Table of Contents

---

I. INTRODUCTION

II. BASIC FEATURES OF THE MODEL

III. ASSUMPTIONS

IV. SOME PRELIMINARY ILLUSTRATIVE RESULTS

V. LIMITATIONS OF THE MODEL

A. DATA

B. ASSUMPTIONS

VI. CONCLUSION

I. INTRODUCTION

1. This paper summarizes work in progress by the Secretariat on a model simulating international banana trade. No results are presented at this stage, as this work is still in progress and the model has not yet been validated.

II. BASIC FEATURES OF THE MODEL

2. FAO has been working with Professor Thomas Spreen of the Institute of Food and Agricultural Science, University of Florida, Gainesville, on a model simulating international banana trade between major exporting and importing regions in the world. It is intended to use this model to project international banana trade in 2010. The model is of the spatial equilibrium type and uses the General Algebraic Modelling System (GAMS) software.

3. The model attempts to determine the optimal prices and quantities in the supply and demand regions. These prices and quantities are determined where a measure of aggregate importer profit is maximized. This measure is equal to total imported value minus total exported value and transportation costs. The rationale for this type of objective function lies in the oligopolistic nature of the banana trade. Approximately 80 percent of international banana trade is controlled by a small number of large, vertically integrated, international banana marketing companies. The sourcing of bananas from the supply areas is made on the basis of prices charged in those areas or, in the case of the companies' own production, on costs of production, transportation costs and tariff policies in importing countries.

III. ASSUMPTIONS

4. The model uses the period 1997-99 as a baseline. This period was chosen in order to be consistent with the projection work undertaken by FAO for other commodities. However, for bananas this base period is not entirely satisfactory since it covers two different banana import regimes in the European Community (EC). A new regulation entered into force in 1999 but the model ignores this change and assumes that the regime in place in 1997-98 continues in 1999.

5. In terms of policy in the EC, the model considers that by 2010 the EC will have established a tariff-only import regime for bananas. The EC has repeated several times that there would be a tariff-only system by January 2006 at the latest. The model assumes that in 2010 bananas imported from countries of Africa, the Caribbean and the Pacific (ACP countries) enter duty free, whereas the tariff for bananas of other origins is Euro 200 per tonne.

6. The model assumes that the exchange rate of the Euro to the dollar will slightly increase from its current level of about 0.91 to parity in 2010.

IV. SOME PRELIMINARY ILLUSTRATIVE RESULTS

7. Preliminary results from the model indicate that under the above assumptions, world exports of bananas would increase by 18 percent from 1997-99, reaching some 13.9 million tonnes in 2010. The annual growth rate of 1.5 percent would be lower than in the period 1987-89 to 1997-99 (5 percent). Most of this growth would be in Latin American countries, in particular in Ecuador and to a lesser extent Costa Rica and Colombia. Exports from Africa and Caribbean countries would decrease, reflecting the loss of the tariff quota reserved for the ACP countries in the EC when it adopts a tariff-only system. The fall in exports would be more marked for the Caribbean countries than for African countries due to higher production costs in the former. In Asia, the Philippines is projected to increase its exports, taking advantage of the growth of the Asian market.

8. Preliminary results also indicate that under the above assumptions, world banana imports are expected to rise to some 13.6 million tonnes in 2010. The difference between global exports and imports is mainly due to loss in transit. Banana imports in developing countries would increase by nearly 50 percent though this implies a lower annual growth rate than in the period between 1987-89 and 1997-99. Developing country imports would account for approximately one-fifth of world imports. The main growth area would be Asia, in particular East and Southeast Asia due to demographic growth and rising incomes.

9. Imports into developed countries are projected to increase but, at a modest annual rate of only one percent, well below the rate of the previous decade. The largest rise in imports is projected to take place in the United States, mainly due to demographic growth coupled with high purchasing power. In the EC also, the planned transition to a tariff-only import system from 2006 is expected to raise imports. Conversely, Japanese imports are projected to rise only marginally due to demographic factors and the unchanged import regime. Growth in banana imports would also be slow in the economies in transition.

10. As indicated above, these results are very preliminary. They are presented solely as an illustration of the type of output the model will be capable of generating and should be treated with caution.

V. LIMITATIONS OF THE MODEL

A. DATA

11. A key element of the projection work is to obtain accurate data on prices (FOB and CIF), quantities (imports and exports), price elasticities (for supply and demand) and transportation costs over a long time span. Unfortunately much of the data essential to the model are not available. The data concerned include:

• FOB prices: although the Secretariat has consistently sought these data from various sources (e.g. governments, banana experts and exporter associations), the data supplied by respondents is the average unit value of exports. The Secretariat has had to use this as a proxy for FOB prices, as it was impossible to obtain FOB series on a long time scale.
• CIF prices: the same problem applies to CIF prices, for which long series were available only for the United States, Germany, France and Japan. For most importing regions the Secretariat has had to use the average unit value of imports as a proxy for the CIF price.
• Transportation costs: the data available for transportation is far from complete and costs change frequently. The matrix of transportation costs for every exporting region to every importing region had to be estimated from the few figures available.
• Price elasticities of demand and supply: as this type of data is not available or incomplete in the literature these had to be estimated econometrically. Annual imports and exports were regressed on a series of factors such as price, exchange rate and a `time' factor (encompassing a mix of variables such as population growth) using extended time series (1973 to 2000). However, the elasticities calculated for Germany and Japan seemed too low (below 0.3 in absolute value), so elasticities calculated in previous work for these countries were used.
• Non-price factors influencing exports and imports: factors that have an impact on demand in the long run include population growth, income growth and exchange rates, among others. Factors influencing supply include technological progress, diseases and climate. The last two are unpredictable. The regression analyses did not show any robust correlation between these various factors and exports or imports.

B. ASSUMPTIONS

12. The particular assumptions made undoubtedly have an impact on the projections. Obviously the level of the tariff in 2010 in the EC and of the preference given to ACP countries have a strong influence on the results. Moreover, the model assumes that there is parity between the Euro and the US dollar in 2010 but if the dollar decreases below parity then exports from Latin American countries to the EC are likely to be higher and consequently those of ACP countries lower. Also, the model simplifies the baseline period, assuming that the trade policy in force in the EC in 1994-98 was applied in 1999. This assumption might have an impact on the accuracy of the results.

VI. CONCLUSION

13. So far, the model has produced results which still need to be validated by further work. The lack of reliable data on prices, transportation costs, elasticities and non-price factors affecting supply and demand is a major constraint. Data from the Sub-Group on all these elements are needed if the Secretariat is to overcome the limitations of the model and produce robust and useful results. The Sub-Group is invited to provide the Secretariat with relevant data. It is hoped that by mid-2002 a more robust projection to 2010 can be achieved.