 13 February 1993, the Group has closely followed EC policy changes to assess possible future market developments. In order to assist the Group in analyzing the latest changes in policy the following paper, based on an econometric model developed at the Institut National de la Recherche Agronomique (INRA), has been prepared to provide the Group with some analytic background regarding the most likely effects of the EC import regime as applied from 1 July 2001. A summary of developments in the EC banana import policy since 1993 is provided in document CCP: BA/TF 01/CRS.9.

2. Many other importing countries maintain various forms of banana import regimes. In order to put other import regimes into perspective for the Group, CCP: BA/TF 01/08 provides a summary of those import policies.

II. THE NEW BANANA IMPORT REGIME IN THE EC (AGREEMENT OF APRIL 2001)

3. The new banana import regime in the EC is a two-step process towards a tariff-only regime that should enter into force no later than 1 January 2006. During the transition period 2001-05, bananas will be imported into the EC under a tariff-rate quota system through import licenses distributed on the basis of past trade. This transitional tariff-rate quota system became effective from 1 July 2001, and is itself divided into two phases (phase I through 31 December 2001 and phase II). This document presents results for the year 2005 that corresponds to the last year of the transitional tariff-rate quota system, prior to the entry into force of the tariff-only system in 2006. Table 1 gives an overview of the new trade regime.

A. MAIN PROVISIONS OF STEP I

4. In step I, phase I, a bound tariff-rate quota (quota A) and an autonomous tariff-rate quota (quota B) are set at respectively 2.2 million tonnes and 353 000 tonnes. The tariff-rate quotas A and B are managed as one (quota A/B) and are not allocated among country suppliers. The tariff applied to banana imports within the quota A/B is 75 Euro per tonne with a tariff preference of 75 Euro per tonne granted to ACP bananas. Corresponding import licenses are divided between traditional operators A/B (83 percent of quota) and non-traditional operators A/B (17 percent). An additional tariff-rate quota of 850 000 tonnes (quota C) has been introduced which is open to all country suppliers. The tariff applied within quota C is 300 Euro per tonne with a tariff preference of 300 Euro per tonne granted to ACP bananas.

5. The import licensing system is still largely managed on the basis of historical references. As indicated above, 83 percent of the A/B quota is allocated to traditional operators. However the definition of traditional and non-traditional operators has changed with respect to previous regulations.¹

6. The most significant change for traditional importers was the use of 1994-96 as the base reference period for imports and the elimination of category B licenses. The change for non-traditional importers is the use of the 1994-96 reference period plus the need to have imported 1.2 million Euro or more during those years in order to qualify for licenses.²

7. In step I, phase II, the provisions applying to phase I will continue except the following changes. First, the autonomous tariff-rate quota B will be set at 453 000 tonnes (an increase of 100 000 tonnes with respect to phase I). Second, the additional tariff-rate quota C will be set at 750 000 tonnes (a decrease of 100 000 tonnes with respect to phase I) and it will be reserved for bananas of ACP origin.

8. The share of import licenses to traditional operators A/B will continue to be allocated following the procedure defined for phase I. Import licenses should thus be distributed through 31 December 2003 based on the 1994-96 reference volume for each qualified traditional operator. Thereafter the share of import licenses to traditional operators A/B would be allocated based only on usage of licenses issued since the beginning of phase II. This second stage should be effective "as soon as possible", subject to Council and European Parliament approval and to adoption of the waiver of Article XIII of the GATT 1994 needed for the management of the autonomous tariff-rate quota C.

B. STEP II

9. In step II, the transitional tariff-rate quota system described above will be replaced by a definitive tariff-only system, effective no later than 1 January 2006.

III. A QUANTITATIVE ASSESSMENT OF THE TRANSITIONAL TARIFF-RATE QUOTA REGIME

10. This section presents the results of simulations of the introduction of the regime based on an econometric model of the world banana market. A brief outline of the model is presented as Annex 2 to this paper.

