78613 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development Nabil Chaherli and John Nash World Bank Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development Nabil Chaherli and John Nash © 2013 International Bank for Reconstruction and Development / International Development Association or The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with Rights and Permissions external contributions. The findings, interpretations, and The material in this work is subject to copyright. Because conclusions expressed in this work do not necessarily re- The World Bank encourages dissemination of its knowl- flect the views of The World Bank, its Board of Executive edge, this work may be reproduced, in whole or in part, for Directors, or the governments they represent. noncommercial purposes as long as full attribution to this The World Bank does not guarantee the accuracy of the work is given. data included in this work. The boundaries, colors, denom- Any queries on rights and licenses, including subsidiary inations, and other information shown on any map in this rights, should be addressed to the Office of the Publisher, work do not imply any judgment on the part of The World The World Bank, 1818 H Street NW, Washington, DC 20433, Bank concerning the legal status of any territory or the en- USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. dorsement or acceptance of such boundaries. Contents Contents iii 1.3 The future: How can LAC help feed 2.2.3 The anatomy of Latin America and the the world? 24 Caribbean’s agriculture export growth 38 Figures iv 1.3.1 Removing the constraints: priorities for 2.2.4 Sector composition of Latin America Tables vii the future of Latin America and the Caribbean’s and the Caribbean’s agricultural trade 41 Boxes vii sustainable agricultural trade 25 2.2.5 Origins of Latin America and the Acknowledgments ix Caribbean’s agricultural exports 43 2.2.6 Destination markets 44 2.3 Diversification and moving up the quality ladder 46 2.3.1 Changes in export concentration: Is Latin Chapter 1. Chapter 2. America and the Caribbean’s agricultural trade Introduction and Summary 1 Latin America and the Caribbean’s Recent becoming more diversified? 46 Performance in Agricultural Trade 29 1.1 Latin America and the Caribbean’s recent 2.3.2 Value adding and moving up performance in agricultural markets: 2.1 Agricultural production and consumption the value chain 51 overall good news 3 patterns 30 2.4 What has been driving the dynamos? 58 1.2 The enabling environment for agricultural trade: 2.2 World and Latin America and the Caribbean trade: trends and changes 33 2.4.1 Argentina 58 potential constraints and what can be done to overcome them 8 2.2.1 World merchandise trade: general 2.4.2 Brazil 63 1.2.1 Trade policy 8 patterns 33 1.2.2 Infrastructure and logistics 13 2.2.2 Structure of trade in Latin America and the Caribbean 34 Figures Figure 1.1. Shares (left axis) and contribution to growth by origin of Latin America and the Caribbean’s agricultural exporters, 1995 and 2009 (percent) 4 Chapter 4. Chapter 3. Figure 1.2. Shares (left axis) and contribution to Assessing Latin America and the Caribbean’s The Enabling Environment for Agricultural growth by export destination, 1995 and 2009 Contribution to Global Food and Feed Trade: Potential Constraints and What Can (percent) 5 Security in 2050 121 Be Done to Overcome Them 69 4.1 Challenges of sustainable food security Figure 1.3. Top 10 Latin America and the Caribbean 3.1 External environment: barriers to exports from agricultural exports to developed and developing toward 2050 122 Latin America and the Caribbean countries 70 economies, 2009 6 4.2 Demand dynamics and supply-side 3.2 Trade policies of Latin America Figure 1.4. Argentine grain production, constraints 123 and the Caribbean countries 74 1979–2010 (millions of tons) 7 4.2.1 Demand dynamics 123 3.3 Trade agreements: the Latin spaghetti bowl 77 Figure 1.5. Market Access Overall Trade 4.2.2 Supply-side constraints 125 Restrictiveness Indices for agricultural exports 3.4 Logistics and trade facilitation 80 4.3 Methodology used for the analysis 127 by region, 2009 (percent) 9 3.4.1 How important is trade facilitation? 81 4.4 Agricultural exports outlook to 2050: Figure 1.6. Relative rates of assistance by region, 3.4.2 Agricultural trade logistics in Latin the current-path case 128 1965–2009 10 America and the Caribbean: benchmarking and breaking it down 89 4.5 Emerging market growth and the reshaping Figure 1.7. Real effective exchange rates, of the global food economy 133 1980–2010 (2005 = 100) 11 3.4.3 Drilling down: What are the most important logistics and trade facilitation challenges 4.6 Food security, trade, and climate change 136 Figure 1.8. Increase in exports from an at the subregional and country levels? 98 improvement in soft infrastructure to the levels 4.7 The global grid, Latin America and the of Organization for Economic Co-operation and 3.5 Conclusions and policy implications 115 Caribbean’s business environment, Development countries 15 and the food connection 141 3.5.1 Global trade policy 115 Figure 1.9. Logistics cost as a percentage 4.8 Green growth and agricultural exports in Latin 3.5.2 For Latin America and the Caribbean 116 of food product value, 2004 (percent) 16 America and the Caribbean 144 3.5.3 Policy recommendations for logistics Figure 1.10. Trade facilitation: comparing Latin 4.9 Summary and conclusions 149 and trade facilitation 117 America and the Caribbean with other regions 18 References 153 Figure 1.11. Cost to export and import—global comparison, 2011 ($ per container) 20 Figure 1.12. Pineapple supply chain from Costa Rica to St. Lucia 21 Figure 2.1 Agricultural net production value index: Figure 2.15. Latin America and the Caribbean’s Figure 2.26. Gini coefficient for Latin America Latin America and the Caribbean and the world, share of world agricultural exports 40 and the Caribbean’s agricultural exports by origin, 1980–2009 (1980 = 100) 31 1994–2008 50 Figure 2.16. Shares of growth (left axis) and Figure 2.2. Share of world net production contribution to growth (right axis) by product Figure 2.27. Latin America and the Caribbean’s value in 2004–06 international dollars, group (percent) 41 share of world agricultural exports, 1994–2008 1980–2009 (percent) 31 (percent) 52 Figure 2.17. Contributions to export growth by Figure 2.3. Evolution of world trade composition, product category and subregion, 1995–2009 Figure 2.28. Agriculture’s share of global and Latin 1990–2009 (US$ millions) 33 (percent) 42 America and the Caribbean trade, 1994–2008 53 Figure 2.4. Latin America and the Caribbean’s Figure 2.18. Share of top 10 product groups in total Figure 2.29. Latin America and the Caribbean’s share of world trade by group, 2009 (percent) 34 agricultural exports, 2001 and 2009 (percent) 43 share of world exports by stage of production, 1994–2008 (percent) 55 Figure 2.5. Merchandise trade as a share of GDP, Figure 2.19. Shares of growth (left axis) and (percent) 34 contribution to growth (right axis) by origin of Latin Figure 2.30. Latin America and the Caribbean’s America and the Caribbean’s agricultural exports, revealed comparative advantage by type Figure 2.6. LAC’s export structure (U$ billions) 35 1995 and 2009 (percent) 44 classification and stage of production, Figure 2.7. Latin America and the Caribbean’s 1994–2008 56 Figure 2.20. Shares and contribution to growth import structure (percent) 36 according to export destination 45 Figure 2.31. Argentina’s sown area, 1979–2010 Figure 2.8. LAC’s intraregional trade structure (hectares) 62 Figure 2.21. Top 10 Latin America and the Caribbean 2001, 2008, 2009 (US$ billions) 36 agricultural exports to developed and developing Figure 2.32. Argentina’s grain production, Figure 2.9. Share of Latin America and the Caribbean economies, 2009 46 1979–2010 (millions of tons) 62 intratrade in total trade by group (percent) 37 Figure 2.22. Latin American and the Caribbean’s Figure 2.33. Share in Brazilian GDP (percent) Figure 2.10. Annual growth rates of Latin America export product diversification, 1995–2009 and value (R$) of Brazilian agricultural GDP, and the Caribbean’s exports (percent) 37 (Herfindahl-Hirschman Index) 48 1994– 2010 63 Figure 2.11. Selected indicators on the Figure 2.23. Concentration of Latin America and Figure 2.34. Share in Brazilian GDP (percent) composition of trade, 2009 (percent) 38 the Caribbean’s agricultural exports by destination, and value (R$) of Brazilian agribusiness GDP, Figure 2.12. Value of Latin America and the 1995–2009 (Herfindahl-Hirschman Index) 48 1994–2010 63 Caribbean’s agricultural exports, 1995–2009 Figure 2.24. Share of the top 3, top 5, and top 10 Figure 2.35. Trade balance—total and Brazilian (US$ billions) 39 countries in total Latin American and the Caribbean agribusiness: 1989–2011 ($ billion) 64 Figure 2.13. Latin America and the Caribbean’s agricultural exports, 1995 and 2008 (percent) 49 Figure 3.1. Market Access Overall Trade share in world agricultural exports (percent) 39 Figure 2.25. Concentration of Latin America Restrictiveness Index for Latin America and the Figure 2.14. Agricultural and food merchandise and the Caribbean agricultural exports by origin, Caribbean exports by sector (percent) 71 trade as share of agricultural GDP (percent) 40 1995–2009 49 Figure 3.2. Market Access Tariff Trade Figure 3.14. Regional comparisons of physical exports, 2010–50 (percent) 133 Restrictiveness Index for Latin America and the infrastructure and business environment 93 Figure 4.8. Business-as-usual volumes of Latin Caribbean exports by region (percent) 72 Figure 3.15. Global comparison of rural access to America and the Caribbean exports, 2010 Figure 3.3. MA-OTRI for agricultural products roads, 2003 (percent) 94 and 2050 (thousands of metric tons) 133 by region (percent) 73 Figure 3.16. Cost to export and import—global Figure 4.9. Shares in global trade in the optimistic Figure 3.4. Relative rates of assistance comparison, 2011 ($ per container) 97 harmonious rebalancing scenario 134 at the regional level, 1965–2009 75 Figure 3.17. Cost to export and import—regional Figure 4.10. Net exports of fruits and vegetables Figure 3.5. Nominal rates of assistance comparison, 2011 ($ per container) 97 under the rebalancing scenario in 2050 for agricultural exportables and import (thousands of metric tons) 134 Figure 3.18. Relationship between a city’s substitutes in Latin America and the Caribbean, distance from Port Manta and the delivered Figure 4.11. Net exports of animal products under 1965–2009 76 cost of wheat 107 the rebalancing scenario in 2050 Figure 3.6. Real effective exchange rates, (thousands of metric tons) 135 Figure 3.19 Relationship between freight rates and 1980–2010 (2005 = 100) 77 connectivity, container shipping, Caribbean 109 Figure 4.12. Net exports of cereals under the Figure 3.7. Logistics costs in Latin America rebalancing scenario in 2050 Figure 3.20 Liner Shipping Connectivity and the Caribbean and in Organization (thousands of metric tons) 135 Index, 2011 111 for Economic Co-operation and Development Figure 4.13. Net exports for selected Latin America countries, 2004 81 Figure 4.1. Structure of the International Model and the Caribbean countries under the rebalancing for Policy Analysis of Agricultural Commodities Figure 3.8. Increase in exports from an scenario in 2050 (thousands of metric tons) 136 and Trade model 128 improvement in hard infrastructure to the Figure 4.14. Shares in global trade in a pessimistic Organization for Economic Co-operation and Figure 4.2. Global meat production in a current-path world view by commodity group (percent) 139 Development level (percent) 86 case (thousands of metric tons) 130 Figure 4.15. Latin America and the Caribbean Figure 3.9. Increase in exports from an Figure 4.3. Global cereal production in current-path aggregate exports by commodity improvement in soft infrastructure to the case (thousands of metric tons) 130 (thousands of metric tons) 139 Organization for Economic Co-operation and Figure 4.4. Business-as-usual meat food demand Development level 87 Figure 4.16. Cereals exports from Latin America (kilograms per capita) 131 and the Caribbean under the pessimistic scenario Figure 3.10. Ad-valorem tariff-cut equivalents 88 Figure 4.5. Business-as-usual cereal food demand (thousands of metric tons) 140 Figure 3.11. Typical logistics chain for agriculture (kilograms per capita) 131 Figure 4.17. Fruits and vegetables exports from Latin exports 90 Figure 4.6. Business-as-usual net agricultural America and the Caribbean under the pessimistic Figure 3.12 Logistics cost as percentage of food export patterns in 2050 scenario (thousands of metric tons) 140 product value, 2004 (percent) 91 (thousands of metric tons) 132 Figure 4.18. Meat exports from Latin America Figure 3.13. Average logistics costs as percentage Figure 4.7. Business-as-usual annual growth rates and the Caribbean under the pessimistic scenario of sales, by sales volume 92 for Latin America and the Caribbean commodity (thousands of metric tons) 141 Figure 4.19. Predicted impact of better business and Tables Table 4.2. Annual per capita consumption growth logistics on Latin America and the Caribbean’s net rates (2000–10) for meat and milk (percent) 124 Table 1.1 Logistics Performance Index international, exports 143 regional, and income group comparisons 19 Table 4.3. Annual average growth in GDP between Figure 4.20. Annual growth rates in exports as 2010 and 2050 (percent) 129 Table 1.2. Summary of logistics challenges faced by a result of better business and logistics 143 Latin America and the Caribbean subregions 22 Table 4.4. Business-as-usual share of global Figure 4.21. Country impact of better business and agricultural trade, 2010 and 2050 (percent) 132 Table 1.3. Regional shares in world net exports in logistics by main commodity groupings 144 business-as-usual and alternative scenarios Table 4.5. Shares in world net exports in business- Figure 4.22. cereals net exports in 2050 for 2050 (percent) 25 as-usual and alternative scenarios for 2050 under a green growth scenario (percent) 152 Table 2.1. Latin America and the Caribbean’s (thousands of metric tons) 147 contribution to total crop area and yields 32 Figure 4.23. meat net exports in 2050 under Table 2.2. Product and destination Boxes a green growth scenario diversification 51 (thousands of metric tons) 147 Box 2.1. The soybean production chain Table 2.3. Trends in revealed comparative in Argentina 62 Figure 4.24. Fruits and vegetables net exports advantage (RCA) by stage in Latin America under a green growth scenario in 2050 Box 2.2. Rural finance innovation has improved and the Caribbean and the rest of the world 58 (thousands metric tons) 148 competitiveness 65 Table 3.1. Gravity model estimates 84 Figure 4.25. Annual growth rates in Latin America Box 3.1. Logistics challenges in wine exports and the Caribbean exports under a green growth Table 3.2. Agricultural exports by product type 85 from Mendoza 104 scenario in 2050 148 Table 3.3. Logistics Performance Index international Box 3.2. Better use of information and communication Figure 4.26. Impact of green growth on countries in regional comparison, 2010 data 95 technology could increase competitiveness in Latin America and the Caribbean in 2050 149 Table 3.4. Latin America and the Caribbean agricultural production and marketing 114 subregional comparison on the Logistics Performance Index, 2010 96 Table 3.5. Subregional Logistics Performance Index score trends 96 Table 3.6 Breakdown of over costs in beef imports into Chile 103 Table 3.7 Summary of logistics challenges facing Latin America and the Caribbean subregions 112 Table 4.1 Annual percentage change in GDP 123 viii Acknowledgments T his report is a product of the Agriculture and Rural De- velopment Cluster of the Sustainable Development Department in the World Bank’s Latin America and the and Florencia Liporaci (World Bank); Alberto Portugal-Pe- rez and Esteban Ferro (World Bank); Benjamin Mandel (U.S. Federal Reserve); Ernesto A. O’Çonnor (FAO/TCI); An- Bank), Ian Gillson (World Bank), Will Martin (World Bank), and Alberto Valdes (Research Associate, Catholic Univer- sity of Chile) and guidance from Francisco Ferreira (World Caribbean Region. It was prepared by a team led by Nabil tônio Márcio Buainain, Junior Ruiz Garcia, and Pedro Abel Bank), Jamele Rigolini (World Bank), and Maurice Schiff Chaherli and John Nash, under the overall guidance of Au- Vieira (FAO/TCI); Aparajita Goyal (World Bank); and Hiau (World Bank). gusto de la Torre, Ethel Sennhauser, and Laurent Msellati. Looi Kee (World Bank). Michelle Friedman and Carmine The team acknowledges with gratitude the generous Background analyses and papers were prepared by Simla Paolo de Salvo provided excellent research support, and financial support provided by the Spanish Trust Fund for Tokgoz, Prapti Bhandary, and Mark Rosegrant (Internatio- Andrea Patten and Erika Salamanca helped organize and Latin America. nal Food Policy Research Institute); Jordan Schwartz, Gwy- format the report. The report has benefited greatly from neth Buchanan Fries, Darwin Marcelo, Adam Allan Stern, the wise counsel of our peer reviewers Marc Juhel (World The Port of Salvador in All Saints Bay The Port of Salvador in All Saints Bay, Bahia, handles both cargo and cruise ships. Photo: Mariana Ceratti / World Bank ix Chapter 1. Introduction and Summary Buying fresh produce at the market Buying food at the market. Photo: Curt Carnemark / World Bank 1.1. With the global population expected to exceed 9 global leader in increasing food exports in the past countries worldwide, the eight LAC countries in billion by 2050, food security—how to produce decade. the global sample (Argentina, Brazil, Chile, Co- enough food of sufficient quality and make it ac- 1.3. China has sent shockwaves through the global lombia, the Dominican Republic, Ecuador, Mexico, cessible and affordable for consumers around the food market since it entered the World Trade Orga- and Nicaragua) displayed a revealed comparative world—is one of the most important challenges of nization (WTO), shifting from a net food exporter to advantage2 in agricultural production of 2.2 on our time. The United Nations estimates that global a net food importer, notwithstanding its 8th posi- average, well above the 1.0 global average.3 LAC’s food demand will double by 2050, with much of tion. With 20 percent of the world’s population and high potential for scaling up its agricultural output that growth in developing countries. The world will income rising more than 10 percent over the last owes largely to its natural endowments, especial- have 2.3 billion more people, and given the deep decade and projected to grow at 5.6 percent over ly land and water. Of the 445.6 million hectares of transformation of growth trajectories in low-in- the next four,1 China is demanding not only agri- land potentially suitable for sustainable expan- come countries, they will be increasingly affluent, cultural products but also different food items, a sion of cultivated area, about 28 percent is in LAC, with demands for more, different, and better food. pattern affecting Brazil and the rest of Latin Amer- more than in any region but Sub-Saharan Africa.4 But how will that additional demand be met? With- ica. New trends in agricultural and food trade are Accessibility considerations magnify this poten- out large increases in exports to world markets, also emerging today as a result of positive devel- tial: the region has 36 percent of the 262.9 million the recent food crisis could be just an omen of opments in transportation, logistics, and informa- hectares of such land situated within six hours of continual future crises. tion and communication technology. At the same the closest market. Further, this potential is not 1.2. On top of changes in food demand, production time, poor agricultural and trade policies continue confined to Brazil and the powerhouse countries patterns will adjust to climate change and techno- to hold back further improvements in global agri- in the Southern Cone. In expansion potential as a logical innovations. As demand and supply both cultural trade rules, and food export restrictions percentage of area, Bolivia, Belize, and Venezuela shift, countries will have to rely more on the inter- remain a serious threat to long-term food security, all rank higher than Brazil and the Southern Cone national food trading system to move food from as evidenced by some nations’ behavior during countries (excluding Uruguay), and Nicaragua and countries with a surplus to those with a deficit, and the recent food price crisis. Colombia come close. LAC is also well endowed in trade patterns will need to change—perhaps dra- 1.4. While countries in Latin America and the Caribbean renewable water resources, with about a third of matically. We can already see evidence of these (LAC) are quite heterogeneous in their production transformations. Just 20 years ago, the world’s potential, overall they are well equipped to con- 2 The revealed comparative advantage is an index used in international economics to calculate a country’s relative advantage or disadvantage top 10 food exporters did not include a develop- tribute to meeting this challenge. LAC has always in a class of goods or services as evidenced by trade flows. It most ing nation. Today, Brazil (5th) and China (8th) are maintained a strong comparative advantage in commonly refers to an index introduced by Bela Balassa (1965)[AQ: Please add reference.]: RCA = (Eij/Eit) / (Enj/Ent), where E = exports; i = members of the club, bringing substantial chang- agricultural production, as indicated not only by country index; n = set of countries; j = commodity index; and t = set of es in the world food market. In less than 30 years, its position as a net food exporter but also by its commodities. A comparative advantage is “revealed” if RCA > 1. If RCA Brazil has turned itself from a net food importer to high comparative advantage. In a study of many is less than unity, the country is said to have a comparative disadvan- tage in the commodity or industry. an agricultural trade powerhouse, emerging as the 3 Anderson and Valdes 2008 1 Dadush and Shaw 2011. 4 Deininger and others 2011. 2 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development the 42,000 cubic kilometers worldwide. Per capita, and future opportunities. Chapter 4 looks into the sent about 23 percent of the region’s exports and LAC has the highest endowment of renewable wa- future to examine how climate change, superim- 10 percent of global trade ter among developing regions, though some sub- posed on expected demographic and economic 1.7. Over 1995–2009, export growth averaged 8 regions in LAC face higher than average scarcity.5 trends, could affect LAC’s agricultural trade oppor- percent a year. Temperate products (cereals, oil- 1.5. But how great is LAC’s potential contribution? What tunities. seeds, and livestock products) accounted for key obstacles could prevent the region from fulfill- more than half of this growth. Seafood and fruits ing its promise, and what will it take to overcome 1.1 and vegetables made up around 15 percent, fol- them? How can LAC do well while doing good? That Latin America and the Caribbean’s recent lowed by processed products like beverages and is, how can it leverage its increased trade oppor- performance in agricultural markets: tobacco. Of course, this pattern varies by subre- tunities to boost rural growth and reduce pover- overall good news gion, with, for example, fruits and vegetables the ty? This report begins to answer these questions. dominant contributor in Mexico and the Andean While we focus primarily on exports from LAC 1.6. Over the last two decades, there has been much region. Almost all LAC countries contributed to the countries—“feeding the world”—we also consider good news for agriculture and agricultural trade in export growth, but Brazil made the largest con- food imports, “feeding LAC.” Following this chap- LAC. Although trade in agricultural products has tribution by far (more than 35 percent), followed ter’s summary of our main messages, the report is declined as a percentage of overall trade world- by the Southern Cone (around 30 percent; figure organized in three chapters. Chapter 2 looks at re- wide, its value has grown substantially. The LAC 1.1). Except Colombia, the region’s largest export- cent developments in global agricultural trade and, region has captured an increasing share of this ers have all increased their global market shares. particularly, at how LAC food exports have evolved growing market and currently holds a much larger Among the second tier of exporters, Peru, Ecuador, in relation to those of other regions. Chapter 3 con- portion of world trade in agriculture (13 percent, Paraguay, and Uruguay have also increased their siders the role of the enabling environment—do- up from about 8 percent in the mid-1990s) than market share. Central American and Caribbean mestic, regional, and external trade policies and in minerals and metals (8 percent) and manufac- countries, except Costa Rica and Guatemala, have logistics—in shaping the region’s trade patterns tures (3 percent). Agriculture and food now repre- maintained or lost their market shares. 5 Bruinsma 2009. Chapter 1. 3 1.8. Both primary and processed products have con- tributed meaningfully to export growth. However, a study for this report showed that LAC exporters have been tilting their specialization from up- stream industries to downstream (more highly Figure 1.1. Shares (left axis) and contribution to growth (right axis) by origin of Latin America processed). LAC appears to be deepening trade and the Caribbean’s agricultural exporters, 1995 and 2009 (percent) in processed products more quickly than other regions, benefiting from these higher value-added Share 1995 Share 2009 Contribution to growth products. 40 45 1.9. Further, LAC has diversified its exports by country of destination.6 The concentration of LAC export 35 40 products increased on average over 1995–2009. But behind this regional trend lie two tendencies. 35 30 Many major exporters of traditional tropical prod- 30 ucts have diversified exports, while producers of 25 temperate products have become less diversified, 25 especially over the last few years, due largely to 20 20 the food price spike and consequent policy re- sponses. The first category includes Colombia, 15 15 Costa Rica, Guatemala, Ecuador, and Mexico; the 10 second includes Brazil, Argentina, Uruguay, Par- 10 aguay, and Bolivia. To some extent, the diversifi- 5 5 cation of destination markets insulates LAC from price shocks emanating from country-specific 0 0 demand fluctuations. LAC countries that have in- Southem Cone Brazil Andean region Mexico Central America Caribbean creased their product concentrations are more exposed to price shocks in these markets, though Source: Computations by authors based on UN COMTRADE data. this is less of a concern for larger economies like 6 A notable exception is Mexico, which continues to export mainly to its North American Free Trade Agreement partners, reflecting its location- al and climatic comparative advantage, as well as the agreement. 4 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development Brazil, with their highly diversified export baskets outside agriculture. 1.10. Though the EU and the United States remain LAC’s most important destinations—accounting for a combined 45 percent of LAC’s exports in 2009, down from 57 percent in 1995—developing countries are becoming the most dynamic des- Figure 1.2. Shares (left axis) and contribution to growth (right axis) tination for the region’s exports (figure 1.2). Over by export destination, 1995 and 2009 (percent) 1995–2009, China and the rest of the world, with a combined 30 percent of the market share, con- Share 1995 Share 2009 Contribution to growth tributed 36 percent of the growth of exports from 35 30 the region, nearly the 38 percent contribution of 30 the EU (20 percent) and the United States (18 per- 25 cent). Also, the composition of the basket traded 25 20 with developed economies tends to differ (figure 20 1.3). While developed economies imported pri- 15 marily fruits, animal fodder, coffee, beverages, 15 10 and seafood from LAC, products from the soybean 10 complex (seeds, oil, and cake), meat, and sugar 5 5 represented almost 60 percent of the trade with developing economies. 0 0 EU-27 US-Canada LAC (intratade) Rest of the world Asia China (without China) Source: Computations by authors based on UN COMTRADE data. Chapter 1. 5 1.11. This report’s in-depth look at Argentina and Brazil Figure 1.3. Top 10 Latin America and the Caribbean agricultural exports to developed identifies looming logistics and policy issues that and developing economies, 2009 threaten to derail these locomotives of agricultural Developing economies, total exports $77.2 billion growth and some policy choices that have contrib- uted to their success and that might be worth emu- Others; 20 Meat; 15 lating. In Argentina, macroeconomic and structural adjustment in the early 1990s created a propitious Seafood; 2 environment for agricultural growth that laid the groundwork for the subsequent production and Tobacco; 2 export boom. Figure 1.4 shows the real take-off to Oilseeds; 14 Beverages; 3 date from the 1997/98 season. Trade reforms in 1991 lowered export taxes and encouraged tech- Fruits; 3 nology transfer by lowering barriers to importing Sugar; 13 Cereals; 8 technology embedded in inputs. They also en- Vegetable oils; 9 couraged the development of a competitive farm Animal fodder; 10 services industry and attracted investment that improved the infrastructure for moving and stor- Developed economic total exports: $75.4 billion ing grains. Innovative commercial arrangements emerged to attract nontraditional financing into Fruits; 15 Others; 22 the sector, take advantage of economies of scale, and vertically integrate the production chain to improve efficiency. As a result, aggregate factor Animal fodder; 12 productivity growth in this sector—1.1 percent a Wood; 4 year in agriculture and 0.9 percent in livestock— Food was higher than in others. Much more than is gen- preparations; 5 Coffee; 10 erally understood, the export-driven expansion of Meat; 5 agricultural production after 1990 boosted em- Vegetable; 6 Beverages; 7 ployment and value added in upstream and down- Oilseeds; 6 Seafood; 7 stream industries, more than import-substituting industries that traditionally have received high protection, like the auto industry. Source: Computations by authors based on UN COMTRADE data. 6 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 1.12. In recent years, however, some of these reforms— particularly trade policies—have been partially re- versed, shifting relative production incentives. The uncertainty and high export tax equivalent have induced farmers to reduce the area planted with corn and wheat and expand the area planted with Figure 1.4 Argentine grain production, 1979–2010 (millions of tons) soybeans, undermining production sustainability. Export restrictions on beef and milk have slowed Com Wheat Soybean Total these sectors’ development. Agricultural growth has continued, stimulated by extremely high inter- 100,0 national prices, but the sector’s full potential has 90,0 gone unrealized. 80,0 70,0 60,0 50,0 40,0 30,0 20,0 10,0 0,0 1979/80 1981/82 1983/84 1985/86 1987/88 1989/90 1991/92 1993/94 1995/96 1997/98 1999/00 2001/02 2003/04 2005/06 2007/08 2009/10 e Source: Ministerio de Agricultura y Ganadería Agritrend and Fundación Producir Conservando (est). Note: 2009/10 data are estimated Chapter 1. 7 1.13. Further increases in production and exports will but many other private companies, universities, reaching the port). Other important potential bot- depend on resolving policy issues and improving and state research institutes also played import- tlenecks are a deficit of rural storage capacity (es- logistics and infrastructure, as most of the current ant parts. EMBRAPA is credited by many with de- timated at 7–20 percent in static capacity terms) infrastructure was completed in the 1990s, with veloping the soil enhancement technology that and inadequate port capacity. little improvement in the 2000s. Argentina shows transformed the vast area of the cerrado from an that both technical innovation and innovation in agricultural wasteland to one of the country’s most 1.2 commercial organizations can be important driv- productive areas. Further, the recent expansion The enabling environment for agricultural ers of competitiveness in the right policy environ- of agricultural production in no way compares to trade: potential constraints and what can ment. the dominant predatory pattern of the 1960s and be done to overcome them 1.14. As in Argentina, Brazil’s rapid growth in production 1970s, when growth was sustained by the con- and exports was stimulated by macroeconomic tinual incorporation of new land into production 1.17. LAC clearly has done very well in global markets stability and sector reforms put in place in the ear- through deforestation, with cut-and-burn, shifting, for food and agricultural products. But could it do ly to mid-1990s. These included trade liberaliza- and extensive production systems. It is based better? What will it take for LAC to and maximize tion (including the elimination of export taxes) to mostly on high investments and the application its contribution to meeting future food demands? improve incentive structure; virtual elimination of of advanced cultivation techniques, making it less This report considers from several angles how im- direct government purchase (including marketing land intensive and more sustainable. proving both external and internal enabling envi- boards); privatization of important state-owned 1.16. Yet the geographic diversification of Brazilian ag- ronments can support growth in productivity and enterprises; and deregulation of markets for sugar- riculture during the last 35 years—and the leg- trade. cane, wheat, and coffee. Agriculture’s share of pub- acy of a closed economy, which did not require lic spending fell from 5.65 percent in the 1980s to efficient links to external markets—has created 1.2.1 2.11 percent in 1995–99, but its composition im- some bottlenecks to the sector’s competitiveness, Trade policy proved. Although considerably less interventionist particularly for grain crops, which will need to be 1.18. In the external environment, as measured by the than in the past, government agricultural policy loosened for Brazil to continue to supply a large Market Access Overall Trade Restrictiveness Indi- continues to be activist in some areas, including share of world markets. The country’s transport ces (MA-OTRIs) calculated for this report, LAC ag- rural finance. Commercial banks are required by efficiency remains inferior to that of Argentina ricultural exports face fairly high market access law to lend 25 percent of their sight deposits to ag- and the United States, its two main competitors, barriers, particularly to low-income countries and riculture. And the government has put in place two because of the fairly large average distance (more South Asia. On average, agricultural exports from rather innovative programs to help farmers with than 1,000 kilometers) between ports and produc- LAC face barriers (including nontariff and tariff) finance and price risk management. er areas in the Center-West. The high dependence higher than those from any other region except 1.15. In addition to policy reform, technological innova- on road transport accounts for 60 percent of the to- East Asia and the Pacific (figure 1.5). Further, a tion plays a huge role in Brazil. The federal research tal transported cost, exacerbated by the excessive comparison of tariff indices with the MA-OTRIs institute, EMBRAPA, was the most significant actor, number of transshipments (three or more before 8 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development shows that the most significant barriers are non- tariff barriers (NTBs). Manufactured products from LAC face lower barriers, indicating that agricultural exports suffer from an anti-agricultural bias in the external trade regime. The restrictions facing LAC agricultural exports even to other LAC countries Figure 1.5. Market Access Overall Trade Restrictiveness Indices are high. This suggests that—at least in agricul- for agricultural exports by region, 2009 (percent) tural products—regional agreements have not lowered the barriers, corroborating one conclusion HI EAP ECA LAC MENA SAS SSA of the discussion of regional trade agreements be- 120,0% low. 100,0% 80,0% 60,0% 40,0% 20,0% 0,0% ALL HI EAP ECA LAC MENA SAS SSA Note: Hi = high income; EAP = East Asia and Pacific; ECA = Transition Europe and Central Asia; MENA = Middle East and North Africa; SAS = South Asia; SSA = Sub-Saharan Africa Source: Authors’ calculations, based on Kee, et. al., 2009 Note: Each bar is an index of the barriers to exports from the region represented by the bar to the region or group of countries named below the bar. Chapter 1. 9 1.19. In their own trade policies, LAC countries have made great strides since the 1960s and 1970s, when highly protectionist trade policies and ex- change rate regimes promoted industry-led de- Figure 1.6. Relative rates of assistance by region, 1965–2009 velopment. This created in LAC and most other developing countries a strong anti-export and an- Africa LAC Asia ti-agriculture incentive structure. Relative rates of assistance show the protection of manufacturing 30 compared with that of agriculture, with negative values indicating an anti-agricultural bias (figure 20 1.6). In LAC, the overall incentive structure has 10 been close to neutral since the early 1990s. By contrast, some developing regions (including Af- 0 rica) still maintain a net taxation of agriculture, 1965-69 1970-74 1975-79 1980-84 1985-89 1990-94 1995-99 2000-04 2005-09 while others have moved to the agricultural subsi- -10 dization model of the high-income countries. This does not imply, however, that there is no need for -20 further reform in LAC. The overall neutral structure -30 masks a greater protection of import substitutes than of exportables, creating an anti-export bias -40 for agricultural production. Nonetheless, this dif- ference has greatly diminished since the 1980s, -50 indicating that this anti-export bias has lessened. Source: Anderson and others for 1965–2004; own calculations for 2005–09, based on updated database for Anderson While biases and distortions persist in some LAC and others (http://go.worldbank.org/5XY7A7LH40). countries, the overall incentive structure is fairly Note: Five-year weighted averages with value of production at undistorted prices as weights. LAC countries in the conducive to an efficient agricultural supply re- study were Argentina, Brazil, Chile, Colombia, the Dominican Republic, Ecuador, Mexico, and Nicaragua. The 2005–09 sponse to higher prices and appropriate invest- relative rate of assistance for Africa was heavily influenced by several countries that provided high positive protection ments. to agriculture (particularly Ethiopia), but this is not representative of the continent as a whole. A majority of countries had negative relative rates of assistance, as in earlier periods. 10 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 1.20. An emerging—or rather reemerging—issue for the region’s agricultural exports is the potential for Dutch disease effects from the boom in com- modity prices and recent hydrocarbon and mineral discoveries. As Krueger and others’ large study of agricultural policy underscores, macroeconomic policy in many countries greatly influences the incentive structure for agricultural production. Ex- Figure 1.7. Real effective exchange rates, 1980–2010 (2005 = 100) change rate policy has often implicitly taxed the sector. In the 2000s, good macroeconomic policy Brazil LAC (mean) Colombia Mexico Argentine Chile in many LAC countries generally maintained real exchange rates at levels much more stable than 250,000 in the past, avoiding large appreciations (figure 1.7). In recent years, however, exchange rates 200,000 have begun to appreciate in important exporters (particularly Brazil and Colombia), threatening the 150,000 sector’s competitiveness. This trend may become more pronounced as production from the new dis- 100,000 coveries ramps up, making good management of 50,000 the boom critical for agricultural (and other) trade. 0,000 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Source: For Argentina, authors’ calculations using data from the Bank for International Settlements; for the others, data from the International Monetary Fund database. Chapter 1. 11 1.21. Preferential trade agreements (PTAs) affect both processed and higher value-added products than trade regimes, this bias remains significant in the external trade environment and each member in commodities. The gravity model distinguishing some countries. Argentina, a major food exporter, country’s own trade regime. Since the 1992 North product groups clearly demonstrates this: PTAs are imposes export taxes and quantitative controls, American Free Trade Agreement, Latin American positively associated with exports of all product with considerable adverse consequences for the countries have negotiated and notified to the WTO groups but more so for agroindustrial goods than sector and the global food trade system. The mo- almost three dozen PTAs, both bilateral and mul- for others. It appears that more recent agreements tivations behind this policy are understandable: tilateral.7 As long as the Doha Round negotiations have had more positive impacts than earlier ones, these taxes make up a substantial part of the gov- remain stalled, PTAs are the only game in town for such as Mercosur, and that PTAs can reduce NTBs. ernment’s revenue (rising from about 1 percent of negotiating mutual trade barrier reduction. Mexi- And one thing is clear from theory and practice: GDP in 2004 to 4.1 percent in 2011) and keep do- co and Chile have been most prolific in this area: PTAs yield larger benefits when member countries mestic prices low for consumers when internation- each has agreements with several LAC countries, have lower trade barriers with partners outside the al prices spike. Yet quantitative controls produce as well as with the United States, the EU, and some preferential area, because this reduces potential no revenue, contribute to policy uncertainty, and, Asian countries (Chile has a PTA with China). Many trade diversion. along with taxes, reduce domestic production in of these agreements go beyond tariff reductions 1.22. Improving the trade environment. Clearly, the glob- the medium term, potentially raising prices. Export to other trade issues, including some relevant to al trade reform agenda is highly relevant, especial- controls are one explanation for the recent drop in agriculture, such as sanitary and phytosanitary ly for agricultural products, for which trade barriers Argentina’s beef production. And they can create (SPS) measures. Virtually all agreements have a remain much higher than for manufactured goods. the need for further controls, as in Ecuador, where phase-in period of progressively reducing tariffs Given agricultural trade’s importance to LAC and export bans had to be accompanied by price con- and subjecting more products to tariff reduction or LAC’s importance as a world food supplier, it is in trols and government purchases to support pro- elimination. By any measure, most of these agree- everyone’s best interest to lower the barriers as ducers. Further, if several major exporters impose ments liberalize agricultural trade less than non- quickly as possible. And as we saw comparing tar- export taxes simultaneously, the effect on inter- agricultural trade. Other research confirms the MA- iffs with NTBs, the agenda should accord high pri- national prices will at least partly offset the first- OTRIs cited above: notwithstanding the spaghetti ority to NTBs. Global gains from implementing the round impact of the taxes in lowering domestic bowl of agreements among the LAC countries and proposals on the table in the Doha Round could prices in those countries. In any case, alternative with extraregional partners, agricultural trade bar- produce gains of $160 billion a year—and even instruments could meet these objectives at lower riers remain fairly high. But in some cases, PTAs higher true gains from reducing the uncertainty costs than either taxes or controls. We hope future have had important positive effects, more so in associated with gaps between bound and applied trade negotiations will address disciplining export tariffs.8 taxes and controls, but until then, countries can 1.23. While LAC countries have substantially reduced act unilaterally to limit their use. 7 Although most of these agreements have the term “free trade” in their names, even when they are fully phased in, trade is not completely the anti-export and anti-agricultural biases in their 1.24. But the LAC region comprises more than big export- free. It is more accurate, then, to refer to them generically as PTAs, ers. Numerous countries—especially the small though when referring to specific agreements, we use their official economies of Central America and the Caribbean— designation, free trade agreement (FTA). 8 Martin and Matoo 2011. 12 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development are net food importers and impose tariffs or NTBs to address issues not handled well in WTO com- ports, especially in South Asia, the Middle East, on food imports, especially items also produced mitments9—particularly, to reduce the effects of and North Africa. locally. These countries should consider the costs NTBs, as Chile has with its bilateral agreements. 