40956 Trade Note July 12, 2007 Economic Partnership Agreements: Does Preferential Access of Non-LDC African Countries Increase? Negotiations on Economic what the alternative to an EPA The Partnership Agreements (EPAs) could be for the non-LDCs. World Bank between the European Union and To maintain pressure to come Group the African, Caribbean, and to agreement, the European www.worldbank.org Pacific (ACP) states were started Union has not put forward any in 2000 and aim at establishing alternative offers even though reciprocal free trade agreements the Cotonou Agreement states (FTAs) to be World Trade that those countries opting Organisation (WTO) consistent. against signing an EPA should While negotiations have end up no worse than under the advanced to the third phase in current trade preferences. International most of the regions, progress has been glacial and no agreements as While the exact legal Trade of this writing have been signed. interpretation of that clause is Department Several contentious points debated, the alternative to an complicate the negotiations, EPA will lie between the two including the internal extremes of an EPA with organization in each of the six 100% product coverage and regions that are negotiating with (presumably) liberal rules of By Mombert Hoppe the EU, agreement on special and origin, and preferences being differential treatment, and the granted under the Generalised highly debated issue of whether System of preferences (GSP). related aid-for trade should be For a preferential trade included legally binding in the agreement to conform to WTO EPAs. norms, it will have to cover "substantially all trade" which These notes summarize recent research on global One key issue for non-Least is usually understood as trade issues. They reflect Developed Countries (non-LDCs) covering at least 90% of the solely the views of the author and not necessarily the views is the impact that alternatives to total current trade flows. of the World Bank Group or an EPA would have on market During the Doha Round, the its Executive Directors. access and preferences in the EU. EU put forward the offer to This is not an issue for LDCs grant quota- and duty-free because they will continue to access to its market for all benefit from quota- and duty-free LDCs for 97% of products. It access to the EU market under is unclear if an EPA would Trade Note 32 the Everything but Arms (EBA) cover all imports into the EU initiative. It is unclear, however, or replicate Cotonou-type Trade Note July 12, 2007 coverage. In early April, the EU offered increase to 4.5 percent of exports, a quota and duty free access to the notional gain of EUR 107 million. If European market for all products except EPAs were concluded with simple rules sugar and rice to EPA signatories. The of origin, however, preferences could be transition period for rice would be brief much higher because this could make while full quota and duty free access for several export products competitive in sugar would be phased in until 2015. the European market. For example, less South African exports of a range of restrictive rules of origin for apparel "globally competitive" products are products in the US under AGOA have exempted from this market access offer. led to an increase in clothing exports The proposal, however, has been from about USD 250 million in 2000 to criticized by France and some other more than USD 800 million in 2004. European states, as threatening the compromise found on the reform in the The value of preferences is the amount sugar and banana sector and as of import duties that exporters did not weakening the EU's position before the have to pay because of preferential tariff WTO with regard to the banana dispute. rates available to them. Its calculation is In any case, an EPA would also require a based on actually observed trade flows. phased market opening for EU products into the ACP countries, even though the Two caveats are in order. First, this phasing could last several years. calculation omits the fact that some of these trade flows might only take place Current EU trade preferences for non- because of the preferences. In that case, LDCs in Africa amounted to 3.9 percent the whole export value would be of their exports to the EU or EUR 782 additional value that is transferred to the million in 2005. Assuming that Cotonou exporter because preferences are preferences expire and no new granted. This tends to bias downward the preferences will be in place by that time, calculated value of the preference. 