A. OVERVIEW OF RESULTS

11. Analysis with the model shows that the quota A/B should be binding. It also shows that a tariff preference of 75 Euro per tonne granted to ACP bananas within the quota A/B would not be sufficient to allow supplies from this origin to compete with dollar zone bananas within the quota A/B (assuming that the quota A/B is open, which seems to be the case in phase I but not in phase II, to all country suppliers). Finally, the model shows that the quota C tariff of 300 Euro per tonne would be too high to allow dollar banana imports within the quota C (assuming that the quota C is open, which seems to be the case in phase I but not in phase II, to bananas of all origins with a tariff preference of 300 Euro per tonne for ACP bananas). Once the quota A/B is filled, it would be more profitable for dollar zone suppliers to export on Rest of World (RoW) markets than to incur a levy of 300 Euro per tonne within the quota C. Over-quota imports would be zero as over-quota tariffs are prohibitive (assuming that over-quota tariffs are 680 Euro per tonne for non-ACP countries and 380 Euro per tonne for ACP countries).

B. EC MARKET SUPPLYING

12. Simulation results suggest that the group of ACP countries would not be able to exhaust a quota C level set at 850 000 tonnes as in phase I. However the fill rate would be 100 percent for a quota C level set at 750 000 tonnes as in phase II (Table 2).

13. In scenario S-II which corresponds to an application of phase II assumed from January 2002, exports of ACP countries to the EC would be equal to 760 523 tonnes in 2005, i.e. 8.4 percent above the base period 1996-98 levels.³ Internal EC production would increase by 6.2 percent from 750 671 tonnes in 1996-98 to 797 090 in 2005. Exports of dollar zone suppliers to the EC would be equal to the binding level of the quota A/B (2.653 million tonnes). This represents an increase of 239 400 tonnes (9.9 percent) with respect to base period 1996-98 quantities (2.414 million tonnes), but an increase of only 140 000 tonnes (5.6 percent) with respect to the 1999 level (2.513 million tonnes). Total ACP exports to the EC would represent 18.1 percent of EC consumption in 2005, i.e. the same share than in 1996-98. Internal EC production and dollar zone countries would represent 18.9 percent and 63.0 percent of the EC market. If the quota A/B were maintained at 2.553 million tonnes over the whole period 2001-05 (scenario S-I), total ACP exports to the EC would be greater (787 212 tonnes in 2005). In that case, the share of dollar zone countries would be smaller (61.7 percent) and the shares of ACP countries and EC internal production would be greater (respectively 19.0 percent and 19.3 percent).

14. Columns 4 and 5 of Table 2 illustrate the sensitivity of simulation results to the size of the tariff-rate quota A/B. An increase in the level of the quota A/B, say from 2.553 million tonnes to 2.653 million tonnes, leads to a decrease in EC imports from ACP countries. The decrease in ACP exports to the EC is lower than the quota A/B increase so that total consumption in the EC increases and average CIF prices in the EC decrease. Internal EC production does not change when the size of the quota A/B varies because of the income support scheme. Since the original CMOB in 1993, EC producers are guaranteed a minimum income through a deficiency payment applicable to a volume of up to 854 000 tonnes. In this paper, we assume that the sum of the FOB price and of the direct aid is constant so that the "effective" price taken into account by EC producers in their profit-maximizing programme does not change when the quota A/B level varies.

C. ACP AND DOLLAR ZONE COUNTRY EXPORTS

15. Distribution of export changes varies significantly among ACP countries (Table 2). In scenario S-II, exports of Côte d'Ivoire and Cameroon to the EC would increase by respectively 22 percent (from 168 410 tonnes in 1996-98 to 205 466 tonnes in 2005) and 17 percent (from 146 490 tonnes in 1996-98 from 171 374 tonnes in 2005). EC imports from the Windward Islands and Jamaica would decrease by 5.7 percent from 230 953 tonnes to 217 802 tonnes. In the same way, Table 3 shows distribution of total export changes among dollar zone suppliers. In scenario S-II, Ecuador would export 4.841 million tonnes in 2005, i.e. about 793 000 tonnes, or 19.6 percent, more than in the base period. The world's largest exporter would profit from increased access to the EC market thanks to a quota A/B level set a 2.653 million tonnes as well as an increased demand in RoW (Rest of the World) markets, in particular in the United States. Exports of other Central American countries would increase by about 7.6 percent for Costa Rica and Guatemala, and by about 12 percent for Panama and Honduras.