1.26. For LAC countries’ agricultural sectors to stay com- of responding to price movements in international Some ways to use PTAs include: petitive, it is important to appropriately manage markets with policies that insulate their domestic • Removing the exemption of agricultural prod- the real exchange rate to minimize Dutch disease. economies while exacerbating international price ucts from the “general tolerance” or de minims Here, Chile is instructive. Notwithstanding large volatility. These policy responses include reducing exceptions in rules of origin, so that producers of revenue increases from copper in recent years, its tariffs on food imports when prices are high and agricultural products (primary and processed) real exchange rate has not appreciated as much raising them when prices fall. Such policies not could take as much advantage of low-cost im- as that of other countries, due largely to its mac- only magnify world price movements but also are ported inputs as producers in other sectors can. roeconomic policies, including a restrained fiscal inefficient for the country involved, because they A second-best alternative would be to exclude response during the commodity boom and its use encourage overconsumption and underproduction only especially sensitive agricultural products of stabilization and sovereign funds. The threat of when prices are high and vice versa. To the extent without excluding the whole sector, as many Dutch disease magnifies the importance of nation- that traders and processors anticipate such ad- PTAs currently do. al innovation and competitiveness policy. Here, justments, they can adjust the timing of their own • Improving the agreements’ treatment of SPS policy should focus on incentives for technology storage and import behavior, resulting in sharp issues. This could include clarifying the rules generation and adoption that are fairly neutral to- import flow fluctuations and supply chain conges- under the multilateral SPS agreement to im- ward specific products or sectors,10 rather than on tion. A better solution would lower tariffs perma- prove transparency or, even better, committing what Justin Lin (2012) calls comparative advan- nently, reducing the anti-export bias that persists countries not to impose more stringent protec- tage–defying strategies, which single out new in- in the current trade and support regimes, as shown tion than that recommended by international dustries for special favors. above, as well as benefiting poor consumers. An- scientific organizations. Harmonization and other option, implemented by Mexico and Brazil, mutual recognition of standards would also 1.2.2 is to ramp up safety net payments to compensate enhance trade. Some of these issues might be Infrastructure and logistics the poor when food prices rise. Nonetheless, it is handled through current committees and work- 1.27. In addition to trade policy, the quality of logistics clear from the frequency of ad hoc tariff reductions ing groups. and infrastructure critically influences trade’s that strong political pressures encourage this re- • Harmonizing PTAs through gradually converging enabling environment. Portugal-Perez and Fer- sponse when food prices spike. But this should be their commitments. ro (2012) estimate the potential importance for considered a policy of last resort. • Exploring agreements with countries with espe- LAC’s agricultural trade of improving logistics and 1.25. While working within the multilateral system for cially high trade barriers for LAC agricultural ex- several kinds of infrastructure. The study distin- further reforms, LAC countries (and countries guishes the effects of “hard infrastructure,” “soft in other regions) could take more advantage of 9 This section draws from Shearer and others (2009), which discusses the opportunities provided by negotiating PTAs these recommendations and others in considerable detail. 10 Sinnott and others 2010. Chapter 1. 13 infrastructure” (institutions and regulations), and 1.29. Upgrading LAC’s soft infrastructure to OECD levels days required to export.11 Using these variables’ would increase agricultural exports 158 percent,12 estimated impacts, it carries out a simulation of a much larger effect than on manufactured exports the effect if all LAC countries improve these indi- (figure 1.8). Even though improving soft infrastruc- cators to the levels of Organization for Economic ture has less impact for total exports than does Co-operation and Development (OECD) countries. improving hard infrastructure, it is overwhelming- 1.28. The average increase in LAC exports from improv- ly important for agricultural exports. Across LAC, ing hard infrastructure to OECD levels is 130 per- the average impact on agricultural exports would cent for total exports, 157 percent for industrial equal a tariff reduction of 79.3 percent in the des- exports, and 49 percent for agricultural exports. tination importing countries. For many countries, Clearly, the benefit of this improvement is greater the tariff concessions needed for such export lev- for industrial exports than for agricultural exports. els are more than 100 percent, which would be Across LAC, the average impact on agricultural ex- equivalent to exporters receiving an import subsi- ports would equal a tariff reduction of 24.7 percent dy from trading partners! in the destination importing countries. 12 The large effect of facilitation is not due to our assumption of a linear effect in the model. Because these effects at first blush seemed extremely large, we tested for the possibility of diminishing returns to trade facilitation by including a squared term for each trade 11 It used a gravity model and a novel factor analysis approach to over- facilitation variable. The coefficient on each squared term was positive come problems with multi-collinearity that are common to this kind of (negative for days to export), indicating increasing rather than dimin- econometric estimation due to the high correlation across countries ishing returns to trade facilitation. Thus, though large, the results of in the quality of many logistics-related variables. our simulations do not have an upward statistical bias. 14 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 1.30. Further, this study found that some logistics is- sues matter more to particular kinds of products. Exports of heavier products, such as industrial and “bulk” agricultural items, depend more on hard infrastructure, whereas time-sensitive products depend more on soft infrastructure. For agricul- tural exports overall, and for all countries, this soft Figure 1.8. Increase in exports from an improvement in soft infrastructure to the levels of infrastructure is much more important than hard Organization for Economic Co-operation and Development countries infrastructure. Total Exports Industrial Exports Agro Exports 1.31. The big picture is that trade logistics—both hard and soft infrastructure—matter a lot for agriculture 800% and deserve to be at or near the top of trade policy 700% priorities. But to transform this overarching policy message into an actionable agenda requires see- 600% ing how close the region is to best practice else- 500% where to assess its potential for improvement and looking at logistics at a more granular level, both 400% more country-specific and more focused on spe- 300% cific logistics and facilitation measures. Another paper for this report used a case study approach 200% and value chain analysis to look in more detail 100% at specific logistics and infrastructure problems faced by particular countries and regions, espe- 0% ARG BOL BRA CHL COL CRI ECU GTM GUY HND JAM MEX NIC PAN PER PRY SLV URY VEN LAC cially for agricultural trade. Where aggregate indi- cators were available, it benchmarked LAC’s perfor- Source: By Ferro and Portugal-Perez (2012) mance against that of other regions and countries. The objective was to diagnose priority areas for improvement. 1.32. LAC’s poor infrastructure is a major factor underly- ing its consistently poor global competitiveness. The World Economic Forum’s Growth and Busi- ness Competitiveness Index and the World Bank’s Chapter 1. 15 investment climate assessments, for instance, have found that most surveyed firms regard poor infrastructure as a main obstacle to the operation and growth of their businesses. One measure of particular interest to agriculture—the Rural Access Index, which measures the percentage of the rural population living within 2 kilometers of an all-sea- son road13—shows LAC lagging behind East Asia Figure 1.9. Logistics cost as a percentage of food product value, 2004 (percent) and middle-income countries along this dimen- 35 sion. Inadequate access to the road network trans- 32 lates into increased costs, losses, and delays; con- 30 sequences are especially severe for perishable 27 26 goods. Food logistics costs for Peru, Argentina, and 25 23 Brazil are greater than 25 percent of product value, 20 while Chile, a regional leader in logistics, has costs 20 18 Percentage of about 18 percent, still double that of the OECD (figure 1.9). 15 9 10 5 0 Peru Argentina Brazil Colombia Mexico Chile OECD Source: Gonzales and others 2008. 13 A road that is passable year-round by the existing means of rural transport, normally a pick-up truck or truck without four-wheel drive. 16 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 1.33. On the production side, small firms, which make tion with customs agencies and other inspection up the majority of firms in LAC countries and are and regulatory agents operating at borders and the region’s employment and growth engines, ports. As a result of these characteristics, smaller also suffer disproportionately from high logistics producers and local agriculture traders are often costs.14 Perishable agricultural products have heavily affected by poor-quality roads and un- unique characteristics that require specialized competitive trucking services. By contrast, large logistics systems, including remote production shippers benefit from integrated supply chains, zones, temperature control, and special sanitary greater access to the primary trade corridors, and inspection procedures. Because of the time sensi- better berth access at ports. tivity of perishable agricultural goods, bottlenecks 1.34. On average, LAC performs better than only Sub-Sa- in the logistics system directly impact the quality haran Africa in physical infrastructure (figure and quantity of goods delivered. For nonperishable 1.10). Even among LAC countries, there is great products, delays often result in increased logistics variability: Panama and Chile have infrastructure expenses for labor, fuel, and storage, as well as levels that reach those of OECD countries, whereas fees or fines for delays and demurrage. Remote the region’s landlocked countries are the worst per- production zones incur higher costs and greater formers. LAC also underperforms in its business losses for the first actors along the supply chain, environment, which is only half as good as that of the farmers themselves. Most perishable products OECD countries and better than only Sub-Saharan cannot be easily consolidated with other types of Africa and South Asia. Within the region, the best cargo, including other refrigerated cargo. SPS sys- performer is Chile and the worst is Venezuela. tems are necessarily complex, involving coordina- 14 Schwartz and others 2009. Chapter 1. 17 1.35. The Logistics Performance Index (LPI) shows that Figure 1.10. Trade facilitation: comparing Latin America and the Caribbean with other regions LAC’s logistics performance fares poorly compared with that of high- and upper middle-income coun- Physical infraestructure by Region tries, though reasonably well with that of other de- veloping regions.15 As seen in table 1.1, LAC’s over- 8 all LPI score of 2.74 (on a 5-point scale) is similar mean of p_tf_infraestructure to those of Europe and Central Asia and East Asia 6 and the Pacific. LAC performs poorly compared with the upper middle-income group and many 4 Asian countries, including China (3.5), Thailand (3.3), Indonesia (2.8), and Singapore (4.1). 2 0 SSA LAC ECA SAS EAP MNA OECO Business Environment by Region 8 6 mean of p_tf_business 4 2 15 The LPI provides both quantitative and qualitative evaluations of a 0 country in six areas: (a) efficiency of the clearance process (speed, sim- SSA LAC ECA SAS EAP MNA OECO plicity, and predictability of formalities) by border control agencies; (b) quality of trade and transport-related infrastructure (ports, railroads, roads, and information technology); (c) ease of arranging competitive- ly priced shipments; (d) competence and quality of logistics services Source: Portugal-Perez and Ferro (2012) (transport operators, customs brokers); (e) ability to track and trace consignments; and (f) timeliness of shipments. 18 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 1.36. The LPI also illustrates that overall logistics perfor- mance has improved in LAC region, though more so over 2007–10 than over 2010–12. Mexico, the Southern Cone, and Andean countries have made the most progress, while the Central America and Caribbean subregions have fallen back since 2010. Table 1.1 Logistics Performance Index international, regional, and income group comparisons 1.37. In all business survey-based reviews, LAC per- forms considerably worse than OECD standards Region LPI Customs Infraestructure International Logistic Tracking & Time of export and import costs. The required export Shipments Competenc Tracing liness and import procedures include the costs for docu- Europe & Central 2.74 2.35 2.41 2.92 2.6 2.75 3.33 ments, administrative fees for customs clearance Asia and technical control, customs broker fees, termi- LAC 2.74 2.38 2.46 2.7 2.62 2.84 3.41 nal handling charges, and inland transport. The Do- East Asia & Pacific 2.73 2.41 2.46 2.79 2.58 2.74 3.33 ing Business indicators reveal that LAC’s average MENA 2.6 2.33 2.36 2.65 2.53 2.46 3.22 cost to export a container is $1,257 and that the South Asia 2.49 2.22 2.13 2.61 2.33 2.53 3.04 cost to import one is $1,546 (figure 1.11). These Sub-Saharan Africa 2.2 2.18 2.05 2.51 2.28 2.49 2.94 costs are lower than in Sub-Saharan Africa, Eastern High income: all 3.55 3.36 3.56 3.28 3.5 3.65 3.98 Europe, and South Asia, though still higher than in Upper MIC 2.95 2.49 2.54 2.86 2.71 2.89 3.36 other developing regions, such as East Asia, and (except LAC) the OECD average. Within LAC, costs to export a Lower MIC 2.59 2.23 2.27 2.66 2.48 2.58 3.24 container are lowest in Central America and high- Low income 2.43 2.19 2.06 2.54 2.25 2.47 2.98 est in the Andean region, at $1,720 to export and $1,951 to import. Note; MIC = middle income; MENA = Middle East and North Africa Source: World Bank 2011. Chapter 1. 19 1.38. Developing an infrastructure and logistics strat- egy. Quantitative estimates of potential cost re- ductions show substantial heterogeneity in how transport and logistics costs affect LAC countries, depending on the shares of different types of ag- riculture exports and imports. However, supply chain analyses indicate that logistics costs gen- Figure 1.11. Cost to export and import—global comparison, 2011 ($ per container) erally constitute a very high proportion of the fi- nal price of food products (see figure 1.12 for an Cost to import Cost to export example in which land and ocean transport and port costs were found to account for 43 percent OECD high income of the final retail price of pineapples imported into St. Lucia from Costa Rica). So, heterogeneity not- Sub-Saharan Africa withstanding, port efficiency gains, road haulage Eastern Europe & Central Asia improvements, expedited customs clearance and border crossings, better inventory practices, and South Asia increased capacity and competition in storage and warehousing could reduce logistics costs 20–50 Latin America & Caribbean percent. This could mean a permanent 5–25 per- Middle East & North Africa cent reduction in the baseline cost of food and agriculture imports—and increased profits for ex- East Asia & Pacific porters. 500 1.000 1.500 2.000 2.500 1.39. A trade supply chain is only as strong as its weak- US$ est link: poor performance in just one or two areas can have serious repercussions for overall com- Source: World Bank LPI 2012. http://go.worldbank.org/7TEVSUEAR0 petitiveness. The multidimensionality of logistics necessitates a coordinated strategy, developed with input from public and private sector stake- holders alike, for improvements to result in lower costs, higher trade volumes, and increased reli- ability and competitiveness. Countries can ensure that all logistics constraints are identified and 20 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development monitored by designating a national logistics en- tity to coordinate efforts and manage dialogue. For Figure 1.12. Pineapple supply chain from Costa Rica to St. Lucia example, through the Plan Mesoamericana, as well as other regional organizations, Central America is Total Transport Total Transport - Retail - profits, logistics Wholesale-profits, Import Duties Land Ocean + Port and other costs logistics and other costs moving toward strategic regional infrastructure planning, but it has encountered considerable dif- Price Decomposition ficulties harmonizing procedures and standards. Chile, Argentina, and Brazil are also engaging in re- + Other Costs Retail Profits gional coordination to improve intraregional trade. Land Transport 2 Land Transport 1.40. An analysis of the breakdown in food types sug- + Storage & Handling gests that for net food importers, costs associated 1.8 Ocean Transport with refrigerated cargo capacity and services are 15% Duties 1.6 the critical bottlenecks, as meat, fish, and dairy + Port SL 1.4 15% represent the largest share of all food imports by value (26 percent). On the other hand, for LAC 1.2 17% Other Costs countries that are net food exporters, bulk storage, Consolidated Cost 1 US $ /lb handling, and transporting are the primary con- Ocean Transport 0.8 Wholesale Land Transport cerns, because on a weighted-average basis dry Port Miami Producer Price 35% + Port CR bulk items make up the largest share of food im- 0.6 ports by value (31 percent). Thus, Organization of 0.4 Eastern Caribbean States (OECS) island countries, 8% 0.2 for example, should work on reducing the costs of 10% 0 refrigerated containerized traffic. Peru, Brazil, Bo- Fam Gate (CR) CR FAS Ocean to Miami Miami Port Miami Warehouse Miami FOB Price Ocean to St. Lucia St. Lucia Port Wholesale Distribution Retail Profits Final Decomposition livia, and Colombia, however, would benefit from improving the importing and distribution process for dry bulk goods. (See table 1.2 for a diagnosis of critical constraints organized by major country and region.) COSTA RICA (CR) ST. LUCIA (ST) Source: Authors, freight forwarder and shipper interviews, and World Bank (2008). Chapter 1. 21 Table 1.2. Summary of logistics challenges faced by Latin America and the Caribbean subregions   Central America and Mexico Caribbean Andean region Southern Cone • Low road density and access, • Empty backhaul in the DR • Low road density and access, • Relatively high access to roads but especially in rural areas (NIC, GTM, especially in rural areas (PER, COL, high transportation costs (general HND) BOL) problem) • Losses while merchandise is in • Poor quality of roads, especially in • Inadequate road maintenance transport due to spoilage or rural areas (PER, COL, BOL, ECU) investment has led to poor-quality breakage (NIC, GTM, HND, CRI) • Low speed circulation due to heavy roads, especially provincial and rural • High domestic transport costs (NIC, traffic and congestion in roads roads (general problem) GTM, HND, CRI, MEX) leading to ports (PER) • Traffic congestion near border points Land transport • Lack of truck competition due to • Lack of appropriate infrastructure in delays border crossings and thus monopolies (NIC, GTM, HND, CRI, land border-crossing points (general increases costs (BRA, ARG) MEX) problem) • Low competition among truck • Lack of enforcement and adequate • High transportation costs (general companies in some areas (south BRA) trucking regulations for vehicle and problem) • High inventory costs due to the lack driver operations (NIC, GTM, HND, of sufficient warehousing capacity CRI, MEX) (BRA) • Lack of appropriate infrastructure in • Rail freight transport is limited land border-crossing points (C.A.) (general problem) • Empty backhaul due to trade laws • Increasing operating costs of trucking (MEX) services (BRA) • Inefficient port entry procedures such • Maritime and logistics costs depend • Higher maritime transport costs than • Lack of intermodal transfer terminals as long waiting times for processing heavily on countries’ geographical Southern Cone countries (general (general problem) of documentation for imports (NIC, location within the Caribbean problem) • Declining competition, high entrance Maritime transport GTM, HND, CRI, SLV) • On average, maritime transport • Lack of intermodal transfer terminals barriers, and collusive behavior • Bottlenecks due to deterioration of costs to the Caribbean are much (general problem)  among existing players (BRA) access roads to ports and delays in higher than in other regions in LAC • Bottlenecks at ports due to a lack of the loading process caused by a lack (especially OECS) container handling capacity (BRA, of port assets such as cranes (NIC, • Economies of scale disadvantage ARG) GTM, HND, CRI, SLV)  require small islands to consolidate cargo (especially OECS) 22 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development   Central America and Mexico Caribbean Andean region Southern Cone • Burdensome customs inspections • Long time for customs clearing • Inefficiencies at customs clearance • Lack of coordination between and heavy fines (NIC, GTM, HND) in comparison with LAC regions (general problem) customs and border management • Delays due to heavy traffic and (especially OECS) • Delays due to heavy traffic and agencies (general problem) congestion (NIC, GTM, SLV, HND, CRI) • Lack of information technology and congestion (general problem) • Traffic congestion near border points • Lack of coordination between electronic documentation (especially • Lack of coordination between delays border crossings (general Border procedures   customs and border management OECS) customs and border management problem) agencies (NIC, GTM, HND) • Informal payments (general problem) agencies (general problem) • Informal payments (general problem) • Delays and increased costs due to • Significant difference of customs • Informal payments (general problem) • Inventory costs and losses of profit onerous sanitary and phytosanitary performance within the Caribbean from incurred delays of 24 hours at controls, such as duplicate region (OECS are worst performers) the border crossing (general problem) fumigation procedure (NIC, GTM, SLV, HND) • Informal payments (NIC, GTM, SLV, HND, CRI, MEX) Chapter 1. 23 1.3 fiber, distortions in agricultural and trade policies, as in 2010 for meats, decline sharply to 13 per- The future: How can LAC and investments in irrigation and infrastructure cent for fruits and vegetables, and increase to 16 help feed the world? (the baseline or “business-as-usual” scenario), percent for cereals. Argentina and Brazil would LAC’s share in global trade will probably grow in expand their trade in cereals, while Mexico would 1.41. This report concludes with a look at the future. all four main food categories in the model—cere- increase its dependence on the world market. With the need to increase food, fiber, and fuel pro- als, oilseeds, meats, and fruits and vegetables. Except for Chile, LAC countries would see their po- duction about 80 percent by 2050 to meet global By 2050, in this scenario LAC would be supplying tential for fruits and vegetables exports curtailed demand, how will LAC contribute? A background more than a third of meat exports, a third of fruit significantly, with Brazil hit the hardest. These study for this report by the International Food Poli- and vegetable exports, half of oilseeds exports, shares would reflect an important switch in trade cy Research Institute assessed global and region- and about a tenth of cereal exports (table 1.3). In a status for the region. This pessimistic scenario un- al drivers of LAC food exports through 2050.16 The “better business and logistics” scenario with great- derscores trade’s importance as a climate change drivers are natural or human-induced factors that er emerging market growth, agricultural trade lib- adjustment mechanism. Protectionism could pose directly or indirectly cause changes in food de- eralization, investment in irrigation and improved a food security threat by preventing trade from mand and supply in domestic and global markets agricultural technology, better infrastructure, and compensating for surplus disappearance in devel- and in the global trading system. The major issues lower marketing costs (proxied by lower wedges oped countries. This is important for Argentina and explored are whether the world can feed itself into between border and domestic prices), LAC could Brazil, two countries called on to increase their ce- the future, the role that LAC could play, and how play an even greater role in meeting global food, real exports to regions facing large grain deficits. that role might change depending on develop- fiber, and biofuel demand. In two other simulated ments in climate change and other drivers. scenarios—“green growth” (with more efficient 1.42. The background study found that if current trends water and fertilizer use and earlier development continue in income and population growth, use of efficient biofuel technologies) and “harmonious of technology and resources to produce food and rebalancing” (with higher growth of incomes and meat consumption but lower population growth in developing countries, higher productivity growth, 16 The study used IMPACT, a global multimarket, partial equilibrium model that provides long-term projections of global food supply, and earlier adoption of efficient biofuels)—LAC’s demand, trade, prices, and food security, balancing water availability share would grow lesser but still be greater than in and uses within economic sectors at the global and regional levels. 2010 for all product groups. IMPACT uses 281 “food-producing units,” which represent the spatial intersection of 115 economic regions and 126 river basins. The model 1.43. The report also considered a more pessimistic but generates projections for agricultural crop area and crop yields as a plausible forecast. This future pathway includes function of global market drivers (such as commodity demand and prices) and local availability of water resources. Crop area and yields less rapid growth in emerging markets, high pop- also depend on the projected rate of exogenous (nonprice) growth ulation growth, and a wetter and warmer climate. trends, labeled intrinsic growth rates. A detailed description can be LAC’s share in 2050 would stay almost the same found in Rosegrant and others (2008). 24 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 1.3.1 Removing the constraints: priorities for the future of Latin America and the Caribbean’s sustainable agricultural trade 1.44. Consonant with this report’s findings, the back- Table 1.3. Regional shares in world net exports in business-as-usual ground study underscored infrastructure’s impor- and alternative scenarios for 2050 (percent) tance as a potential barrier to increasing produc- tion and trade in LAC. Improving LAC infrastructure Business as Harmonious Pessimistic Green Better business and participation in the global grid could vastly 2010 usual rebalancing view growth and logistics improve agricultural production efficiency and vol- Latin America and the Caribbean ume, essential to meet rising global food demand, Meat 30 36 30 29 33 41 especially in developing markets. Improving its Cereals 8 11 9 16 9 13 business environment and logistics could provide Fruits and vegetables 25 34 33 13 33 38 LAC with much greater opportunities for meet- Oilseeds 42 50 49 55 50 56 ing global food and fiber needs. LAC could thus Developed capture an estimated 5–15 percent more market Meat 38 36 39 29 40 33 share than under a business-as-usual scenario. Cereals 63 45 45 33 47 44 LAC would be capturing this extra market share Fruits and vegetables 17 19 23 13 22 18 from developed countries in meat and oilseeds Oilseeds 31 29 30 9 30 25 and from other developing countries in fruits and Rest of developing (non–Latin America and the Caribbean) vegetables and cereals. LAC’s export profile could Meat 32 28 31 42 27 26 also be restructured in this scenario, with a great- Cereals 29 44 46 51 44 43 er increase in exports of bulk commodities and Fruits and vegetables 58 47 44 74 45 44 of more processed items, such as soybean oil Oilseeds 27 21 21 36 20 19 instead of beans. Exports of soybean meal would grow slower, as higher livestock production in LAC Source: Tokgoz, Bhandary and Rosegrant (2012) requires less meal to be available for trade. 1.45. Another potential constraint to ramping up produc- tion worldwide to meet future demand is inefficient water use. Agriculture consumes about 70 percent of the world’s freshwater supply. While LAC cur- rently has fairly abundant supplies, water scarcity Chapter 1. 25 poses problems for increasing agricultural produc- duction change, moving food from countries where 1.47. Global trade reform in biofuels is especially im- tion in some parts of the region and in other parts it is produced efficiently to food-deficit countries portant in ensuring that LAC can sustainably ramp of the world. The tension between rapidly rising will require new trade patterns. Second, on a year- up its contribution to the global food supply while natural resource consumption and environmental to-year basis, greater weather variability will cre- minimizing global greenhouse gas emissions. sustainability will be a critical pressure point over ate short-term local shocks to food supply that will Liberalizing trade in biofuels could increase sec- the next decades in all regions, including LAC. Pric- require rapidly adjusting food trade to avoid short- tor competition, improve efficiency, lower costs, ing resource use and adopting more sustainable ages. The recent precipitous food price increases and enable the world’s most efficient producers practices in water and land management could demonstrated that when shortages arise, coun- to expand their share of the biofuels market. For help LAC green its growth and agricultural exports. tries tend to react with “beggar thy neighbor” trade example, producing a liter of ethanol from sug- If all regions use water more efficiently (as in a policies that insulate domestic consumers and arcane in Brazil requires only about half the land green growth scenario), LAC’s comparative advan- producers from international price movements. In area needed to produce the same liter from corn tage from its abundant resources would shrink so doing, they increase global price volatility and in the United States. Transferring production from (as reflected in a reduced market share compared shift the adjustment costs to others. Such actions the United States to Brazil would thus reduce the with some other scenarios) but its current share included increases in export barriers (including in amount of land diverted from growing food. But would increase. some LAC countries), which amplified the spike. currently, biofuel promotion policies and trade 1.46. Final constraints, as argued above, are trade bar- Export bans accounted for an estimated 40 per- barriers distort international trade patterns and riers. Maintaining an open, efficient trading sys- cent of the world price increase for rice and 25 impede this shift while imposing large costs on the tem is critical if LAC is to continue—and even in- percent of that for wheat. Less commonly under- populations of the countries employing them. Of crease—its contributions to feeding the world. And stood, the ad hoc reductions in import barriers in course, to fully realize these benefits, Brazil would the looming threat of climate change magnifies many countries had a similar effect, reducing price need to expand its production without deforesting the importance of increasing the trade system’s fluctuations domestically while magnifying inter- land, but as argued earlier, the country has plenty flexibility—for two reasons. First, in the long run, national price variability. of degraded pasturelands that could be used more as patterns of comparative advantage in food pro- productively for these crops. 26 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development Chapter 1. 27 Chapter 2. Latin America and the Caribbean’s Recent Performance in Agricultural Trade Wheat Close up of wheat. Photo: Yosef Hadar / World Bank 2.1. Latin American and the Caribbean (LAC) agriculture diversification has changed in origins, destina- has shared in the large trade expansion of the past tions, and quality and sophistication. And section several decades. While trade in agricultural prod- 2.4 discusses what has driven the rapid growth of ucts accounts for just 7 percent of world trade, LAC exports in the two main growth engines, Argentina has contributed to deepening global food market and Brazil. integration and has steadily increased its mar- ket share. LAC agro-food exports have more than 2.1 Agricultural production and doubled in real terms and tripled in nominal terms consumption patterns over the past 15 years, significantly boosting de- veloping countries’ contribution to agro-food trade 2.3. LAC’s aggregate net agricultural output17 grew to growth. This section looks at features of this import- $262 billion (2004–06 prices) in 2009. This repre- ant growth. In particular, how has LAC’s export struc- sented average growth of 2.9 percent a year since ture changed? Has the export sector become more 1980. Yet LAC output performance, while above specialized, focusing on particular groups of goods, the 2.4 percent world average, has been lagging or has it diversified as it has grown? Has there been well behind China’s 5.0 percent (figure 2.1). LAC in- an upgrading process in this export growth? Where creased its market share of world production from is LAC’s food surplus heading? How homogeneous 11 percent in 1980 to 12.6 percent in 2009, about has LAC’s export growth been in products, desti- the same rate as India’s and Africa’s. But this per- nations, and origins? The answers have important formance has been unremarkable when compared implications for the global consequences of LAC’s with that of China, which had about the same share export expansion and for the future growth of the as LAC in 1980 but moved to almost a quarter of region’s agro-food export subsectors. world production in 2009 (figure 2.2). 2.2. This chapter reviews changes in the production, consumption, and trade of key agricultural com- 17 The analysis uses data for net production value (2004–06 1,000 modities since 1990 and summarizes trends international dollars) from FAOSTAT (2012). Value of gross production in production and consumption of agricultural has been compiled by multiplying gross production in physical terms by output prices at farm gate. Value of production thus measures products in LAC, benchmarking the region’s per- production in monetary terms at the farm-gate level. Value of net formance against that of other regions (section production is obtained by subtracting from production data inter- 2.1). Section 2.2 looks at LAC’s overall trade per- mediate uses within the agricultural sector (seed and feed). Value of production in constant terms is derived using a base period. Constant formance, including the evolution of its trade price series are used to show how the quantity or volume of products composition, and then focuses on its agricultural has changed. U.S. dollar figures for value of gross production are con- verted from local currencies using official exchange rates as prevailing trade. Section 2.3 assesses how agricultural trade in the respective years. 30 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 2.4. LAC agricultural production is spatially concen- Figure 2.1 Agricultural net production value index: Latin America trated, with Brazil, Argentina, and Mexico alone and the Caribbean and the world, 1980–2009 (1980 = 100) producing more than 70 percent of the region’s United States Latin America India Europe Africa China Rest of Asia Oceania agricultural output since 1980 (see figure A1.1 in and Canada and the Caribbean annex 1). The top five countries accounted for 82 450 percent of the LAC total in 2009. Brazil is the only 400 country to have significantly increased its share 350 of the region’s total value of agriculture production 300 (growing from 36 percent in 1980 to 48 percent in 250 200 2009). Value shares for Peru, Chile, and Paraguay 150 saw a slight increase while shares of Argentina, 100 Mexico, Colombia, and the countries of Central 50 America and the Caribbean all fell. For most crops, 0 average yields in LAC countries are above world average yields (table 2.1). 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Source: Compiled by authors from FAOSTAT (2012). Figure 2.2. Share of world net production value in 2004–06 international dollars, 1980–2009 (percent) 1980 2009 35 30 25 20 15 10 5 0 United States Africa Latin America China India Rest of Asia Europe Oceania and Canada and the Caribbean Source: Compiled by authors from FAOSTAT (2012). Chapter 2. 31 2.5. LAC was home to 590 million people in 2010,18 with more than half in just Brazil (33 percent) and Mex- ico (19 percent). Only 20 percent of the population lives in rural areas. Changes in food consumption patterns are important for food trade prospects in the region. The commodity composition of con- sumption is affected not only by taste and culture but also by prices and per capita incomes. People Table 2.1. Latin America and the Caribbean’s contribution to total crop area and yields (percent) in high-income LAC countries—Brazil, Mexico, Ar- gentina, Chile—tend to consume more than 2,950 Latin America and the Caribbean’s Ratio of Latin America and the Caribbean Commodity or group share of world crop area (2010) yields to world average yields calories a day,19 more than the world average of Rice 4 92 2,780. For lower income LAC countries—Guatema- Wheat 4 90 la, Bolivia, Haiti—the figure was less than 2,200. Maize 19 78 Where do these calories come from? According to Other grains 3 111 Pardey Wood and Herford (2010), LAC diets have Sorghum 9 229 changed substantially since the early 1960s. Soybeans 45 100 Maize, once accounting for about 16 percent of Sugarcane 46 113 the average calorie intake, has been displaced Sunflower 17 124 by sugar as the primary calorie source, as in the Vegetables 5 129 West. Beans, potatoes, bananas, and cassava are Tropical and 18 104 secondary sources, overtaken by soybean oil and subtropical fruits poultry. More generally, the share of daily calorie Temperate fruits 6 135 intake coming from animal products has grown substantially to more than 20 percent. Source: Tokgoz and others 2012. 18 Population data is derived from UN (2010) as downloaded from FAOSTAT (2012). 19 Consumption here is defined as the per capita calorie equivalent per day of [production + imports – exports + change in stocks]. 32 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 2.2 World and Latin America and the Caribbean trade: trends and changes 2.6. The massive increase in global trade over the past decades is well known. While falling as a percent- age of all trade, agriculture has participated in that trade expansion, and LAC agriculture has captured Figure 2.3. Evolution of world trade composition, 1990–2009 (US$ millions) a growing market share. Agricultural products Fuels and mining products Manufactures 2.2.1 18000 World merchandise trade: general patterns 16000 2.7. World merchandise trade grew from $3,395 billion 14000 in 1990 to $15,763 billion in 2008 (figure 2.3) be- 12000 fore falling sharply to $12,177 billion in 2009.20 10000 But global merchandise trade flows rebounded 8000 strongly in 2010, up 22 percent from the year be- 6000 fore. The rise in the volume of goods exports was 4000 the largest on record, enabling world trade to re- 2000 turn to its pre-2009 level but not to its long-term 0 trend.21 1990 1995 2000 2008 2009 Source: Calculations by authors based on UN Commodity Trade Statistics.. 20 WTO 2011. 21 WTO 2011. Chapter 2. 33 2.8. Over 1990–2008, manufacturing and fuel and mining products trade expanded at 9 percent and Figure 2.4. Latin America and the Caribbean’s share of world trade by group, 2009 (percent) 12 percent a year, respectively, with agricultural export growth at 7 percent a year. In 2008, agri- 14 culture trade represented 9 percent of total trade, down from 13 percent in 1990, while fuel and min- 12 ing product trade represented 23 percent, up from 10 15 percent. World agricultural trade in 2009 was 8 $1,169 billion, up from $415 billion in 1990 but 13 6 down from $1,340 billion in 2008. 4 8 5 2.2.2 2 3 Structure of trade in Latin America 0 Agricultural and food Minerals and metals Manufacturing Total and the Caribbean 2.9. LAC increased its exports from $330 billion in Source: Calculations by authors based on UN Commodity Trade Statistics.. 2001 to $860 billion [in 2009], before falling to $675 billion during the global financial crisis. In Figure 2.5. Merchandise trade as a share of GDP, (percent) 2009, LAC trade made up 5 percent of world trade, but agriculture held a disproportionate share. Imports Exports While $1 of every $8 of world agricultural trade comes from a LAC country, only $1 of every $33 of 20 world manufactures trade does (figure 2.4). 19 2.10. LAC is not as integrated with the rest of the world 18 as are other regions, as measured by the trade to 17 GDP ratio (averaging 42 percent for LAC and 53 16 percent for the world).22 However, the combined 15 importance of merchandise exports and imports for the region increased from 32.5 to 37 percent 14 2001 2008 2009 over 2001–08 before falling to 34 percent in 2009 (figure 2.5). Source: Calculations by authors based on UN Commodity Trade Statistics. 22 World Bank 2011b. 34 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 2.11. Although manufacturing product exports continue to represent the most important export group for LAC, their share dropped from 57 percent in 2001 to 43 percent in 2008, rebounding to 44 percent in 2009 (figure 2.6). Minerals and metals (including petroleum) saw the largest increase, from 24 per- cent of the total to 37 percent, before dropping to 34 percent in 2009. Agricultural and food products slightly increased their share of the total, from 19 percent in 2001 to 20 percent in 2008, and bene- Figure 2.6. LAC’s export structure ($ billions) fited substantially from high food prices in 2009 to Agriculture and food Minerals and metals Manufacturing reach 23 percent. 1000 2.12. Manufacturing is also the region’s most important 900 import group, despite its share dropping from 76 800 percent in 2001 to 68 percent in 2008 before re- 700 365 bounding to 72 percent in 2009 (figure 2.7). The 600 293 share of agricultural products remained stable 500 400 over the decade, at 9–10 percent of the total. Im- 319 300 227 ports of minerals and fuels surged from 15 percent 188 200 in 2001 to 23 percent in 2008 before dropping to 100 80 173 153 18 percent in 2009. 63 0 2001 2008 2009 Source: Calculations by authors based on UN Commodity Trade Statistics.. Chapter 2. 35 2.13. The value of intraregional trade in LAC increased from $50 billion in 2001 to $160 billion in 2008 Figure 2.7. Latin America and the Caribbean’s import structure (percent) before falling to $115 billion in 2009 (figure 2.8). Agriculture and food Minerals and metals Manufacturing Of every $1 of intratrade between LAC countries, about 20 cents were agricultural or food-related 100 and 53 cents were in manufactured products. 90 80 2.14. Between 2001 and 2009, about 17 cents of ev- 70 68 72 ery $1 of agricultural exports from LAC went to 60 76 another LAC country (figure 2.9). The share of this 50 40 intratrade for minerals and metals dropped from 30 17 cents to 13 cents. LAC countries bought more 20 23 18 15 manufacturing products from each other over the 10 9 9 10 period, the share rising from 15 cents for each $ 0 2001 2008 2009 exported to 22 cents. Source: Calculations by authors based on UN Commodity Trade Statistics. Figure 2.8. LAC’s intraregional trade structure 2001, 2008, 2009 (US$ billions) Agriculture and food Minerals and metals Manufacturing 160 140 120 82 100 80 62 60 45 40 28 29 20 14 31 25 0 11 2001 2008 2009 Source: Calculations by authors based on UN Commodity Trade Statistics. 36 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 2.15. LAC exports of minerals and metals grew the most over 2001–08, at 22 percent annually, followed Figure 2.9. Share of Latin America and the Caribbean intratrade in total trade by group (percent) by agriculture, at 15 percent. Manufacturing, the largest item in value, grew 10 percent. But the Agriculture and food Minerals and metals Manufacturing global financial crisis led to the sharpest and deep- 25 est trade contraction since the Great Depression, 20 with world merchandise imports falling 36 percent between the end of 2007Q4 and 2009Q2.23 The cri- 15 sis affected all three groups, with minerals hit the 10 most (down 29 percent) and agricultural exports hit the least (12 percent; figure 2.10). 5 0 2001 2008 2009 Source: Calculations by authors based on UN Commodity Trade Statistics. Figure 2.10. Annual growth rates of Latin America and the Caribbean’s exports (percent) 2001-08 2008-09 30 22 20 15 10 10 0 Agriculture and food Minerals and metals Manufacturing -10 -12 -20 -20 -30 -29 -40 Source: Calculations by authors based on UN Commodity Trade Statistics. 23 World Bank 2010. Chapter 2. 37 2.16. One characteristic distinguishing LAC trade from global trade is the importance of commodities (ag- riculture, mineral, and mining). While commodity trade represents only 32 percent of world trade, it accounts for about 57 percent of LAC trade (figure 2.11). The emphasis on commodities is even more Figure 2.11. Selected indicators on the composition of trade, 2009 (percent) pronounced in trade between LAC and developing market economies. While commodity exports rep- Agriculture and food Minerals and metals Manufacturing resent only 45 percent of LAC’s trade with devel- oped economies, the share jumps to 70 percent for trade with developing countries. Because de- Total world trade 9 22 68 veloped and developing economies do not tend to differ in their total share of commodity imports Developed market economies' imports from world 10 22 68 (about a third of total imports), this new pattern of trade reflects new South-South commerce pat- Latin America and Caribbean's exports 20 25 55 terns. to developed market economies Latin America and Caribbean's 23 34 2.2.3 exports to world 44 The anatomy of Latin America and the Developing market economies' Caribbean’s agriculture export growth 9 26 65 imports from world 2.17. Exports of agricultural goods from LAC rose near- Latin America and Caribbean's exports ly threefold over 1995–2009, with the fastest 28 42 30 to developing market economies growth in the 2000s (figure 2.12). Agricultural exports peaked at $170 billion in 2008. Prices for 0 10 20 30 40 50 60 70 80 90 100 commodities also peaked in 2008, and when ac- Source: Calculations by authors based on UN Commodity Trade Statistics. counting for price fluctuations through the Food and Agriculture Organization’s Food Price Index, 2008 no longer stands out. Instead, it falls in line with the pace of increasing agricultural exports since 1995. In 2009, LAC exported $153 billion in agricultural goods, highlighting agriculture’s im- 38 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development portance to the region’s GDP. In 2009, agriculture Figure 2.12. Value of Latin America and the Caribbean’s agricultural exports, accounted for about 11 percent of GDP on average 1995–2009 (US$ billions) across the region. LAC increased its agricultural ex- ports market share from 11 percent in 2001 to 13 Exports of LAC LAC Exports, price deflated, 2002-2004 average=100 percent in 2009 (figure 2.13). 180 160 140 120 100 80 60 40 20 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Source: UN Commodity Trade Statistics and authors’calculations. Figure 2.13. Latin America and the Caribbean’s share in world agricultural exports (percent) 14 13 13 12 12 11 11 10 2001 2008 2009 Source: UN Commodity Trade Statistics and authors’ calculations. Chapter 2. 39 2.18. Agricultural and food exports account for about Figure 2.14. Agricultural and food merchandise trade as share of agricultural GDP (percent) 60 percent of LAC agricultural GDP, while imports account for less than 30 percent (figure 2.14). In Imports Exports world merchandise trade, about 10 cents of every 70 $1 are agriculture-related. This group of commodi- ties represents a higher share for LAC trade, almost 60 23 cents for every $1 exported by the region, and 50 this difference has been growing (figure 2.15). 