2008, it is possible that these countries would only have access to preferences Second, it is impossible to identify from under the GSP. Other things equal, this the data those who actually benefits from would reduce the value of their the preferential rates. The issue is that preferences to 0.5 percent of their tariffs create a wedge between the exports or EUR 103 million, a loss of producer price in the exporting country, EUR 679.1 including transport costs to the export market, and the consumer price in the Alternatively, if EPAs with Cotonou- importing country. In this note, we type rules of origin were to be in place assume that the exporter actually by 2008, the value of preferences could benefits as he continues to sell the goods in the destination market for the higher 1Another possibility that has been discussed consumer price but does not have to pay would be to include the non-LDCs as GSP+ the import duties. He hence receives a beneficiaries. This would increase the eligibility net transfer of the value of preferences. for preferences of their exports and could be temporarily justified under the "vulnerability" Depending on the market power of the criterion as happened in the case of the Andean importer, the exporter, or an countries. intermediary, it could be also be the 2 Trade Note July 12, 2007 importer or the intermediary who benefit value of preferences for each of the 13 from the reduction in tariff rates. If sub-Saharan non-LDC countries, taking market power of the importer is high, he into account its current preference can continue to pay the producer price to utilization.3 This allows us to correct for the exporter (keeping his income that proportion of products that currently constant as compared to the situation do not claim preferences because of without trade preferences) but sell at the problems with rules of origin, lack of higher consumer price in the importing knowledge of the available preferences, country, keeping the difference in price or other reasons.4 Ad-valorem as a profit. Similarly, the intermediary equivalents (AVEs) for specific tariffs could benefit from the lower tariff rate are estimated at a flow basis (i.e. by by buying from the producer at producer exporter, product, and preference status). prices and selling to the final reseller at We calculate the difference between consumer prices. In this situation, the duties that would have been paid if all intermediaries income increases while imports entered under MFN and duties both income for the exporter and final that would have been paid if exporters importer stay constant (as compared to could chose the best preferential rate (in the situation without tariff preferences). a small number of cases, GSP rates are This tends to bias upward the `value of lower than Cotonou rates). This preferences' because the calculated gain difference is reported as value of may overstate the actual transfer to the actually claimed tariff preferences. In a exporting country (Olarreaga and Özden, second step, we calculate the value of 2005). preferences if only GSP preferences were accessible. For comparison, we Using a new dataset the EU has recently also present the hypothetical value of made available,2 we have calculated the preferences if EPAs with duty- and quota-free access and Cotonou-type rules of origin were to be concluded (i.e. 2While an earlier version of preferential data for 2002 distinguished between different preferential setting all EU tariffs to zero and keeping schemes that were used by the importers utilization rates constant). (Brenton and Ikezuki, 2005), this is no longer the case in the current dataset. It is likely that this results from the fact that many products are eligible for preferences under various preferential schemes. When analyzing utilization rates for the different