D. QUOTA A/B RENT AND TARIFF EQUIVALENT

16. Scenario S-II would lead the average CIF price in the EC to decrease by about 30 Euro per tonne, from 593 Euro in 1996-98 to 560 Euro in 2005. It would lead the average FOB price in dollar zone countries to decrease by about US$20 per tonne, from US$292 in 1996-98 to US$272 in 2005. The quota rent (the difference between the world price and the higher EC price arising from the existence of the trade restriction) on dollar zone banana imports in the EC would decrease from 239 Euro per tonne in 1996-98 to 182 Euro per tonne in 2005.⁴

17. Although quota A/B rent varies with the Euro/US$ exchange rate, it is sensitive to policy and exogenous variables as well. An increase in EC demand should increase rents as would increases in dollar zone supplies. Similarly, a lower A/B quota would raise A/B rents (see Table 4 for effects of demand and supply shifters).

18. According to Table 4, a tariff equivalent of about 182 Euro per tonne in 2006 with a tariff preference of the same amount granted to ACP bananas would be "equivalent" to the transitional tariff-rate quota system of phase II (scenario S-II). A tariff equivalent of that order of magnitude would keep the average CIF price in the EC at its 2005 level (about 560 Euro per tonne). It would leave dollar zone exports to the EC and those from producer and ACP countries largely unaffected in 2006 (relative to scenario S-II results).

19. Of course, this result is contingent upon parameter choice and policy assumptions in the analysis. In particular, it appears that a tariff of that order of magnitude would be underestimated (i.e. less than "equivalent") if the Euro weakens vis-à-vis the US dollar (relative to a parity of 1 in 2005 as in scenario S-I and S-II). Furthermore the "equivalence" applies strictly for the year 2006 alone. As productivity and production increases are likely to be greater in the dollar zone than in ACP countries, if EC policy makers were to decide to avoid further competitive shifts in supply based on productivity changes the model implies that the tariff equivalent would have to be increased to permit ACP suppliers to maintain a long-term market share comparable to that of 2005.

20. Obviously the analysis incorporates certain simplifying assumptions, and the empirical results are therefore subject to several caveats. In particular, a pertinent and complete analysis of the distribution of the quota A/B rent needs careful modelling of all the operators involved in the banana industry and of all the aspects of the market structure, including operator strategies and expectations.⁵ Unfortunately no data are available to perform such a modelling exercise since only country data are available. This is an important point because the new licensing import scheme is essentially a system of company quotas.

IV. GENERAL CONCLUSIONS

21. From a country point a view, the transitional import regime in the EC may largely be viewed as a continuation of a managed market with two tariff-rate quotas and an import licensing system based on past trade. The choice of the 1994-96 reference period does not necessarily take into account the dynamic changes and investments that have taken place since that period by many operators. From Table 5 and despite many uncertainties about data, it clearly appears that the EC and world market shares of one multinational company (Chiquita Brands International) have substantially decreased since 1992 while another (Dole Food Company) has significantly increased its EC and world market shares. Changing market shares are obviously difficult to evaluate and may reflect not only variations in investment activities, but may instead be attributable to other causes, i.e. outbreaks of banana disease, bad weather, strikes by workers and shipping and operating disruptions. Furthermore, taking 1992 as reference point may be somewhat misleading insofar as banana exporters from the dollar zone began to increase their shipments to the EC in the years immediately preceding the original 1993 CMOB, in anticipation of the new trade regime to come.

22. Welcoming the agreement reached in April 2001, both the United States and the EC recognized that they had shared objectives, i.e. "to reach an agreement on a WTO-compliant system, to ensure fair and satisfactory access to the European market for bananas of all origins and all operators, and to protect the vulnerable ACP producers." Simulation results suggest it will be crucial for these countries to enhance their competitiveness during the transitional tariff-rate quota period in anticipation of the tariff-only system.

23. Since the original CMOB in 1993, EC producing regions have benefited from income support in the form of direct aids. In the scenarios presented in section III., it was assumed that the effective price (FOB price per tonne plus direct aid per tonne) considered by EC producers remained unchanged at its 1996-98 level. In scenario S-II, EC territories would supply around 797 000 tonnes (+ 46 419 tonnes with respect to 1996-98) at a price 30 Euro lower than the average 1996-98 price, implying extra compensation of about 24 million Euro relative to 1996-98. This corresponds to an extra compensation of around 30 Euro per tonne of bananas.⁶ With respect to 1996-98, EC producers would also lose from the suppression of the category B of operators. The growth rate of EC territory supply is positive over the eight-year period 1993-2000 (Annex 1).⁷ It is likely that this favourable trend is a consequence of the income support scheme. Lack of reliable data prevents evaluation of the extent to which EC producing regions have used this income support scheme to reduce unit production costs and improve their competitiveness.