40 30 20 10 0 2001 2008 2009 Source: UN Commodity Trade Statistics and authors’ calculations. Figure 2.15. Latin America and the Caribbean’s share of world agricultural exports Share of Agriculture in World Exports Share of Agriculture in LAC Exports 25 20 15 10 5 0 2001 2008 2009 Source: UN Commodity Trade Statistics and World Bank staff calculations. 40 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 2.2.4 Sector composition of Latin America and the Caribbean’s agricultural trade 2.19. To get a sense of trends and changes in trade, it is important to measure the contributions of differ- ent product groups. Agricultural products can be separated into five groups: tropical products; tem- Figure 2.16. Shares of growth (left axis) and contribution to growth perate products; seafood, fruits, and vegetables; (right axis) by product group (percent) other processed products; and other agricultural products, such as wood and cork. Tropical products Share 1995 Share 2009 Contribution to growth consist mainly of traditional developing-country 50 60 products, such as coffee, cocoa, tea, nuts, spices, 45 textile fibers (mostly cotton), and sugar and con- 50 40 fectionary products. The temperate group includes 35 highly protected temperate zone products grown 40 30 in many industrial countries, such as meats, milk 25 30 and milk products, grains, and oilseeds. 20 20 15 10 10 5 0 0 Temperate products Seafood, fruits, and Tropical products Other processed Other agricultural vegetables products products Source: UN Commodity Trade Statistics and World Bank staff calculations. Chapter 2. 41 2.20. Over 1995–2009, agricultural exports grew 8 per- cent a year. This growth can be disaggregated by Figure 2.17. Contributions to export growth by product category and subregion, product group. Temperate products accounted for 1995–2009 (percent) most of the growth, at nearly 45 percent, followed by seafood, fruits, and vegetables (around 15 OAP OPP SFV TEM TRO percent)and processed products, such as bever- ages and tobacco (around 14 percent). Other ag- 100 ricultural products, including wood, cork, and live animals, contributed 10 percent. Tropical goods 90 contributed least to the region’s agricultural export 80 growth, most likely because the industry for these products is more mature than for the other prod- 70 uct groups, limiting room for growth. Of course, this pattern varies by subregion—for example, with 60 seafood, fruits, and vegetables the dominant con- tributor in Mexico and the Andean region (figure 50 2.17). 40 30 20 10 0 Brazil Southern Cone LAC Andean region Central Mexico Caribbean America Source: UN Commodity Trade Statistics and authors’ calculations. Note: TRO = tropical products; TEM = temperate products; SFV = seafood, fruits, and vegetables; OPP = other processed products; OAP = other agricultural products. LAC = Latin America and Caribbean. 42 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 2.21. Using a more disaggregated scheme of grouping temperate products, meat products increased their market share more than oilseeds products, and both increased their shares much more than any other group in the 2000s (figure 2.18). 2.2.5 Origins of Latin America and the Caribbean’s Figure 2.18. Share of top 10 product groups in total agricultural exports, 2001 and 2009 (percent) agricultural exports 2.22. Which countries are LAC’s major exporters? And 2001 2009 what are the predominant spatial patterns of the 12 region’s exports? As expected, Brazil, Argentina, 10 and Mexico feature prominently in most aspects of agricultural exports—but not uniformly so. These 8 countries accounted for 65 percent of the region’s 6 2009 agricultural exports, up from 59 percent in 1995. Brazil is the biggest exporter, with around 4 36 percent of the LAC total (figure 2.19). Argenti- 2 na, Chile, Uruguay, and Paraguay combined for an- other 35 percent while Mexico and the countries in 0 the Andean region added about 10 percent each. Animal Meat Oilseeds Fruits Sugar Co ee Fats Beverages Cereals Fish and fodder and oils crustaceans Most of the action over 1995–2009 came from the Southern Cone countries and Brazil, contributing Source: UN Commodity Trade Statistics and authors calculations. about 75 percent of the export growth. More specif- ically, the region’s largest exporters all increased their market share, except Colombia. Among the second tier of exporters, Peru, Ecuador, Paraguay, and Uruguay also increased their market share. Central American and Caribbean countries tended to maintain or lose market share, except Costa Rica and Guatemala. Chapter 2. 43 2.2.6 Destination markets 2.23. The past two decades have seen important chang- es in the destination of LAC’s exports. Though the EU and the United States remain LAC’s most im- portant destinations—accounting for a combined 45 percent of LAC’s exports in 2009 but down from Figure 2.19. Shares of growth (left axis) and contribution to growth (right axis) by origin of Latin America and the Caribbean’s agricultural exports, 1995 and 2009 (percent) 57 percent in 1995 (figure 2.20)—developing countries are becoming the most dynamic desti- Share 1995 Share 2009 Contribution to growth nation for the region’s exports. Over 1995–2009, China and the rest of the world, with a combined 30 40 45 percent of the market share, accounted for almost 35 40 36 percent of the growth in LAC exports, nearly the 30 35 combined 38 percent contribution of the EU (20 30 percent) and the United States (18 percent). 25 25 20 20 15 15 10 10 5 5 0 0 Southern Cone Brazil Andean region Mexico Central America Caribbean Source: Computations by authors based on UN COMTRADE data.. 44 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 2.24. Trade composition with developed economies tends to differ. While developed economies im- ported primarily fruits, animal fodder, coffee, bev- erages, and seafood from LAC, products from the soybean complex (seeds, oil, and cake), meat, and sugar accounted for almost 60 percent of the trade with developing economies (figure 2.21). Figure 2.20. Shares and contribution to growth according to export destination Share 1995 Share 2009 Contribution to growth 35 30 30 25 25 20 20 15 15 10 10 5 5 0 0 EU-27 US-Canada LAC (intratrade) Rest of the Asia (without China world China) Source: Computations by authors based on UN COMTRADE data. Chapter 2. 45 Figure 2.21. Top 10 Latin America and the Caribbean agricultural exports 2.3 to developed and developing economies, 2009 Diversification and moving up the quality ladder To developing economies, total exports: $77.2 billion Others; 20 Meat; 15 2.3.1 Changes in export concentration: Is Latin Seafood; 2 America and the Caribbean’s agricultural Tobacco; 2 trade becoming more diversified? Oilseeds; 14 2.25. Specializing in products in which a country has Beverages; 3 comparative advantage maximizes trade benefits. Fruits; 3 And geographical proximity and cultural and lin- Cereals; 8 Sugar; 13 guistic affinity create similar advantages in trading with a limited number of partners. However, putting Vegetable oils; 9 Animal fodder; 10 all the eggs in one basket increases risks in an un- certain world. Concentration in products exposes To developed economies, total exports: $75.4 billion the exporting economy to price fluctuations, and concentration in destination markets exposes it to Fruits; 15 importer countries’ business cycles or changes in Others; 22 tastes, whereas a more diversified basket reduc- es these risks. There are thus tradeoffs between Animal advantages of specialization (concentration) and fodder; 12 Wood; 4 advantages of diversification. 2.26. One strand of research has characterized sector Food preparations; 5 diversification along the development path. Imbs Coffee; 10 Meat; 5 and Wacziarg (2003), using data on sector-lev- el employment and value added and covering a Vegetable; 6 Beverages; 7 wide cross-section of countries at various levels of Oilseeds; 6 Seafood; 7 disaggregation, provide evidence that economies tend to grow through two stages of diversification. At first, sector diversification increases, but be- Source: Authors’ calculations based on UN Commodity Trade Statistics. 46 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development yond a certain level of per capita income, the sec- 2.27. Given these concerns and tradeoffs, it is worth two tendencies. The countries that have historical- tor distribution of economic activity starts concen- looking empirically at how LAC exports have ly based their agriculture on one or two traditional trating again. In other words, sector concentration evolved in product and market diversification crops for export (coffee, bananas, pineapple, and follows a U-shaped pattern in relation to per capita or concentration. We do this using the Herfind- so forth) have also made great efforts over the income. Klinger and Lederman (2004, 2005) fol- ahl-Hirschman Index (HHI).24 Of interest are three past two decades to encourage local producers lowed with a similar result on export data. They HHIs: the HHI-product, the HHI-origin, and the and exporters to diversify so that they can reduce built on Hausmann and Rodrik (2003) to explore HHI-destination. their dependency on the main products. Most of a causal link from market failures to insufficient • The HHI-product is calculated using the share of the countries that depended heavily on tradition- diversification. Essentially, they find that opening the value of each two-digit-level product export al exports have diversified their export basket, as new export markets is an entrepreneurial gamble with respect to all agricultural exports for one reflected in the considerably lower values of the that, if successful, is quickly imitated. The inability LAC exporter country and period. HHI-product in 1995 than in 2009. By contrast, of export entrepreneurs to keep private the bene- • The HHI-origin considers the share of the value countries highly specialized in the production fits of their activity leads to a classic public-good of all agricultural exports from one country with of cereals, oilseeds, or meats (Brazil, Argentina, problem. Klinger and Lederman show that poor respect to all agricultural exports from LAC. Uruguay, Paraguay, and Bolivia) saw their share institutions appear empirically to compound the • The HHI-destination compares the share of ag- of these products in total exports increase. This problem, supporting the Hausmann-Rodrik view. ricultural exports to a group of countries with has translated into a greater concentration of their Cadot, Carrere, and Strauss-Kahn (2011) using total agricultural exports for one LAC exporter agricultural export baskets (greater values of the a methodology different from that of Klinger and and period. HHI-product). Lederman (2005), explore the evolution of export 2.28. Taking the inverse of the HHI provides an equiv- diversification patterns along the economic devel- alent in numbers of products, destination, or ori- opment path. They find a hump-shaped pattern gins. The HHI measures by origin, destination, and of export diversification similar to what Imbs and products provide different movements for concen- Wacziarg (2003) found for production and employ- tration figures over 1995–2009. The main findings ment. Low- and middle-income countries diver- on the HHIs are as follows. sify mostly along the extensive margin, whereas 2.29. LAC’s agricultural export structure is becoming high-income countries diversify along the inten- more concentrated (figure 2.22). The concentra- sive margin and ultimately re-concentrate their tion of LAC exports increased consistently over exports on fewer products. This hump-shaped 1995–2009, but behind this regional trend are pattern aligns with the conjecture that countries travel across diversification cones, as discussed 24 The HHI is calculated as follows: given a group divided into n catego- in Schott (2003, 2004) and Xiang (2007). ries, the share of each n category with respect to the group is squared, and the HHI results from the sum of the squared shares. Hence, the index varies from 0 (highly diversified) to 1 (highly concentrated). Chapter 2. 47 Figure 2.22. Latin American and the Caribbean’s export product diversification, 1995–2009 2.30. LAC’s agricultural export destination structure is (Herfindahl-Hirschman Index) becoming more diversified. Figure 2.23 presents the destination country concentration of LAC’s 0,35 total agricultural exports to its trading partners. 0,3 R = 0,7726 The country concentration increased over 1995– 2000, after which the concentration decreased un- 0,25 til 2009, reflecting an expansion to more markets. 0,2 LAC countries, especially those that had depend- 0,15 ed less on a few traditional markets, generally 0,1 focused on diversifying the destination markets. Although a few of the region’s countries increased 0,05 their dependence on certain markets, most re- 0 duced their dependence in comparison with 1995 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 (especially Paraguay, Bolivia, and Costa Rica). Source: UN Commodity Trade Statistics and World Bank staff calculations. 2.31. Agricultural exports by country origin are becom- ing more concentrated. Figures 2.24 and 2.25 present the origin country concentration of LAC’s Figure 2.23. Concentration of Latin America and the Caribbean’s agricultural exports by total agricultural exports. The origin concentration destination, 1995–2009 (Herfindahl-Hirschman Index) of LAC exports remained fairly stable over 1995– 2000, after which the concentration increased 0,12 until 2005. Exports stabilized again in the most 0,1 recent period but at a higher level than in the late 1990s. 0,08 R = 0,8598 0,06 0,04 0,02 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Source: UN Commodity Trade Statistics and World Bank staff calculations. 48 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development Figure 2.24. Share of the top 3, top 5, and top 10 countries in total Latin American 2.32. The finding that LAC’s agricultural exports exhibit and the Caribbean agricultural exports, 1995 and 2008 (percent) different trends before and after 2000 is con- firmed using the Gini coefficient, an alternative 1995 2008 way to measure changes in concentration by 100% 90% measuring export equality in each period (figure 80% 2.26).25 LAC’s Gini coefficient remained fairly sta- 70% ble until 2000, after which it increased from 0.69 60% to 0.73. A smaller number of countries account for 50% 40% the increased size of LAC’s exports after 2000. 30% 20% 10% 0% Top 3 Top 5 Top 10 Source: UN Commodity Trade Statistics and World Bank staff calculations Note: Top 3 = Brazil, Argentina, and Mexico; Top 5 = Top 3 + Chile and Colombia; Top 10 = Top 5 + Ecuador, Costa Rica, Peru, Uruguay, and Guatemala for 1995 or Paraguay for 2009. Figure 2.25. Concentration of Latin America and the Caribbean agricultural exports by origin, 1995–2009 0,25 R = 0,9486 0,2 0,15 0,1 0,05 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 25 It is defined as: Gini = 1 – (1/n)*SUM{cumshare (i–1) + cumshare(i)} where i is a country’s rank if there are 25 countries (i = 1 is smallest and n = 25 is largest), and cumshare(i) is the cumulative share of exports of Source: UN Commodity Trade Statistics and World Bank staff calculations. the ith country. Chapter 2. 49 2.33. The product and destination trends over 1995– 2009 reflect the strong growth of Brazilian and Argentinean exports with an increasing concen- tration in products and a higher diversification in export destination, a pattern shared by Chile, Uruguay, and Paraguay (table 2.2). Brazil and Ar- gentina had the most market equivalents and con- sistently increased the number of destinations Figure 2.26. Gini coefficient for Latin America and the Caribbean’s over the period. Mexico, however, kept a limited agricultural exports by origin, 1994–2008 number of destinations, owing to its strong inte- 0,74 gration with its North American Free Trade Agree- R = 0,9871 ment partners (the United States and Canada), 0,73 and tended to have a stable and fairly large set of 0,72 products exported. Between these two extremes, 0,71 the other LAC countries exhibited various patterns 0,70 in export diversification. Guatemala and Ecuador simultaneously increased their number of prod- 0,69 ucts exported and markets reached; Colombia and 0,68 Peru reduced their span of destination but slightly 0,67 increased product diversification. 0,66 0,65 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Source: UN Commodity Trade Statistics and World Bank staff calculations. 50 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 2.34. Even though LAC’s overall agricultural exports showed greater market diversification but further Table 2.2. Product and destination diversification product concentration, there were great disparities among LAC countries. While the Southern Cone Product Product Destination Destination countries deepened their specialization in pro- diversification diversification diversification diversification Product- Product- equivalent equivalent equivalent equivalent destination index destination index ducing meat, cereals, and oilseeds and reached number (1995) number (2009) number (1995) number (2009) equivalent (1995) equivalent (2009) (1) (2) (3) (4) (5) = (1)*(3) (6) = (2)*(4) more markets, countries in Central America and Latin America 4.1 3.4 11.7 17.0 48.3 56.9 the Andean regions reached more markets and di- and the Caribbean versified from their original concentration in a few Brazil 3.8 2.7 19.0 23.8 72.9 64.3 traditional products, especially coffee, bananas, Argentina 2.1 1.7 15.8 27.1 33.6 45.0 and tropical fruits. Peru 3.0 3.3 12.4 11.3 37.7 36.9 Chile 3.2 2.9 9.4 10.1 30.4 29.0 2.3.2 Guatemala 2.2 3.2 5.8 6.7 12.6 21.4 Value adding and moving up the value chain Colombia 2.2 3.2 7.3 6.5 15.8 20.7 2.35. Over 1994–2008, the share of world agricultural Uruguay 2.6 1.6 7.0 12.4 18.2 20.3 Ecuador 1.8 2.0 5.4 8.8 9.8 17.8 exports supplied by LAC countries increased by Bolivia 3.3 2.0 6.3 6.8 20.6 13.7 about half, from 8 percent to almost 12 percent Costa Rica 2.3 2.3 5.4 5.9 12.3 13.5 (figure 2.27).26 Paraguay 2.7 1.2 3.7 9.9 10.0 11.6 Mexico 3.3 3.4 1.4 1.7 4.7 5.7 Source: UN Commodity Trade Statistics and World Bank staff calculations. Note: The equivalent number is a theoretical value that represents the number of categories (product or destination markets) of identical size that would lead to the degree of export concentration exactly equal to the one observed. It is defined as the inverse of the Hirschman-Herfindahl Index. Calculating product differentiation through the equivalent number distinguishes for each country the equivalent number of exported goods of equal importance (within the agricultural exports sector) leading to the same concentration of agricultural exports. The equivalent number takes values between 0 and +∞. The larger the index value, the greater the export diversification in products or in destinations. Columns 5 and 6 are basically the product equivalent number multiplied by the destination equivalent number for 1995 and 2008, respectively. The index captures developments in both the product and destination space. The data in columns 1–4 have been sorted according to the product destination index in column 6. 26 This statistic was tabulated using the aggregates of a bilateral prod- uct-level trade dataset described below and includes trade between countries within LAC. Chapter 2. 51 2.36. At the same time, the contribution of agricultural products to overall LAC exports fell by a third, or about 10 percentage points, in line with the shrink- ing global share of agricultural trade (figure 2.28). Dividing the agriculture share for LAC by that of the rest of the world yields a measure very close to the Figure 2.27. Latin America and the Caribbean’s share of world agricultural exports, revealed comparative advantage (RCA) index by 1994–2008 (percent)\ Balassa (1965); the flatness of the ratio of the two 14% series (that is, LAC’s agriculture share is always about twice that of the rest of the world) indicates that LAC’s overall RCA in agricultural goods re- 12% mained fairly stable. Thus, overall LAC agricultural exports consolidated global market share without Share of World Agricultural Exports (%) 10% a corresponding deepening in specialization. 8% Sugars, sugar preparations and honey (6) Fixed vegetable fats and oils, crude, refined or fractionated (42) Meat and meat preparations (1) 6% Textile yarn, fabrics and related products (65) Oil seeds and oleaginous fruits (22) 4% Fish and crustaceans and preparations (3) Feeding stu for animals (8) Co ee, tea, cocoa, spices and 2% Manufactures (7) Vegetables and fruit (5) 0% 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Source: Mandel (2012), based on UN Commodity Trade Statistics. 52 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 2.37. But how could LAC’s share of agricultural markets rise and agriculture’s share in LAC and global ex- ports fall more or less in line with each other? One explanation could be that productivity (and com- petitiveness) in LAC improved across the board in agriculture and nonagriculture sectors. Recent Figure 2.28. Agriculture’s share of global and Latin America and the Caribbean trade, 1994–2008 research found that more productive and compet- itive exporters expand their market share by sell- Rest of World LAC ing better products, raising the question of wheth- er quality upgrading could be behind the rise in 35% LAC market share for agricultural products.27 In a Share of LAC Exports Accounted for by Agricultural Products 33% 31% background paper for this study, Mandel (2012) 29% addresses this question by measuring agricultural 27% export differentiation at the level of detailed prod- 25% ucts and finds limited evidence of differentiated 23% varieties within agriculture product groups. In oth- 21% 19% er words, consistent with conventional wisdom, 17% agricultural exports are fairly homogeneous, and 15% consequently there is limited space for increasing 13% value added by producing the same products but 11% of better quality. Mandel then investigates a sec- 9% ond type of product upgrading, the manufacture of 7% 5% product groups with fairly high degrees of process- 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 ing and value added, to see whether interproduct upgrading accounts for part of LAC’s recent market Source: Mandel (2012), based on UN Commodity Trade Statistics. share increases. This study quantifies upstream Note: LAC = Latin America and the Caribbean. and downstream specialization by defining which industries are more input-intensive. It uses two methods to allocate industries to stages of pro- 27 See Manova and Zhang (2012) and Harrigan and Baldwin (2011). Chapter 2. 53 duction, each with assumptions about agricultural growth contribution of products in the two stages period, stage 1 goods classified by net forward industries’ input-output structure. maintained the 60/40 split between upstream and flow added 1.7 percentage points to LAC’s share 2.38. Conditional on the definitions of upstream and downstream export sales throughout the sample. of world agricultural exports, similar to the 1.8 per- downstream industries, as well as on global trends 2.39. The second categorization scheme takes a more centage points contributed by stage 1 products. in specialization, LAC exporters have been tilting holistic approach of the input-output structure of their specialization from upstream to downstream production by assigning stages to the entire set of (that is, more highly processed) industries. Specif- industries from an input-output table, both agricul- ically, LAC’s specialization in early-stage products tural and nonagricultural. It utilizes the concept of is growing more slowly than in the rest of the world, forward flow, which is the amount of output from and its specialization in stage 2 products is grow- upstream stages used as inputs to downstream ing more rapidly. First, the textual descriptions of stages minus the reverse flows of outputs from agricultural export products using Standard Inter- upstream stages used as inputs to downstream national Trade Classification (Rev. 4) categories stages. Industries are assigned to stages to max- are used to infer their level of processing accord- imize the overall forward flow across all industries ing to a set of keywords. For example, categories in the economy.28 These categories are applied to containing the words “prepared” or “processed” LAC export products. Interestingly, and in contrast are assumed to be farther downstream in the val- with the keywords classification, the net forward ue chain of production than those containing the flow classification scheme indicates that LAC is words “raw” or “fresh.” Using these categories, already more specialized in downstream products LAC agriculture exports are more specialized in than in upstream products, though the extent of less processed, upstream stages of production downstream specialization has been declining and there have been no consequential changes (see figure 2.29). Over the course of the sample in specialization across stages over 1994–2008. 28 The ideal choice of input-output data to measure the net forward flow Figure 2.29 illustrates that both stage 1 and stage of LAC exports would be a set of industry-level matrices for each LAC 2 categories contributed meaningfully to LAC mar- country. Unfortunately, input-output matrices at a level of detail close ket share and LAC market share growth during to the Standard International Trade Classification 4-digit products used in the keywords classification are not publicly available for any that period. Stage 1 products accounted for 4.8 LAC country. For this analysis, two other input-output tables, for the percent, or roughly 60 percent, of LAC’s agricultur- United States and Thailand, are used to approximate the structure of al market share in 1994, growing 2.3 percentage production in LAC. For each country, we use the input-output table to define stages of production, with industries allocated to stages to points thereafter. Stage 2 products accounted for maximize the economywide value of the net forward flow expression 3.2 percent of LAC’s market share in 1994, grow- above. The key data used to define the production stages are the 2002 benchmark input-output tables from the Bureau of Economic Analysis ing 1.2 percentage points thereafter. The relative and the 2005 input-output table for Thailand. 54 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 2.40. The extent to which LAC exporters specialize in up- Figure 2.29. Latin America and the Caribbean’s share of world exports by stage of production, stream or downstream products can be examined 1994–2008 (percent) by computing a Balassa RCA index of the following form for the nominal exports (X) for each country Keywords classification (k) and each stage: 14%  X kstage   stage   X  Share of World Agricultural Exports (%) 12% RCAkstage =  Xk  10%   X  8% Stage 2 2.41. This index is simply the country’s global market 6% share in a stage divided by that country’s global 4% market share of total agricultural exports. Accord- 2% Stage 1 ing to classification of stages by keywords, be- cause LAC’s market share of stage 1 products is 0% greater that its market share of stage 2 products, 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 LAC’s stage 1 RCA will be greater than 1 and its Forward flow classification using a U.S. input-output table stage 2 RCA will be less than 1; this is illustrated by the hatched lines in figure 2.30. The market 14% share of stage 1 products grew in line with their Share of World Agricultural Exports (%) 12% proportion, which made the LAC RCA in each stage 10% fairly stable over the sample period. Using the key- words classification, the estimated degree of spe- 8% cialization is consistent with a standard narrative Stage 2 6% about resource-abundant developing countries: 4% their level of output depends heavily on primary goods with lower value added and sophistication. 2% Stage 1 Moreover, this classification does not indicate that 0% specialization in stage 1 products has given rise 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 to an increasing proportion of exports of stage 2 products. Such a trend would be reflected in the in- Source: Mandel (2012), based on UN Commodity Trade Statistics. creasing RCA of stage 2 products and the decreas- Chapter 2. 55 ing RCA of stage 1 products. Using the forward flow classification based on the U.S. input-output table, the slow growth of stage 2 products relative to the level of their share means that stage 2 RCA fell from more than 1.45 in 1994 to 1.2 in 2008. Thus, LAC specialization in downstream products according to this classification was high and decreasing. In figure 2.30, this high level of stage 2 market share Figure 2.30. Latin America and the Caribbean’s revealed comparative advantage by type is almost exactly the opposite of LAC RCA as mea- classification and stage of production, 1994–2008 sured by the keyword classification. But despite the different findings in the relative values of RCA 1.5 indices, the dynamics of comparative advantage Stage 1 (keywords) 1.4 by production stage remained consistent for LAC Stage 2 (forward flow 1.3 - U.S. I -O table) Advantage in Stage 1 & 2 Products over the study period. Using various structures of LAC's Revealed Comparative input-output relations, LAC increased its special- 1.2 ization in upstream industries and reduced it for Stage 1 (forward flow 1.1 - Thai I -O table) downstream industries. 1 0.9 Stage 2 (forward flow - Thai I -O table) 0.8 Stage 1 (forward flow - U.S. I -O table) 0.7 Stage 2 (keywords) 0.6 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Source: Mandel (2012), based on UN Commodity Trade Statistics. 56 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 2.42. The study also compared trends in LAC RCA in each rate of specialization deepening of exporters con- stage of production with RCA trends in the rest of ditional on being a LAC country. the world. A common keywords classification and 2.44. The estimated RCA trends for agricultural export common (U.S.- and Thai-based) forward flow clas- products over 1994–2008 are shown in table sification to export products from all countries 2.3. For stage 1, the estimates for α 2 are nega- was uniformly applied. An important reason to tive and significant regardless of the classification compare LAC trends with those of other exporters scheme. This means that LAC RCA in stage 1 grew is to control for global trends in demand that can more slowly than RCA for stage 1 industries in the vary by stage. For example, if China increases its rest of the world. Similarly, the point estimates for imports of raw materials from all sources, all ex- LAC’s specialization deepening in stage 2 are pos- porters would appear more specialized in stage 1 itive, though only statistically significant for two products; this would not be very informative about of the three classifications. This indicates a faster the value chain (production) dynamics that we rate of specialization in stage 2 industries than seek to describe. The changes in RCA over time that in the rest of the world. Together, these results for each stage and classification scheme are mea- suggest that despite trends in LAC specialization sured with the following panel regression: toward upstream products, the composition of LAC K exports relative to other exporters has tilted to- stage RCAkt = α 0 + α1 yeart + α 2 yeart *LAC + ∑δ k + ε kt stage k =1 ward downstream products. In contrast with α 2 , the estimates for α 1 are sensitive to the classifi- 2.43. where LAC is a dummy for LAC countries and δ k cation scheme. While the keywords classification is an exporter fixed effect. α 1 can be interpreted indicates that overall specialization in stage 2 in- as the rate of specialization deepening in a given creased, both net forward flow classifications indi- stage, while α 2 can be interpreted as the relative cate the opposite. Chapter 2. 57 2.45. In this section, we analyzed the composition of LAC’s specialization patterns among products with different degrees of processing. Using various methods to classify products as either upstream or downstream, we show that both types of prod- ucts have contributed meaningfully to LAC’s recent increases in market share. Controlling for global Table 2.3. Trends in revealed comparative advantage (RCA) by stage in Latin America specialization trends among upstream and down- and the Caribbean and the rest of the world stream products, LAC exporters are found to have deepened specialization in downstream products. Dependent variable: RCA by exporter-stage-year Keywords Net Forward Flow (U.S.) Net Forward Flow (Thai) Stage 1 Stage 2 Stage 1 Stage 2 Stage 1 Stage 2 2.4 What has been driving the dynamos?29 Year -0.008** 0.005** 0.008** -0.011** 0.008** -0.007** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) 2.46. LAC’s two biggest powerhouses, Argentina and Brazil, together accounted for around two-thirds of Year * LAC -0.010* 0.008** -0.007* 0.001 -0.009** 0.010** the region’s export growth over 1995–2010. Both (0.004) (0.003) (0.003) (0.004) (0.003) (0.003) have lessons for other countries within LAC and outside the region. Exporter FE Yes Yes Yes Yes Yes Yes Observations 3,559 3,565 5,567 3,558 3,564 3,566 2.4.1 R-squared 0.79 0.77 0.75 0.77 0.73 0.73 Argentina 2.47. Because of its favorable climate and natural en- Source: Mandel 2012. dowments of land and water, Argentina has a strong * significant at 5 percent; ** significant at 1 percent. FE = Fixed Effect; comparative advantage in temperate agricultural products and has long been a major producer and exporter of grains and meat. Nonetheless, recur- ring macroeconomic mismanagement and an- 29 This section is drawn principally from two papers sponsored by the FAO/CP program—O’Çonnor (2011) and Buanain and others (2011)— and from Regúnaga 2010, commissioned as a background paper for this study. 58 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development ti-export and anti-agricultural biases—created by 2.49. Other reforms were aimed specifically at the agri- • Strengthening the biotechnology programs in import-substitution policies and export taxes and cultural sector:30 the National Institute of Agricultural Technolo- restrictions—led to fairly stagnant production in • Eliminating export taxes for most agricultural gy31 and SENASA to improve food health issues. the five decades leading up to the 1990s. products. Export taxes were reduced from 41 • Promoting (public and private) the use of no-till 2.48. In the early 1990s, however, a package of reforms percent for soybeans in May 1989 to 20 percent practices. dramatically changed the policy environment for in March 1990 with the new administration. In 2.50. These reforms encouraged technology transfer agriculture. This was followed by a rapid expan- 1992, a 1.5 percent export tax earmarked for by lowering barriers to importation of technology sion of agricultural area—grain and soybean area the National Institute for Agricultural Technology embedded in inputs. It also spurred a competitive increased by about half from crop years 1995/96 was eliminated. farm services industry and attracted investment to 2005/06—and much more rapid expansion of • Eliminating government intervention through that improved the infrastructure for moving and production, which increased around 225 percent the National Grain Board. storing grains. Innovative commercial arrange- (figures 2.31 and 2.32). General economic re- • Promoting future markets for grains and oil- ments emerged to attract nontraditional financing forms that affected agriculture and other sectors seeds. into the sector, take advantage of economies of included: • Opening the economy, including the reduction scale, and vertically integrate the production chain • Stabilizing domestic prices associated with the of import taxes on agricultural inputs and the to improve efficiency. As a result, aggregate factor free convertibility of the currency and other elimination of import taxes on capital goods. productivity increased—1.1 percent a year in agri- sound macroeconomic and fiscal policies im- Elimination of quantitative restrictions on im- culture and 0.9 percent in livestock—much more plemented since 1991. ports. than in other sectors, which increased 0.6 percent • Implementing tax reform, including the elimina- • Creating the National Seed Institute (for intellec- a year over 1990–2006.32 tion of taxes on production and trade. tual property rights and seed quality controls) 2.51. In recent years, however, some of these reforms— • Passing new legislation promoting foreign in- and the new Seed Law to adapt to the Interna- particularly trade policies—have been partly re- vestment (no discrimination, bilateral invest- tional Union for the Protection of New Varieties of versed, shifting relative production incentives. ment agreements with the main countries). Plants intellectual property principles. • Eliminating quantitative restrictions on trade, • Developing biotechnology policy to promote the 31 The National Institute of Agricultural Technology was created in 1956 price controls, and other nontariff barriers. development and use of genetically modified to promote and strengthen the development of agricultural research and extension and accelerate the benefits of these core functions: • Privatizing public services, through the sale con- seeds, including the creation of the CONABIA modernizing and improving agricultural companies and rural life. cession to the private sector of port and country (Biotech and Biosafety National Organization) It depends on the Ministry of Agriculture, Livestock, and Fisheries, elevators, ports, highways, railways, energy, within the Ministry of Livestock, Agriculture, and with operational and financial autarky. The institute’s main objective is to contribute to the competitiveness of agriculture, forestry, and and other government-owned firms. Fishing. agribusiness across the country within a framework of ecological and social sustainability, generating information and technologies for processes and products and putting them in the service of rural producers through the extension system. 30 Regúnaga 2010. 32 Coremburg (2008), as cited in O’Çonnor (2011). Chapter 2. 59 Restrictions on corn and wheat exports combined on soybean exports in Rosario could translate into with export taxes now create an effective tax rate a tax of 45 percent on production in Salta.34 higher than the soybean export tax. The uncer- 2.52. Further increases in production and exports will tainty and high level of export tax equivalent have depend on improving logistics and infrastructure. induced farmers to reduce the area planted with Indeed, while most of the current infrastructure corn and wheat and to expand the area planted was built in the 1990s with little improvement in with soybeans. This is not good for soil rotation the 2000s, production has grown from 67 million and conservation and increases the risk of plant tons in 2000/01 to around 102 million in 2010/11 diseases associated with monoculture. The real (see figure 2.32). impact on farms is much higher than the nominal 2.53. Export prohibitions on beef have likewise led to rate would suggest, because the prices at the farm large reductions in the herd, from 56 million head level are much lower than at the ports.33 The wedge to 47 million in the last four years. Restrictions on between the farm-gate and free-on-board price is milk exports and commercial controls have held higher for products that are packaged or processed production at around 10 billion liters since 2000— off-farm than for those shipped in bulk. The export after it increased 70 percent over the previous taxes thus create a special bias against these nine years—despite a booming demand in world high-value products and their associated process- markets that generated rapid growth in production ing industries, which can be significant sources in Brazil and Uruguay in the 2000s. of rural employment. For example, the nominal tax rate of 10 percent on apple exports translates into a tax rate of 27 percent on apple production. And because the effective rate at the farm gate is magnified for regions with higher transport costs to the ports, these taxes discriminate against the development of the regions far from the ports. For example, under reasonable assumptions about prices and transport costs, a tax rate of 35 percent 33 This is because the latter includes all the expenses needed for the transportation, conditioning, packaging, refrigeration, and other expenditures associated with the added value. 34 Regúnaga 2011. 60 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development Figure 2.31. Argentina’s sown area, 1979–2010 (hectares) 2.54. The export-driven expansion of agricultural pro- duction after 1990 created significant employ- Corn Wheat Soybean Total ment and value added in upstream and down- Devaluation and 35.000.000 Export Taxes commodities prices stream industries (box 2.1). Studies comparing and low Export Tax Cuts boom (2002-2011) 30.000.000 international (1991) this job-creation potential to that of import-sub- prices ('80s) 25.000.000 stituting industries traditionally receiving high 24,3% Ha 20.000.000 -5,70% 31,4% protection—like the auto industry—conclude that 15.000.000 agriculture is far superior. 10.000.000 5.000.000 0 1979/80 1981/82 1983/84 1985/86 1987/88 1989/90 1991/92 1993/94 1995/96 1997/98 1999/00 2005/06 2007/08 2009/2010 e 2001/02 2003/04 Source: O’Çonnor 2011. Note: e = estimated Figure 2.32. Argentina’s grain production, 1979–2010 (millions of tons) Corn Wheat Soybean Total 100,0 90,0 80,0 70,0 60,0 million tn 50,0 40,0 30,0 20,0 10,0 0,0 2009/2010 e 1979/80 1981/82 1983/84 1985/86 1987/88 1989/90 1991/92 1993/94 1995/96 1997/98 1999/00 2005/06 2007/08 2001/02 2003/04 Note: e = estimated Source: Ministerio de Agricultura y Ganadería, Agritrend and Fundación Producir Conservando (est). Chapter 2. 61 Box 2.1. The soybean production chain in Argentina Box table 1. Employment in the Argentine soybean chain, 2003/04 The soybean production chain, with many different stages and participants, strongly (number of jobs) influences employment and regional consumption, especially for service providers. In recent years, the agroindustrial chain has become more sophisticated and has promot- Steps / Participants Minimun estimate Maximun estimate ed the development of several specialized industries. Some of them are high-tech, like Seeds 1,250 1,250 biotechnology and precision farming, and have created opportunities for new and more Fertilizers and 1,878 10,500 specialized jobs. Chemicals Agricultural 19,350 23,000 Several studies have estimated employment in the agribusiness sector and in specific MAchinery value chains. Bisang and Sztulwark (2007) studied employment in the soybean chain, Primary producers 91,498 108,500 from input providers to crushers (but excluding exports and such related services as fi- Contractors 37,700 56,500 nancing, quality services, and technical services). Using various assumptions, they esti- Country dealers - 36,961 55,035 mated a range of maximum and minimum jobs for each step and total direct employment coops of 215,000–304,000 jobs (box table 1). Most jobs are associated with total soybean Carriers 19,000 38,000 acreage (14.5 million hectares) and production (32 million tons in 2003/04). Crushers 7,000 11,000 Using similar assumptions, our estimates for the 2007/08 crop (16.6 million hectares TOTAL JOBS 214,641 303,786 and 46.2 million tons) range from 266,000 jobs to 381,000. This is several times the di- rect and indirect employment effects in protected manufacturing industries, such as the Most of the employment in the soybean value chain involves small and medium-size automotive and spare parts value chain (from auto parts to distributors). Following an firms (primary producers, contractors, country dealers, and carriers) based in towns and approach similar to Bisang and Sztulwark’s, Castillo and others (2007) estimated total di- small cities in the interior. Despite the concentration of primary production and crushing, rect employment in 2006 at 90,000: 57,600 for auto parts and vehicles, 18,400 for spare the recent evolution of the industry organization has been positive, providing opportuni- parts, and 14,000 for official distributors. ties for the emergence and development of local services and input providers, members of an emerging middle class who live in small cities and towns in the main producing areas. The regional distribution of firm owners and employees has been important to social and regional development, stimulating local demand for products and services and im- proving regional welfare. Source: Regúnaga 2010. 62 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 2.4.2 Figure 2.33. Share in Brazilian GDP (percent) and value (R$) of Brazilian agricultural GDP, 1994– 2010 Brazil 2.55. Brazil is the world’s third-largest food exporter, af- Agricultural GDP (R$ billion) Share in Brazilian GDP (%) ter the United States and the EU. Even in a period of rapid growth for the economy as a whole, there 8% was no discernible declining trend in agriculture’s 200 7% contribution over 1994–2010 (figure 2.33); al- 6% though highly variable from year to year, the share 150 5% R$ billion remained around 6–8 percent. The share of agri- 4% business is much higher than that of primary ag- 100 3% ricultural production and increased rapidly in the 2% 2000s (figure 2.34), as did exports from the sector 50 1% (figure 2.35). 0 0% 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Source: Buainain and others (2011), based on Ipeadata. Figure 2.34. Share in Brazilian GDP (percent) and value (R$) of Brazilian agribusiness GDP, 1994–2010 Share in Brazilian GDP (%) Agribusiness GDP (R$ billion) 35% 900 30% 800 700 25% R$ billion 600 20% 500 15% 400 10% 300 200 5% 100 0% 0 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Source: Buainain and others (2011), based on Cepea/USP/CNA 2011. Chapter 2. 63 2.56. As in Argentina, this rapid growth was stimulated by macroeconomic stability and sector reforms put in place in the early to mid-1990s, which in- cluded trade liberalization (including elimination of export taxes); virtual elimination of direct gov- ernment purchase (including marketing boards); and deregulation of markets for sugarcane, wheat, and coffee. Overall public spending in agriculture fell from 5.65 percent of federal spending in the Figure 2.35. Trade balance—total and Brazilian agribusiness: 1989–2011 ($ billion) 1980s to 2.11 percent in 1995–99, but its compo- sition improved as research spending increased. Trade Balance Brazilian Trade Balance Agribusiness Brazilian Overall, Brazil’s Producer Support Estimate (as estimated by standard Organisation for Econom- 70 ic Co-operation and Development methodology) 60 50 rose from 0.92 in 1995–97 (indicating a net tax US$ Billion 40 of 8 percent on agriculture) to 1.03 (a 3 percent 30 protection rate) in 2010.35 While overall support 20 10 levels are quite low, much of that support is pro- 0 vided through relatively distortionary instruments -10 (price supports, credit subsidies, and directed -20 credit). The potential negative effects, however, 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 are minimized by limiting the support per farmer and targeting some payments (for example, mak- Source: Buainain and others (2011), based on AgroStat. ing purchases from “family farmers” at relatively high prices). While considerably less intervention- ist than in the past, government agricultural policy continues to be activist in some areas, including rural finance. Commercial banks are required by law to lend 25 percent of their sight deposits to agriculture. In addition, the government has put 35 OECD 2011. 