schemes, trade flows were often counted more than once in the denominator while they could only be allocated to the numerator of one of the utilization rates. This possibly led to non-representative low take-up rates for some schemes and it is likely that the EU tries to avoid this problem in the new dataset. This new dataset only contains five different 3We treat trade flows in the unknown category kinds of import categories: Imports entering as not receiving preferences. under MFN with zero rates, imports entering 4The costs of the documentation requirements to under MFN with non-zero rates, preferential proof that the rules of origin are satisfied is often imports facing zero duties, preferential imports estimated at three percent of the goods' value, facing positive (but reduced) duties, and imports casting doubt on the economic value of such whose status is unknown. small preference margins. 3 Trade Note July 12, 2007 respectively to nearly zero. For Kenya, a Figure1:Preferences as share of total shift to GSP preferences would reduce exports preferences from 9.7 to 3.9 percent.6 Botsw ana Cameroon If EPA negotiations were to be Congo successful and quota- and duty- free Gabon access were to become available to non- Ghana LDCs, preferences would amount to 4.5 Ivory Coast percent of total exports. Compared to Kenya current Cotonou preferences, EPAs Mauritius would provide only a marginal increase, Namibia Nigeria because only few products are omitted Seychelles 49% by the generous access Cotonou offers.7 Sw aziland Zimbabw e The actual value of preferences under an EPA, however, would depend on the Total rules of origin that were to be agreed. 0% 5% 10% 15% 20% 25% While the 4.5 percent figure assumes identical utilization rates as under the GSP Cotonou EPA current regime, inclusion of non- note: all values in EUR1000, calculation of preferences restrictive rules of origin in EPAs could takes current utilization rates into account increase that value to 5.2 percent Source: EU preference data, TRAINS, Bank staff calculations assuming constant trade flows. The World Bank has advocated non- Overall, actually claimed preferences for restrictive rules of origin--such as the all 13 sub-Saharan non-LDCs5 represent 10% value-added rule proposed by the 3.9 percent of their exports to the EU. If Blair commission and/or a simple only GSP preferences were available, change of tariff heading rule--that could this value would fall to 0.5 percent of open a range of new export opportunities total exports. The fall would be for beneficiary countries, thereby substantial for all countries but increasing the value of preferences. It is particularly high in the case of very difficult to estimate the potential Mauritius, the Seychelles, and Swaziland value of these changes in where the value of preferences would competitiveness, however, but they fall from about 23, 16, and 50 percent could potentially be very large. 5South Africa is not included in the analysis as it 6While a recent paper by the ODI (ODI, 2007) is only an ACP country since 1998 and is not a only does a number count of tariff lines that signatory to the commercial provisions of the would see increases in tariff rates if Cotonou Cotonou Agreement, which foresees the preferences were withdrawn, the numbers negotiation of EPAs. While South Africa has presented in this paper support the relative signed an FTA with the EU in 1999, its partners importance of preferential access for various in the customs union (SACU) are supposed to exporting countries. sign an EPA. Both arrangements are likely to 7 The value of preferences under the EBA create barriers within SACU. South Africa has initiative would be similar, showing the limited expressed interest to join the SADC EPA but this additional coverage that EBA preferences offer has been rejected by the EU. over Cotonou preferences 4 Trade Note July 12, 2007 It is also important to note that the Seychelles, and Swaziland) and exceed importance of preferences strongly five percent of exports for another four differs by country. This is mainly due to countries. Nonetheless, while it is differences in the composition of exports unclear who