24. Looked at from a supplier point of view, the econometric model implies that the interim EC regime is unlikely to completely safeguard the market shares of Caribbean suppliers, as their share is expected to decrease, at least by about five to fifteen thousand tonnes between 2001 and 2005 (Table 2). Compared to the base period West African producing countries could be expected to benefit. However, the level of quota C could limit West African growth. Furthermore, as their historical base is small, licenses would have to be purchased to enter additional fruit.

25. There are no certainties that either phase II will enter into force on 1 January 2002 or a tariff-only regime will begin on 1 January 2006. The setting of the appropriate tariff is likely to be a point of considerable discussion up to that later date. In light of the importance of that tariff level to world banana trade the Sub-Group may wish that the Secretariat undertake further modelling analysis on this concept.

Table 1 - The new banana trade regime in the European Community
(agreements of April 2001)

Sources: European Commission, Commission Regulation (EC) 896/2001 of 8 May 2001; FruiTrop, n° 79, April 2001, page 6; WTO, 2001, WT/DS27/58, 2 July 2001.

Table 2 - Banana imports in the European Community (tonnes)

⁽¹⁾ Source: ODEADOM (Office de développement de l'économie agricole des départements d'outre-mer (2001) and FAO (2001)
⁽²⁾ In scenario S-II, ACP country exports to the EC are slightly greater than the quota C level of 750 000 tonnes (+ 10 500 tonnes). There is no significant change in the results if we constrain ACP country exports to be strictly equal to the quota C level

Table 3 - Total banana exports by dollar zone countries, Central and South America (tonnes)

⁽¹⁾ Source: FAO (2001)

Table 4 - Sensitivity of the quota A/B rent to policy parameters and/or exogenous variables

⁽¹⁾ assumes that the Euro/US$ exchange rate remains at 0.85 (0.85 Euro = US$1) instead of 1
⁽²⁾ assumes in addition to ⁽¹⁾ that demand shifters in the EC are multiplied by 2
⁽³⁾ assumes in addition to ⁽²⁾ that supply shifters in the dollar zone are increased by 2 percent per year

Table 5 - World and EC market shares of banana companies

Sources: ⁽¹⁾ Ledemé F. quoted in FruiTrop, n° 62, October 1999
                ⁽²⁾ Van de Kasteele A., February 1998, from various sources

ANNEX I - BANANA SUPPLY IN THE EUROPEAN COMMUNITY (thousand tonnes) ⁽¹⁾

Source: European Commission - 2000 CD Eurostat 5-2001

⁽¹⁾ From 1991 to 1994 - 12 EC; from 1995 - 15 EC

ANNEX II

The simulations presented in this document are carried out with an updated version of a single-commodity, multi-country partial equilibrium model of the world banana market. The model is detailed in Guyomard et al. (1997, 1999a, 1999b). In summary, it consists in demand (import) and supply (export) equations written as constant-elasticity functions. Import CIF prices in importing zones and export FOB prices in exporting zones are linked by transportation costs and constant-margin equations. Market-clearing equations guarantee the supply-demand equilibrium on "relevant" markets. At this stage, it is worthwhile to note that the number of market-clearing equations depends closely on the EC import policy. For example, in the case of the 1993 CMOB and the 1994 FAB with a binding MFN quota, two market-clearing equations had to be specified, one for the EC market to determine CIF prices and in the EC as well as FOB prices in EC territories and ACP countries, one for the Rest of the World (RoW) to determine CIF and FOB prices on RoW import-export markets. Value and volume of bilateral trade flows (i.e. exports of supplier j to importer i and imports of purchaser i from exporter j) are based on EUROSTAT (COMEXT) and FAO (FAOSTAT) data. FOB and CIF unit values are derived from these volume and value data. Base period data used for model initialization and calibration correspond to the 1996-98 average.

Supply (export) and demand (import) functions include time shifters. Growth trends of supply and demand were estimated from data over the past 15 years. These growth trends were then purified from price-trend impacts assuming independence between price and time effects. This assumption implies that policy changes have no effect on the magnitude of supply and demand shifters.