64 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development in place two innovative programs to help farmers Box 2.2. Rural finance innovation has improved competitiveness with finance and price risk management (box 2.2). The Rural Producer Note (Cédula do Produtor Rural, CPR) was created by Law nº 8.929, Increasingly in recent years, Brazil’s support pro- August 22, 1994. It is a private bond (promissory note) issued by farmers, farmers’ as- grams are oriented toward environmental goals: sociations, and cooperatives. Through this note, producers or farmers’ associations sell farmers qualify for some payments only if they forward their production at a fixed price and agree to deliver the production according to respect agricultural zoning laws, and support is specifications defined in the note for quality, quantity, data, and place of delivery. The planned for environmentally friendly cultivation note is endorsable and, if duly registered, fully negotiable. For countries at early stages practices. Even without special subsidies, a very of the development of commodity exchange and derivatives markets—such as Brazil— large fraction of Brazil’s land is cultivated using no- investor acceptance of producers’ notes, commodity and commercial papers, and other or low-till technology. financial assets usually requires the intermediation (intervention) of a well-known finan- 2.57. While Brazil has a large commercial farming sector cial institution. Bank of Brazil is so far the only financial institution operating with CPR. that has received much government attention, it It is not only helping its clients issue and sell CPR but also guaranteeing the note itself. also has many small family farms. Reforms in the The bank charges its clients for the endorsement and carries out regular auctions of CPR. mid-1990s initiated programs aimed at this sector. Apparently, CPR has enormous potential to mobilize resources to finance agricultural ac- The most important of these is the National Pro- tivities. It can be competitive with alternative financial assets and attract foreign investors gram to Strengthen Family Farms. The program in- familiar with this type of operation. So far, however, its extent has been very limited, due cludes a number of subprograms, including a Har- to its expense and the fact that producers bear all price risk. vest Guarantee Program, which provides farmers In 1996, the government introduced options contracts, which may provide in the me- a minimum payment in case of catastrophic crop dium run an efficient instrument for hedging against price risks. To introduce the option loss due to weather. The guarantee covers only contract as an alternative for minimum price, the government, through the Companhia farmers in the poorest regions. The government Nacional de Abastecimento, or National Supply Company, is taking the role of the specu- also started another weather insurance program lator, technically called writer or seller of the option contract. While the buyer has the right for family farms in 2004, with wider coverage. to exercise the option to buy or sell the commodity, the seller is the opposite party, which must stand ready to honor the contract in case the buyer decides to exercise the option. In any case, the seller gets the premium paid for the option contract. The supply company is selling options contracts for rice, maize, and wheat. Chapter 2. 65 2.58. In addition to policy reform, technological innova- savannah land with high potential for sustainable sponsible for 60 percent of the total volume of soy- tion played a huge role in the Brazilian experience. production. Under the right production systems, beans moved in the country, rail for 33 percent, The federal research institute, EMBRAPA, was its exploitation would have minor negative envi- and waterways for 7 percent, while trucks were the most significant actor in this, but many oth- ronmental impacts. Besides, as Brazilian agricul- used to transport other grains. Silveira and others er private companies, universities, and research tural production is very heterogeneous, with many (2005) highlight the stark differences among the institutes also played important parts. EMBRAPA producers far from the technological frontier, good main soybean-exporting countries in transport is credited by many with developing the soil en- opportunities for efficiency gains and productivity efficiency and show that Brazil’s remains inferior hancement technology that transformed the vast increases abound, in systems from sugar cane to to those of Argentina and the United States, its two area of the cerrado from an agricultural wasteland cattle breeding and even traditionally extractive main competitors—because of the large average to one of the country’s most productive areas. crops. distance (more than 1,000 kilometers) between 2.59. The recent expansion of agricultural production 2.61. Notwithstanding the technical potential, sustain- ports and producer areas in the Center-West. Ac- differs from the dominant predatory pattern of the able growth depends fundamentally on the avail- cording to estimates by Abiove, the high depen- 1960s and 1970s, when growth was sustained by ability of four additional factors: technology to dence on road transport accounts for 60 percent the continual incorporation of new land into pro- sustain a vigorous innovation process and, in par- of the total transported costs,37 exacerbated by duction through deforestation.36 The occupation of ticular, to cope with the threat of climate change; the excessive number of transshipments (three or the new frontier moved away from the traditional institutions to ensure appropriate incentives and more before reaching the port). cut-and-burn, shifting, and extensive production conditions for sustainable growth; infrastructure to 2.63. Another difference between main worldwide grain systems. It instead uses high investments and reduce production costs, ensure favorable market export countries relates to the storage facilities advanced cultivation techniques that are more access, and create better conditions for integrat- network. Today’s deficit in bulk storage dynamic sustainable and environmentally friendly than tra- ing small and medium-size farmers in dynamic capacity in Brazil is an estimated 36 percent. With ditional occupation patterns, in which downgraded agrochains; and microeconomic incentives and an estimated 5 percent annual expansion of Bra- subsistence crops areas were replaced by pasture financial capital to fund investments at the farm zilian agriculture by 2012, the need to expand port after three to five years of cultivation. level. capacity is estimated to be 560,000 tons a year,  2.60. Can Brazil maintain its position as the foremost de- 2.62. The geographic diversification of Brazilian agri- due to the need to increase moving capacity to veloping country food supplier to world markets? It culture during the last 35 years—and the legacy 46.5 million tons of grains.38 In Brazil, the storage has the resource base to do so. Brazil has 90–150 of a closed economy that did not require efficient network includes many intermediate storage facil- million hectares of unused arable land (a conser- links to external markets—has created some bot- ities, private or owned by cooperatives, outside of vative estimate) to sustain the increase in agri- tlenecks to agriculture competitiveness (partic- farm properties. According to Oliveira (2007), only cultural production without further deforestation. ularly for grain crops) that the country will need 11 percent of Brazilian grain storage facilities are This comprises mostly degraded pastures and to resolve to continue supplying a large share of world markets. According to Barbosa and Noguei- 37 Silveira and others 2005. ra Junior (2007), in 2005, road transport was re- 36 Silveira and others 2005. 38 Borges and others 2006. 66 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development located on-farm (40 percent in Argentina and 60 costs stem mostly from nondelivery or expensive percent in the United States), with 53 percent in measures to meet delivery schedules. Resolving towns, 32 percent in rural villages and 5 percent this infrastructure deficit is critical for the compet- in ports. These off-farm facilities have substantial itiveness of Brazilian agribusiness. This situation capacities, ranging from 30,000 to 60,000 met- creates investment opportunities for the private ric tons. This generates economies of scale but sector, but providing incentives to undertake them higher transaction costs between producers and will require addressing regulatory and institutional intermediates and losses due to an excessive issues. number of transshipments. Additionally, more than 60 percent of the Brazilian logistics system is in the hands of a few international corporations, like Cargill, Dreyfuss, and ADM.39 The poor quality of Brazil’s roads, the deficit of rural storage capac- ity (estimated at 7–20 percent in static capacity terms), and the seasonal nature of exports reduce Brazilian grain competitiveness, especially when prices are low. 2.64. Brazil places 61st on the Logistic Performance Index, which takes into account not only direct transportation costs—freights, ports, and han- dling charges, fees, and side payments—but also induced costs, “hedging for the lack of predict- ability and reliability of the supply chain.”40 These 39 These efforts would be counteracted by reducing the dynamic/static capacity relationship from seven a year to four, on average, through an identity conservation system based on tests by the Cartagena Protocol on Biosafety to control live transgenic transboundary flow between grain exporters and importers (Silveira and others 2005). Imposing unfriendly export procedures leads to overregulation rents, inade- quate market structure, and a lack of incentive to invest, resulting in low-quality, fragmented services and frauds, though requiring new regulations. This consolidates the vicious logistic circle characterized by the World Bank study (Arvis and others 2007). 40 Arvis and others 2007, 16. Chapter 2. 67 Chapter 3. The Enabling Environment for Agricultural Trade: Potential Constraints and What Can Be Done to Overcome Them Rancher herding cattle. Brazil. Photo: Thomas Sennett / World Bank 3.1. A region’s or country’s relative trade performance the tariffs levied on exports of manufactured prod- pacts into price effects, using the import demand provides a summary measure of its global mar- ucts from most LAC countries are lower than those elasticities in Kee and others (forthcoming). This ket competitiveness. Chapter 2 looked at Latin on agricultural exports. This creates an effective methodology can be used to estimate the real ef- America and the Caribbean’s (LAC’s) performance anti-export bias in the external environment for fect of tariffs without the effect of NTBs. This mea- indicators. But we also want to understand the agriculture. For some countries, the difference can sure is reported as the Market Access Tariff Trade underlying determinants of competitiveness, the be substantial. Restrictiveness Index (MATTRI). most constraining factors, and the policy levers 3.3. But tariffs are not the whole story. Actual protec- that might overcome these constraints. Of course, tion rates are sometimes lower than tariffs would many factors determine competitiveness—in- indicate because of “water” in the tariffs or exemp- cluding natural endowments and policies and tions and sometimes higher because of nontariff investments that make on-farm and in-factory barriers (NTBs). To take these factors into account, production more or less efficient—some of which a background study for this report computed a we considered in chapter 2. In this chapter, we Market Access Overall Trade Restrictiveness Index analyze determinants of trade competitiveness (MA-OTRI).41 The MA-OTRI summarizes the impact focusing on four areas directly connected to trade of other countries’ trade policies on each country’s policy: external market access (section 3.1); exports. The index answers the following question: countries’ own trade policies, which may create What is the uniform tariff that, if imposed by all anti-agricultural and anti-export biases (section trading partners on exports of country c instead 3.2); regional trade agreements (section 3.3); and of their current structure of protection, would trade logistics and infrastructure (section 3.4). We leave exports of country c at their current level? conclude in section 3.5 with policy recommenda- The partial equilibrium MA-OTRI is the weighted tions derived from the analysis. sum of protection levels (including tariffs and the ad valorem tariff equivalent of domestic support 3.1 payments and other NTBs) facing the exports of External environment: barriers country c in other countries, where weights are to exports from Latin America and the given by the elasticities of import demand in other Caribbean countries countries and their imports from country c. To ob- tain the ad valorem tariff equivalent of NTBs, the 3.2. An important determinant of the overall incentive study first estimated the quantity impact of NTBs structure for production and trade is the level of on imports and then translated the quantity im- barriers that a country’s exports face in trade part- 41 See Kee and others (2009) for a detailed description and literature ners’ markets. As with many developing countries, survey of trade restrictiveness indices. 70 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development Figure 3.1. Market Access Overall Trade Restrictiveness Index for Latin America and the Caribbean exports by sector (percent) LAC-All LAC-AG LAC-MF 120,0% 100,0% 80,0% 60,0% 40,0% 20,0% 0,0% ALL HI EAP ECA LAC MENA SAS SSA Each bar is an index of the barriers to exports from the Latin America and the Caribbean to the region named below the bar. Note: Hi = high income; EAP = East Asia and Pacific; ECA = Transition Europe and Central Asia; MENA = Middle East and North Africa; SAS = South Asia; SSA = Sub-Saharan Africa AG = agricultural exports; MF = manufacturing exports; ALL = all exports Source: Authors’ calculations, based on Kee, et. al., 2009 Chapter 3. 71 3.4. Figures 3.1 and 3.2 show the trade restrictiveness indices (indicating market access barriers) fac- ing exports from LAC to regions on the horizontal axis, while figure 3.3 indicates the barriers facing exports from each region (the bars) being import- Figure 3.2. Market Access Tariff Trade Restrictiveness Index for Latin America ed into each region or country grouping on the and the Caribbean exports by region (percent) horizontal axis. Several conclusions can be drawn from these figures: LAC-All LAC-AG LAC-MF • As comparing the MA-TTRI with the MA-OTRI 60,0% demonstrates, NTBs have a substantial effect— in many cases, more than doubling the effects 50,0% of tariffs alone. For this reason, we focus on the MA-OTRI in the subsequent discussion. 40,0% • LAC agricultural exports face fairly high mar- 30,0% ket access barriers, particularly in low-income countries and South Asia. LAC manufactured 20,0% products face lower barriers, indicating that ag- ricultural exports suffer from an anti-agricultural 10,0% bias in external trade. 0,0% ALL HI EAP ECA LAC MENA SAS SSA Each bar is an index of the barriers to exports from the Latin America and the Caribbean to the region named below the bar. Note: Hi = high income; EAP = East Asia and Pacific; ECA = Transition Europe and Central Asia; MENA = Middle East and North Africa; SAS = South Asia; SSA = Sub-Saharan Africa AG = agricultural exports; MF = manufacturing exports; ALL = all exports Source: Authors’ calculations, based on Kee, et. al., 2009 72 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development • Globally, on average LAC’s agricultural exports face higher barriers (green bar above “All” in fig- ure 3.3) than those of any other region except East Asia and the Pacific. While tariff barriers are relatively low to agricultural trade within LAC, the overall restrictions facing LAC exports to oth- Figure 3.3. MA-OTRI for agricultural products by region (percent) er LAC countries are relatively high (the green bar above LAC), indicating significant NTBs. This HI EAP ECA LAC MENA SAS SSA corroborates the point made earlier that at least in agricultural products, regional—even intra- 120,0% regional—agreements have not succeeded in reducing barriers to low levels. 100,0% • LAC agricultural exports face the most severe 80,0% market access restrictions in South Asia, fol- lowed by the Middle East and North Africa, the 60,0% EU, Europe and Central Asia, LAC, Sub-Saharan 40,0% Africa, and East Asia and the Pacific. The ex- tremely high restrictiveness in South Asia main- 20,0% ly results from the high tariffs and NTBs levied by India on Brazilian exports of raw cane sugar 0,0% ALL HI EAP ECA LAC MENA SAS SSA (with tariff of 60 percent) and undenatured eth- yl alcohol (with tariff of 150 percent). LAC coun- Each bar is an index of the barriers to exports from the region represented by the bar to the region or group of tries face significant market access restrictions countries named below the bar. on their main agricultural exports of sugar, ba- Note: Hi = high income; EAP = East Asia and Pacific; ECA = Transition Europe and Central Asia; MENA = Middle East and nanas, and rice, which are among the most pro- North Africa; SAS = South Asia; SSA = Sub-Saharan Africa tected agricultural products in the world market. Source: Authors’ calculations, based on Kee, et. al., 2009 • Some of the Caribbean states (especially Guy- ana, St. Lucia, Dominica, and Cuba) face the highest barriers—50–55 percent. Among the major exporters, Uruguay faces the highest bar- riers (close to 50 percent), followed by Brazil Chapter 3. 73 (about 35 percent). For country details, see the assistance42 to agriculture yearly since the mid- annexes. 1960s. It found very negative rates in all develop- ing regions in the 1960s and 1970s (figure 3.4). 3.2 However, in the 1980s many countries embarked Trade policies of Latin America on structural reforms that began steadily eroding and the Caribbean countries the anti-export and anti-agricultural biases. After several decades of reforms, the incentive frame- 3.5. An economy’s overall incentive structure for ag- work is now perhaps the least distorted in the riculture depends on its own policies as well as world. Meanwhile, many countries in Africa have those of its trading partners. LAC, along with oth- continued to tax the sector, while developing coun- er developing regions, followed an industrial-led tries in Asia have followed the industrialized world import-substitution strategy in the 1960s and in subsidizing agricultural production, though not 1970s, with high trade barriers protecting do- to the same degree as high-income countries. LAC mestic manufacturing industries and implicit or stands out as the region in which the overall incen- explicit taxation of agriculture. This created a sig- tive structure has been close to neutral since the nificant anti-agricultural—and anti-export—bias early 1990s. in these countries’ incentive structures. A careful multicountry study calculated the relative rates of 42 Relative rates of assistance measure the support through trade and exchange rate policy and budget support accorded to agriculture, relative to similar support accorded to nonagriculture in the economy. 74 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 3.6. But this is not the whole story. In many develop- ing countries, trade policies and support programs Figure 3.4. Relative rates of assistance at the regional level, 1965–2009 have been much more protective of import-substi- tute products than of exportables. The Anderson AFRICA LAC ASIA study calculated the nominal rates of assistance to these two classes of goods and found a large 30 and persistent difference in almost all regions, creating an anti-export bias within agricultural 20 policy (figure 3.5). LAC is no exception. Further, while nominal rates of assistance on exports 10 have become less negative over time, those for import-substitute products have grown more posi- 1965-69 1970-74 1975-79 1980-84 1985-89 1990-94 1995-99 2000-04 2005-09 0 tive. Nonetheless, in LAC, this difference has great- ly diminished since the 1980s, indicating that the -10 anti-export bias has been reduced. On balance, while biases and distortions persist in some LAC -20 countries—mainly favoring import substitutes over export crops—the overall incentive structure -30 is relatively conducive to an efficient agricultural supply response to higher prices and appropriate -40 investments. -50 Source: Anderson and others for 1965–2004; authors’ own calculations for 2005–09, based on updated database for Anderson and others (http://go.worldbank.org/5XY7A7LH40). Note: The 2005–09 relative rate of assistance for Africa was heavily influenced by several countries (particularly Ethiopia) that provided high positive protection to agriculture, but this is not representative of the continent. A majority of countries had negative relative rates of assistance, as in earlier periods. Five-year weighted averages with value of production at undistorted prices as weights. Latin America and Caribbean countries in the study were Argentina, Brazil, Chile, Colombia, the Dominican Republic, Ecuador, Mexico, and Nicaragua. Chapter 3. 75 3.7. As Krueger and others’ large study of agricultur- al policy underscores, macroeconomic policy in many countries strongly influences the incentive structure facing agricultural production. Exchange Figure 3.5. Nominal rates of assistance for agricultural exportables and import substitutes rate policy has often been used to implicitly tax the in Latin America and the Caribbean, 1965–2009 sector. In the 2000s, good macroeconomic policy in many LAC countries generally maintained real Import-competing Total Exportables exchange rates at levels much more stable than in the past, avoiding large appreciations. In recent 30,0 years, however, with booming commodity prices and recent resource discoveries, exchange rates 20,0 have begun to appreciate in important exporters (figure 3.6), threatening the sector’s competitive- 10,0 ness. This trend may become more pronounced as resource production ramps up. 0,0 1965-69 1970-74 1975-79 1980-84 1985-89 1990-94 1995-99 2000-04 2005-09 -10,0 -20,0 -30,0 Source: Anderson and Valdes for 1965–2004; authors’ own calculations for 2005–09, based on updated database for Anderson and others (http://go.worldbank.org/5XY7A7LH40). Note: Five-year weighted averages with value of production at undistorted prices as weights. Latin America and Caribbean countries in the study were Argentina, Brazil, Chile, Colombia, the Dominican Republic, Ecuador, Mexico, and Nicaragua. 76 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 3.3 Trade agreements: the Latin spaghetti bowl43 3.8. Trade agreements affect both a country’s own trade policies and the external barriers facing its exports. Following the 1992 North America Free Trade Agreement (NAFTA), the countries of Lat- Figure 3.6. Real effective exchange rates, 1980–2010 (2005 = 100) in America have negotiated and notified to the World Trade Organization (WTO) almost three doz- Brazil LAC (mean) Colombia Mexico Argentine Chile en preferential trade agreements (PTAs).44 Mexi- 250,000 co and Chile have been most prolific in this area, each having agreements with several countries in 200,000 LAC as well as with the United States, the EU, and some Asian countries (Chile has a PTA with Chi- 150,000 na). Many of these agreements go beyond tariff reductions and address other trade issues, includ- 100,000 ing some relevant to agriculture, such as sanitary and phytosanitary (SPS) measures. Virtually all 50,000 agreements have some kind of phase-in period of 0,000 progressively reducing tariffs and subjecting more products to tariff reduction or elimination. 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 3.9. By any measure, these agreements liberalize agri- Source: Authors’ own calculations using data from Bank for International Settlements database for Argentina; IMF cultural trade less than nonagricultural trade. The database for others. General Agreement on Tariffs and Trade requires that PTAs eliminate tariffs on “substantially all 43 This section draws on Estevadeordal and others (2008). 44 Although most of these agreements have the term “free trade” in their names, even when they are fully phased in, trade is not completely free. It is thus more accurate to refer to them generically as PTAs, though when referring to specific agreements, we use their official designation, free trade agreement. Chapter 3. 77 trade” within a “reasonable length of time.” In a ucts (or inputs) from a country outside the PTA and details of each country’s PTAs, see table A2.3 in sample of PTAs in LAC, the agreements completed re-export products (or products made with the im- annex 2.) Many WTO Plus agreements reside not their implementation with shares of duty-free tar- ported inputs with only negligible value added) to in the PTA’s basic text but in ad hoc agreements iff lines of 97 percent, compared with their initial a partner in the PTA, taking advantage of the low signed during its negotiation. But even some of share of only 30 percent.45 This represents sub- duty. But these rules generally contain an exclu- these do not take full advantage of the opportuni- stantial liberalization. But on average, the percent- sion granting preferential treatment for products ty. For example, in an ad hoc agreement accom- age of products made duty-free at the beginning of made with some imported inputs, as long as these panying the Colombia–U.S. free trade agreement, the agreement is lower for agriculture, the phase- inputs do not exceed a threshold percentage of the Colombia commits to recognizing as equivalent the in period is longer, and the percentage of duty-free good’s final value. (for most PTAs, 7–10 percent). U.S. inspection system for meat and poultry and products remains lower—even at the end of the However, many agreements exclude agricultural accepting the United States Department of Agri- transition.46 A few, however—notably, Chile’s PTAs products from this “general tolerance rule”—that culture/Food Safety and Inspection Service Export with the United States, Mexico, and China—do is, they do not allow agricultural products to ben- Certificate, notwithstanding U.S. outbreaks of mad come close to complete liberalization. efit from the de minimis exemption. cow disease. (The Peru–United States agreement 3.10. The dairy and sugar subsectors are particularly 3.12. SPS standards have become more stringent, a also did this.) Yet it did not provide similar benefits noteworthy for their very limited liberalization. In trend driven by consumer concern over chemicals for Colombia, which had been trying for two years general, PTAs with countries in Asia and Europe in the food chain and transmission of food-borne to get the United States to accept regionalization, tend to have more exceptions to liberalization, disease, as well as their potential for providing non- allowing meat exports to the United States from though this is not the case in the Chile–China and transparent protection that is less constrained by regions of Colombia that had been declared free of Transpacific (Chile–New Zealand–Singapore– WTO rules than other measures. As a result, these foot-and-mouth disease. Brunei) free trade agreements. Many products standards are perhaps the most serious barrier to 3.13. A few agreements do include reciprocal commit- exempt from duty-free status are subject to tar- agricultural trade worldwide, making PTAs’ handling ments on other technical barriers to trade in agri- iff-rate quotas, whereby some limited quantity of them highly relevant. A PTA provides an oppor- culture. NAFTA and the U.S. agreements with Chile may be imported at a low or zero rate, with imports tunity for partners to go beyond the minimal WTO and Panama, as well as Mexico’s agreements with over this quota subject to a higher—often the most measures, particularly by operationalizing some Costa Rica, Bolivia, and Nicaragua, commit both favored nation—rate. core principles of the SPS Agreement: harmoniza- parties to respecting the principle of national 3.11. Most of these PTAs also have different rules of tion, equivalence, regionalization, evaluation of risk treatment on marketing requirements for packag- origin for agriculture. All PTAs have rules of origin and appropriate protection level, and transparency. ing, grading, and sizes. NAFTA and Mexico’s oth- that exclude from preferential treatment goods not Yet, in a sample of recent PTAs involving LAC coun- er agreements mentioned here also establish a made with local inputs in the country exporting tries, 48 percent contained no provisions beyond committee or working group to review how these them, so that exporters cannot import cheap prod- the basic WTO commitments, with only 52 percent standards affect trade and to resolve problems categorized as “WTO Plus.”47 (For a summary of the arising from them. Chile’s agreement with the Unit- 45 Fulponi and others 2011. ed States is the only one in which a Latin American 46 Shearer and others 2009. 47 Shearer and others 2009. 78 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development country receives a detailed commitment for mutu- which delay their full implementation. Determining point in the preferential tariff in a free trade agree- al recognition of a specific program— in this case, effects on welfare is also complicated by the fact ment, inducing a reduction of almost 0.2 percent- the grading program for beef. that even if a PTA clearly increased trade, it may age point in the most favored nation tariff in the 3.14. Most PTAs call for phasing out export subsidies on have actually reduced welfare if the trade was di- subsequent year.49 This paper did not, however, agricultural products, with the conspicuous excep- verted—that is, if imports previously sourced from look separately at agricultural products, which of- tions of those with European countries, Israel, the a very efficient producer outside the PTA now come ten receive special treatment in PTAs. Republic of Korea, and Japan, and those signed from a less efficient PTA member because tariffs 3.17. A few in-depth studies have looked at PTAs’ ef- in the 1990s with Mercosur and Central America. were lowered with PTA partners. This potential for di- fects on specific agricultural products. After NAFTA Many, however, do not specify a timetable for the version is one reason trade economists prefer multi- took effect, corn exports from the United States to phase out. Slightly more than a third of the agree- lateral trade liberalization (WTO) and are skeptical of Mexico boomed, as expected.50 But, contrary to ments regulate trade-distorting domestic support preferential agreements. In general, the higher the predictions that low-cost imports would crush the (for example, from government budgets) to the trade barriers with countries outside the preferential domestic growers, production in Mexico (mostly agricultural sector. Also, more than a third of the area, the more serious the problem of trade diver- white corn) also grew substantially.51 This out- agreements (all those involving the United States sion. Conversely, the more globally competitive the come arose partly because the imports, rather and most involving Mexico and Chile) include pro- trade partners within the agreement, the more trade than competing with domestic production, were visions for safeguard actions for some products, will increase, not be diverted. absorbed by robust demand for yellow corn to use allowing the importing country, when faced with 3.16. One recent Organization for Economic Co-opera- as livestock feed; Mexican consumers consumed high import levels of a product, to impose addition- tion and Development (OECD) study of three fairly more meat, and exports of beef from Mexico to the al, temporary protection. Products most common- long-standing PTAs—Mercosur, AFTA [the ASEAN United States grew from around $274,000 in 1993 ly mentioned in safeguards are beef, pork, dairy, Free Trade Agreement], and the Common Market to $2 million in 1995 to $89 million by 2007.52 poultry, fruits, and vegetables. for Eastern and Southern Africa—concluded that While this was dwarfed by the growth in U.S. meat 3.15. So, have PTAs liberalized trade in agricultural they stimulated agricultural trade among mem- exports to Mexico, it nonetheless increased Mex- products and improved member states’ welfare? bers and that the increased trade was not due ican use of feed corn substantially. And it is also Notwithstanding the large number of PTAs, few ro- to trade diversion. On the other hand, the agree- clear that the agreement stimulated exports of bust (that is, generally applicable) answers have ments did not increase trade with nonmember fruits and vegetables from Mexico to the United emerged from research attempting to answer this countries,48 a fact the study attributes at least in States. The effect of NAFTA-induced liberalization question. This is due to several factors: the highly part to poor infrastructure and supply constraints. specific nature of each agreement; difficulty es- Thus, the agreements have limited value. Another 49 Estevadeordal and others 2008. tablishing a counterfactual (what would have hap- interesting study of PTAs’ impact in LAC concluded 50 It is not entirely clear to what extent this can be attributed to NAFTA, pened to trade without the agreement) and thus that they have been building blocks to multilateral because the government reduced barriers much faster than required causality; and the lengthy phase-in periods of many liberalization, with a reduction of one percentage under the agreement. agreements (especially for agricultural products), 51 World Bank/IMCO 2007a. 48 Korinek and Melatos 2009. 52 Shearer and others 2009. Chapter 3. 79 in the beans market was less significant, due to is, in which specific national market) the product trade facilitation mean higher producer prices and the nature of bean demand in Mexico.53 And in is sold; what matters is the world market price. So greater competitiveness in world markets. But giv- the sugar market, NAFTA’s liberalizing impact was commodities blocked from one country by high tar- en food’s importance to the poor, improving logis- attenuated, or at least delayed, by disputes and iffs or NTBs are still saleable to buyers in other coun- tics and facilitation can also help reduce poverty countermeasures.54 tries at the same price. For differentiated products, by keeping prices lower for imported foods. This is 3.18. A case study of the Dominican Republic–Central on the other hand, the price and the ability to sell the especially critical in an era of very high prices for America Free Trade Agreement (CAFTA-DR) and the product may depend heavily on market characteris- many staples. United States showed that the agreement great- tics, so accessing the markets of specific countries 3.21. In the past, when formal agricultural trade barriers ly expanded fruit exports from the United States could be quite important. Second, tariff structures were high, logistics costs were often considered and likewise greatly expanded exports from the in most countries are escalated—that is, tariffs on of secondary importance. But now that many of CAFTA-DR countries to the United States of fresh more processed products are higher—so there is these formal barriers have been reduced—though fruits and vegetables, some sweeteners, nursery more space for PTAs to improve access for the more not nearly as much for agriculture as for manufac- products and flowers, and dairy products. On the processed goods. The results of the gravity model tures—these “behind the border” and international other hand, domestic political pressure from U.S. reported in subsection 3.4.1 suggest this: while all transport costs are increasingly being recognized sugar producers resulted in severe constraints on subcategories of agricultural exports show a signif- as impediments to trade. For LAC in particular, lo- liberalization in that market, so the agreement is icant positive effect of PTAs, the two with the high- gistics and transport costs are two to three times unlikely ever to have a significant impact on trade. est coefficients are processed products and meats. greater than tariffs and duties. Average import tar- 3.19. On balance, these case studies confirm the MA- (The fact that bulks have a slightly higher coefficient iffs for LAC have declined since the early 1990s OTRIs: notwithstanding the spaghetti bowl of agree- than fruits and vegetables seems odd, however.) and remain at an average of about 11 percent. ments among the LAC countries and with extra- National average logistics costs, however, repre- regional partners, agricultural trade barriers remain 3.4 sent a share of product value of 18–32 percent, relatively high for LAC exports, including within the Logistics and trade facilitation well above the OECD benchmark of around 9 per- region. For both intra- and extraregional trade, NTBs cent (figure 3.7). The international maritime and are more significant than tariffs. PTAs will likely be 3.20. Agricultural products’ competitiveness depends road haulage components alone can total about more important for differentiated products (pro- on many factors, from the field to the port. But one 20 percent of the free-on-board value of goods. By cessed products, fruits and vegetables, meats) critical area is the efficiency of the related logistics the time products are transferred, handled, stored, than for standardized commodities—for two rea- systems and the ability to connect effectively and and distributed domestically, the logistics compo- sons. First, for commodities, the price received by reliably to global supply chains. This section over- nent of the delivered good is often more than 50 the exporter does not depend much on where (that views the logistics challenges facing LAC countries percent of the final price to consumers, twice the and corresponding recommendations for how to levels of OECD countries. tackle them. Much of this report focuses on export 53 World Bank/IMCO 2007b. products, for which low logistics costs and good 54 World Bank/IMCO 2007c. 80 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 3.22. In this section, we look at the importance of trade facilitation and logistics from two perspectives. Subsection 3.4.1 takes a macro or “bird’s-eye” look at the importance of these factors in agricultural trade. It uses an augmented gravity model of trade to quantify the impact of various trade facilitation indicators on countries’ export performance, with a focus on LAC agricultural exports. It also looks Figure 3.7. Logistics costs in Latin America and the Caribbean and in Organization broadly at what improvements would be most use- for Economic Co-operation and Development countries, 2004 ful for agricultural and nonagricultural trade and for particular kinds of agricultural products. And final- OECD LAC ly, relying on the findings from the gravity model, it quantifies what impacts on trade would result from 30 improving LAC policies and infrastructure to OECD Productivity gain (percent) 25 standards. Subsection 3.4.2 then delves deeper at 20 a more micro level to look at specific challenges 15 faced by different LAC countries, benchmarking 10 them against each other and against other re- 5 gions. For this, the analysis relies on a variety of 0 micro-level studies and surveys, including some Losses to markets. Logistics cost as a Inventory levels Proportion of goods not proportion of value of very detailed studies of supply chains in Latin reaching markets products American countries. Subsection 3.4.3 concludes (primary sector) this chapter with policy implications. Source: Guasch, 2011 3.4.1 How important is trade facilitation?55 3.23. Trade facilitation consists of policies to reduce trade costs so that firms can expand their export 55 This section is drawn largely from a background paper for this study, “Trade Facilitation in Latin America and Agricultural Exports,” by Esteban Ferro and Alberto Portugal-Perez, both of the World Bank. Chapter 3. 81 volumes and reach new markets. We use two ag- pare the benefits of investment or policy reform ment and quality and is derived from four primary gregate indicators developed by Portugal-Perez along both dimensions. variables compiled from the World Economic Fo- and Wilson (2012) to measure the trade facilita- 3.25. Constructing trade facilitation indicators. Recent rum: quality of ports infrastructure, quality of air- tion environment in each country. The first indica- empirical research has used many trade facilitation ports infrastructure, quality of road infrastructure, tor quantifies the quality of the physical infrastruc- indicators at the country level and with extensive and quality of railroad infrastructure. ture, such as ports and roads, also known as hard geographic coverage to estimate their impact on 3.27. As a proxy for soft infrastructure, business and reg- infrastructure. The second indicator measures the trade. From an econometric point of view, including ulatory environment measures the development overall quality of the regulatory and business en- variables of trade facilitation, measuring similar level of regulations and transparency. The synthet- vironment, or soft infrastructure. In addition, we aspects on the right-hand side of a model, such as ic indicator is derived from five simple indicators use the number of days to export as a proxy for a gravity specification, can be conducive to multi- from the World Economic Forum and one indicator the overall logistics in each country. The number collinearity. To circumvent multicollinearity, we can from Transparency International. The World Eco- of days to export is a good measure of both in- reduce the dimension of the data by aggregating nomic Forum indicators are public trust for govern- side-the-border and at-the-border obstacles and highly correlated indicators into a single indicator. ments, irregular payments in exports and imports, controls for possible complementarities between With this in mind, a background study for this report irregular payments in public contracts, measures hard and soft infrastructure. borrowed two aggregate trade facilitation indicators to combat corruption, and favoritism of government 3.24. Using the estimates in the gravity model described derived by Portugal-Perez and Wilson (2012)57 us- to well-connected firms. The Transparency Interna- below, we simulate the impact of enhanced trade ing factor analysis: physical infrastructure and tional indicator is government transparency. facilitation, measured by higher indicators, on ex- business and regulatory environment. Both indica- 3.28. Gravity model. The study used a gravity model ports expansion in each LAC country. The results tors are standardized on a scale ranging from 0 to 1, to assess the impact of trade facilitation on trade are clear: if LAC countries could achieve the trade with 1 being the best-performing country. flows.58 The gravity model of international trade facilitation levels of the average OECD country, the 3.26. As a proxy for hard infrastructure, physical infra- flows is a common approach to model determi- region could increase agricultural exports around structure measures overall infrastructure develop- nants of bilateral trade flows. The standard gravity 140 percent and total exports nearly 100 percent. 58 Gravity models are mathematical models based on an analogy with This finding is consistent with a broad range of oth- 57 Portugal-Perez and Wilson (2012) derive four new aggregate indicators Newton’s gravitational law and used to account for aggregate human er studies that have shown the value of both hard of trade facilitation from a wide range of primary indicators using behaviors related to spatial interaction, such as migrations, traffic, and and soft infrastructure.56 By distinguishing soft in- factor analysis, a statistical modeling technique that explains the trade flows. The gravity model states that the volume of trade can be correlation among a set of observed variables through an unobserved estimated as an increasing function of the national incomes of trading frastructure from hard infrastructure we can com- “common factor.” The new aggregate indicators contain the informa- partners and a decreasing function of the distance between them. Al- tion of a wider range of individual indicators than previous studies and though the model’s ability to predict trade flows has been popular for 56 On the value of institutional quality in improving trade performance, focus on factors related to the hard or soft dimension of trade facilita- several decades, only recently has it been supported with theoretical see Levchenko (2004), Anderson and Marcoullier (2002), Ranjay and tion. The indicators are derived for 101 countries over 2004–07, greater foundations. These studies show that estimates can be derived from Lee (2003), and Depken and Sonora (2005). On the role of infrastruc- and more recent coverage than previous indicators. The indicators are various theoretical frameworks, such as the Ricardian, Heckscher–Oh- ture and business environment, see Limao and Venables (2001), derived from a pool of 20 primary indicators collected from different lin, and increasing returns to scale models. Theoretical foundations Wilson and others (2003, 2005), Djankov and others (2006), Francois sources: Doing Business, the World Development Indicators, the World for estimating gravity equations were developed by Anderson (1979) and Manchin (2007), and Portugal-Perez and Wilson (2012). Economic Forum, and Transparency International. and enhanced in Anderson and van Wincoop (2003, 2004). 