actually benefits from the to the EU. For example, the ratio of preferences, the value of preferences for actually claimed preferences to exports non-LDCs would fall to 0.5 percent, a ranges from 0.1 percent for Nigeria, notional loss of EUR 679 million, if no which mainly exports oil to the EU, to new trading arrangements were in place 49.5 percent for Swaziland that mainly in 2008, when Cotonou preferences exports sugar to the EU which benefits expire. Alternatively, if EPAs with from large preferences (see table 1). Cotonou-type rules of origin were to be in place by 2008, the value of While the importance of preferences as a preferences could increase to 4.5 percent share of exports varies between of exports, a notional gain of EUR 107 countries, more detailed analysis shows million. that in each case, two or three products account for the bulk of claimed There still exists the possibility that the preferences (table A.2).Even though EU will continue to grant these these products differ by exporting preferences even in violation of WTO country, among the most important ones laws, but that might politically difficult ­ are frequently fish, edible vegetables, and reportedly several states stand ready meat preparations, sugar, and cocoa to challenge any unilateral arrangements (preparations). This is also true at the under WTO rules. What happens aggregate level. Taking all export flows strongly depends on whether any kind of together, 6 products account for 78 trading agreement can be reached and percent of the total value of preferences. implemented within the next 12 months. The most important ones are sugar (36 percent of preference value), edible EPAs are not only about preferences, fruits and nuts (12 percent), and however. More importantly, the preparations of meat, fish or crustaceans negotiations offer the opportunity to (12 percent). strengthen regional integration and foster economic growth. A larger internal market and harmonized policies will Conclusion increase the efficiency of resource allocation and investment incentives. The total value of EU preferences ACP countries could also use the claimed by African non-LDCs, negotiations to address domestic expressed as a share of exports, is constraints to growth that inhibit small--3.9%--but could potentially be resources to flow from low-productivity important if that margin (which is quite sectors to high-productivity sectors. substantial for some countries) is a Necessary reforms could possibly be necessary condition for these countries' supported by additional aid-for-trade that exports to successfully compete in the should be used to address supply-side European market. Preferences amount to constraints. a substantial share of total exports for three countries (Mauritius, the 5 Trade Note July 12, 2007 On the other hand, signing EPAs would References: also mean that companies in the EU Brenton, Paul and Ikezuki, Takako would gain preferential market access to (2005) `The Value of Trade ACP markets. Under these agreements, Preferences for Africa', in however, liberalization of the ACP countries vis-à-vis the EU would be Newfarmer, R (ed) Trade, Doha phased in over a long time horizon. and Development: A Window on While liberalization would provide new the Issues, World Bank, competitive pressures on certain Washington DC. companies in ACP countries, it would also discipline prices, drive productivity Overseas Development Institute (ODI), improvements, and lead to increased 2007. "The Costs to the ACP of efficiency and growth. This would Exporting to the EU under the contribute to making ACP countries GSP", London. more competitive in the global market. Olarreaga, Marcelo and Özden, Çaglar, If no agreement came into place in 2008, "AGOA and Apparel: Who the direct costs to exporting non-LDCs Captures the Tariff Rent in the in terms of foregone preferences were generally manageable while there are Presence of Preferential Market small but non-trivial gains in the value Access?". The World Economy, of preferences to be obtained from Vol. 28, No. 1, pp. 63-77 signing EPAs. However, the implementation of a `bad' EPA--an EPA that does not address domestic policy constraints--or the non- implementation of an agreement would remove the leverage for domestic policy reforms that the European Union still holds and would impose much higher long-term costs on ACP countries because of another missed opportunity for structural reforms. This would be the real drama of unsuccessful EPA negotiations. 