References for the model

Guyomard H., Mahé L.-P., Tavéra C., Trochet T., 1991, Technical change and EC-US agricultural trade liberalisation. Journal of Agricultural Economics, 42(2), 119-137.

Guyomard H., Herrard N., Laroche C., Le Mouël C, 1997, Analyse théorique et empirique du fonctionnement de l'Organisation Commune de Marché de la banane dans l'Union européenne. INRA-ESR, Rennes, Rapport pour la Commission européenne, Direction Générale VI, Bruxelles.

Guyomard H., Laroche C., Le Mouël C., 1999a, An economic assessment of the Common Market Organization for bananas in the European Union. Agricultural Economics, 20, 105-120.

Guyomard H., Laroche C., Le Mouël C., 1999b, Impacts of the Common Market Organization for bananas on European Union markets, international trade and welfare. Journal of Policy Modeling, 21(5), 619-631.

_______________________________

1 Traditional operators are "economic agents [...] established in the Community during the period for determining their reference quantities who have, for their own account, purchased a minimum quantity of bananas originating in third countries from the producers or, where applicable, produced, consigned and sold such products in the Community." Operations defined above are called primary imports. "Traditional operators A/B means traditional operators who have carried out the minimum quantity of primary imports of third country bananas and/or non-traditional ACP bananas in accordance with the definitions in Article 16 of Regulation (EEC) 404/93, as amended by Regulation (EC) 1637/98. Traditional operators C means traditional operators who have carried out the minimum quantity of traditional ACP bananas in accordance with the definitions in the above mentioned Article 16, as amended by Regulation (EC) 1637/98."

2 "The reference quantity for each traditional operator A/B shall be established [...] on the basis of the average of primary imports of third-country bananas and/or non-traditional ACP bananas during 1994, 1995 and 1996 taken into account for 1998 for the purposes of administering the tariff quota for imports of third-country bananas and non-traditional ACP bananas [...]. The reference quantity for each traditional importer C [...] shall be established on the basis of the average of primary imports of traditional ACP bananas during 1994, 1995 and 1996 carried out for 1998 as traditional quantities of ACP bananas [...]." Non-traditional operators means "economic agents established in the Community at the time of their registration who, (a) have been engaged independently and on their own account in the commercial activity of importing into the Community fresh bananas [...] in the two years immediately preceding theyear in respect of which registraiton is sought, (b) by virtue of this activity, have imported produce to a declared customs value of 1 200 000 Euro or more duing the period referred to in point (a) and (c) do not have a reference quantity as a traditional operator under the tariff quota for which they are applying for registration [...].

3 In scenario S-II, ACP country exports to the EC are slightly greater than the quota C level of 750 000 tonnes (+ 10 500 tonnes). There is no significant change in the results if ACP country exports are constrained to be strictly equal to the quota C level.

4 The quota rent on dollar zone banana imports in the EC is calculated as follows: average CIF price in the EC - transportation costs between the EC import market and the dollar export zone - average "normal" commercial margins - average FOB price in the dollar zone. It is worthwhile to remember that the EC import market clears in Euro while the RoW market clears (to a large extent) in US$.

5 Three multinational firms account for about 70 percent of the world import-export banana market and most national markets in the EC are dominated by a small number of firms/operators, including these three multinational firms. This suggests that the perfect competition assumption is questionable. However this does not imply automatically that the world banana import-export market is not perfectly competitive. For example, studies have shown that the German banana market cannot be characterized by the exercise of market power despite the low number of firms that compete in that market (the four-firm concentration ratio would be greater than 80 percent for Germany).

6 In 1996-98, the unit direct aid granted to EC producers was around 260 Euro per tonne (three-year weighted average). It increased significantly in 1999 (297 Euro per tonne) and 2000 (383 Euro per tonne). In part, this is due to the increase in historical reference earnings used for compensation calculation (from 592.9 Euro per tonne for the years 1993 to 1997 to 622.5 Euro per tonne in 1998 and 640.3 Euro per tonne in 1999 and 2000).

7 However it is worth noting that quantities supplied in 1991-92 (about 703 000 tonnes) were significantly higher than volumes marketed in the four first years of the CMOB (1993 to 1996).