82 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development formulation includes measures of market size (that nic similarities). The study augmented the basic were available from the World Bank’s World Devel- is, GDP, population), measures of remoteness (dis- model to include tariffs as well as the trade facilita- opment Indicators. As mentioned, the aggregate tance and adjacency), and measures of kinship tion indicators and some additional controls. More trade facilitation indicators for each exporter were (regional trade arrangements and language or eth- precisely, it estimated the following specification: taken from Portugal-Perez and Wilson (2012) and the days to export in each country were taken from ln ( X ij ) = β 0 + φ1 ln( Hard Infrai ) + φ2 ln( Soft Infrai ) + φ3 ln( Export daysi ) + β1 ln (1 + tij ) + the World Bank’s Doing Business survey. β 2 ln(GDPpci ) + β3 ln( Populationi ) + β 4 ln ( Dis tan ceij ) + β5 PTAij + β 6 LACi + 3.31. Estimation and results. Table 3.1 reports esti- mates for the gravity model defined by equation β 7 REER ij + β8 Borderij + β9 Languageij + β10 Colonyij + β11Common Colonizerij + τ j + ε ij (1), with each regression having three dependent variables—total exports, industrial exports, and 3.29. where Xij is country i’s exports to country j; Hard 3.30. Data. The dataset covers 100 exporting coun- agricultural exports—to evaluate the differenti- Infrai and Soft Infrai are our two infrastructure indi- tries and 137 importing countries over the period ated impact of each type of export to changes in cators for country i with higher values representing 2005–08, as most of our trade facilitation indica- the trade facilitation environment. In column 1, better performance; Export daysi is the number of tors only cover this period.60 Aggregate trade flows both hard and soft infrastructure have positive days it takes a standard cargo to move from the were compiled from the United Nations Commod- and significant coefficients, as improving both factory gate to the ship; tij, are tariffs on imports of ity and Trade Database (COMTRADE), whereas areas of trade facilitation would result in greater country j from country i; REER is the real exchange applied tariffs were compiled from the UNCTAD exports. The coefficient on the number of days to rate; PTAij is a dummy equal to 1 when countries i Trade Analysis and Information System (TRAINS) export is negative and significant, as a reduction and j have an active preferential trade agreement; database. Core gravity variables, such as bilater- in the time that a shipment takes from factory re- LAC is a dummy equal to 1 when the export is a LAC al distances, colonial ties, and common language lease to board shipment is associated with higher country; Borderij, Common Languageij, Colonyij, dummies were obtained from the Centre d’Etudes exports. The coefficient of the real exchange rate and Common Colonizerij are dummy variables that Prospectives et d’Informations Internationales (REER) is positive, as depreciation in the currency are equal to one if countries i and j, respectively, website. Compiled from Penn World Tables, the real is associated with exports surge. The positive and have common border, common language, same exchange rate is defined as the exporter’s nominal significant coefficient for the LAC dummy means colonizers, and common colonizers post-1945. exchange rate deflated by the country’s consumer that, all else equal, LAC countries export more than Finally, is a vector of importer-specific fixed ef- price index over the importer’s nominal exchange the average country. fects, whereas Ɛij is a random error term.59 rate deflated by its consumer price index. Other 3.32. The remaining traditional variables in gravity mod- relevant variables, such as GDP and population, els have signs and magnitudes consistent with the literature. Indeed, higher tariffs, as well as longer 59 A complete specification would also require fixed effects for exporters 60 Because there is almost no variation across time for the explana- distances between partners, discourage trade. to control for multilateral resistance terms, but their inclusion would tory variables, we decided to aggregate data across time and do a By contrast, the trade volume is higher between wipe out the effect of exporter-specific variables, including our vari- cross-section analysis in which the variation comes from the differ- partners in a regional trade agreement, as well as ables of interest, hard and soft infrastructure. ence between countries in trade flows and explanatory variables. Chapter 3. 83 between richer and more populous countries. Con- tiguous partners, countries with a common official Table 3.1. Gravity model estimates language, and countries having had a common col- onizer or a colonial relationship are also likely to Gravity Model Estimates trade more intensively.   (1) (2) (3)  Dependent Variable ln(total_exp) ln(indus_exp) ln(agro_exp) 3.33. Estimates for specifications explaining industrial ln (hard infra) 1.077 1.304 0.408 exports and agricultural exports reported in col-   [0.067]*** [0.070]*** [0.109]*** umns 2 and 3 of table 3.1 are fairly similar to es- ln (soft infra) 0.163 0.12 1.079   [0.052]*** [0.055]** [0.080]*** timates explaining total exports in column 1. Yet ln (1+tariff) -1.028 -1.646 -1.998 the greater difference is in the magnitude of coef-   [0.341]*** [0.349]*** [0.341]*** ficients for trade facilitation, as hard infrastructure PTA (d) 0.235 0.276 1.296   [0.064]*** [0.067]*** [0.096]*** has a greater impact on industrial exports, whereas REER 0.021 0.024 -0.011 soft infrastructure has a greater impact on agricul-   [0.004]*** [0.005]*** [0.007] tural exports.61 In addition, increasing the number LAC (d) 0.568 0.171 1.571 of days to export has a higher impact on agricul-   [0.053]*** [0.055]*** [0.078]*** ln (exp gdpcap) 1.026 1.224 0.432 tural exports than on industrial exports, because   [0.022]*** [0.023]*** [0.033]*** most agricultural products are perishable. Prefer- ln (exp population) 1.273 1.297 1.180 ential tariff agreements have a greater impact on   [0.012]*** [0.013]*** [0.019]*** agricultural exports, whereas real exchange rates ln (distance) -1.348 -1.460 -1.021   [0.032]*** [0.032]*** [0.049]*** have a greater impact on industrial exports. The Border (d) 0.916 0.900 1.130 LAC coefficient is higher for agricultural exports   [0.126]*** [0.131]*** [0.154]*** than for industrial exports, which indicates LAC’s Language (d) 0.610 0.560 0.837   [0.058]*** [0.058]*** [0.083]*** comparative advantage in agriculture.62 Colony (d) 0.422 0.422 0.895   [0.107]*** [0.114]*** [0.135]*** 61 The coefficients estimated in the industrial and agricultural regres- Common Colony (d) 0.835 0.964 0.347 sions (that is, columns 2 and 3 of table 3.1) are statistically different   [0.084]*** [0.083]*** [0.117]*** from each other with the exception of days to export. Constant -12.857 -13.190 -10.012 62 As there is no control for zero-trade in these regressions, any existing   [0.538]*** [0.591]*** [1.591]*** selection bias would tend to be greater in the agroexport regression, Observations 12,948 12,814 9,641 because only 70 percent of the possible importer-exporter combina- R-squared 0.776 0.782 0.524 tions actually trade, whereas for total and industrial exports almost Robust standard errors in brackets clustered by exporter-importer pairs. 94 percent of the possible country pairs have nonzero trade flows. We * significant at 10%; ** significant at 5%; *** significant at 1%. (d) for dummy variable. All regressions tested for the seriousness of this bias. A partial correction for selection include importer fixed effects. and firm heterogeneity bias shows that the results are nearly identical to the ones reported—evidence that the bias is small for our sample. 84 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 3.34. Table 3.2 shows the results of a gravity model of trade in which the dependent variables are exports of several subcategories of agricultural products.63 Table 3.2. Agricultural exports by product type As the dependent variables are exports of agricul- tural goods, we include two additional controls— Agricultural Exports by Product hectares of arable land and total renewable water   (1) (2) (3) (4) (5) (6) sources available in a country—as proxies for   Total Agro Meat Bulks Fruits & Veg Agro-ind Other Agro factor endowments intensively employed in agri- ln (hard infra) 0.778 0.855 0.931 0.942 0.829 0.299 cultural production.64 As a benchmark, we report   [0.113]*** [0.162]*** [0.163]*** [0.154]*** [0.132]*** [0.138]** estimates for total agricultural exports in column 1 ln (soft infra) 1.011 0.628 0.586 0.845 0.693 1.017 of table 3.2. Estimates show that trade facilitation   [0.080]*** [0.117]*** [0.120]*** [0.138]*** [0.091]*** [0.108]*** is relevant for exports of all agricultural categories. ln (arable land) 0.198 0.246 0.743 0.072 0.049 -0.062 Indeed, the coefficients of hard and soft infrastruc-   [0.029]*** [0.038]*** [0.042]*** [0.041]* [0.031] [0.038] ture are positive and statistically significant in all ln (water reserves) 0.225 0.338 0.124 0.092 0.138 0.297 regressions, whereas number of days to export is   [0.022]*** [0.028]*** [0.032]*** [0.028]*** [0.023]*** [0.027]*** always negative and statistically significant. The ln (1+tariff) -2.089 -1.554 0.587 -0.145 -2.300 -1.930 results are fairly consistent with previous findings:   [0.347]*** [0.621]** [0.576] [0.567] [0.434]*** [0.426]*** exports of heavier products such as industrial PTA (d) 1.044 0.844 0.778 0.719 1.262 0.714 products and “bulks” depend more on the quali-   [0.099]*** [0.126]*** [0.136]*** [0.130]*** [0.100]*** [0.119]*** ty of hard infrastructure, whereas time-sensitive REER -0.011 -0.048 0.019 -0.012 -0.020 -0.032 products depend more on soft infrastructure and   [0.007] [0.012]*** [0.010]* [0.009] [0.009]** [0.009]*** clearly on number of days to export. For example, LAC (d) 1.290 1.114 1.071 1.407 1.336 1.097 exports of meats, a highly perishable product, are   [0.083]*** [0.122]*** [0.131]*** [0.116]*** [0.088]*** [0.106]*** extremely sensitive to time delays. Observations 9,297 5,383 5,377 5,495 7,220 6,514 3.35. Simulations. The study also assessed the impact R-squared 0.541 0.431 0.415 0.437 0.531 0.432 of a potential improvement of trade facilitation Dependent variable is the natural logarithm of exports for the group of products in each column header. Robust standard errors in brackets clustered by exporter-importer pairs. * significant at 10%; ** significant at 5%; *** significant at 1%. (d) for dummy variable. All regressions include importer fixed effects. All regressions include all gravity variables but are omitted from the table due to space constraints. 63 The subcategories of agricultural products are defined by the follow- ing harmonized system categories. Meats: HS 02-04. Bulks: HS 10-12. Fruits and vegetables: HS 06-08. Agro industry: HS 16-24. Other agro: HS: 05, 09, 13–15. 64 Arable land was taken from the World Development Indicators and total renewable water source from FAO Aquastat. Chapter 3. 85 indicators for each LAC country to the level of the average OECD on total exports, industrial exports, and agricultural exports by simulating the gravity models estimated in table 3.1. 3.36. The average OECD measure of hard infrastructure in our sample is 0.77, compared with the average LAC hard infrastructure measure of 0.35. The ex- ercise performed for figure 3.8 is to assume that Figure 3.8. Increase in exports from an improvement in hard infrastructure to the Organization each country in LAC improves its quality of hard in- for Economic Co-operation and Development level (percent) frastructure to receive a mark of 0.77. Every coun- try in LAC would benefit from such an improve- Total Exports Industrial Exports Agro Exports ment. The average increase in LAC exports from an improvement in hard infrastructure to the level 600% of OECD countries is 106 percent for total exports, 500% 133 percent for industrial exports, and 28 percent for agricultural exports. While the specific quan- 400% titative estimates cannot be taken too seriously, clearly the benefit of an improvement in hard in- 300% frastructure would be greater for industrial exports 200% than for agricultural exports. 100% 0% ARG BOL BRA CHL COL CRI ECU GTM GUY HND JAM MEX NIC PAN PER PRY SLV URY VEN LAC Source: Background paper for this study by Ferro and Portugal 86 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 3.37. Figure 3.9 shows the results of an improvement in the soft infrastructure indicator to the level of the OECD average. The average soft infrastructure measure for LAC is 0.31, whereas the average grade for the OECD countries is 0.76. Achieving a grade in soft infrastructure as good as that avail- Figure 3.9. Increase in exports from an improvement in soft infrastructure to the Organization able in OECD countries will more than double LAC’s for Economic Co-operation and Development level agricultural exports. Again, the specific numbers cannot be taken too seriously, but the exercise Total Exports Industrial Exports Agro Exports indicates that even though the impact of improv- ing soft infrastructure is not as strong as that of 800% improving hard infrastructure in total exports, it 700% proves to be overwhelmingly important for agricul- tural exports. These estimates illustrate trade facil- 600% itation’s importance. However, two points should 500% be kept in mind. First, the gravity model does not consider any cost variable; therefore, we cannot 400% base any policy recommendation that would flow 300% from a cost-benefit analysis on this type of model. Clearly, the cost of improving, for example, Boliv- 200% ia’s hard infrastructure to the level of the average 100% OECD country will be exorbitant and bring dimin- ishing returns to investment. The main purpose 0% of this exercise is to identify trade constraints in HND GTM PAN ECU MEX BOL JAM ARG GUY COL CHL BRA URY VEN PER PRY LAC SLV NIC CRI LAC countries relative to other countries. Second, the large effect of trade facilitation on trade does Source: Background paper for this study by Ferro and Portugal-Perez not result from mistaking the relationship between facilitation and trade as linear. Because these ef- fects seemed quite large, we tested for the possi- bility of diminishing returns to trade facilitation by including a squared term for each trade facilitation variable. The coefficient on each squared term was Chapter 3. 87 positive (negative for days to export), indicating increasing rather than diminishing returns to trade facilitation. Thus, though large, our simulations’ re- sults are statistically conservative. 3.38. Another way to measure this effect is to esti- mate the tariff concessions that LAC countries would need to obtain from their trading partners to achieve a level of agricultural exports equal to those reached with the improvement in the trade Figure 3.10. Ad-valorem tariff-cut equivalents facilitation measures simulated above. To illus- trate how these counterfactuals are estimated, TF Infrastructure TF Business suppose that an investment in port infrastruc- 400% ture of an exporter country leads to a 1 percent increase in the hard infrastructure indicator. This 350% leads to a change in trade flows of about β ˆ 300% hard percent, according to the gravity estimates.65 The 250% same change in trade flows would be brought 200% about if all importers were to cut the tariffs applied 150% to imports from the country by an equivalent value ˆ β ˆ hard / βtariffs . Therefore, the latter ratio rough- 100% ly represents the “ad-valorem tariff-cut equivalent” 50% or “ad-valorem equivalent” of a 1 percent change 0% in the cost of export procedures inferred from grav- ARG BOL BRA CHL COL CRI ECU GTM GUY HND JAM MEX NIC PAN PER PRY SLV URY VEN LAC ity model estimates. Source: Background paper for this study by Ferro and Portugal-Perez 65 For notation purposes, let ˆ β be the estimated elasticity of imports x with respect to the variable X entering in the gravity equation. In the case of Doing Business days to export, the estimates should be nega- tive. 88 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 3.39. The results are reported in figure 3.10. Expansion the region is to best practice elsewhere to assess of agricultural exports from improving hard infra- the potential for improvement and looking at logis- structure quality to the level of the average OECD tics at a more granular level, both more focused country would also be feasible with a reduction on specific logistics and facilitation measures and of 13.5 percent in tariffs in importing countries. more country-specific. In this subsection, we com- If the improvement is made in soft infrastructure pare LAC countries to other regions and countries quality, the ensuing expansion of agricultural ex- in overall performance and identify priority areas ports would be feasible with a tariff reduction of for improvement according to subregion and coun- 68.7 percent. However, for many countries the try. tariff concessions needed to achieve the levels of 3.41. The logistics chain has many links, and inefficien- exports achieved by improvements in soft infra- cies in any of them fall disproportionately on poor structure are more than 100 percent, which would consumers and small producers. There are many be equivalent to an import subsidy from trading aspects to logistics and trade facilitation—ser- partners. vices, physical infrastructure, and procedures. The main logistics components discussed in this chap- 3.4.2 ter are illustrated in figure 3.11. Agricultural trade logistics in Latin America 3.42. And, while we saw in subsection 3.4.1 that LAC and the Caribbean: benchmarking and does not stack up well to the OECD countries, there breaking it down66 is a wide range among LAC countries. For example, food logistics costs for Peru, Argentina, and Brazil 3.40. As we saw in subsection 3.4.1, the big picture is are greater than 26 percent of product value, while that trade logistics—both hard and soft infrastruc- Chile, a regional leader in logistics, has costs of ture—matter a lot for agriculture and deserve to be about 18 percent, still double that of the OECD (fig- at or near the top of the trade policy priority list. But ure 3.12) to transform this overarching policy message into an actionable agenda requires seeing how close 66 This subsection is drawn largely from a background paper by Gwyn Fries, Jordan Schwartz, and Adam Stern. Chapter 3. 89 3.43. Although the logistics burden affects consumers, importers, and the competitiveness of LAC’s firms across the region, the impact is greatest on the region’s poor consumers and on small producers. Food expenditures make up a large part of the disposable income of the region’s poor. Across in- come levels in LAC, food is the primary purchase of households, accounting for 20–30 percent of all Figure 3.11. Typical logistics chain for agriculture exports spending. The poor, however, spend up to 70 per- cent of household income on food. Further, about half the region’s poor are now urban dwellers, no 1. Inland transport 2. Points of import/export 3. Maritime transport longer able to rely on subsistence farming. 3.44. Small firms, which are the majority of firms in Trucking from farm gates Cargo handling, storage, LAC countries and the region’s employment and Services and production zones to consolidating and Ocean shipping: ports/airports forwarding bulk and containerized growth engines, also suffer disproportionately from high logistics costs.67 Perishable agricultural products have unique characteristics that require Roads, Rail Roads, Silos, Ports: berth, cranes, Border crossing: lanes, specialized logistics systems, including remote Infrastructure Refrigeration Storage, apron, wharf, container scanners, storage area and Warehousing storage production zones, temperature control, and spe- cial sanitary inspection procedures. Because Transit regulation, Customs, inspections, of the time sensitivity of perishable agricultural municipal taxes, and phytosanitary clearances Container security, Procedures letters of credit and fumigation insurance goods, bottlenecks in the logistics system directly affect the quality and quantity of goods delivered. For nonperishable products, delays often result in Source: Adapted from Schwartz and others (2009).. increased logistics expenses for labor, fuel, and storage, as well as fees or fines for delays and de- murrage. Remote production zones incur higher costs and greater losses for the first actors along the supply chain, the farmers themselves. Most perishable products cannot be consolidated easily with other types of cargo, including other refriger- 67 Schwartz and others 2009. 90 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development ated cargo. SPS systems are necessarily complex, involving coordination with customs agencies and other inspection and regulatory agents operating at borders and ports. As a result of these character- istics, smaller producers and local agriculture trad- ers are often most heavily affected by poor-quality roads and uncompetitive trucking services. By contrast, large shippers benefit from integrated supply chains, greater access to the primary trade Figure 3.12 Logistics cost as percentage of food product value, 2004 (percent) corridors, and better berth access at ports. 3.45. While food products moving from producer to con- 35 sumer incur a number of costs outside the logis- 32 30 tics chain, such as profits, customs duties, and 27 26 23 taxes, the transport and logistics components of 25 20 the cost of delivered food products can be stagger- Percentage 20 18 ingly high. For some products, domestic logistics 15 costs in LAC may be the largest single cost ele- 9 ment of a good’s final price. While there are import- 10 ant variances by subregion and type of firm, LAC’s 5 logistics costs are very sensitive to the firm’s size. 0 For small mills, markets, and retailers of foods in Peru Aregntina Brazil Colombia Mexico Chile OECD towns and secondary cities around LAC, the impli- cations are clear: domestic logistics costs can total Source: Gonzales and others 2008. more than 42 percent of the price of a firm’s sales (figure 3.13). By comparison, larger firms spend 15–18 percent of sales on logistics. This is driven by such factors as lack of access to warehousing, storage, and transfer facilities and the quality of the infrastructure and trucking services that link rural markets, smaller towns, and secondary cities to large production and consumption centers.68 68 Schwartz and others 2009. Chapter 3. 91 3.46. Transport costs tend to rise and fall with food pric- es. For example, while commodity prices rose sharply in the decade leading to the 2008–09 global recession, freight rates as a share of the value of food products remained fairly steady. Over 2000–07, trucking costs relating specifi- cally to food products imported into LAC rose 40 Figure 3.13. Average logistics costs as percentage of sales, by sales volume percent, while maritime costs doubled and air shipment costs of high value food products rose Inventory Management & Warehousing Transport & Distribution even more.69 Even though transport rates dropped sharply during the last recession due to reduced 45 demand for transport and logistics services, the 40 cost of logistics has historically risen along with 35 12.7 international price changes for major food indices. 30 3.47. LAC logistics performance in a global perspective. Most LAC countries have significantly improved 25 5.4 their trade facilitation environments. However, 20 exporting firms in the region still face high trade 15 29.4 6.8 8 costs. According to the Global Competitiveness 6.6 8.7 11 10 19.6 Report 2011–2012, the two persistent challenges 5 8.3 11 10.3 that affect the region are poor infrastructure devel- 8.8 6.9 0 opment and weak institutions with high costs from Less than US$ 5 M to US$ 10 M to US$ 50 M to US$ 100 M to US$ 300 M to More than lack of physical security. The World Bank’s Doing US$ 5 M US$ 10 M US$ 50 M US$ 100 M US$ 300 M US$ 500 M US$ 500 M Business 2012 places LAC well behind the global Source: Centro Logistica de Latinoamerica, 2007. best practice for trading across borders. According to the survey, the region takes, on average, nearly twice the number of days to export than high-in- come OECD economies. 3.48. Figure 3.14 depicts LAC’s performance relative to other regions as measured by physical infrastruc- ture and business environment. LAC’s physical 69 UNECLAC 2012. 92 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development infrastructure indicator is better only than Sub-Sa- haran Africa’s. Even among LAC countries, how- ever, there is great variability—Panama and Chile have infrastructure levels that reach those of OECD Figure 3.14. Regional comparisons of physical infrastructure and business environment countries, whereas the region’s landlocked coun- tries are the worst performers. The region also un- Physical Infrastructure by Region derperforms in its business environment indicator. .8 According to the indicator depicted on the right panel of figure 3.14, LAC’s business environment .6 is only half as good as that of OECD countries, and better only than those of Sub-Saharan Africa and .4 South Asia among developing regions. The region’s .2 best performer is Chile. 3.49. LAC’s inadequate infrastructure is a major factor 0 underlying its consistently poor performance in SSA LAC ECA SAS EAP MNA OECD global competitiveness. The World Economic Fo- rum’s Growth and Business Competitiveness Index Business Environment by Region and the World Bank’s investment climate assess- .8 ments, for instance, have found that the majority of surveyed firms identify poor infrastructure as a .6 main obstacle to their operation and growth. One measure of particular interest to agriculture—the .4 Rural Access Index, which measures the percent- .2 age of the rural population living within 2 kilome- ters of an all-season road70—shows LAC lagging 0 behind East Asia and other middle-income coun- SAS SSA LAC ECA EAP MNA OECD tries along this dimension (figure 3.15). Inade- quate access to the road network translates into Note: EAP = East Asia and Pacific; ECA = Transition Europe and Central Asia; MENA = Middle East and North Africa; SAS = South Asia; SSA = Sub-Saharan Africa increased costs, losses, and delays; consequenc- Source: Logistics Performance Index 2010. http://go.worldbank.org/OLLQVB8NQ0 es are especially severe for perishable goods. 70 A road that is passable year-round by the existing means of rural transport, normally a pick-up truck or truck without four-wheel drive. Chapter 3. 93 3.50. There is evidence of an infrastructure gap in road quantity (road density) and quality (percentage of paved roads) between LAC and other regions since 1990. In 2005, less than 25 percent of the road network in an average LAC country was paved, far lower than the non-LAC middle-income country average of close to 50 percent. This gap has likely Figure 3.15. Global comparison of rural access to roads, 2003 (percent) affected the logistics performance and compet- itiveness of the LAC region. LAC’s road density in Percentage 2005 had hardly increased since the early 1980s, and it is therefore well below that of non-LAC mid- East Asia dle-income countries. Road density in the Andean 94 region is below the LAC-wide norm.71 Europe & Central Asia 75 3.51. The Logistics Performance Index (LPI) shows that LAC’s logistics performance fares poorly compared South Asia 58 with those of high- and upper middle-income countries, though reasonably well in comparison Latin America & the Caribbean 54 with those of other developing regions.72 As ta- ble 3.3 shows, LAC’s overall LPI score of 2.74 (on Middle East & North Africa 34 a 5-point scale) is similar to those of Europe and Central Asia and East Asia and the Pacific. When Sud-Saharian Africa 30 comparing LAC’s LPI score with those of other income groups, we see that LAC is closest to the 20 30 40 50 60 70 80 90 100 lower middle-income (2.59) group. LAC performs Source: Roberts and others 2006. 71 Calderón and Servén 2010. 72 The LPI provides both quantitative and qualitative evaluations of a country in six areas: efficiency of the clearance process (speed, sim- plicity, and predictability of formalities) by border control agencies; quality of trade- and transport-related infrastructure (ports, railroads, roads, and information technology); ease of arranging competitive- ly priced shipments; competence and quality of logistics services (transport operators, customs brokers); ability to track and trace consignments; and timeliness of shipments. 94 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development poorly compared with the upper middle-income group and many Asian countries, including China (3.5), Thailand (3.3), Indonesia (2.8), and Singa- pore (4.1). Table 3.3. Logistics Performance Index international regional comparison, 2010 data income group International Performance Tracking and Infrastructure competence Timeliness shipments Region or Logistics Logistics Customs tracing Index Europe and Central Asia 2.74 2.35 2.41 2.92 2.6 2.75 3.33 Latin America and the Caribbean 2.74 2.38 2.46 2.7 2.62 2.84 3.41 East Asia and the Pacific 2.73 2.41 2.46 2.79 2.58 2.74 3.33 Middle East and North Africa 2.60 2.33 2.36 2.65 2.53 2.46 3.22 South Asia 2.49 2.22 2.13 2.61 2.33 2.53 3.04 Sub-Saharan Africa 2.42 2.18 2.05 2.51 2.28 2.49 2.94 High income: all 3.55 3.36 3.56 3.28 3.5 3.65 3.98 Upper middle income (except LAC) 2.95 2.49 2.54 2.86 2.71 2.89 3.36 Lower middle income 2.59 2.23 2.27 2.66 2.48 2.58 3.24 Low income 2.43 2.19 2.06 2.54 2.25 2.47 2.98 Source: World Bank Logistics Performance Index. http://go.worldbank.org/OLLQVB8NQ0 Chapter 3. 95 3.52. The LPI also reveals that logistics performance is quite heterogeneous within LAC. Table 3.4 shows Mexico obtaining the highest LPI score (3.05), Table 3.4. Latin America and the Caribbean subregional comparison while the Caribbean scores lowest (2.46). While on the Logistics Performance Index, 2010 the regional average for clearance time with physi- cal inspection is 3.5 days, for instance, it takes 13 Infrastructure International Performance Tracking and competence days in Venezuela, 5.5 days in Brazil, and 2 days shipments Subregion Timelines Logistics Logistics Customs in the Dominican Republic. Another example is a tracing Index typical charge for a 40-foot export container or a semitrailer, which costs $1,500 in Bolivia, $884 in Mexico 3.05 2.55 2.95 2.83 3.04 3.28 3.66 Mexico, and $383 in Panama. Taking a closer look Southern Cone 2.98 2.60 2.75 2.89 2.89 3.08 3.66 at individual indicators, we see no big surprises. Central America 2.76 2.47 2.42 2.53 2.65 2.85 3.61 Poorly performing regions tend to maintain their Andean region 2.71 2.33 2.46 2.75 2.57 2.74 3.34 low ranking across indicators. Mexico is a top per- Caribbean 2.46 2.09 2.09 2.64 2.27 2.60 2.96 former and scores particularly well when it comes Regional average 2.74 2.38 2.45 2.70 2.62 2.84 3.41 to logistics competence and infrastructure. 3.53. The LPI also illustrates that overall logistics per- Source: World Bank Logistics Performance Index (LPI). http://go.worldbank.org/OLLQVB8NQ0 formance has improved in the LAC region, though more so over 2007–10 than over 2010–12. Mex- ico and the Southern Cone and Andean countries Table 3.5. Subregional Logistics Performance Index score trends have made the most progress, while the Central America and Caribbean subregions have fallen Subregion 2007 2010 2012 back since 2010 (table 3.5). Mexico 2.87 3.05 3.06 Southern Cone 2.81 2.98 2.96 Andean region 2.56 2.71 2.73 Central America 2.56 2.76 2.72 Caribbean 2.33 2.51 2.41 Regional average 2.63 2.80 2.78 Source: Authors’ calculations based on the World Bank Logistics Performance Index (LPI). http://go.worldbank.org/OLLQVB8NQ0 96 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 3.54. In all business survey-based reviews, LAC per- forms considerably worse than OECD standards Figure 3.16. Cost to export and import—global comparison, 2011 ($ per container) when it comes to export and import costs. Costs associated with required export and import pro- Cost to import Cost to export cedures include the costs for documents, admin- OECD high income istrative fees for customs clearance and technical Sub-Saharan Africa control, customs broker fees, terminal handling Eastern Europe & Central Asia charges, and inland transport. The Doing Business South Asia indicators reveal that LAC’s average cost to export Latin America & Caribbean a container is $1,257, and the average cost to im- Middle East & North Africa port one is $1,546 (figure 3.16). These costs are East Asia & Pacific lower than their equivalents in Sub-Saharan Africa, 500 700 900 1.100 1.300 1.500 1.700 1.900 2.100 2.300 2.500 Europe and Central Asia, and South Asia, though US$ still higher than those of other developing regions, such as East Asia and the Pacific, and the OECD av- Source: World Bank LPI 2012. http://go.worldbank.org/OLLQVB8NQ0 erage. 3.55. Within LAC, costs to export a container are low- Figure 3.17. Cost to export and import—regional comparison, 2011 ($ per container) est in Central America and highest in the Andean region, with cost to export at $1,720 and cost to Cost to import Cost to export import at $1,951 (figure 3.17). Andean Pact Mexico Caribbean Southern Cone LAC Average Central America 1.000 1.100 1.200 1.300 1.400 1.500 1.600 1.700 1.800 1.900 2.000 US$ Source: World Bank LPI 2012. http://go.worldbank.org/OLLQVB8NQ0 Chapter 3. 97 3.4.3 Below is a brief description of the various compo- vide duty assessments, transit control, and risk Drilling down: What are the most important nents of export procedures: management. logistics and trade facilitation challenges at • Customs is the authority responsible for collect- 3.58. Central America and Mexico. As global indicators the subregional and country levels? ing customs duties and controlling the flow of show, Central America performs poorly in logistics goods in and out of the country. Customs proce- infrastructure and services. Primary logistics chal- 3.56. LAC subregions demonstrate tremendous hetero- dures include the completion of required docu- lenges for agricultural trade in Central America in- geneity in logistics burdens, with different food mentation and payment of duties. clude the overall poor quality of the road network, and agricultural products and different trading • Customs brokers are experienced with local including principal trade corridors, secondary and modalities. This subsection identifies key logistics customs regulations and trade practices and are rural roads, and overly burdensome export proce- challenges that impact agricultural trade, focusing thus able to assist with the customs process by dures. Many ports are inefficient due to poor man- on those that available literature and expert opin- preparing customs documentation, calculating agement and lack of infrastructure and equipment. ion have identified and measured over the past 20 duties, and communicating with customs au- Mexico performs relatively well, though room years. The challenges presented here are by no thorities. for improvement exists in rural connectivity and means exhaustive, but they suggest some priority • SPS measures are taken to ensure that goods transport costs. areas. entering a country are safe for consumers, and 3.59. Road quality. Approximately 40 percent of all food 3.57. Logistics constraints can be loosely organized into that exported products are free of contaminants imports into LAC are handled using road transport. ground transport, maritime transport, and export and abide by The Codex Alimentarius Commis- Moreover, most domestic shipments and a large procedures. Ground transport and maritime trans- sion (the international framework for food safe- share of inputs to food exports are transported by port correspond roughly to the hard infrastructure ty, operated by the Food and Agriculture Organi- road. Poor road quality is therefore emerging as the categories in the previous discussion, though they zation and the World Health Organization). SPS biggest threat to affordable and reliable delivery of include some aspects of logistics services. Export procedures include documentation, sampling, goods in LAC. High postharvest losses result from procedures correspond to soft infrastructure. By testing, and fumigation (and sometimes treat- lack of road maintenance in the region, especially ground transport, we refer to the transport of goods ment and quarantine). on secondary and access roads (and from lack of by trucks from farm gates and production zones to • Information and communication technology cooling capacity). Access even to local markets is ports and airports. Maritime transport refers to the enables users to access, create, and transmit limited, not to mention regional and national mar- transport of goods from the port of the exporting information, promoting virtual integration of kets.73 country to the port of the importing country. While supply chains. Information and communication 3.60. Central American countries suffer from some of the the cost of shipping may not appear to have much technology services include Internet-based worst roads in LAC. The Global Competitiveness In- policy relevance, reforms that increase a country’s tools for transport and logistics services and connectivity can actually reduce these costs. Fi- assist customs authorities with their clearance nally, border procedures refer to the various proce- operations through computer systems that pro- dures of exporting products outside LAC countries. 73 Guasch 2011. 98 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development dex illustrates this clearly. 74 Figure A2.5 in annex Guatemala, the local trucking industry’s aged fleet, which are determined, in turn, by road quality and 2 displays country-by-country ratings. Of the 142 discrimination of foreign trucks, and lack of ade- access.78 countries surveyed, none in Central America rank quate regulations for vehicle and driver operations 3.65. In Mexico, transportation costs are significant. In in the top 40 in road quality, while some perform result in transport interruptions for about 30 per- the case of tomato exports to the United States, especially poorly, such as Honduras, Nicaragua, cent of firms, which consequently incur losses of for instance, transportation costs represent 15.3 and Costa Rica. an average of 1.6 percent of sales.76 percent of the selling price in Chicago, while tar- 3.61. In Costa Rica, logistics costs represent 24 percent 3.63. Poor road quality leads to high domestic transport iffs represent only 2.3 percent. Transaction costs of the price of domestically produced processed costs. Supply chain analyses of grain imports into associated with the commercialization of toma- foods. Among firms surveyed, 80 percent of re- Nicaragua and Honduras reveal that within total toes from the state of Sinaloa (located in Mexico’s spondents identified road quality as one of the transport costs, domestic transport costs repre- northwest) represent 34.1 percent of the final three main impediments for their business. The sented the largest share as a percentage of goods’ price of the good sold in Chicago. Domestic trans- country’s poor-quality road network also caus- final price. For corn chains particularly, domestic port costs alone from Sinaloa to Mexico City repre- es direct losses from delays in shipments and transport costs in Nicaragua are higher than the sent 17.9 percent of the final price of tomatoes.79 breakage of 8–12 percent of the sales value of U.S. transport, ocean transport, and other logistics 3.66. Export procedures. Poor interagency coordination exported goods. While Costa Rica has 30 percent costs combined (30 percent versus 18 percent). of border management procedures, including doc- more paved roads per worker than the next most This analysis also suggests that transportation umentation, monitoring, and inspection, puts LAC densely paved country in LAC, only 32 percent of and logistics costs for wheat, rice, and corn repre- at a disadvantage just as monitoring and traceabil- its paved roads are of good quality.75 sent a large share of the final price of goods sold ity requirements for international trade are becom- 3.62. The impact of transport interruptions is particu- at wholesale, ranging from 29 percent to 48 per- ing stricter. At entry points, multiple agencies have larly significant for firms whose  main market is cent.77 overlapping mandates to protect domestic securi- within Central America: they face a loss of about 3.64. These supply chain analyses also suggest that ty and inspect for microbiological contaminants, 5.3 percent of their consignment value. In Hondu- market size and distance are not main drivers of pests, diseases and public health risks, illegal ras, close to 40 percent of local firms incur losses transport costs; infrastructure quality and compe- drugs, and money. Where information capture and while merchandise is in transport due to spoilage tition among transport providers contribute more. inspection responsibilities are not coordinated, or breakage resulting from poor road quality and Surveys of local mills point out that the cost differ- they are multiplied, sometimes adding three days lack of enforcement of trucking regulations. In ence results mainly from trucks’ turnaround times, to total transit times. Unique sanitary inspection 74 This index combines quantitative and qualitative measures and is the 78 The majority of food producers and traders in Central America are largest poll of its kind; in 2011 insights from over 15,000 individuals required to use unpaved roads, some of which become inaccessible to were collected to examine their business operating environment. Re- vehicles of five tons or more during the rainy season. Since only pick- spondents were asked, “How would you assess roads in your country up trucks are useful during the rainy season, food transport from field (from 0 to 7, with 7 being the best possible score)?” 76 Schwartz and others 2009. to market becomes slow and costly. 75 Schwartz and others 2009. 77 LCSSD Economics Unit 2010. 79 Kjöllerström 2004. Chapter 3. 99 procedures make this duplication especially tax- customs clearance and fumigation costs repre- broker or losses that may result from exposure to ing on the times and costs for agricultural trade. sent the most important subcomponent, together open air while samples are taken. Central American countries, in particular, may be with silage and warehousing costs at the mill for 3.70. Reforms in border procedures can incease effi- losing international buyers for fresh agricultural wheat and corn imports into both Honduras and ciency. In El Salvador, for instance, an ongoing exports because of the uncertainties and time de- Nicaragua. These costs occupy the largest share World Bank project has been attempting to consol- lays that result from this poor interagency coordi- as a percentage of the final wholesale price of idate all border control agencies into a single office nation at the border. these goods. In Nicaragua, milled rice importers and implement information and communication 3.67. Customs in many LAC countries still impose heavy criticize the SPS process they are required to go technology systems. This project has already re- fines for minor errors. In the Port of Santo Tomás through, whereby a sample of their imported prod- sulted in improved agency coordination and in- in Guatemala, for instance, customs inspection uct is sent to be examined in Managua and takes creased efficiency. In Mexico, customs operations consists of a light indicator system that randomly up to a day for results to arrive. This often results were improved by implementing a single window. determines the cargo to be inspected. The system in a delay of up to 48 hours, at a cost of $110 per These examples show that measures can be taken was designed to allow for a 15 percent inspection metric ton per day while the product is held up at to improve border procedures, which in turn facili- rate, but the actual rate is 90–100 percent. Con- the border.81 tate trade and increase competitiveness. sequently, containers spend 7 days beyond the 3.69. An Honduras dairy export supply chain analysis 3.71. Port infrastructure and services. Maritime shipping 5-day free storage period, after which the shipper found that by the time a refrigerated dairy con- costs are related to logistics conditions through must incur storage charges. In addition, customs tainer undergoes a duplicate inspection at the El the connectivity of ports and the time in port. In finds that about 50 percent of the containers have Salvador border, logistics expenses may increase Central America, time of port stay is the import- discrepancies, which translates into the shipper by up to $900—and the time in transit from two ant driver of transportation costs. A comparison having to pay a bond for the amount of the discrep- days to nine. One exporter reported that phytosan- of total ship costs for a port stay of a similar ship ancy. Until the bond is paid and the administrative itary inspection duplication at the Amatillo border between Limón, Costa Rica, and Cartagena, Co- process to resolve the discrepancy is finalized, with El Salvador repeatedly added seven days and lombia, highlights time’s importance. The Port of the shipper has to continue paying the storage $400 in additional expenses to the logistics costs Limón posts a $28 per 20-foot-equivalent unit charges.80 to export a container. In addition, dairy exporters cost advantage over Cartagena based solely on 3.68. SPS procedures for imports often imply unneces- and customs brokers report charges of $400– port fees to cargo and vessel. However, this cost sary delays and administrative hassles that can $500 per lot to take a sample, transport it to the advantage disappears when time is considered in raise overall logistics costs and make product de- laboratory in San Salvador, and perform the rele- the cost equation, leaving Limón with a cost per livery unreliable. A supply chain analysis of grain vant sanitary exams; this estimated cost increase 20-foot-equivalent unit $111 higher than that for imports from the United States to Nicaragua and does not factor in increased fees for the customs Cartagena. Honduras illustrates that reception at the port and 3.72. Within Central America, quality of port infrastruc- ture varies greatly. While Panama is considered to have world-class port infrastructure, Nicaragua 80 Schwartz and others 2009. 81 LCSSD Economics Unit 2010. 100 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development and Costa Rica exhibit some of the lowest scores port-of-entry requirements and procedures. For ly for intraregional trade. In addition, as noted in when it comes to quality of port infrastructure (see instance, rather than shipping grain imports from chapter 2, Argentina and Brazil will face problems figure A2.6 in annex 2, which shows the ratings of New Orleans through Puerto Cortes on the Atlantic from insufficient port and storage capacity if their each Central American country and Mexico). 82 side of Honduras and then by land to Nicaragua, high growth rates continue uninterrupted. 