6 Trade Note July 12, 2007 Annex: Table A.1: Comparison of Cotonou, GSP, and EBA Preferences for 13 African Non-LDCs in 2005 Actually Actually Potential Preferences claimed Exports Potential Preferences claimed prefs under if only GSP preferences* used for preferences if only GSP preferences* EPA (as % used (as % (as % of COMEXT percent calculation under EPA used (GSP&ACP) of exports) of exports) exports) TOTAL missing note Botswana** 2,383,430 25,219 319 23,556 1.1% 0.0% 1.0% 2,383,947 0.0% Cameroon 1,958,817 121,241 2,718 59,647 6.2% 0.1% 3.0% 1,960,381 0.1% Congo 293,385 1,206 577 1,206 0.4% 0.2% 0.4% 294,698 0.4% Gabon 546,179 6,881 5,315 6,881 1.3% 1.0% 1.3% 547,536 0.2% Ghana 958,415 42,441 6,178 41,458 4.4% 0.6% 4.3% 973,995 1.6% 1 Ivory Coast 1,949,361 152,888 17,796 102,491 7.8% 0.9% 5.3% 1,955,971 0.3% Kenya 946,670 92,015 36,930 91,706 9.7% 3.9% 9.7% 954,631 0.8% Mauritius 1,120,824 259,415 11,652 259,415 23.1% 1.0% 23.1% 1,125,148 0.4% Namibia 929,521 54,772 5,349 52,654 5.9% 0.6% 5.7% 950,789 2.2% 2 Nigeria 8,319,340 11,638 7,496 11,637 0.1% 0.1% 0.1% 8,343,282 0.3% Seychelles 259,097 41,086 756 41,086 15.9% 0.3% 15.9% 260,745 0.6% Swaziland 110,628 54,719 744 54,718 49.5% 0.7% 49.5% 115,792 4.5% 3 Zimbabwe 367,444 35,953 7,560 35,947 9.8% 2.1% 9.8% 390,606 5.9% 4 Total 20,143,111 899,474 103,390 782,401 4.5% 0.5% 3.9% 20,257,523 0.6% note: all values in EUR1000, calculation of preferences takes current utilization rates into account, "potential preferences under EPA " assumes Cotonou-type rules of origin Missing trade in which products and why missing 1) special lines (HS12, 99); 2, 3) annex2 (HS8); 4) annex2 (HS8) and special lines *Due to the treatment of bananas in the analysis, total preferences claimed are approximately EUR 32m larger in 2005 (Ivory Coast 28m, Cameroon 4m) ** Comparing these values to the 2002 values, one should notice that Botswana's value of preferences as a share of total exports has decreased substantially. This is due to an immense increase in exports of diamonds from EUR 240m in 2002 to EUR 2.3b in 2005 Source: EU preference data, TRAINS, Bank staff calculations 7 Trade Note July 12, 2007 Table A.2: Comparison of Commodity Protocol and Cotonou Tariff Preferences With Basic GSP Preferences for African non-LDCs (2005) Prefs under Prefs under GSP (as % GSP&ACP HS- of exports) (as % of change chapter Description (a) exports) (b) (a)-(b) Botswana 2 Meat and edible meat offal 0.0% 1.0% -1.0% 61 Art of apparel & clothing access 0.0% 0.0% 0.0% 85 Electrical machinery and equipment 0.0% 0.0% 0.0% top3 0.0% 1.0% -1.0% Total 0.0% 1.0% -1.0% Cameroon 8 Edible fruit and nuts; peel of citrus 0.0% 2.5% -2.5% 76 Aluminium and articles thereof 0.0% 0.3% -0.3% 18 Cocoa and cocoa preparations 0.0% 0.1% -0.1% top3 0.0% 2.9% -2.8% Total 0.1% 3.0% -2.9% Congo 24 Tobacco and manuf. tobacco 0.0% 0.2% -0.1% 3 Fish & crustacean, mollusc & other 0.1% 0.2% -0.1% 44 Wood and articles of wood 0.0% 0.0% 0.0% top3 0.2% 0.4% -0.2% Total 0.2% 0.4% -0.2% Gabon 44 Wood and articles of wood 0.7% 0.9% -0.1% 3 Fish & crustacean, mollusc & other 0.2% 0.4% -0.1% 39 Plastics and articles thereof 0.0% 0.0% 0.0% top3 1.0% 1.3% -0.3% Total 1.0% 1.3% -0.3% Ghana 16 Prep of meat, fish or crustaceans 0.0% 1.6% -1.6% 7 Edible vegetables and certain roots 0.0% 1.3% -1.3% 18 Cocoa and cocoa preparations 0.2% 0.5% -0.3% top3 0.3% 3.5% -3.2% Total 0.6% 4.3% -3.7% Ivory Coast 8 Edible fruit and nuts; peel of citrus 0.1% 2.2% -2.1% 18 Cocoa and cocoa preparations 0.6% 1.6% -0.9% 16 Prep of meat, fish or crustaceans 0.0% 0.8% -0.8% top3 0.7% 4.6% -3.9% Total 0.9% 5.3% -4.3% Kenya 7 Edible vegetables and certain roots 2.6% 3.8% -1.2% 6 Live tree & other plant; bulb, root 0.9% 2.8% -1.9% 20 Prep of vegetable, fruit, nuts or o 0.2% 1.1% -0.9% top3 3.8% 7.7% -4.0% Total 3.9% 9.7% -5.8% 8 Trade Note July 12, 2007 Mauritius 17 Sugars and