3.73. As relative efficiency declines at a port, cargo is a three-day journey, Nicaraguan importers choose 3.76. Ground transport. A heavy reliance on road trans- chased away to other ports. This, in turn, sends to ship their goods through the Panama Canal to port, coupled with a poor record of maintenance a port into a downward spiral, affecting cargo ag- Puerto Corinto on the Pacific Ocean—a trip that and rehabilitation, characterizes the region. Al- glomeration and diminishing connectivity or reg- takes 12 days—all to avoid unnecessary adminis- though a number of South American countries ularity of ocean service. In 2005, approximately trative hassles related with entry at Puerto Cortés have doubled their length of paved roads in the 60,000 containers of Costa Rican origin or destina- and crossing the land border. These bottlenecks in- past three decades (Argentina and Brazil tripled tion—equaling 100,000 20-foot-equivalent units, clude waiting times for the receipt of the bill of lad- their road network), the regional transport system or 15 percent of the country’s container cargo— ing from the United States (3–5 days), processing as a whole did not improve. Not only did the main traveled an extra 200 kilometers over poor-con- of import documentation (5–7 days), and delays existing highways receive the bulk of the invest- dition roads to avoid the Port of Limón’s conges- at the El Guasaule border point (up to 2 days). ment, but there was insufficient emphasis on con- tion and inefficiency and to seek better services 3.75. Southern Cone and Brazil. Within LAC, the Southern nectivity between countries.84 Main roads receive through Panama’s ports. This represented about Cone is the best performer across many indicators, most of the paving investment and are usually in $1,500 of extra road haulage fees per container particularly in export and import procedures for good condition, while provincial and rural roads are for Costa Rican importers and exporters, for a year- agricultural products. Chile is a regional bench- often unpaved or simply neglected and in need of ly total of $70–$100 million in additional trucking mark for the exportation of fresh agricultural prod- rehabilitation.85 Due to the decline in road main- costs. In the case of Limón, firms surveyed men- ucts, including highly perishable products like fish tenance investment since the 1980s, merely 25 tioned as main bottlenecks the deterioration of and grapes. Argentina, Uruguay, and some south- percent of Brazil’s road network in 2007 was rated access roads and delays in the loading process ern states of Brazil also show fairly strong perfor- caused by the availability of only one crane. Cos- mance, while Paraguay seems to lag. For Southern 84 FAO 2008. In Brazil, highways in good condition are concentrated in the south and southeastern areas. ta Rica’s liner shipping connectivity is below the Cone countries, the literature singles out high costs 85 The only quality indicator widely available for road quality—the per- potential of a country with its income level, which for trucking services as the primary constraint for centage of paved roads in the total road network—is not a completely both causes and results from the Costa Rican car- ground transport, though Brazil has room for im- accurate or appropriate measure for a number of reasons: many countries in the region do not keep full count of their unpaved roads, go slipping away to Panama.83 provement with road network density and rural distorting the measurement; many paved roads are poorly maintained 3.74. The supply chain analyses of grain imports into connectivity. Specific examples from this region and thus of poor quality; many nonpaved roads are well maintained Central America reveal challenges concerning point to a common problem throughout LAC: high and thus of high quality; sometimes it is appropriate not to pave a road. Another measure of road quality is the international roughness costs are exacerbated by “over costs,” or logistics index, which evaluates pavement roughness by calculating the ratio of 82 Quality of port infrastructure is based on business executives’ percep- costs that are not actual, paid-out expenses. These a standard vehicle’s accumulated suspension motion divided by the tions of their country’s port facilities. distance traveled by the vehicle during the measurement. But data for include time delays, losses, and bribes, particular- 83 Schwartz and others 2009. this indicator are very limited. Chapter 3. 101 as in good or very good condition. Consequently, transporting low-cost products, such as seeds, but higher levels of congestion on the road network, the operating costs of trucking services in Brazil these independent truck drivers struggle to secure fewer scale economies in shipping, and higher have increased 10–30 percent, depending on the a steady volume.88 emissions from transport. region.86 3.79. The need for additional storage is also a conse- 3.81. Customs clearance and border crossings. A 2006 3.77. Dependence on ground transport increases trans- quence of the shortage of efficient intermodal study of intraregional food trade quantified the port costs for commodities from rural production transfer terminals. For instance, Brazil’s warehous- logistics costs associated with meat imports into zones in Brazil. A recent corn supply chain anal- ing shortage alone is about 40 million tons a year. Chile and soybean imports into Brazil from Par- ysis conducted in western Bahia state showed If Brazil were to double its number of intermodal aguay. Besides the anticipated logistics costs, how a multimodel transport system, using the transfer terminals from the current 250, it could re- this study revealed inefficiencies and added cost San Francisco waterway as opposed to relying duce its total inventory and warehousing as much burdens resulting from losses, delays, and brib- solely on ground transport, could reduce per ton as $1 billion a year. Warehousing shortages in the ery. Those costs amounted to 20–25 percent of transport costs at least 9 percent. Though large ex- region can be mitigated through third-party logis- the total domestic shipping, inventory, and clear- porters have been aware of the benefits for more tics providers (3PLs), which have played a growing ance costs—or a 25–35 percent cost over the than a decade, costly investments in small river role in providing complex logistics and transport “legitimate” logistics costs. These costs represent port improvements and storage capacity appear to services.89 around 20 percent of the total costs incurred in the make the switch cost-prohibitive. As it stands, bulk 3.80. Rail freight transport in the region is limited. In import of soybeans into Brazil and beef into Chile. shipment trucks are in short supply during harvest Brazil, only 20 percent of freight is shipped by rail, The unnecessary logistics costs related to beef season, making shipments unreliable.87 and most railways are concentrated in the south are equivalent to $14.01 per ton, 93 percent of it 3.78. Competition among trucking companies rep- and southeastern parts of the country; in Chile, the ($13.02) related to the public sector. Table 3.6 lists resents a particular challenge. Brazil, for in- figure is 5 percent. In these countries, the vast ma- these over costs. stance, can be seen as different countries inside jority of freight is transported by road (62 percent one country due to its varying infrastructure and in Brazil and 92 percent in Chile).90 The chronic markets. While the south and southeast regions lack of investment in transport infrastructure oth- of Brazil have a relatively large number of provid- er than roads translates into lack of competition ers, promoting competition and lower costs, the among modes, fewer alternatives for shippers, Central-West, North, and Northeast Regions lack providers and infrastructure, leading to higher 88 Langley and Capgemini, 2012. transportation costs. In these areas, many farm- 89 89 Langley 2012: In a 2012 study of 3PLs, a large percentage of survey respondents claimed to use 3PLs in LAC for warehousing (63 percent), ers employ drivers who use their own trucks for international transportation (84 percent), domestic transportation (69 percent), freight forwarding (65 percent), and customs brokerage (45 percent; see annex 1). In 2010, 3PL revenues in LAC were $27.5 billion 86 Schwartz and others 2009. (of global 3PL revenue of $541.6 billion). 87 LCSSD Economics Unit 2012. 90 FAO 2008. 102 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 3.82. Some inland border crossings in the Southern Cone are poorly operated, leading to time delays similar to those common in Central America. This, in turn, translates into congestion problems at many border crossings in the region. For instance, Argentina’s Paso de los Libres border crossing (the busiest in South America) experiences severe traf- fic problems: it takes 30–36 hours for most trucks to cross the border.91 According to Brazilian firms surveyed for the LPI, deficiencies in the country’s customs performance present greater challenges Table 3.6 Breakdown of over costs in beef imports into Chile to business operations than those related to infra- structure or to logistics operators. 3.83. According to the LPI, customs is the weakest com- Over costs $ per ton Percent ponent of overall logistics performance for LAC’s Informal payment five largest agricultural exporters (Brazil, Argen- Inventory costs 7.93 56 tina, Chile, Mexico, and Colombia). Chile outper- Losses due to delays forms all countries in customs and infrastructure Inefficiencies at customs clearance 4.41 32 indicators but falls behind Brazil in the four other Delays in collection of payments 0.57 4 categories. Interestingly, poor customs perfor- mance does not seem to lead to delays and unre- Source: CARANA Corporation 2006. liability for the region’s largest exporters. Though customs indicators are relatively weak, most of the countries receive relatively high scores for timeli- ness. This may be related to years of experience exporting large volumes of agricultural products. 91 Schwartz and others 2009. Chapter 3. 103 Box 3.1. Logistics challenges in wine exports from Mendoza During many months, containers may travel an alternative route over the Andes Moun- Wine trade logistics exemplify some key constraints for Southern Cone exports. Wine tains to the Port of Valparaiso located just beyond Santiago de Chile on the Pacific Ocean. exportation from the Mendoza region in Argentina is a booming business. Over 1995– Although this route cuts actual travel time in half, and the distance is a mere third of the 2010, the free-on-board value of total wine exports from the region increased an aver- other, the comparable cost and high possibility of extended wait times (even exceeding age of 19 percent annually, reaching more than $631 million in 2010.92 In recent years, 10 hours for processing) at the border make this option unpopular. Seasonal and geo- 25–35 percent of total wine exports have been destined for the United States. graphical factors play a role as well. Though the road is reported to be in good condition, Most wine leaves Argentina through the Port of Buenos Aires and arrives at port in New heavy snowfall makes it impassable in the winter months, and the Andes Mountains York or Texas. One thousand boxes, each with 12 bottles of wine, travel 1,200 kilometers slow truck speeds, increasing travel times. Since 2007, the governments of Argentina over land from Mendoza to the port in approximately 15 hours. The inland transport seg- and Chile have been discussing the $3 billion construction of a cross-county low-altitude ment costs $2,500 per trip, an estimated 7.3 percent of the free-on-board price of most tunnel, including the rehabilitation of a railroad network from the 1800s, referred to as wines. Though the distance is much greater, maritime freight costs from the Port of Bue- the Corredor Bioceánico Aconcagua. Several private sector businesses avidly support the nos Aires to New York are only slightly higher than ground transport, at $3,000. project, which would address a major geographical barrier to trade among Southern Cone Congestion and the ever-present threat of workers’ strikes make the Port of Buenos countries and Brazil. Aires unreliable and inefficient. To address this constraint, Mendoza inaugurated a “dry Taking into account these constraints, exporting wine from Mendoza should take 25 port” in 2007, a multimodel customs zone where exporters can process documentation days from the moment a sale is made until the wine arrives at its destination in the United prior to leaving the region, and long before arriving at the port. A dry port, of which there is States, but inefficiencies along the supply chain can easily double export times. another example in Mexico, helps facilitate export procedures but does little to address Source: Fundación ProMendoza, various publications. port bottlenecks related to operations, infrastructure, and technical capacity for mari- time-bound goods. 92 Fundación ProMendoza. 104 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 3.84. Maritime transport. Leading global shipping carriers since 1995, the increase in trade has increased bulk goods decreased, reducing the average ship- are using regional shipping lines to overcome the port congestion, forcing many ship owners to can- ping costs for imports and exports.97 existing sabotage restrictions and regulations prev- cel ship calls.95 3.87. Andean region. Poor rural connectivity is a com- alent in South America, especially on its east coast. 3.86. Port modernization has reduced average export mon logistics challenge for Andean countries. In While capacity supply expanded over 2000–08, and import costs. The concessioning of general Colombia and Peru, recent investments have at- effective competition fell, and high entrance barri- cargo port terminals throughout the region starting tempted to address challenges in port infrastruc- ers and collusive behavior among existing players in the early 1980s liberalized labor practices and ture and border management integration. Land- in the market prevail, as evidenced by the almost privatized port services.96 This, in turn, drove com- locked Bolivia has a fairly high-quality national complete absence of new entries in the market. petition among private stevedores, improved man- road network, while Colombia has notoriously poor 3.85. Port efficiency affects not only the cost of ocean agement of stevedores, and led to investments in roads, due partly to its history of violent conflict shipping but also the costs incurred by cargo own- container-handling equipment by port operators and a decentralization program that complicated ers and consignees.93 A poorly run or otherwise in- and investments in shoreside equipment by ste- earlier improvement efforts. efficient port will serve as a tax on cargo because vedore companies. These changes boosted pro- 3.88. Ground transport. Andean countries exhibit poor storage, warehousing, inventory, and demurrage ductivity and lowered fees for port users, prompt- road network quality and density. Venezuela and charges are accrued when ports hold up cargo and ing many exporters to use the newly privatized Ecuador retain the largest percentage of paved delay delivery. Brazilian ports, for example, suffer ports rather than roads, vertically integrated ports, road network in the region, with an index of oc- from serious dredging backlogs and lack contain- or more efficient neighboring ports. This was the cupation of 101 kilometers of roadway for each er handling capacity and post-Panamax facilities. case for Chilean fruit exporters after port conces- 1,000 square kilometers of area in Venezuela, and The low container capacity presents a significant sioning in Chile started in the 1980s, as well as 159 kilometers of roadway for each 1,000 square bottleneck for the sector: four of five main ports in for Brazilian soy exporters and Colombian banana kilometers of area in Ecuador. Peru and Bolivia, on Brazil reached their capacity in 2007. According and coffee exporters. In Argentina, the conces- the other hand, contain the lowest percentage of to ship owners, Brazil’s ports in 2007 lacked a to- sioning of the country’s four main ports starting paved road network, with approximately 50 kilo- tal of 1.1 million square meters. To maintain their in 1990 significantly improved port efficiency, re- meters of roadway for each 1,000 square kilome- current operating levels, 5.4 million square meters duced port charges, and increased cargo volume. ters of land area. By the density measure, Peru and will have to be added by 2012.94 Despite Brazil’s Shipping charges for containers, grain, and other Colombia have the fewest kilometers of roads per continued port investments of around $1 billion inhabitant.98 A series of three World Bank–funded 93 Various indicators are measured when evaluating port capacity and technical efficiency, including the number of terminals, open office 95 UNCTAD 2011. 97 Estache 1996: By 1995, the port of Buenos Aires saw a capacity hours 24/7, quay operation 24/7, reception and delivery 24/7, total 96 Concessions are long-term contracts between the government and a increase of 250 percent and a productivity increase of close to 400 port area, storage capacity, reefer points, number of berths, average private investor that bundle investment and service provision. Unlike percent. Modernizing ports has thus facilitated the growing quantities berth length, maximum draught, and maximum berth length. privatization, concessions do not entail sale of assets; they last for a of food products being exported from Argentina. 94 Schwartz and others 2009. limited time and entail government supervision. 98 FAO 2008. Chapter 3. 105 projects have improved an estimated 15,000 kilo- ing resulted from the longer time required for the However, domestic trucking costs from transport- meters of rural roads in Peru over the past decade. product to be transported.100 ing the wheat flour to other cities are more signifi- 3.89. Rural roads in the Andean countries are especial- 3.91. Peru also has one of the lowest road densities in cant. The delivered cost to a city like Ambato adds ly important. In this region, a large percentage of LAC, increasing transportation costs. As a result of another 20–25 percent onto the product’s cost. the population lives in rural areas, many of which its underdeveloped road network, transportation The large price difference is a result of road quality are extremely poor. Therefore, the lack of adequate costs are almost double those in Chile, and the and the ability of trucks to make a return trip with- road infrastructure in rural areas makes trade es- country’s average circulation speed is among the in a day when traveling to and from Quito, even pecially difficult, driving up transportation and lowest in the world, at 17 kilometers per hour. In though Quito is further from the port than Ambato. production costs and reducing export products’ addition, logistics costs in Peru are high, account- Calculations for 11 cities in Ecuador show a strong competitive potential.99 ing for 34 percent of total operating costs, com- correlation between domestic transport costs and 3.90. A dairy production study in rural Peru (2008) con- pared with 25 percent in Brazil, 20 percent in Mexi- flour price, meaning that higher domestic trans- cluded that the price structure of dairy products co, and 8–10 percent in OECD countries.101 port costs in Ambato, for example, explain the flour is directly related to poor road conditions. Local 3.92. A supply chain analysis of wheat flour imports into price difference relative to Quito.102 roads and footpaths often become impassable Ecuador reveals that by the time the wheat arrives 3.93. The findings of an analysis of Ecuadorian wheat during the rainy season (June through August), at the mill, logistics costs constitute more than transport costs illustrate that distance may not hampering transportation efforts. Further, roads 30 percent of the mill’s purchase price. Once the be a central driver of costs. The trend line in figure are usually maintained by local farmers who lack cargo is unloaded in Ecuador, the cost of domestic 3.18 shows a slight upward incline in the relation- adequate road maintenance equipment. Due to transportation to Quito is minimal due to the high ship between distance and cost. Yet, even with poor road quality, milk producers reported a loss degree of competition in the Quito market and the only 11 data points, the high dispersion is evident, of milk amounting to 1.5 percent; in addition, sour- availability of good roads linking Quito to the coast. as are the exceptions to the trend. 100 FAO 2008. 99 FAO 2008. 101 Fretes-Cibils and others 2007. 102 Schwartz and others 2009. 106 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 3.94. Maritime transport. Maritime costs may be much higher for the Andean countries than for the South- ern Cone countries. A study assessing maritime transport costs for food imports shipped from the EU found that Brazil and Argentina had clear cost advantage (with marine transport costs at 4.9 per- cent and 6.53 percent of free-on-board value, re- Figure 3.18. Relationship between a city’s distance from Port Manta spectively), while Peru and Ecuador exhibited the and the delivered cost of wheat highest costs (with 9.97 percent and 7.47 percent of free-on-board value).103 65 3.95. Recent port reforms are changing the story for the Andean region’s underperforming ports. Colom- 60 bia’s public ports, once inefficient and costly to Price/kg wheat flour (US$ cents) run, experienced a 45 percent increase in general 55 cargo throughput a year after the concessioning of their ports in the early 1990s.104 In addition, container movements per ship hour increased 50 from 7 to 52, and the average time for cargo to sit at the port dropped to 2 days from 30+ days be- 45 fore private operations. In Cartagena, average ship waiting times dropped from 10 days to 0, and ship 40 turnaround time fell from 72 hours to 24. These im- 0 100 200 300 400 500 600 700 provements were accompanied by a reduction of Distance to Manta (km) the terminal handling charge, from $984 to $222, and prompted many global shipping lines to begin Source: Schwartz and others 2009. including port calls in Colombia. In Peru, imple- menting electronic information processing at the 103 Kjöllerstrom 2004. 104 Gaviria 1998. Chapter 3. 107 Port of Callao in Peru in 2008 reduced cargo clear- 3.98. The degree of competition among service provid- ance time about 70 percent.105 ers is a related factor driving shipping costs. In a 3.96. Caribbean. Maritime transport is the protagonist of study on ocean shipping rates for a sample of 189 the logistics story in the Caribbean. In this region, routes within the Caribbean basin, Wilmsmeier and countries’ geographical location has played a key Hoffmann (2008) show statistical evidence that role in trade and logistics costs. While some coun- around two-fifths of the variance of the freight rate tries, such as Jamaica and the Bahamas, have lev- can be explained by the number of carriers oper- eraged their geographical positioning in the main ating on the given route. That is, the oligopolistic East-West global trade routes to increase trade market structures in many Caribbean routes due to and develop their ports, smaller Caribbean islands dispersed markets with low trade volumes imply that are more remote from main trade routes suf- higher costs for shippers. fer from higher transportation costs and have not 3.99. Countries’ positions within the network have a been able to develop their ports.106 Countries in more significant impact than distance. Figure key shipping routes become regional hubs and 3.19 shows a statistically significant correlation enjoy lower maritime freight costs than other Ca- between connectivity and ocean freight rates, ribbean destinations.107 which has important implications for the price of 3.97. Shipping costs to the Caribbean are considerably delivered food imports. Since the islands’ shipping higher than in other areas worldwide (13 percent structures provide for little direct service, goods of the free-on-board price versus the world aver- are often consolidated in Miami and shipped on age of 6.64 percent). Following the reduction of small carriers with relatively infrequent services tariff barriers, transport costs are one of the most that travel to multiple islands on each voyage. The significant NTBs. Exports from CARICOM to the Unit- results are low economies of scale in shipping; ed States in 2003 averaged freight and insurance infrequent port calls; and large numbers of mid- charges ranging from 7.6 percent (Jamaica) to 16 dlemen buying, repackaging, and reselling pro- percent (Antigua) as a percentage of the free-on- duce. Further, because volumes of imports to and board price.108 exports from the Caribbean islands are relatively small, importers and exporters typically attempt 105 Schwartz and others 2009. to consolidate cargo to reduce shipping costs. 106 Sanchez and Wilmsmeier 2009. However, this is still more costly than shipping full 107 Jha 2005: Countries importing higher volumes of goods typically incur containers. lower freight rates and, in turn, lower prices for their imports and lower costs for exports. 108 Sanchez and Wilmsmeier 2009. 108 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 3.100. The maritime transport component represents a significant part of transport costs in the Caribbe- an. Pineapples imported into St. Lucia are grown in Costa Rica and consolidated in Miami. With all that traveling and handling, the producer price of the pineapple itself accounts for only about 10 percent of the final delivered price, whereas trans- Figure 3.19 Relationship between freight rates and connectivity, container shipping, Caribbean port (land and ocean) costs and handling reach 43 percent. Further, warehousing, consolidation, and 1 retail and wholesale profits collectively account 0.9 for 33 percent—about half of which is also logis- 0.8 tics. Because of the consolidation in Miami and the 0.7 need for two ocean movements, ocean shipping represents a significantly large part of the trans- 0.6 Freight port costs: 3.5 times as much as the producer Freight 0.5 rate index price for the pineapple itself.109 0.4 3.101. Costs for exports and imports are further in- creased by higher logistics costs arising from inef- 0.3 ficiencies in customs. Studies of customs clearing 0.2 conclude that CARICOM experiences considerably 0.1 longer average times for customs clearing than 0 other regions. The lack of information technolo- 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 gy and electronic documentation in customs, for Connectivity index instance, was identified as a common constraint among Caribbean countries. At the same time, Source: Sanchez and Wilmsmeier 2009. there is also a significant difference in customs performance within the Caribbean region. Be- cause many agricultural products are perishable and time sensitive, delays are particularly costly. Importing goods into an Organizaion of Eastern 109 Guasch 2011; this is an assessment of a supply chain analysis of pineapples imported into St. Lucia. Chapter 3. 109 Caribbean States (OECS) country takes on aver- glut and shortage. This study also mentioned con- appears to be variation within the LAC subregions. age 19 days and requires seven documents, while nectivity as a challenge in the region. For instance, Within the Caribbean region, Trinidad and Tobago exporting goods takes an average of 17 days. All in one reported case, Belize could not secure the appears much more connected than the rest of the OECS countries require that a bill of lading, invoice, necessary transport for a 10,000-ton corn ship- region’s countries. Within Central America, it ap- certificate of origin, and customs declaration be ment to Guyana and could not ship the product pears that Guatemala is the most connected coun- completed to import goods.110 due to lack of reliable and cost-effective maritime try while Nicaragua is the least connected. 3.102. CARICOM ports vary widely in efficiency of move- transport services to that part of the region. ments of 20-foot-equivalent units, storage and 3.104. According to the Liner Shipping Connectivity In- handling, and customs. Gaps exist between ports dex, Caribbean countries, on average, are the with gantry cranes, which achieve high levels of least integrated into the liner shipping network, quay-side efficiency, and smaller ports with mo- when compared with other LAC countries (figure bile cranes or no cranes at all, reducing a port’s 3.20).112 This is especially unfortunate for Caribbe- productivity considerably. The ports in Guyana an countries, given that maritime transport is their and Suriname, for instance, lack sufficient in- most important freight transport mode. Mexico frastructure, navigational aids, and regulatory and the Southern Cone countries score highest on framework, increasing the overall costs and time this dimension. Transshipment ports are excluded for services calling in these ports, which leads to from the subregional comparison as they handle higher maritime transport costs.111 Lower levels mostly foreign traffic, which requires specialized of port efficiency lead to higher import costs and equipment and processes; consequently, they are reduced export returns, hurting countries’ compet- largely independent of the national economy.113 itiveness. Even when excluding transshipment ports, there 3.103. CARICOM countries also suffer from a number of logistics challenges that reduce their exporters’ 112 The Liner Shipping Connectivity Index is an indicator for the supply competitiveness. A 2011 study carried out by the of liner shipping services and can be considered a proxy for trade Caribbean Integration Support Program identified accessibility. It is calculated by five components: number of ships; 20-foot-equivalent unit capacity; number of shipping companies; that poor postharvest capability, including lack of number of services; and maximum vessel size. The index aims at cap- refrigerated trucks, and the absence or poor quali- turing a country’s level of integration into the existing liner shipping ty of distribution and cold chain facilities are lead- network. The higher the index, the easier it is to access a high-capacity and frequency global maritime freight transport system and thus ing to high levels of product loss and periods of effectively participate in international trade. 113 Kingston, for instance, handles a very high number of 20-foot-equiv- alent units as a result of it being used by ZIM Line as a principal 110 Sanchez and Wilmsmeier 2009. transshipment hub. Still, Jamaica’s GDP per capita is lower than that 111 Sanchez and Wilmsmeier 2009. of many other LAC countries. 110 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 3.105. Table 3.7 summarizes the logistics challenges fac- ing the various LAC subregions, broken down by the three logistics subsectors: ground transport, maritime transport, and border procedures. The challenges presented here are not exhaustive; the table focuses on challenges that available litera- Figure 3.20 Liner Shipping Connectivity Index, 2011 ture and expert opinion have identified in recent years. Though individual countries vary in the Mexico Caribbean logistics challenges they face and the degree of Southern Cone Transshipments those challenges, this table presents a summary Andean Region of logistics performance by region. LAC Average Central America Caribbean 5 Central America 12 LAC Average 21 Andean Region 23 Transshipments 28 Southern Cone 29 Mexico 36 0 5 10 15 20 25 30 35 40 Source: UNCTAD 2011. Chapter 3. 111 Table 3.7 Summary of logistics challenges facing Latin America and the Caribbean subregions   Central America & Mexico Caribbean Andean region Southern Cone • Low road density and access, • Empty backhaul in the DR • Low road density and access, • Relatively high access to roads but especially in rural areas ((NIC, GTM, especially in rural areas (PER, COL, high transportation costs (general HND) BOL) problem) • Losses while merchandise is in   • Poor road quality, especially in • Inadequate road maintenance transport due to spoilage or rural areas (PER, COL, BOL, ECU) investment has led to poor road breakage (NIC, GTM, HND, CRI) quality, especially of provincial and rural roads (general problem) • High domestic transport costs (NIC,   • Low speed circulation due to heavy • Traffic congestion near border GTM, HND, CRI, MEX) traffic and congestion in roads points delays border crossings and leading to ports (PER) thus increases costs (BRA, ARG) Land transport • Lack of truck competition due to   • Lack of appropriate infrastructure • Low competition among truck monopolies (NIC, GTM, HND, CRI, in land border-crossing points companies in some areas (south MEX) (general problem) BRA) • Lack of enforcement and adequate   • High transportation costs (general • High inventory costs due to the lack trucking regulations for vehicle problem) of sufficient warehousing capacity and driver operations (NIC, GTM, (BRA) HND, CRI, MEX) • Lack of appropriate infrastructure     • Rail freight transport is limited in land border-crossing points (General problem) (C.A.) • Empty backhaul due to trade laws     • Increasing operating costs of (MEX) trucking services (BRA) 112 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development   Central America & Mexico Caribbean Andean region Southern Cone • Inefficient port entry procedures • Maritime and logistics costs • Higher maritime transport costs • Lack of intermodal transfer such as long waiting times for depend heavily on countries’ than in Southern Cone countries terminals (general problem) processing of documentation for geographical location within the (general problem) imports (NIC, GTM, HND, CRI, SLV) Caribbean Maritime transport • Bottlenecks due to deterioration • On average, maritime transport • Lack of intermodal transfer • Declining competition, high of access roads to ports and delays costs to the Caribbean are much terminals (general problem) entrance barriers, and collusive in the loading process caused by a higher than in other regions in LAC behavior among existing players lack of port assets such as cranes (especially OECS) (BRA) (NIC, GTM, HND, CRI, SLV)   • Economies of scale disadvantage   • Bottlenecks at ports due to a lack require small islands to consolidate of container-handling capacity cargo (especially OECS) (BRA, ARG) • Burdensome customs inspections • Long time for customs clearing • Inefficiencies at customs clearance • Lack of coordination between and heavy fines (NIC, GTM, HND) in comparison to LAC regions (general problem) customs and border management (especially OECS) agencies (general problem) • Delays due to heavy traffic and • Lack of information technology • Delays due to heavy traffic and • Traffic congestion near border congestion (NIC, GTM, SLV, HND, and electronic documentation congestion (general problem) points delays border crossings Border procedures CRI) (especially OECS) (general problem) • Lack of coordination between • Informal payments (general • Lack of coordination between • Informal payments (general customs and border management problem) customs and border management problem) agencies (NIC, GTM, HND) agencies (general problem) • Delays and increased costs due • Significant difference in customs • Informal payments (general • Inventory costs and losses of profit to onerous SPS controls such as performance within the Caribbean problem) from incurred delays of 24 hours duplicate fumigation procedure region (OECS are worst performers) at the border crossing (general (NIC, GTM, SLV, HND) problem) • Informal payments (NIC, GTM, SLV,       HND, CRI, MEX) Chapter 3. 113 Box 3.2. Better use of information and communication technology the meat is distributed and thus follow the stock’s evolution and the animals’ sanitary could increase competitiveness in agricultural production and situation. Because tracking is required by quality standards in many international val- marketing ue chains, the use of TRAZ.AR strengthened cattle farmers’ competitiveness in the global The rapid spread of information and communication technology (ICT) in developing meat market.114 Moreover, through this program, farmers improved reproduction selec- countries offers an opportunity to tackle some of the problems afflicting LAC agricultural tion, reduced animal stress, and improved sanitation. Galiani and Jaitman (2010) find economies to increase productivity, overcome logistics problems, and enhance global that farmers in the TRAZ.AR program were less affected by a severe drought experienced competitiveness. during the period and were able to sell at better prices. Increasing farmers’ access to markets Strengthening financial services and mitigating risk ICT can reduce the costs of acquiring relevant market information for potential buyers, Mobile financial applications have facilitated financial transactions, such as bill pay- prices of inputs and outputs, and alternative marketing channels. The efficiency gains ment and money transfer among bank accounts, and have reduced the cost of providing from introducing ICT can translate into an increase in the welfare of agricultural producers. public services to the rural poor. In Colombia, coffee growers are increasingly perform- Camacho and Conover (2010) found that the farmers who received price and weather in- ing financial transactions, including input purchases, by mobile phones.115 Subsidies in formation through text messages (SMS technology) in Boyacá, Colombia, faced lower dis- the conditional cash transfer programs in Colombia and Mexico (Familias en Acción and persion in the expected crop price and a large reduction in crop loss. Beuermann (2010) Oportunidades) are now being transferred through electronic payments to individual and examines the impact of an intervention that provided at least one public (satellite) pay- collective bank accounts,116 as are payments to farmers from the Procampo program of phone to villages in rural Peru that previously had neither landlines nor cellphones. Re- area-based support payments. ICT has been shown to influence risk faced by agricultural sults show a sizable increase in agricultural profitability and incomes as the improved households. To the extent that ICT improves communications among the social network, information access increased farmers’ bargaining power. . it can increase the effectiveness of informal insurance arrangements. ICT also facilitates the transmission of information on potential shocks. This was the goal of the intervention Supporting logistics and quality control in Boyacá, Colombia, in which farmers were provided with price- and weather-related in- ICT can help farmers meet the escalating demands in distribution and quality control posed by modern supply chains. An interesting example is the TRAZ.AR program in Ar- 114 IDB 2011. gentina. This program provided small and medium-size cattle farmers with Internet-con- 115 IDB 2011. nected software that allowed them to track each animal from the time of its delivery until 116 IDB 2011. 114 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 3.5 Conclusions and policy implications formation through text messages.117 An improved system of alerts for natural disasters is 3.5.1 also a good example of ICT’s role in reducing risks. In response to the 2010 earthquake, Global trade policy the Chilean government initiated an SMS earthquake alert system program that, by 2012 would be incorporated in all mobile phones in the country. 3.106. This chapter’s findings underscore the relevance of the global trade reform agenda, especially in ag- Improving the adoption of agricultural technologies ricultural products, for which trade barriers are still significantly higher than for manufactured goods. One of the main channels for promoting adoption of agricultural technologies is agri- And NTBs are more important than tariffs, especial- cultural extension. Traditional ICT, such as television and regular radio broadcasts, has ly for agriculture. Given agricultural trade’s impor- long been used to support the extension service. Governments around the world are now tance to LAC, and LAC’s importance as a world food incorporating new versions of extension services that are supported by modern forms of supplier, it is in everyone’s best interest to reduce ICT, such as voice-based information systems, SMS, and e-learning. ICT can improve the remaining barriers as quickly as possible. And as diffusion of private information on technologies. The adoption of productive technologies we saw comparing tariffs with NTBs, the remaining can be sped up with increased communication between farmers and other technologi- agenda should focus on NTBs. cal adopters. Anecdotal evidence from the Huaral Valley in Perú indicates that installing 3.107. The looming threat of climate change magnifies telecommunication information centers improved the distribution of water from irrigation the importance of increasing the trade system’s sources and helped the communities coordinate its use in times of water scarcity.118 flexibility, for two reasons. First, in the long run, as Source: Goyal and González-Velosa 2012. patterns of comparative advantage in food produc- tion change, moving food from countries where it is produced efficiently to food-deficit countries will require new trade patterns. Second, on a year-to- year basis, greater weather variability will create short-term local shocks to food supply that will require rapid food trade adjustment to avoid short- 117 Camacho and Conover 2010. ages. One of the lessons of the recent precipitous 118 IDB 2011. increases in food prices is that when shortages arise, countries tend to react with “beggar thy neighbor” trade policies that insulate domestic Chapter 3. 115 consumers and producers from international price tion in the sector, thus helping improve efficiency, 1 percent of GDP in 2004 to 4.1 percent in 2011) movements. In so doing, they increase global price bring down costs, and enable the world’s most ef- and serve to keep domestic prices low for consum- volatility and shift the adjustment costs onto oth- ficient producers to expand their share of the bio- ers in periods of international price spikes. Yet the ers. Such actions include increases in export bar- fuel market. For example, producing a liter of eth- quantitative controls produce no revenue, and in riers, which are now widely understood to have anol from sugarcane in Brazil requires only about the medium term, these controls, along with taxes, amplified the spike. Export bans accounted for an half the land area needed to produce the same reduce domestic production, potentially raising estimated 40 percent of the world price increase liter from corn in the United States. Transferring prices. Export controls are one explanation for the for rice and 25 percent of the increase for wheat.119 production from the United States to Brazil would reduction in Argentina’s beef production in recent But, less well understood, the ad hoc reductions in thus reduce the amount of land diverted from years. Further, if several major exporters impose import barriers in many countries had a similar ef- growing foods. But currently, biofuel promotion export taxes simultaneously, the effect on inter- fect—they reduced price fluctuations domestical- policies distort international trade patterns and national prices will at least partly offset the first- ly while magnifying international price variability. impede this shift—and at the same time impose round impact of the taxes in lowering domestic The impact of these import tariff reductions has large costs on the populations of the countries prices in each of those countries. In any case, al- not been quantified, but given their frequency, employing them. Of course, to fully realize these ternative instruments may meet these objectives the aggregate impact could have been quite sig- benefits, Brazil must ramp up its production with- at lower costs than either taxes or controls. nificant. Neither of these measures is effectively out deforesting land, but the country has plenty of 3.110. Other countries that are net food importers should disciplined under current WTO rules. degraded pasturelands that could be used more also consider the costs of responding to price 3.108. The three-way connection among climate change, productively for these crops. movements in international markets with policies food trade, and global trade policy is also import- that insulate their domestic economies while ex- ant. Part of the solution to reducing greenhouse 3.5.2 acerbating international price volatility. These gas emissions from transportation will inevitably For Latin America and the Caribbean policy responses include reducing tariffs on food involve increasing substitution of biofuels for gas- imports when prices are high and raising them 3.109. While LAC countries have come a long way in re- oline and diesel. However, while the quantitative when prices fall. Such policies not only magnify ducing the anti-export and anti-agricultural biases significance is debated, increased cultivation of world price movements but also are inefficient in their trade regimes, the biases remains signif- crops for biofuels will clearly have a greater and for the country involved, because they encourage icant in some countries. Argentina, a major food greater impact on land available for food crops. To overconsumption and underproduction when pric- exporter, imposes export taxes and quantitative minimize this tradeoff, crops for biofuels must be es are high and vice versa. To the extent that trad- controls, with considerable adverse consequenc- grown in the most land-efficient way possible. And ers and processors anticipate such adjustments, es for the sector and the global food trade system. here is where the trade connection comes in: lib- they can adjust the timing of their own storage The motivations behind this policy are under- eralizing trade in biofuels could increase competi- and import behavior, resulting in sharp import flow standable: these taxes are a substantial part of fluctuations and supply chain congestion. A better the government’s revenue base (rising from about solution would be to permanently lower tariffs, 119 Martin and. Andersen 2010. 116 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development which would reduce the anti-export bias in the cur- clarifying the rules under the multilateral SPS important as revenues from recent resource dis- rent trade and support regimes, as shown above, agreement to improve transparency. Even bet- coveries begin to increase in countries like Brazil, and benefit poor consumers. Another option is to ter, the agreements could commit countries not Argentina, and Colombia. Here, Chile is instructive. ramp up safety net payments to compensate the to impose protection more stringent than the Notwithstanding large revenue increases from poor when food prices rise. Nonetheless, it is clear measures recommended by relevant interna- copper in recent years, its real exchange rate has from the frequency of ad hoc tariff reductions that tional scientific organizations. This could be es- not appreciated as much as that of other coun- strong political pressures encourage this response pecially useful in implementing the concept of tries, due largely to its macroeconomic policies, in times of spiking food prices. But this should be regionalization of risk. Harmonization and mutu- including a restrained fiscal response during the considered a policy of last resort. al recognition of standards would also enhance commodity boom and use of stabilization and sov- 3.111. While working within the multilateral system for trade. Some of these issues might be handled ereign funds.122 The threat of Dutch disease also further reforms, countries in LAC, as well as in without negotiating new agreements through magnifies the importance of national innovation other regions, could take better advantage of the current committees and working groups. policy. Here, policy should focus on incentives for opportunities provided by negotiating PTAs to ad- • Although the process would undoubtedly be technology generation and adoption that are fairly dress some issues not handled well in WTO com- complex, there could be great benefits in harmo- neutral with respect to products or sectors,123 rath- mitments,120 particularly to reduce the effects of nizing PTAs by gradually converging their com- er than on what Justin Lin (2012) calls compara- NTBs, as Chile has with its bilateral agreements. mitments. A first step would be to agree on the tive advantage–defying strategies, which single Some issues that could be addressed in PTAs in- scope of the efforts (all sectors, all PTAs; specific out new industries for special favors. clude: sectors, only some PTAs; and so forth), followed • Removing the exemption of agricultural prod- by establishing a technical group to study how 3.5.3 ucts from the “general tolerance” or de minimis to improve the selected PTAs and make them Policy recommendations for logistics and exceptions in rules of origin, so that producers of more effective in expanding trade both inside trade facilitation agricultural products (primary and processed) and outside the region.121 Given the high trade 3.113. Overall policy guidance. First and foremost, as can take as much advantage of low-cost import- barriers facing LAC agricultural exports to some we saw in subsection 3.5.1, improving logistics ed inputs as can producers in other sectors. A developing regions, it could be worthwhile to and trade facilitation is hugely important. Further, second-best alternative would be to exclude explore agreements with those countries, espe- some logistics issues matter more to particular only especially sensitive agricultural products, cially in South Asia, the Middle East, and North kinds of products. Exports of heavier products, without excluding the whole sector, as many re- Africa. such as industrial products and “bulks,” depend gional trade agreements currently do. 3.112. For LAC countries’ agricultural sectors to stay com- more on the quality of hard infrastructure, where- • Improving the treatment of SPS issues in the petitive, they must manage the real exchange rate as time-sensitive products depend more on soft agreements. At a minimum, this could include to minimize Dutch disease. This will become more 120 This section draws from Shearer and others (2009), which discusses 122 Sinnott and others 2010. these recommendations and others in considerable detail. 121 Estevedeordal and others 2009. 123 Sinnott and others 2010. Chapter 3. 117 infrastructure and clearly on number of days to ex- lower costs, higher trade volumes, and increased losses and damages, and increase producer and port. For example, exports of meats, a highly per- reliability and competitiveness. Countries can consumer surplus. In addition, road maintenance ishable product, are extremely sensitive to time ensure that all logistics constraints are identified is cost-efficient: the present value of maintaining a delays. For agricultural exports overall, and for all and monitored by designating a national logistics road regularly is less than rehabilitating it once ev- countries, this soft infrastructure is much more entity to coordinate efforts and manage dialogue. ery 10 years. As LAC enters a period of fiscal con- important than hard infrastructure. Improving the Through the Plan Mesoamericana, as well as other straint, it may be worth reiterating the importance region’s performance to OECD levels would double regional organizations, Central America is moving of road quality and its effects on food prices. agricultural exports. toward strategic regional infrastructure planning 3.118. Policy guidance for customs and border re- 3.114. Quantitative estimates of potential cost reductions but has encountered considerable difficulties har- form—improve intraregional trade through better indicate substantial heterogeneity in the impact monizing procedures and standards. Chile, Argen- cross-border collaboration and greater coordina- of transport and logistics costs on LAC countries, tina, and Brazil are also engaging in regional coor- tion between national customs services and har- depending on the relative shares of various ag- dination to improve intraregional trade. monizing regional SPS standards. In Central Amer- ricultural exports and imports. However, overall, 3.116. An analysis of the breakdown in food types sug- ica, across Mercosur, and across the U.S.-Mexico reducing logistics costs from port efficiency gains, gests that for net food importers, costs associated borders, for instance, customs harmonization to improving road haulage, expediting customs with refrigerated cargo capacity and services are facilitate the transport of goods and to reduce clearance and border crossings, improving inven- the critical bottlenecks, as meat, fish, and dairy clearance costs could be addressed as an issue tory practices, and increasing capacity and com- represent the largest share of all food imports by of regional integration and trade facilitation. For petition in storage and warehousing could reduce value (26 percent). For LAC net food exporters, countries that are partners in trade agreements, logistics costs 20–50 percent. This could perma- however, bulk storage, handling, and transport are the duplication of phytosanitary inspections can nently reduce the baseline cost of food and agri- the primary concerns, as, on a weighted-average be avoided by initiating a common set of testing cultural imports about 5–25 percent and increase basis, dry bulk items make up the biggest share procedures. These may be applied only once on profits for exporters.124 of food imports by value (31 percent). Thus, OECS one side of a border crossing, using officials from 3.115. A trade supply chain is only as strong as its weak- island countries, for example, should work on re- both countries. This could save considerable time est link: poor performance in just one or two ar- ducing the cost of refrigerated containerized traf- and costs, particularly for perishable goods.127 eas can have serious repercussions for overall fic. Peru, Brazil, Bolivia, and Colombia, on the other 3.119. Improve customs coordination with phytosanitary competitiveness.125 The multidimensionality of hand, would benefit from improving the importing services. Customs clerks are often not available logistics necessitates a coordinated strategy, de- and distribution process for dry bulk goods.126 when phytosanitary inspectors are available and veloped with input from public and private sector 3.117. Policy guidance for ground transport and distribu- vice versa, making import and export approvals stakeholders alike, for improvements to result in tion—improve road quality. Better road mainte- doubly complicated and time-consuming. Ad- nance would lower vehicle operating costs, reduce dressing this frustration may require extending 124 Schwartz and others 2009. 125 World Bank, Doing Business 2012. http://www.doingbusiness.org/ 126 Schwartz and others 2009. 127 Schwartz and others 2009. 118 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development the operating hours of both agencies, colocating economies of scale, and lower prices in the ship- operators suffer from a prisoner’s dilemma when offices, or even hiring additional staff, but it is par- ping of their food products. individually considering their incentives for port ticularly important for reducing the cost of traded 3.123. Introduce modern port-operating practices. LAC investments. Many of these investments require a foods.128 ports can be found in a heterogeneous state of public commitment and regulatory function to co- 3.120. Employ risk-based selectivity processes for in- evolution. At the most problematic levels, many ordinate the shared commitment and benefits.131 spections. This will help ensure fair selection of Central American ports remain mired in 1980s 3.125. Encourage consolidation or coordination of small inspections, minimize opportunities for corrup- practices. They are yet to introduce private ter- private operators. While decentralization and com- tion, and reduce the need for high-percentage in- minal operators, which bring modern electronic petition have helped many of the region’s ports, spections. Haiti reduced export time by one day by tracking of containers, links to global shipping atomization is a regressive curve of problems for introducing risk-based inspections. networks, investments in labor-reducing and port operations. Several island countries in the Ca- 3.121. Harmonize customs standards for subregions. time-saving gantry cranes, transtainers, and other ribbean, as well as Guyana and Belize, have many Particular logistics-related challenges to imports modern cargo handling equipment. The vessels small private terminals, often controlled by verti- and exports can best be addressed at the region- that call those ports generally carry their own cally integrated operators. Most of these countries al or subregional level. In Central America, for in- “gear” or cranes and are old, small, and inefficient. do not have a unifying port authority, a national stance, customs harmonization to facilitate the The cost of poor shipping services willing to call transport plan, an active port regulator, or any way transport of goods and to reduce the costs of doing these inefficient port conditions is passed on to of monitoring private operators’ anticompetitive so is a regional issue.129 consumers and local producers.130 practices. Given the importance of cargo agglomer- 3.122. Policy guidance for maritime service and port re- 3.124. Anticipate growth and invest in landside and water- ation, planned development, competition in carrier form—develop coherent port strategies. Agricul- side capacity. In concessioning and decentralizing services, and access of third-party cargo to port ture and food ocean shipping costs are affected by the region’s ports, many LAC countries washed facilities, governments may benefit from estab- port efficiencies, port capacity, cargo agglomera- their hands of all port-related investments. While lishing or strengthening the regulatory oversight tion, and the level of connectivity and competition private operators made quick gains in efficien- of these facilities and, in some cases, encouraging in the global liner shipping network. Countries with cy and turnaround with “superstructure” invest- their consolidation.132 coherent port development strategies that link to ments— such as cranes, handling equipment, inland networks, allow for cargo agglomeration, and refrigerated storage—larger investments in provide for fast turnaround of large vessels, and greater yard capacity, deeper channels, wider use antitrust regulations to ensure competition turning basins, on-dock railroads, and better land- among carriers can benefit from faster services, side access for roads was beyond the capacity and contractual commitments of individual private operators. Moreover, ports with multiple private 128 Schwartz and others 2009. 131 Schwartz and others 2009. 129 Schwartz and others 2009. 130 Schwartz and others 2009. 132 Schwartz and others 2009. Chapter 3. 119 Chapter 4. Assessing Latin America and the Caribbean’s Contribution to Global Food and Feed Security in 2050 133 133 This chapter is drawn largely from a background paper describing the results of research carried out for this report by Simla Tokgoz, Prapti Bhandary, and Mark Rosegrant of the International Food Policy Research Institute, “Forces shaping present and future agricultural trade trends in Latin America and the Caribbean: alternative scenarios.” Trucks and bulldozer on roadway. Brazil. Photo: Thomas Sennett / World Bank 4.1. Food, fiber, and fuel production needs to increase 4.5. Message 4: Agriculture consumes approximately dynamics (volatility in food and energy markets, about 80 percent by 2050 to meet global demand. 70 percent of the world’s freshwater supply. Water population growth, urbanization, and income How will Latin America and the Caribbean (LAC) ad- scarcity poses problems for increasing agricul- growth). There will be billions of new customers, dress regional and global food and fiber security tural production in several parts of the world, and and new market opportunities are set to emerge. concerns? This chapter summarizes the assess- climate change poses challenges to increasing Yet there are many threats to a sufficient response ment of global and regional drivers of LAC food ex- agricultural production. The tension between rap- by the supply side to meet this growing and chang- ports—natural or human-induced factors that di- idly rising natural resource consumption and envi- ing demand. Agricultural production systems in rectly or indirectly cause changes in demand and ronmental sustainability will be a critical pressure many countries are neither resource-efficient nor supply of food, in domestic and global markets point over the next decades. Pricing resource use producing to their full potential. The stock of natu- and in the global trading system. This assessment and adopting more sustainable practices in water ral resources like land, water, and energy is shrink- contributes to the debate on whether the world can and land management could help LAC green its ing relative to demand. This requires their use to feed itself and the role that LAC could play. growth and agricultural exports, though it could become more efficient to reduce environmental 4.2. Message 1: If current trends continue in income come with a lower increase in its market shares. impacts and preserve the planet’s productive ca- and population growth, use of technology and re- 4.6. Message 5: With few exceptions, LAC’s global mar- pacity. Moreover, as one of the major contributors sources to produce food and fiber, distortions in ket share in all four major food categories is forecast to climate change, the agricultural sector must agricultural and trade policies, and investments in to be higher than at present under all scenarios con- reduce greenhouse gas emissions substantial- irrigation and infrastructure, LAC’s share in global sidered here—in most cases substantially higher. ly. These changes are reflected in the prices of trade will probably be even bigger than it is today. The exceptions are a very small drop in LAC’s share major agricultural commodities. Starting in 2007, 4.3. Message 2: As markets grow, agricultural and of meat and a large drop in fruits and vegetables un- the prices of major grains increased dramatically trade liberalization, investment in irrigation, and der the pessimistic scenario, which includes the po- in real terms and peaked in 2008. Real prices of improved agricultural technology could help LAC tential impacts of global warming and a stable share these commodities declined in 2009 and 2010; play a larger role in meeting global food, fiber, and in meat under the rebalancing scenario. they did not rebound to their previous levels but biofuel demand. increased again in 2011. 4.4. Message 3: A major barrier to increasing produc- 4.1 4.8. This study uses the International Food Policy and tion and trade in LAC is poor infrastructure. Logis- Challenges of sustainable food Research Institute’s (IFPRI’s) International Model tics expenses to move products are greater in LAC security toward 2050 for Policy Analysis of Agricultural Commodities and than in most developed countries. Improving LAC’s Trade (IMPACT) global agricultural model to assess infrastructure and its participation in the global 4.7. Sustainable production of agricultural commod- long-term agricultural supply and demand relations. grid could vastly improve the efficiency and vol- ities and growth of international trade in these This chapter’s main objective is to project the global ume of agricultural production, which is essential goods are challenged as never before by sup- food supply gap, analyze various factors that con- to meet rising global demand for food, especially ply-side constraints (such as climate change and tribute to this gap, and assess LAC countries’ role in developing markets. natural resource exploitation) and by demand-side in filling this gap through agricultural exports. This 122 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development assessment of long-term projections uses the trade outcome analysis conducted in earlier chapters. 4.2 Demand dynamics and supply-side constraints 4.2.1 Demand dynamics 4.9. Demand for agricultural commodities is increasing Table 4.1 Annual percentage change in GDP worldwide, due to population growth in low-income countries that have limited potential to increase Country name 2000 2001 2002 2003 2004 2005 2006 2007 2008 domestic supply and thus are net food importers; Argentina -0.8% -4.4% -10.9% 8.8% 9.0% 9.2% 8.5% 8.7% 7.0% income growth in East and South Asia that allows Brazil 4.3% 1.3% 2.7% 1.2% 5.7% 3.2% 4.0% 5.7% 5.1% consumers to shift to protein-rich goods (meat and Chile 4.5% 3.4% 2.2% 3.9% 6.0% 5.6% 4.6% 4.7% 3.2% dairy products) away from staples (cereals); and Colombia 2.9% 2.2% 2.5% 4.6% 4.7% 5.7% 6.9% 7.6% 2.9% various biofuel initiatives that increase demand China 8.4% 8.3% 9.1% 10.0% 10.1% 10.4% 11.6% 13.0% 9.0% for agricultural feedstocks. Ecuador 2.8% 5.3% 4.3% 3.6% 8.0% 6.0% 3.9% 2.5% 6.9% 4.10. China and India have in recent years had very high Mexico 6.6% -0.2% 0.8% 1.4% 4.0% 3.2% 4.8% 3.2% 1.8% GDP growth rates, with no slowdown expected in India 4.0% 5.2% 3.8% 8.4% 8.3% 9.1% 9.7% 9.1% 7.1% Peru 3.0% 0.2% 5.0% 4.0% 5.0% 8.9% 7.7% 8.9% 9.8% the near future (table 4.1). The Chinese economy Uruguay -1.4 -3.4% -11.0% 2.2% 11.8% 7.6% 4.7% 7.6% 8.9% grew at more than 9 percent annually over the last seven years, with double-digit growth rates Source: World Bank 2009. in most. India’s annual GDP growth rate averaged Note: GDP is computed using 2000 dollars more than 7 percent over the last eight years. LAC countries also had high GDP growth rates in the past decade, though not as high as in China and India. Argentina’s GDP growth rate was more than 7 percent after the 2002 economic crisis. The Bra- zilian GDP growth rate averaged 3.7 percent over the last eight years. Similarly, the Peruvian and Uruguay economies also grew substantially. Chapter 4. 123 4.11. Table 4.2 shows the change in emerging econo- mies’ consumption patterns brought about by in- come growth. China and India have significantly higher growth rates of milk and meat consumption than higher income countries that have already completed this transformation or lower income countries that do not have the means to purchase Table 4.2. Annual per capita consumption growth rates (2000–10) for meat and milk (percent) these protein-rich goods.134 Country Total meata Total fluid milk BRICs Brazil 2.61 1.70 China 1.88 13.14 India 4.76 2.29 Russian Federation 4.87 0.51 Higher income countries Australia 1.27 –2.82 Canada –0.72 –0.53 EU-27 0.49 0.11 United States –0.43 0.52 Lower income countries Egypt –2.18 –0.29 Philippines 1.01 1.27 Peru 0.70 2.24 Source: Calculated by authors using U.S. Department of Agriculture (2011). a. Total meat consumption includes beef, veal, swine, broiler, and turkey. 134 USDA 2011. 124 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 4.12. The other source of demand growth for agricultural increased to meet the new food demand levels? new arable land is in developing countries and feedstocks is the expanding biofuel sector.135 The The three main sources of growth in crop produc- some 80 percent of this amount is found in Latin Renewable Fuel Standard, renewed under the Ener- tion are increased exploitation of worldwide arable America and Sub-Saharan Africa. By contrast, there gy Independence and Security Act of 2007, sets out land and water resources, yield increases, and in- is virtually no spare land available for agricultural new mandates for the use of biofuels in the United creases in cropping intensity. expansion in South Asia and the Middle East and States.136 To achieve this mandate, more agricultural 4.15. Long-term studies suggest that there is as much North Africa, the regions with the highest popula- land may be diverted to the production of first-gen- potentially arable land available as the amount of tion growth. eration biofuels, resulting in a competition between land currently under cultivation. However, a num- 4.17. Productivity growth is the other critical component fuel and food. The European Council has laid out ber of steps, including land reclamation, construc- of agricultural supply increase. A number of key similar initiatives.137 Such directives and biofuel pol- tion, farm infrastructure, and investment capital, factors affect crop yields, including climatic, en- icy initiatives in both the United States and the EU have to be undertaken before the available ara- vironmental, technological, economic, and policy have led to increased production of biofuels and di- ble land is ready for production. In practice, the conditions. The factors that increase crop yields version of agricultural feedstocks for the production amount of arable land worldwide has grown by a include the development of new varieties, tech- of such fuels as ethanol and biodiesel. net average of 5 million hectares a year over the nological diffusion, input use (fertilizer), land im- 4.13. These changes in demand dynamics also affected last two decades.138 This means that it could take provements, adoption of conservation tillage tech- trade patterns of agricultural commodities. China a number of years to prepare all the potential new niques, denser planting, earlier planting, irrigation, and India have become important LAC customers, arable land for agricultural production. pest control, and weed control. The factors that though they are not the only countries that have 4.16. But in the long run, much of the land for future pro- hurt crop yields include land degradation, adverse increased their LAC imports. These developments duction is likely to come from LAC. Deininger and climate conditions, and limited resource condi- have contributed to growth of agribusiness in LAC others (2011) estimate that of the approximately tions. Further, many countries may have reached a and created new opportunities for these economies. 445.6 million hectares of land worldwide poten- yield plateau. However, for many developing econ- tially suitable for sustainable expansion of culti- omies, there is still room for yield growth through 4.2.2 vated area, about 28 percent is in Latin America, increased input use, technological change, and Supply-side constraints more than in any other region except Sub-Saharan better farm management. Africa.139 Accessibility considerations magnify this 4.18. Since agriculture is a major consumer of the 4.14. The long-term increase in food demand driven by potential: the region has 36 percent of the 262.9 world’s fresh water, water scarcity is a critical increasing population and economic growth will million hectares of such land situated within six problem affecting food security. Water scarcity is have to be matched by subsequent increases in hours’ of the closest market. Similarly, Rabobank severe in an increasing number of regions, such food production. But how can food production be (2010) finds that two-thirds of potentially suitable as the Middle East and North Africa and East and South Asia. LAC is also well endowed in renewable 135 Oxfam International 2008; Runge and Senauer 2008. water resources, with about a third of the 42,000 136 EPA 2011. 138 FAOSTAT 2011. cubic kilometers available worldwide. Per capita, 137 EUROPA 2011. 139 Deininger and others 2011. Chapter 4. 125 LAC has the highest endowment in this resource tial? This chapter will provide a selective assess- in for regions with falling productivity, reducing among developing regions, though some subre- ment of global and regional drivers of LAC food ex- their export potential? gions face higher than average scarcity.140 Precip- ports based on the diagnostic analysis conducted 4.23. Key driver 3: Infrastructure and logistics deficit in itation changes that accompany climate change in earlier chapters. The drivers explored are natural LAC. A major barrier to increasing production and will exacerbate water shortages in some parts of or human-induced factors that directly or indirect- trade in LAC is poor infrastructure, as shown in the world while increasing water availability in oth- ly cause changes in demand and supply of food, chapter 3. Logistics expenses to move products ers. As agriculture is the largest user of fresh water, in domestic and global markets and in the global are greater in LAC than in most developed coun- improving irrigation efficiency will be essential for trading system. This assessment thus contrib- tries. Improving LAC’s infrastructure and its partic- sustainable food production as well as for meeting utes to the debate on what will the world will need ipation in the global grid could vastly improve the increased demands for drinking water and indus- to feed itself in the future. Through LAC-focused efficiency and volume of agricultural production, trial needs. lenses and based on the demand dynamics and which is essential to meet rising global food de- 4.19. Climate change adds further pressure to the dra- supply-side constraints discussed above, we will mand, especially in developing markets. matic transformation of global agricultural mar- examine the three directions of change that could 4.24. Key driver 4: Greening growth and trade. Agricul- kets, due to its effect on local temperature and pre- restructure the world food economy for the fore- ture consumes approximately 70 percent of the cipitation conditions. A recent IFPRI report (Nelson seeable future. The stresses and tensions will be world’s freshwater supply. Water scarcity poses and others 2010) projected that world crop prices great, offering not only the richest opportunities problems for increasing agricultural production will increase significantly due to climate change, for LAC to improve its contribution to feeding the in several parts of the world, and climate change income growth, and population growth. It also em- world but also representing the biggest threat to poses growing challenges to increasing agricultur- phasized that climate change adds a further chal- the sustainable management of its resource base. al production. The tension between rapidly rising lenge to food security and agricultural productiv- 4.21. Key driver 1: Emerging market growth and the natural resource consumption and environmental ity. One main message of the report is that “trade “global harmonious rebalancing. As markets grow, sustainability will be a critical pressure point in flows can partially offset local climate change pro- agricultural trade liberalization, investment in ir- the next decades. Pricing resource use and adopt- ductivity effects, allowing regions of the world with rigation, and improved agricultural technology ing more sustainable practices in water and land positive (or less negative) effects to supply those could help LAC play a larger role in meeting global management could help LAC green its growth and with more negative effects.”141 food, fiber, and biofuel demand. agricultural exports. 4.20. With the need to almost double food, fiber, and fuel 4.22. Key driver 2: Climate change, trade, and food secu- 4.25. The remainder of the chapter is structured as fol- production by 2050 to meet global demand, we rity. With climate change adding further pressure lows: ask in this chapter the following questions: What on food production systems to meet the challenge • Section 4.3 outlines the methodology used for role can LAC play? How can it achieve its full poten- of feeding an increasing world population, could this analysis. trade flows partially offset local climate change • Section 4.4 assesses projected long-term global shocks by allowing regions with positive effects to food demand and supply in a business-as-usual 140 Bruinsma 2009. supply those with more negative effects or to step or current-path case scenario. 141 P. xviii. 126 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development • Section 4.5 presents outcomes from a scenario make using large spatial units inappropriate for cereals, oilseeds, oilcakes, vegetable oils, roots emphasizing higher emerging market growth water resource assessment and modeling. and tubers, meats, milk, sugar, fruits, and vegeta- under a more harmonious global rebalancing 4.27. IMPACT generates projections for agricultural crop bles.142 scenario. area and crop yields as a function of global market 4.28. In addition, IMPACT incorporates all Latin American • Section 4.6 presents outcomes from a pessimis- drivers (such as commodity demand and prices) countries either individually or as part of a regional tic view of the world economy characterized by and local availability of water resources. Crop area aggregate. Argentina, Brazil, Chile, Colombia, Ec- low income growth, high population growth, and and yields also depend on projected rate of exog- uador, Mexico, Peru, and Uruguay submodels are climate change. enous (nonprice) growth trends, labeled “intrinsic included individually for all commodities. IMPACT • Section 4.7 looks at outcomes of a greater focus growth rates.” Further, in IMPACT, food security in- also includes the United States, China, India, the on improving LAC’s business environment and dicators, such as the percentage and number of EU, and Middle East and North African countries, logistics. malnourished children younger than age five and which allows for incorporating fast-changing food • Section 4.8 analyzes the consequences of a per capita calorie consumption, are also projected demand dynamics and income growth in these greener growth path on LAC’s agricultural trade based on various indicators. IMPACT uses a system regions. contribution. of supply and demand elasticities, incorporated 4.29. The structure of this model is outlined in figure 4.1, • Section 4.9 summarizes the main findings and into a series of linear and nonlinear equations, to with the following key inputs: concludes the chapter. approximate the underlying production and de- • GDP projections to 2050. mand functions. World agricultural commodity • Base-year data (2010) for supply and demand 4.3 prices are determined annually at levels that clear levels, input, and resource use in agricultural Methodology used for the analysis the international markets. Growth in crop produc- production. tion in each country is determined by crop and in- • Assumptions for different scenarios involving 4.26. IMPACT is a global multimarket, partial-equilibrium put prices, exogenous rates of productivity growth GDP growth rates, population growth rates, and model that provides long-term projections of glob- and area expansion, investment in irrigation, and consumption patterns for high-income coun- al food supply, demand, trade, prices, and food water availability. Demand is a function of prices, tries and the fast-developing economies of security. It is integrated with a Water Simulation income, and population growth and contains five Brazil, the Russian Federation, India, and China Model that balances water availability and uses categories of commodity demand—food, feed, (BRICs); second-generation biofuel agricultural within various economic sectors at the global and biofuel feedstock, crushing demand, and other and trade policy wedges on the demand side; regional scale. IMPACT uses a finer disaggregation uses. In IMPACT, extensive market links exist, re- and fertilizer prices, agricultural productivity of 281 “food-producing units,” which represent the flecting derived demand for feed in livestock and growth rates for crops, productivity growth for spatial intersection of 115 economic regions and dairy sectors, competition for land in production, livestock producers, irrigated area growth rates, 126 river basins out of recognition that significant and consumer substitution possibilities for close and climate change on the supply side. climate and hydrologic variations within regions substitutes. IMPACT is an agricultural model with 142 A detailed description can be found in Rosegrant and others (2008) or 46 crop and livestock commodities, including at www.ifpri.org/book-751/ourwork/program/impact-model. Chapter 4. 127 4.4 Agricultural exports outlook to 2050: the current-path case 4.30. In the current-path case, our future world results from current patterns of economic, social, and environmental activity, with an increasing popu- Figure 4.1. Structure of the International Model for Policy Analysis of Agricultural lation and economic growth driving the long-term Commodities and Trade model increase in food demand. Based on United Nations projections, the world population is expected to Urban growth & Area elasticities grow approximately 34 percent from 2010 to changes in food habitats more than 9 billion in 2050. While this represents Income growth a considerable slowdown in population growth projections Yield elasticities compared with the last 45 years—in which the Populations projections Area and yield growth world population doubled—it marks an increase of some 2.4 billion people on top of this already high Domestic prices number. The majority of this growth will, more- Demand Supply over, take place in developing countries, while the projections projections populations of high-income countries will fall. The Climate world will continue to undergo rapid urbanization Net Trade Change LAC Technology over the next decades. Approximately 67 percent of the world’s population will live in urban areas by Emerging 2050, compared with today’s 52 percent. Average World Trade markets income is expected to grow in the coming decades Source: based on model description in Rosegrant et al (2008). (table 4.3). Income growth rates are also assumed to be much higher in developing countries than in developed countries, with emerging economies having the highest economic growth worldwide. As incomes rise, more consumers move from a cere- al-based diet to an animal protein diet and a higher consumption of fruits and vegetables. For incen- tives on the agricultural supply side, current pat- 128 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development terns of fertilizer use, growth of second-generation the traditionally dominant North-South trade from biofuels, agricultural productivity, and investment developed to developing countries will continue to in irrigation will continue to 2050. lose ground in favor of South-South trade from de- 4.31. What are the consequences of this differential in veloping to developing countries, a pattern largely world population and income growth rates with driven by LAC’s increasing role in trade. In 2050, current patterns of resource use in agricultural LAC is expected to be a bigger player than it is to- Table 4.3. Annual average growth in GDP systems? First, we could witness an important day, with its share of global exports rising from 8 between 2010 and 2050 (percent) shift in the share of global food demand and sup- percent to 11 percent in cereals, from 25 percent ply among regions. For meat and cereals, the world to 34 percent in fruits and vegetables, and from 30 Business Region would need to be producing 65 percent more meat percent to 36 percent in meat (table 4.4). Figures as usual and 50 percent more cereals to meet expected 4.7 and 4.8 show the commodity composition of East Asia and the Pacific 5.4 demand arising from the combined population LAC’s increasing role with export growth rates and Europe and Central Asia 3.5 and income growth. East Asia and the Pacific, the volumes. A broad spectrum of commodities con- biggest meat- and cereal-producing region af- tributes to this growth with—most notably—beef Latin America and the Caribbean 3.7 ter the group of developed countries, will see its and poultry on the animal side and soybeans, Middle East and North Africa 3.8 shares fall. By contrast, LAC will be substantially fruits, and vegetables on the vegetal side. increasing its share of production. Overall, LAC’s South Asia 6.1 contribution to global production growth will be Sub-Saharan Africa 4.9 around 22 percent for meat and 14 percent for ce- Developed 2.5 reals (figures 4.2 and 4.3). Developments on the demand side (figures 4.4 and 4.5) —substantial Developing 4.7 increases in meat and cereal per capita consump- World 3.2 tion and that of other food items—will result in a reconfigured structure of the global agricultural Source: Tokgoz, Bhandary and Rosegrant (2012) commodity trade. Figure 4.6 shows the regional trade flows for cereals, animal products, and fruits and vegetables projected for 2050. LAC will contin- ue to be a leading exporting region for fruits and vegetables and animal products, while Europe and Central Asia is one of the most important surplus regions for cereals along with the developed coun- try group. Also, increasingly over the next decades, Chapter 4. 129 Figure 4.2. Global meat production in a current-path case (thousands of metric tons) 2010 2050 160.000 140.000 120.000 100.000 80.000 60.000 40.000 20.000 0 East Asia and Developed Latin America Europe and South Asia Sub-Saharan Middle East and Pacific and the Central Asia Africa North Africa Caribbean Source: Tokgoz, Bhandary and Rosegrant (2012) Figure 4.3. Global cereal production in current-path case (thousands of metric tons) 2010 2050 1.200.000 1.000.000 800.000 600.000 400.000 200.000 0 Developed East Asia and South Asia Europe and Latin America Sub-Saharan Middle East Pacific Central Asia and the Africa and North Caribbean Africa Source: Tokgoz, Bhandary and Rosegrant (2012) 130 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development Figure 4.4. Business-as-usual meat food demand (kilograms per capita) 2010 2050 99 90 74 74 56 58 50 48 49 39 41 40 32 26 16 19 7 12 East Asia and Europe and Latin America Middle East and South Asia Sub-Saharan Developed Developing World Pacific Central Asia and the North Africa Africa Caribbean Source: Tokgoz, Bhandary and Rosegrant (2012) Figure 4.5. Business-as-usual cereal food demand (kilograms per capita) 2010 2050 215 223 176 171 165 171 152 149 144 121 132 116 118 116 Middle East and East Asia and Europe and South Asia Latin America Sub-Saharan Developed North Africa Pacific Central Asia and the Africa Caribbean Source: Tokgoz, Bhandary and Rosegrant (2012) Chapter 4. 131 Figure 4.6. Business-as-usual net agricultural export patterns in 2050 (thousands of metric tons) a. Cereals b. Animal Products c. Fruits and Vegetables 250000 30000 120000 200000 25000 100000 150000 20000 80000 100000 15000 60000 10000 40000 MENA 50000 EAP SSA 5000 20000 DEV EAP EAP SSA SSA SA SA SA 0 0 0 LAC DEV ECA -50000 MENA ECA LAC -5000 -20000 MENA ECA DEV LAC -100000 -10000 -40000 -150000 -15000 -60000 -200000 -20000 -80000 Source: Tokgoz, Bhandary and Rosegrant (2012). Note: SSA = Sub-saharanAfrica; EAP = East Asia and Pacific; SA = South Asia; MENA = Middle East and North Africa; LAC = Latin America and the Caribbean; DEV = Developed economies; ECA = Easter Europe and Central Asia. Table 4.4. Business-as-usual share of global agricultural trade, 2010 and 2050 (percent) Cereals Fruits and vegetables Meat 2010 2050 2010 2050 2010 2050 Developed countries 63.0 45.0 17.0 19.0 38.0 36.0 Latin America and the 8.0 11.0 25.0 34.0 30.0 36.0 Caribbean Rest of developing 29.0 44.0 58.0 47.0 32.0 28.0 countries Source: Tokgoz, Bhandary and Rosegrant (2012) 132 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development Figure 4.7. Business-as-usual annual growth rates for Latin America and the Caribbean 4.5 commodity exports, 2010–50 (percent) Emerging market growth and the reshaping of the global 10,0% food economy 8,0% 6,0% 4.32. A first alternative to the current-path case explored 4,0% Sorghum in the previous section is an optimistic harmoni- Wheat 2,0% Rice ous global rebalancing scenario in which: 0,0% • The center of economic growth shifts at a more Pork Milk Vegetables Beef Poultry Tropical Fruits Maize Soybeans Rapeseed Soybean Meal Soybean Oil Temperate Fruits -2,0% rapid pace from developed to developing coun- -4,0% tries characterized by higher growth rates in de- -6,0% veloping countries and lower population growth -8,0% rates. -10,0% • As incomes rise, more consumers move from Source: Tokgoz, Bhandary and Rosegrant (2012) a grain-based diet to an animal protein diet but become more concerned about the origin and quality of the products; in particular, consum- Figure 4.8. Business-as-usual volumes of Latin America and the Caribbean exports, ers in high-income countries and BRICs adopt a 2010 and 2050 (thousands of metric tons) healthier diet by reducing their meat consump- tion. Net Exports in 2010 Net Exports in 2050 • Second-generation biofuels start earlier than as- 80.000 sumed in the current-path case. 70.000 • Farmers have better access to inputs, irrigation 60.000 investments are increased, and improvements 50.000 of crop technologies lead to higher growth rates 40.000 of yield adoption. 30.000 20.000 • Policy distortions in world markets are further 10.000 reduced. 0 • Climate change is perfectly mitigated. -10.000 Beef Pork Poultry Milk Rice Wheat Maize Soybeans Soybean Soybean Vegetables Tropical Temperate 4.33. One consequence of these potential develop- -20.000 Meal Oil Fruits Fruits ments in food demand and on the supply side is that LAC remains a big contributor to global food Source: Tokgoz, Bhandary and Rosegrant (2012) Chapter 4. 133 security despite a slightly slower growth in its Figure 4.9. Shares in global trade in the optimistic harmonious rebalancing scenario animal products and cereal exports (figure 4.9). As consumers in the high-income countries and Cereals Animal Products Animal Products BRICs shift to a healthier, less meaty diet, LAC’s Latin America and the Caribbean Latin America and the Caribbean Latin America and the Caribbean exports of fruits and vegetables may compensate Developing Developing Developing for the less pronounced increase in the share of Developed Developed Developed animal products and cereals (figures 4.10–4.12). For fruits and vegetables, the biggest market that 8% 11% 9% could open for LAC is in East Asia and the Pacif- 29% 30% 36% 30% 30% 36% 30% 44% 46% ic. For animal products, LAC could be competing 32% 28% 31% 32% 28% 31% 63% against exports from high-income countries, while 45% 45% 38% 36% 39% 38% 36% 39% countries in Europe and Central Asia increase Share in World BAU-2050 Optimistic Share in World BAU-2050 Optimistic Share in World BAU-2050 Optimistic their cereal exports arising from higher needs in Net Exports in Scenario Net Exports in Scenario Net Exports in Scenario 2010 2010 2010 Sub-Saharan Africa and East Asia. A more harmo- nious rebalancing would translate into lower cere- Source: Tokgoz, Bhandary and Rosegrant (2012). als exports for the region’s main cereal exporters Note: BAU = Business as usual. (Argentina and Brazil); expanded opportunities for Figure 4.10. Net exports of fruits and vegetables under the rebalancing scenario in 2050 fruits and vegetables exports for such countries as (thousands of metric tons) Mexico, Chile, and Argentina; and still higher meat exports for Brazil and Uruguay (figure 4.13). Fruits & Vegetables BAU Optimistic Scenario 150000 100000 50000 SSA SA EAP DEV 0 MENA ECA LAC -50000 -100000 Source: Tokgoz, Bhandary and Rosegrant (2012). Note: BAU = Business as usual. 134 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development Figure 4.11. Net exports of animal products under the rebalancing scenario in 2050 (thousands of metric tons) Animal Products BAU Optimistic Scenario 30000 20000 10000 EAP SSA SA MENA 0 ECA DEV LAC -10000 -20000 -30000 Source: Tokgoz, Bhandary and Rosegrant (2012). Note: BAU = Business as usual. Figure 4.12. Net exports of cereals under the rebalancing scenario in 2050 (thousands of metric tons) Cereals BAU Optimistic Scenario 250000 200000 150000 100000 50000 SSA EAP SA MENA 0 LAC DEV ECA -50000 -100000 -150000 -200000 -250000 Source: Tokgoz, Bhandary and Rosegrant (2012). Note: BAU = Business as usual. Chapter 4. 135 4.6 Food security, trade, and climate change 4.34. The first decade of the 21st century has brought signs of a troubled future for global food securi- ty. The food price spikes have prevented millions of people from escaping extreme poverty.143 The Figure 4.13. Net exports for selected Latin America and the Caribbean countries record prices in 2008 kept or pushed 105 million under the rebalancing scenario in 2050 (thousands of metric tons) people below the poverty line in the short run. Cereals Fruits & Vegetables Animal Products They hit urban poor and female-headed house- holds hardest. While food prices dropped sharply BAU Optimistic Scenario BAU Optimistic Scenario BAU Optimistic Scenario in 2009 with the financial crisis, they quickly re- 45.000 45.000 40.000 18.000 bounded and by early 2011 were almost back to 40.000 16.000 35.000 35.000 14.000 2008 levels. Sudden, unexpected increases in 30.000 30.000 12.000 25.000 10.000 food prices impose particularly severe hardship 25.000 20.000 8.000 on many households because they need time to 20.000 6.000 15.000 15.000 10.000 4.000 adjust to higher prices. The large, initial impact on 10.000 5.000 2.000 5.000 0 0 poverty of a rise in food prices tends to decline as 0 production increases and the income of the poor in Argentina Uruguay Peru Brazil Mexico Caribbean and Central Chile Argentina Ecuador Peru Brazil Chile Ecuador Argentina Brazil Uruguay rural areas rises, but it is usually not large enough to offset the initial negative impact on poverty in the short run. Source: Tokgoz, Bhandary and Rosegrant (2012). 