sugar confectionery 0.0% 17.0% -17.0% 61 Art of apparel & clothing access 0.7% 3.5% -2.8% 16 Prep of meat, fish or crustaceans 0.0% 1.6% -1.6% top3 0.7% 22.1% -21.4% Total 1.0% 23.1% -22.1% Namibia 2 Meat and edible meat offal 0.0% 3.2% -3.2% 3 Fish & crustacean, mollusc & other 0.6% 2.3% -1.7% 16 Prep of meat, fish or crustaceans 0.0% 0.1% -0.1% top3 0.6% 5.6% -5.0% Total 0.6% 5.7% -5.1% Nigeria 3 Fish & crustacean, mollusc & other 0.0% 0.1% 0.0% 18 Cocoa and cocoa preparations 0.0% 0.0% 0.0% 41 Raw hides and skins 0.0% 0.0% 0.0% top3 0.1% 0.1% 0.0% Total 0.1% 0.1% 0.0% Seychelles 16 Prep of meat, fish or crustaceans 0.0% 12.9% -12.9% 3 Fish & crustacean, mollusc & other 0.3% 3.0% -2.7% 40 Rubber and articles thereof 0.0% 0.0% 0.0% top3 0.3% 15.9% -15.6% Total 0.3% 15.9% -15.6% Swaziland 17 Sugars and sugar confectionery 0.0% 46.7% -46.7% 20 Prep of vegetable, fruit, nuts or 0.5% 2.4% -1.8% 7 Edible vegetables and certain roots 0.0% 0.1% -0.1% top3 0.6% 49.2% -48.6% Total 0.7% 49.5% -48.8% Zimbabwe 17 Sugars and sugar confectionery 0.0% 6.0% -6.0% 24 Tobacco and manuf. Tobacco 0.6% 1.3% -0.7% 6 Live tree & other plant; bulb, root 0.3% 0.8% -0.6% top3 0.9% 8.1% -7.3% Total 2.1% 9.8% -7.7% Source: EU preference data, TRAINS, Bank staff calculations 9 Trade Note July 12, 2007 Annex 3: Data assumptions The figures presented in this analysis include most of trade registered in COMEXT, the official trade database of the European Union. Trade flows that are excluded from the analysis are confidential trade and trade in non-confidential tariff lines where the entry- price system is applied. For both we do not know and cannot estimate the applicable tariff rates. Overall, we exclude only 0.5% of exports from African non-LDCs to the EU due to these data constraints (ref. columns 9&10 of table A.1). Trade in tariff lines where the entry-price system is applied takes mainly place in HS chapter 8 (edible fruits and nuts) so the value of preferences for these products is likely to be understated. However, only 7% of HS-chapter-8 imports into the EU from African non-LDCs are excluded. Some additional caveats to the data apply. With regard to the three specific protocols (sugar, beef, bananas), we had to use the following assumptions: · For sugar exports, we assume all export enter under quotas and hence with a zero tariff. This treatment is applied to raw cane sugar and cane or beet sugar (as we cannot distinguish between cane and beet sugar from the data we reasonably assume that only cane sugar is exported from African non-LDCs to the EU). The EU-market price for sugar will be cut by 36 percent over the next four years which means that income from sugar exports will fall strongly. · All beef and veal exports are assumed to fall under the quotas (substantial exports are only registered from countries that are actually granted quotas (minimal exception of Ivory Coast with 50,000 EUR)) and hence the specific element of the ACP tariff rates are reduced by 92% (see protocol 4 of Annex V of the Cotonou Agreement) if preferences have been claimed. · With regard to bananas, we have applied a tariff preference of 300 EUR/ton for 2005 in light of the rapidly changing regulations in the market for bananas. In the year of the analysis, the difference between out of quota MFN and ACP specific tariffs was 300 EUR/ton but also some ACP preferences of 0% tariff existed. Following the market opening and the successful challenge of the regime by some Latin American countries before the WTO, the out of quota MFN tariff (and hence the maximal margin of preference) stood at 176 EUR/t at the beginning of 2007. A substantial amount of the bananas from ACP countries entered with zero duties in 2005 and using a preference margin of 300 EUR/t hence understates the amount of preferences granted in 2005 as the preference margin for bananas entering duty free (as compared to entering out of quota from third countries) would have been much larger (i.e. 680 EUR/t). The reported value, however, has to be used with care as it overstates the potential value of preferences for 2006 and in the future. The preferential margin that is used for its calculation had strongly narrowed. 10