4.35. The factors that caused the price spikes can also Note: BAU = Business as usual. make prices more volatile and thus less pre- dictable. Biofuel mandates, which have boosted demand for grains despite slowing global food demand, have reduced the price elasticity of de- mand for grains. Sharp increases in fertilizer pric- es, linked to energy prices, have made production 143 World Bank 2012a. 136 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development costs more volatile and, to the extent that higher as the global recovery in the second decade of the (2035 rather than 2030), thus increasing demand prices have reduced the use of fertilizers, have 21st century shows signs of stalling amid deteri- for agricultural feedstock. On the supply side, a made yields less stable. Low global stocks have orating financial conditions. Global growth slowed pessimistic view assumes constraints in access contributed to price volatility at time-of-production to 3.9 percent in 2011 and was projected to de- to inputs, lower agricultural productivity growth, shortfalls. Moreover, trade interventions meant to cline further to 3.5 percent in 2012. The strongest and difficulties expanding land area and access to stabilize domestic prices often have had the oppo- slowdown is being felt in advanced economies, water.146 The pessimistic view also includes lower site effect and increased price volatility globally.144 but the worsening external environment and some GDP growth rates and higher population growth 4.36. To these already daunting challenges, climate weakening in internal demand is expected to lead rates for all countries in IMPACT. change adds further pressure, with adverse to lower growth in emerging and developing coun- 4.38. International trade flows provide a balancing weather patterns becoming more frequent and tries as well. This outlook is subject to downside mechanism for world agricultural markets. Coun- more variable. Because food production depends risks, such as a much larger and more protracted tries with a comparative advantage in a crop or on local temperature and precipitation conditions, bank deleveraging in the Eurozone or a hard land- livestock activity can produce it more efficiently any changes require farmers to adapt their prac- ing by key emerging economies.145 But how big are and exchange it for other goods with other coun- tices, and this adaptation requires resources that the food security challenges caused by a combi- tries whose comparative advantage lies else- could be used for other purposes. Farmers every- nation of slow global growth and climate change, where. But comparative advantage is clearly not where will need to adapt to climate change. For a and who will be most affected? fixed. Climate change on the supply side and few, the changes might ultimately be beneficial, 4.37. A second alternative to the current-path case pre- changing consumer preferences on the demand but for many farmers, analyses—for example, Nel- sented in section 4.4 is offered in a pessimistic side alter comparative advantage. While income son and others (2010)—point to major challenges view of the world economy, characterized by low growth itself changes the mix of goods demand- to productivity and more difficulties in managing income growth, high population growth, and cli- ed by consumers, population growth changes risk. The agricultural system will struggle to supply mate change. One main difference between this consumption volume and resulting trade levels. adequate quantities of food to maintain constant scenario and business as usual is that it incorpo- Agricultural trade flows depend on the interaction real prices. And the challenges extend further: to rates climate change. Specifically, it implements between comparative advantage in agriculture national governments, which must provide the the MIROC A1B climate change scenario based on (as determined by agroecological conditions and supporting policy and infrastructure environment, Nelson and others (2010). MIROC A1B assumes and to the global trading regime, which must en- a wetter and warmer future. It also assumes that 146 We also capture the effect of farmers’ difficulty accessing inputs by sure that changes in comparative advantage BRICs increase their meat consumption and be- increasing fertilizer prices in the world market. This is achieved by translate into unimpeded trade flows to balance come more like high-income countries and that increasing the growth rate of fertilizer prices 25 percent. We analyze world supply and demand. But this balancing second-generation biofuels start five years later the importance of crop productivity, lowering the intrinsic productivity growth rates for crop yields using the same rate of change from the might prove not to be a straightforward task today, than assumed in the business-as-usual scenario global rebalancing scenario. We also reduce the productivity growth rate for livestock producers (beef, pork, poultry, lamb, eggs, and milk) by 0.25 units. And we reduce intrinsic area growth rates for irrigated 144 World Bank 2012a. 145 World Bank 2012a. crop area for all crops in all countries by 0.15 units. Chapter 4. 137 resource endowments) and a context-specific set 4.40. Second, climate change and less bright economic 4.41. The findings above are particularly important for of local, regional, national, and international pro- development prospects make LAC adjust its agri- Argentina and Brazil. A global trading system with duction and trade policies. In the combination of cultural trade flows by contributing a bigger share limited protectionism allows comparative advan- changes in agricultural supply and demand condi- of global cereal and oilseed exports and a lower tage to be more fully exploited. Restrictions on tions explored in a pessimistic view, the interest is share of meat exports (figure 4.14). LAC’s share trade risk worsen the effects of climate change by in how the LAC region could contribute to meeting in global trade of fruits and vegetables takes a reducing the ability of producers and consumers the challenge of achieving sustainable food pro- big hit, mainly as a result of changes in weather to adjust. If climate change reduces the productiv- duction by 2050, a challenge made more difficult conditions. Maintaining a business-as-usual ap- ity of certain crops in some regions and does not by a growing world population and increasingly proach allows LAC to increase its contribution from increase productivity adequately in others, trade negative productivity effects from climate change. its current share of 25 percent to about one-third. cannot fully compensate for the global reduction 4.39. First, we found that trade flows are a potentially Climate change and lower income growth reduce in productivity. Given the increasing role that LAC’s important climate change adjustment mecha- this share significantly to a mere 13 percent. Fig- top Argentina and Brazil could be called on to play nism. Production effects of a wetter and warmer ure 4.15 shows that the region’s surpluses in veg- in cereals trade, reducing restrictions becomes a climate result in a dramatic decline in developed etables and tropical fruits are curtailed to a large priority. Trade flows can partially offset local cli- country exports—particularly those from big ag- extent; maize is the only major product that shows mate change productivity effects, allowing world ricultural exporters like the United States. In all an increase in exports. Figures 4.16–4.18 show by regions with positive (or less negative) effects to product categories included in the simulations, commodity the adjustment by LAC countries as a supply those with more negative effects. the developed country market shares fall substan- result of the climate change and economic devel- tially. This contrasts with findings in studies sug- opment prospects in a pessimistic view. Argentina gesting that agricultural exports from developed and Brazil expand their trade in cereals, while Mex- to developing countries could increase as a result ico increases its dependence on the world market of climate change (for example, Parry and others (see figure 4.16). Except for Chile, LAC countries 2004). While developed countries are responsible will see their potential for fruits and vegetables for two-thirds of cereal trade today and expected exports significantly curtailed, with Brazil hit the to have about 45 percent of that trade in 2050, a hardest (see figure 4.17). LAC’s loss is a gain prin- scenario with climate change and lower growth cipally for South Asia, a region that could step into prospects leads to a reversal of contribution, with the trade as a result of climate change yielding the developing countries, including those in LAC more productivity for it than for LAC. Exports of capturing about 67 percent. A pessimistic view livestock products by the biggest LAC producers would also predict significantly lesser flows in oil- will fall almost a third, and Mexico will have to rely seeds, fruits and vegetables, and meat from devel- on the world market to fill its population’s higher oped countries. meat needs. 138 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development Figure 4.14. Shares in global trade in a pessimistic world view by commodity group (percent) Developed LAC Rest of Developing 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% BAU in 2050 Pessimistic BAU in 2050 Pessimistic BAU in 2050 Pessimistic BAU in 2050 Pessimistic (Cereals) (Cereals) (F&V) (F&V) (Meat) (Meat) (Oilseeds) (Oilseeds) Source: Tokgoz, Bhandary and Rosegrant (2012). Note: BAU = Business as usual. Figure 4.15. Latin America and the Caribbean aggregate exports by commodity (thousands of metric tons) Net Exports in 2050 Pessimistic Scenario 80.000 60.000 40.000 20.000 0 -20.000 Beef Pork Poultry Milk Wheat Maize Soybeans Soybean Soybean Vegetables Tropical Temperate Meal Oil and Sub- Fruits -40.000 Tropical Fruits Source: Tokgoz, Bhandary and Rosegrant (2012). Chapter 4. 139 Figure 4.16. Cereals exports from Latin America and the Caribbean under the pessimistic scenario (thousands of metric tons) BAU 2050 Pessimistic Scenario 60.000 50.000 40.000 30.000 20.000 Northern Caribbean South and Central 10.000 America Colombia Peru Mexico America 0 Argentina Brazil Uruguay Central Ecuador Chile -10.000 South -20.000 America -30.000 Source: Tokgoz, Bhandary and Rosegrant (2012). Figure 4.17. Fruits and vegetables exports from Latin America and the Caribbean under the pessimistic scenario (thousands of metric tons) BAU 2050 Pessimistic Scenario 50.000 40.000 30.000 20.000 10.000 0 -10.000 Brazil Mexico Caribbean Chile Argentina Ecuador Peru Uruguay Colombia Northern Central -20.000 and Central South South America America America Source: Tokgoz, Bhandary and Rosegrant (2012). Note: BAU = Business as usual. 140 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 4.7 The global grid, Latin America and the Caribbean’s business environment, and the food connection 4.42. With the important developments in crop and live- stock technology, some argue that the problem of feeding the planet’s population does not arise Figure 4.18. Meat exports from Latin America and the Caribbean under the pessimistic scenario from insufficient food production; the world cur- (thousands of metric tons) rently produces more than enough food to feed the globe’s entire population. Rather, the problem BAU 2050 Pessimistic Scenario is inadequate distribution: food insecurity arises 18.000 simply because the world’s food supply is not dis- 16.000 tributed equitably among the human population, 14.000 12.000 due to distortions from faulty logistics infrastruc- 10.000 ture and inadequate business environments to 8.000 6.000 move goods efficiently. For LAC, this argument has 4.000 some merit. 2.000 0 4.43. Farming developments in Brazil substantiate the -2.000 prioritization of moving goods over producing -4.000 Brazil Argentina Uruguay Chile Peru Ecuador Caribbean Central Northern Colombia Mexico them. In the 1990/91 crop year, Brazil had around and Central South South America America America 9.7 million hectares in cultivation, producing more than 15.4 million tons of cereals. In 2010/11, land Source: Tokgoz, Bhandary and Rosegrant (2012). under cultivation reached 24.2 hectares, while Note: BAU = Business as usual. production grew to 72.2 million tons. According to government data, productivity almost doubled from 1.58 tons a hectare to 3 tons a hectare.147 Thanks to investments in land and research and development, better water management, and the expansion of productive agriculture in the 147 Financial Times, Investing in Brazil Supplement May 17, 2012. Chapter 4. 141 savannah-like cerrado, farmers in Brazil are very diary markups, and regulatory costs. It thus com- 4.46. The importance of providing better logistics will efficient but lose out to competitors for reasons bines the hard and soft infrastructure discussed vary by country and by current comparative ad- not necessarily related to productivity. Brazil and in chapter 3 and cannot distinguish between the vantage. The country results (figure 4.21), when other LAC countries, except for Bolivia, all have effects of the two. The effects of reducing these compared with a business-as-usual scenario, coastlines and seaports that provide outlets to costs through the IMPACT model provide an im- show that Argentina could increase its exports of international markets. However, as chapter 2 dis- portant reality check on the estimates in chapter cereals and meat more than 20 percent and those cusses, the costs of getting products to overseas 3, because the methodology used there cannot of fruits and vegetables about 15 percent. Brazil markets are a major concern, because producers account for market links, supply-side constraints, could increase its exports by 40 percent for cere- tend to be far from their principal markets and face or the effects of climate change. Using IMPACT, we als, 30 percent for fruits, and 20 percent for meat. internal logistics systems that are less developed can check to see, for example, whether water avail- The Caribbean and Central American countries than those of their main agrofood rivals for export ability or the effects of global warming might limit could increase their fruits and vegetables exports markets (mostly developed countries). the ability of LAC countries to respond to reduced around 25 percent, while Colombia (not shown) 4.44. In a better business and trade logistics scenario, logistics costs. could see them rise at least 200 percent. Chile, be- we offer a perspective on LAC country economies 4.45. Reassuringly, and consistent with the findings in cause of its already good logistics platform, sees and trade potentials if institutional reforms and chapter 3, lower marketing margins are found to less pronounced improvements but still could reap infrastructure developments to improve their busi- have a substantial impact on the region, expand- 15 percent more meat exports and 5 percent more ness environment and trade logistics take place,148 ing trade potential with a 100 percent increase fruit and vegetable exports. Peru’s improvement assuming they reach the standards of Organisa- in cereals and about a 20 percent increase in an- in logistics makes its exports of fruits and vegeta- tion for Economic Co-operation and Development imal products and fruits and vegetables over the bles jump 25 percent. The potential for capturing a (OECD) countries by 2050. In the IMPACT model, business-as-usual scenario (figure 4.19). Net LAC higher share in global markets is thus substantial these improvements are incorporated by reducing exports could increase significantly for beef, pork, across a wide range of the region’s commodities the marketing margins—which indicate the wedg- lamb, poultry, milk, maize, other grains, vegeta- and countries. es between border prices and domestic producer bles, temperate fruits, and tropical and subtropical prices—for agricultural products. The marketing fruits (figure 4.20). The change in several import- margin is an aggregate measure of all the factors ant global markets is noteworthy. LAC switches that create a divergence between border and do- from a net importer to a net exporter for rice and mestic prices, including transport costs, interme- significantly improves its wheat trade position. LAC significantly increases its soybean oil exports 148 This is simulated in the IMPACT model by incorporating lower mar- and reduces soybean meal exports, as meal pro- keting margins for LAC countries for the agricultural commodities. Specifically, comparisons are made between the weighted-average duction gets reallocated to meet the higher needs marketing margins in IMPACT for LAC and for the OECD regions. of the livestock industry driven by the increase in Marketing margins in LAC countries are reduced to the average OECD meat exports. levels starting in 2001 and ending in 2050. 142 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development Figure 4.19. Predicted impact of better business and logistics on Latin America and the Caribbean’s net exports BAU Better Business & Logistics 140000 120000 100000 80000 60000 40000 20000 0 Cereals Livestock Fruits and Vegetables Source: Tokgoz, Bhandary and Rosegrant (2012). Note: BAU = Business as usual. Figure 4.20. Annual growth rates in exports as a result of better business and logistics BAU Better Business & Logistics 12,0% 10,0% 8,0% 6,0% 4,0% Soybean 2,0% Meal 0,0% Rice Soybean Milk Maize Poultry Tropical Beef Pork Vegetables Temperate Soybeans -2,0% Oil Fruits Fruits -4,0% -6,0% -8,0% Source: Tokgoz, Bhandary and Rosegrant (2012). Note: BAU = Business as usual. Chapter 4. 143 4.8 Green growth and agricultural exports in Latin America and the Caribbean Figure 4.21. Country impact of better business and logistics by main commodity groupings 4.47. Over the past 20 years, economic growth has lifted more than 660 million people out of poverty and Cereals Meat has raised the incomes of millions more, but growth BAU Better Business BAU Better Business has too often come at the environment’s expense. Due to market, policy, and institutional failures, 60000 25000 Earth’s natural capital tends to be used in eco- 50000 20000 nomically inefficient and wasteful ways, without 40000 15000 sufficient reckoning of the social costs of resource 30000 10000 20000 depletion and without adequate reinvestment in 5000 10000 other forms of wealth. These failures threaten the 0 0 long-term sustainability of growth as well as prog- Argentina Brazil Uruguay Central Brazil Argentina Uruguay Chile Peru Ecuador South America ress on social welfare. Moreover, despite the gains from growth, 1.3 billion people still have no access Fruits to electricity; 2.6 billion still have no access to sanitation; 900 million lack safe, clean drinking BAU Better Business water; and more than 100 million children younger than age five remain underweight.149 Malnutrition 60000 prevalence in 2010 remained highest in low- and 40000 lower middle-income countries (23 percent and 20000 24.6 percent, respectively). Growth has not been 0 Brazil Mexico Caribbean and Chile Argentina Ecuador Peru inclusive enough.150 In this context, the tension Central America between rapidly rising resource consumption and environmental sustainability will continue to be Source: Tokgoz, Bhandary and Rosegrant (2012). one of the next decades’ critical pressure points. Note: BAU = Business as usual. Natural resources and commodities account for 149 World Bank 2012a. 150 World Bank 2012a. 144 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development approximately 10 percent of global GDP, support 4.50. First, it must recognize that, more so than in oth- two gases grew 35 percent, versus 16 percent every sector in the economy, and are a major er regions, land-use change—mainly deforesta- worldwide, with most of the increase attributable contributor to goods trade. For agricultural goods, tion and forest degradation—has been the major to nitrous oxide. Emissions per dollar of agricultur- the interplay of three powerful forces will deter- contributor to greenhouse gas emissions in LAC. al GDP, however, declined 21 percent, compared mine what resources we use, how we use them, Land-use change accounts for 62 percent of total with 15 percent worldwide. This indicates that the and what we pay for them: growing demand; con- emissions in LAC, versus 16 percent worldwide (in region’s growing share of world food markets has strained supply; and increased regulatory and so- 2005 and for CO2 emissions only). Agricultural ex- not been matched by a proportionate increase in cial scrutiny on how we produce agricultural goods pansion remains the main cause of deforestation. emissions. and how we use human, physical, and natural cap- An estimated 85 percent of deforestation in Brazil 4.52. The most important pillar of a strategy to reduce ital in the process. is caused by the creation of new pasture land for agriculture’s environmental footprint is to preserve 4.48. Demand for food and fibers will continue to grow, livestock. Deforestation also threatens the region’s existing forest cover and encourage reforestation particularly in emerging markets. As easy-to-tap vast biodiversity. Of the world’s 10 most biodiverse with native species where feasible. Substantial water resources and high-quality land are deplet- countries, 5 are in LAC: Brazil, Colombia, Ecuador, progress on this is the key to lowering the emis- ed, supply of agricultural goods will come from Mexico, and Peru. This list also includes 5 of the 15 sions trajectory, conserving biodiversity, and harder-to-access, more costly, and probably more countries whose fauna are most threatened with reducing erosion. Success will depend largely on politically unstable environments. Around the extinction.151 discouraging unsustainable livestock production. world, political leadership, regulatory and scien- 4.51. Apart from deforestation, agricultural produc- Some notable successes have been achieved. tific communities, and consumers are gravitating tion itself is a significant source of other import- Costa Rica, for example, has gone from having high to a new consensus based on reconciling the rap- ant greenhouse gases. These include methane deforestation rates to net forest increases. Uru- id growth required to bring developing countries (about 25 times more powerful per metric ton as guay also has achieved a net increase. Deforesta- to their aspired prosperity level and the need to a greenhouse gas than CO2),152 from livestock and tion has dropped significantly in Mexico. But in better manage the environment. Climate change irrigated rice production, and nitrous oxide (about many other parts of LAC, forest degradation rates may be the most contentious issue, but other is- 298 times more powerful), mainly from fertiliz- remain high, though both their extent and impact sues come into view: depletion of global fishing er. A large percentage of methane emissions (62 remain poorly documented. stocks, water scarcity, pollution, and food safety, percent for LAC, compared with 51 percent world- 4.53. Many LAC countries have eliminated the most and more. wide) and almost all nitrous oxide emissions come egregious policies that encouraged deforestation, 4.49. To stay on its trajectory of high output growth and from agriculture. Over 1990–2005 (the latest data such as giving possession or land titles to those poverty reduction and realize its full potential to available), LAC’s agricultural emissions of these who “improve” forests by cutting down trees. In- ramp up production in the face of climate change deed, many have formally banned deforestation, without increasing its environmental footprint, 151 www.usaid.gov/locations/latin_america_caribbean/issues/biodiversi- though enforcement of such bans has been limit- agriculture in LAC must become greener and more ty_issue.html. [AQ: ed. Brazil, the most important country in the fight 152 Data come from the National Council for Air and Stream Improvement “climate smart,” in several significant ways. against deforestation, has developed forest pro- (other estimates differ slightly). Chapter 4. 145 tection policies and projects to counter the rising veloping and developed countries. In line with the move toward specialization in downstream prod- pressure on forests at the expansion frontier and environmental objectives, efficiency in irrigation ucts in the soybean complex. Soybean oil could now has considerable experience in economic ac- increases, as do domestic and industrial use of grow 8 percent annually. tivity compatible with forest sustainability. Brazil’s water resources. Second-generation biofuels are 4.58. In figure 4.26, we provide a more detailed picture rate of deforestation has fallen 80 percent over assumed to start 5 years earlier than in the busi- for the LAC countries. In the green-growth scenar- the past six years, as the government carved out ness-as-usual scenario (2025 rather than 2030), io, Brazil and Argentina increase their meat, fruit, about 150 million acres for conservation—an area thus lowering demand for agricultural feedstocks and vegetable net exports while slightly reducing roughly the size of France. for the first-generation biofuels. We also include their cereal exports. The Caribbean, Central Ameri- 4.54. Apart from reducing and reversing deforestation, the effect of more efficient fertilizer use, mean- ca, and Central South America switch from net ex- other pillars of a climate-smart strategy for agricul- ing less nitrogen per kilogram of food production. porter to net importer of meat with higher domes- ture in LAC include153 We assume effective mitigation, so crop area and tic demand. Central South America and Uruguay • Making production and distribution more effi- yields are not affected by climate change for this increase their cereal net exports. Chile increases cient, by making better use of information and scenario. its meat and fruit and vegetable net exports while communication technology to increase efficien- 4.57. Figures 4.22–4.24 compare the regional net ex- Uruguay exports more of all commodities. cy and reduce postharvest losses and by devel- ports expected under a green-growth scenario 4.59. Overall, in the green-growth scenario, LAC’s share oping more ecological inputs industries. with those under a business-as-usual scenario in 2050 is expected to decline (relative to busi- • Managing water resources more efficiently, for 2050. We see the effect of multiple factors on ness as usual) from 36 percent to 33 percent for ensuring that energy plans make the best use net trade patterns. Higher productivity increases meat, from 34 percent to 33 percent for fruits and of each country’s hydroelectric potential and supply LAC, as well as its net exports, for some vegetables, and from 11 percent to 9 percent for designing dams for dual use (energy and irriga- commodities, including beef, lamb, other grains, cereals. Even in the green-growth scenario, howev- tion) wherever feasible. soybean meal, soybeans, and subtropical fruits. er, LAC’s shares in all these markets are forecast to 4.55. With these resource-related challenges, the ques- Maize net exports decline due to a reduction in increase over actual 2010 levels. tion that arises for LAC is whether a greener path of global demand when second-generation biofuels growth brings expanded or reduced opportunities eliminate some of the need for maize in ethanol for trade. We investigate this question in this part production. For some commodities, lower global of the assessment. prices shrink the market for LAC. If countries can 4.56. In the green-growth scenario, we present a view of meet domestic demand with domestic supply, the world economy that combines sustainability they import less from LAC. This is the case for pork, and environmental goals with development. This poultry, soybean oil, vegetables, and temperate view includes higher GDP growth rates for both de- fruits. Exports of processed products, such as soybean oil, will grow faster than the raw materials from which they are derived, showing a more rapid 153 See World Bank (2012b). 146 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development Figure 4.22. cereals net exports in 2050 under a green growth scenario (thousands of metric tons) BAU Greener Growth 300000 200000 100000 SSA EAP SA MENA 0 LAC DEV ECA -100000 -200000 -300000 Source: Tokgoz, Bhandary and Rosegrant (2012). Note: BAU = Business as usual. Figure 4.23. meat net exports in 2050 under a green growth scenario (thousands of metric tons) BAU Greener Growth 30000 20000 10000 EAP SSA SA 0 MENA ECA DEV LAC -10000 -20000 -30000 Source: Tokgoz, Bhandary and Rosegrant (2012). Note: BAU = Business as usual. Chapter 4. 147 Figure 4.24. Fruits and vegetables net exports under a green growth scenario in 2050 (thousands metric tons) BAU Greener Growth 150000 100000 50000 SSA SA EAP DEV 0 MENA ECA LAC -50000 -100000 Source: Tokgoz, Bhandary and Rosegrant (2012). Note: BAU = Business as usual. Figure 4.25. Annual growth rates in Latin America and the Caribbean exports under a green growth scenario in 2050 BAU Greener Growth 10,0% 8,0% 6,0% 4,0% Soybean 2,0% Meal Rice Wheat 0,0% Beef Pork Poultry Maize Soybeans Soybean Vegetables Tropical Temperate -2,0% Oil Fruits Fruits -4,0% -6,0% -8,0% -10,0% Source: Tokgoz, Bhandary and Rosegrant (2012). Note: BAU = Business as usual. 148 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development 4.9 Summary and conclusions Figure 4.26. Impact of green growth on countries in Latin America and the Caribbean in 2050 4.60. This chapter profiles world agricultural trade in 2050 and quantifies the impacts of shifts in the Cereals Fruits & Vegetables global economy representing various future path- ways. With the rapid growth of the emerging mar- BAU Greener Growth BAU Greener Growth kets contrasting with high unemployment in rich 45.000 countries, and threats of economic weaknesses 40.000 45.000 35.000 40.000 and climate change adding further pressure, the 35.000 30.000 30.000 global economy is experiencing a seismic shift. 25.000 25.000 20.000 20.000 This shift will have important implications for 15.000 15.000 10.000 10.000 production and consumption of agricultural com- 5.000 5.000 0 modities and trade flows between regions and 0 Brazil Mexico Caribbean Chile Argentina Ecuador Peru Argentina Brazil Uruguay Central South and Central countries. By 2050, the collective size of the econ- America America omies deemed emerging will be larger than the Animal Products developed world. At the same time, there could be a marked decline in the economic power of many BAU Greener Growth rich economies in the West.154 By 2050, the world’s population is likely to reach 9 billion. Most of these 20.000 18.000 people are expected to live in developing countries 16.000 and have higher incomes than today, which will re- 14.000 12.000 sult in increased demand for food and feed. In the 10.000 8.000 best of circumstances, the challenge of meeting 6.000 4.000 this demand sustainably will be huge. When one 2.000 accounts for the effects of climate change (high- 0 Brazil Argentina Uruguay Chile Peru Ecuador er temperatures, shifting seasons, more frequent and extreme weather events, flooding, drought) on Source: Tokgoz, Bhandary and Rosegrant (2012). food production, the challenge grows even more Note: BAU = Business as usual. overwhelming. Signs of disturbing prospects for 154 HSBC 2011. Chapter 4. 149 global food security include the 2010 floods in holds across all the scenarios modeled for this markets, high population growth, and an explicit Pakistan and excessive heat and drought in the study, except for LAC’s share of meat and fruits impact of climate change on agricultural produc- Russian Federation resulting in wildfires and pro- and vegetables under a pessimistic view including tivity as a result of a wetter and warmer climate. tectionist actions, including tariff increases and climate change. LAC’s share in 2050 is expected to decline to 29 export restrictions like those implemented by Ar- 4.64. By 2050, in a business-as-usual scenario, LAC percent for meat and to 13 percent for fruits and gentina. would be supplying more than a third of meat ex- vegetables and increase to 16 percent for cereals. 4.61. This research uses a detailed global agriculture ports, a third of fruit and vegetable exports, half of These shares reflect an important switch in trade model to analyze crop growth under perfect miti- oilseeds exports, and about a tenth of cereal ex- status for the region as a result of lower global con- gation and a simulated future climate scenario. It ports. Developed countries would be increasing sumption of meats and the rise of South Asia as a takes advantage of and expands on IFPRI’s cut- their exports of fruits and vegetables but substan- major fruits and vegetables provider due to posi- ting-edge climate modeling expertise to address tially decreasing those of cereals, creating a gap tive impacts of a wetter and warmer climate on the climate change threat in the context of larger that developing countries not in LAC would be fill- productivity. But the pessimistic scenario still un- food security challenges. It provides a compre- ing. Higher needs for animal products and oilseeds derscores trade’s importance as a climate change hensive analysis of the scope of infrastructure as a result of higher demand in East Asia and the adjustment mechanism and the threat that protec- investment in logistics and a better business envi- Pacific could be met only by a higher contribution tionism could pose by not allowing trade to soften ronment in LAC as well as the adoption of greener from LAC. the blow to consumers of grains affected by sur- growth-enhancing mechanisms worldwide. 4.65. Larger food and fiber needs resulting from more plus disappearance in developed countries. This 4.62. The analysis shows LAC’s importance today in sustained emerging market growth under a harmo- is an important result for Argentina and Brazil, two agricultural trade and emphasizes the continuing nious rebalancing could be accommodated with countries that could increase their cereal exports increasing role the region could play in meeting agricultural and trade liberalization, higher invest- to regions facing large grain deficits. global food and fiber needs. Table 4.5 summarizes ment in irrigation, and improved agricultural tech- 4.67. Poor infrastructure is a major barrier to increasing current and future trade outcomes under various nology. In a more harmonious scenario with higher production and trade in LAC. Logistics expenses scenarios, showing LAC’s projected share in global incomes and healthier diets in high-income and to move products are greater than in most devel- net exports of meat, cereals, fruits and vegetables, BRICs countries, LAC continues to play a signifi- oped countries. A focus on improving LAC’s busi- and oilseeds. cant role in meeting global food, fiber, and biofuel ness environment and logistics could provide LAC 4.63. If current trends continue in income and popula- demand, though its expansion of meat and cereals with more opportunities for meeting global food tion growth, use of technology and resources to is not as pronounced as in the business-as-usual and fiber needs. LAC could thereby capture 5–15 produce food and fiber, distortions in agricultural case. percent more market share than under a busi- and trade policies, and current investments in ir- 4.66. In contrast with the optimistic harmonious rebal- ness-as-usual scenario. LAC would be capturing rigation and infrastructure, LAC’s share in global ancing scenario, a more pessimistic plausible this extra market share from developed countries trade will probably be even bigger than it is today. scenario has been considered. In this scenario, in meat and oilseeds and from developing coun- As table 4.5 illustrates, this is a robust result that the future includes less rapid growth in emerging tries for fruits and vegetables and cereals. There 150 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development could also be a restructuring of LAC’s export pro- emerging countries) and global supply conditions file, with a greater increase in exports of bulk com- but also on improving its own supply environment. modities and those of more processed items, such Depending on what driving force exerts the most as soybean oil instead of beans. Soybean meal ex- pressure, LAC will be called on to increase its trade ports are expected to grow less rapidly, as higher to meet the higher food and fiber needs of the rap- livestock production in LAC requires less meal to idly growing East Asia and Pacific region or to step be available for trade. in to fill the lower surpluses generated in devel- 4.68. Agriculture consumes approximately 70 percent oped countries, particularly the United States, for of the world’s freshwater supply. Water scarcity cereals trade. poses problems for increasing agricultural pro- duction in several parts of the world, and climate change poses growing challenges to increasing agricultural production. The tension between rap- idly rising natural resource consumption and envi- ronmental sustainability will be a critical pressure point over the next decades. Pricing resource use and adopting more sustainable practices in water and land management have implications for glob- al supply and resulting trade flows among world regions. Higher income growth globally and more efficient input use patterns lead to slightly lower market share increases for LAC than under the business-as-usual scenario. A greener incentive framework in agriculture worldwide coupled with higher global demand, particularly in developing countries, generates lowers surpluses to be trad- ed. Only developed countries can increase their market share in the better business and logistics scenario for most commodity groupings. 4.69. Overall, LAC’s potential to increase its market share will depend not only on demand trends in fast-growing developing countries (particularly Chapter 4. 151 Table 4.5. Shares in world net exports in business-as-usual and alternative scenarios for 2050 (percent) Harmonious Pessimistic view of Better business and 2010 Business as usual rebalancing the world Green growth logistics Latin America and the Caribbean Meat 30 36 30 29 33 41 Cereals 8 11 9 16 9 13 Fruits and vegetables 25 34 33 13 33 38 Oilseeds 42 50 49 55 50 56 Developed Meat 38 36 39 29 40 33 Cereals 63 45 45 33 47 44 Fruits and vegetables 17 19 23 13 22 18 Oilseeds 31 29 30 9 30 25 Rest of developing (non–Latin America and the Caribbean) Meat 32 28 31 42 27 26 Cereals 29 44 46 51 44 43 Fruits and vegetables 58 47 44 74 45 44 Oilseeds 27 21 21 36 20 19 152 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development References Anderson, J.E. 1979. “A Theoretical Foundation for the sources.worldbank .org/INTTLF/Resources/lpire- Trabajo, Empleo y Seguridad Social. Buenos Aires, Gravity Equation.” American Economic Review 69: port.pdf. 2007. 106–16. Balassa, B. (1965), Trade Liberalisation and “Revealed” Borges, I.C., et. al. 2006. “Impactos do protocolo de Carta- Anderson, J.E, and D. Marcouiller. 2002. “Insecurity and Comparative Advantage. The Manchester School, gena sobre o comércio de commodities agrícolas.” the Pattern of Trade: An Empirical Investigation.” 33: 99–123. doi: 10.1111/j.1467-9957.1965. Sociedade Brasileira de Economia e Sociologia The Review of Economics and Statistics 84: 342– tb00050.xBarbosa, M. Z. and Nogueira Junior, S. Rural (Sober), Fortaleza, July 23–27. www.sober. 52. (2007). Agroindústria da Soja: Competição entre org.br. Brasil e Argentina. Available in: How Much Do Land, Water, and Crop Yields Need to Gravitas: A Solution to the Border Puzzle.” Ameri- can Economic Review 93:170–92. Bisang, R. “ El Desarrollo Agropecuario en las Ultimas Increase by 2050?” In Proceedings of FAO Expert Décadas: Volver a Creer?” in Crisis, Recuperación Meeting on How to Feed the World in 2050. ———. 2004. “Trade Costs.” Journal of Economic Litera- y Nuevos Dilemas. La Economía Argentina 2002- Buainain, A.M., J.R. Ruiz, and P.A. Viera. 2011. “Recent ture 42 (3): 691–751. 2007. CEPAL. Santiago de Chile, 2007. Development Patterns and Challenges of Brazil- Anderson, K and A. Valdés. 2008. Distortions to Agricul- Bisang. R., G. Anlló, et al. “Una Revolución No Tan Silenci- ian Agriculture.” FAO/CP Programme, TCI Division, tural Incentives in Latin America. Washington, DC: osa. Claves para Repensar el Agro en Argentina”. background paper for the current report. World Bank. Desarrollo Económico No. 190-191 Vol 48. Buenos Cadot, Olivier, Céline Carrère, and Vanessa Strauss-Kahn. Anderson, K. (ed.). 2009. Distortions to Agricultural In- Aires, 2008. 2011. “Export Diversification: What’s Behind the centicves: A Global Perspective, 1955- 2007. Was- Bisang, R. y S. Sztulwark. “Tramas Productivas de Alta Hump?” Review of Economics & Statistics 93: hington, DC: Palgrave MacMillan and World Bank. Tecnología y Ocupación. El Caso de la Soja Trans- 590–605. Arvis, J.F., .M. A. Mustra,L. Ojala,B. Shepherd, and D. génica en la Argentina”. In Estructura Productiva Calderón, C., and L. Servén. 2004. “The Effects of Infra- Saslavsky2007. “Connecting to Compete: Trade y Empleo. Un Enfoque Transversal. Ministerio de structure Development on Growth and Income.” Logistics in the Global Economy. http://sitere- References 153 www-wds.worldbank.org/external/default/WD- Coremberg, Ariel (2009). Origen sectorial y factorial tion towards Non-Members?” Discussion Paper SContentServer/IW3P/IB/2004/09/21/0000120 del declive de la productividad argentina. Sem- 868. London School for Economics Centre for 09_20040921105044/Rendered/PDF/WPS3400. inario internacional. Inversión, productividad y Economic Performance. http://cep.lse.ac.uk/pubs/ pdf crecimiento en el nuevo escenario global. 17 de download/dp0868.pdf ———. 2010. “Infrastructure in Latin America.” Policy Re- noviembre de 2009. Cancillería argentina- Pros- Estevadeordal, A., M. Shearer, and K. Suominen. 2009 search Working Paper 5317. World Bank, Washing- perar-CEPAL. “Multilateralizing RTAs in the Americas: State of ton, DC. www-wds.worldbank.org/servlet/WDSCon- Dadush U. and W. Shaw. Juggernaut: How Emerging Pow- Play and Ways Forward.” In Multilateralizing Re- tentServer/WDSP/IB/2010/05/24/000158349_20 ers Are Reshaping Globalization. Washington, gionalism, eds. R. Baldwin and P. Low. Cambridge, 100524133508/Rendered/PDF/WPS5317.pdf. DC: Carnegie Endowment for International Peace UK: Cambridge University Press. CARANA Corporation. 2006. “Supply Chain Analysis of (June 1, 2011). Etache, A., and Jose Carbajo. 1996. “Competing Private Chile, Paraguay and Brazil, Land Shipments of Deininger, K.. D. Byerlee, with J. Linsay, A. Norton, H. Sel- Ports—Lessons from Argentina.” Note 100. World Beef and Soy.”, mimeo od, and M. Stickler. 2011. “Rising Global Interest in Bank, Washington, DC. http://www.ppiaf.org/ http://siteresources.worldbank.org/INTOECS/Resources/ Farmland: Can It Yield Sustainable and Equitable ppiaf/sites/ppiaf.org/files/documents/toolkits/ InstitutionsPerformanceandtheFinancingofInfra- Benefits?” Washington, DC: World Bank. Cross-Border-Infrastructure-Toolkit/Cross-Bor- structureintheCaribbean.pdf. Depken II, C.A., and Sonora, R.J. 2005. “Asymmetric Ef- der%20Compilation%20ver%2029%20Jan%2007/ fects of Economic Freedom on International Trade Resources/Estache%20Carbajo%20-%20Compet- Castillo, V., S. Rojo et al. “La Trama Automotriz, Dinámica Flows.” International Journal of Business and Eco- ing%20Private%20Ports.pdf del Empleo y Trayectorias Laborales desde Me- diados de los Noventa”. In Estructura Productiva nomics 4 (2): 141–55. EUROPA. 2011. EUR – Lex 32009L0028. http://eur-lex. y Empleo. Un Enfoque Transversal. Ministerio de Djankov, S., C. Freund, and C. Pham. 2010. “Trading on europa.eu/LexUriServ/LexUriServ.do?uri=CELEX- Trabajo, Empleo y Seguridad Social. Buenos Aires, Time.” Review of Economics and Statistics 92(1): :32009L0028:EN:NOT. 2007. 166–73. FAO (Food and Agriculture Organization). 2008. “Rural Centro Logistics de Latinoamerica, Bogota, Colombia. Doing Business 2012. “Trading across Borders.” http:// Transport of Food in Latin America and the Caribbe- Benchmarking 2007: Estado de la Logistica en www.doingbusiness.org/reports/global-reports/~/ an.” Agriculture Service Bulletin 155. Rome. ftp:// America Latina. Anexo Maria Rey Logistic Summit media/FPDKM/Doing%20Business/Documents/ ftp.fao.org/docrep/fao/010/y5711e/y5711e00. 2008. Presentation at: http://www.authorstream. Annual-Reports/English/DB12-Chapters/Trad- pdf. com/Presentation/MariaRey-85542-bench- ing-Across-Borders.pdf. FAOSTAT (Food and Agriculture Organization Statistics Di- marking-logistica-latinoamerica-2007-lalc-ma- EPA (Environmental Protection Agency). 2011. Renew- vision). 2011. http://faostat.fao.org/. ria-rey-indicadores-bmk-entertainment-ppt-pow- able Fuel Standards. Washington, DC. www.epa. Ferro, E. and A. Portugal-Perez. 2012. “Trade Facilitation erpoint/ gov/otaq/fuels/renewablefuels/index.htm. in Latin America and Agricultural Exports” Back- CISP (Caribbean Integration Support Program). “Agribusi- Estevadeordal, A., Caroline Freund, and Emanuel Ornelas. ground paper for this report. ness Sector and Targeted Commodity Enterprises.” 2008. “Does Regionalism Affect Trade Liberaliza- 154 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development Francois, J., and M. Manchin. 2007. “Institutions, Infra- America and the Caribbean: How ICTs Can Make a Kee, H. L., A. Nicita and M. Olarreaga. “Estimating trade structure, and Trade.” Working Paper 4152. World Difference.” LCSSD Report 68255. restrictiveness indices”, Economic Journal, 2009, Bank, Washington, DC. Guasch, Jose Luis. 2011. “Logistics as a Driver for Com- vol. 119, p. 172--199. Fretes-Cibils, V., Giugale, M., and J. Newman. 2007. An Op- petitiveness in Latin America and the Caribbean.” Kjöllerström, Monica. 2004. “Liberalización commercial portunity for a Different Peru. Washington, DC: World Inter-American Development Bank Americas Com- agrícola con costos de transporte y transacción Bank. http://books.google.com/books?id=0-y_xY- petitiveness Forum V, October. www.competecar- elevados: evidencia para América Latina.” CE- 1L9yQC&pg=PR15&lpg=PR15&dq=An+Oppor- ibbean.org/content/JL_Guasch_final_Logistics%20 PAL. www.eclac.org/publicaciones/xml/4/20994/ tunity+for+a+Different+Peru&source=bl&ots=- as%20a%20Driver%20for%20Competitiveness%20 DP160.pdf JK5NqXF2h_&sig=RmhrI3J7vOzjVI0hxCk3ZcbEZ- %28english%29.pdf Klinger, B. and D. Lederman 2004. “Discovery and Devel- kE&hl=en#v=onepage&q=An%20Opportunity%20 Harrigan, James and Richard Baldwin (2011). “Zeros, opment: An Empirical Exploration of ‘New’ Prod- for%20a%20Different%20Peru&f=false Quality, and Space: Trade Theory and Trade Evi- ucts.” Mimeo. Fulponi, L., M. Shearer, and J. Almeida. 2011. “Regional dence.” American Economic Journal: Microeco- ———. 2005. “Diversification, Innovation, and Imitation Trade Agreements: Treatment of Agriculture.” OECD nomics. 3, pp. 60–88. off the Global Technology Frontier.” Mimeo. Food, Agriculture, and Fisheries Working Paper 44. Hausmann, R., and D. Rodrik. 2003. “Economic Develop- Korinek, J., and M. Melatos.  2009.  “Trade Impacts of OECD Publishing, Paris. ment as Self-Discovery.” Journal of Development Selected Regional Trade Agreements in Agricul- Gaviria, Juan. 1998. “Port Privatization and Competition Economics 72: 603–33.HSBC. 2011. “The World ture.” OECD Trade Policy Working Paper 87. OECD in Colombia.” Note 167. World Bank, Washington, in 2050: Quantifying the Shift in the Global Econ- Publishing, Paris. DC. http://cdi.mecon.gov.ar/biblio/docelec/bm/ omy.” http://globalconnections.hsbc.com/assets/ Krueger, Anne O., Maurice Schiff and Alberto Valdés ppps/N167.pdf docs/hsbc-the-world-in-2050.pdf. (1988), ‘Agricultural Incentives in Developing Wilson, J., C. Mann and T. Otsuki. 2005. “Assessing the Imbs, J and R. Wacziarg. 2003. “Stages of Diversification.” Countries: Measuring the Effect of Sectoral and Benefits of Trade Facilitation: A Global Perspec- American Economic Review 93 (1): pages 63-86, Economy-wide Policies’, World Bank Economic Re- tive,”, The World Economy 28(6): 841-871,. March. view 2(3): 255-72, September. Gonzales, J., J. Guasch, and T. Serebrisky. 2008. “Improv- Jha, Abhas Kumar, ed. 2005. “Institutions, Performance, Krueger, Anne O., Maurice Schiff and Alberto Valdés ing Logistics Costs for Transportation and Trade and the Financing of Infrastructure Services in the (1991), The Political Economy of Agricultural Pric- Facilitation.” Policy Research Working Paper 4558. Caribbean.” Working Paper 58, World Bank, Wash- ing Policy, Volume 1: Latin America, Volume 2: World Bank, Washington, DC. www-wds.worldbank. ington, DC. Kee, Hiau Looi, Alessandro Nicita, and Asia, and Volume 3: Africa and the Mediterranean, org/external/default/WDSContentServer/IW3P/IB Marcelo Olarreaga. 2009. “Estimating Trade Re- Baltimore: Johns Hopkins University Press for the /2008/03/13/000158349_20080313132618/ strictiveness Indices.” The Economic Journal 119 World Bank. Rendered/PDF/wps4558.pdf (1): 172–99. Landell Mills Development Consultants (2011): “Devel- Goyal, A., and C. González-Velosa. 2012. “Improving Agri- oping an Action Plan for Promotion of a Regional cultural Productivity and Market Efficiency in Latin Agribusiness Sector and Targeted Commodity En- References 155 terprises.” Caribbean Integration Support Program Martin, W., and K. Andersen. 2010. “Trade Distortions and gramação linear. Revista de Economia Agrícola (CISP) Food Price Surges.” American Journal of Agricultur- 54: 33–41. Langley, John, and Capgemini. 2012. “ Third-Party Lo- al Economics (forthcoming). Oxfam International. 2008. “Another Inconvenient Truth: gistics Study: The State of Logistics Outsourcing.” Martin, W., and A. Matoo. 2011. Unfinished Business? The How Biofuels Policies Are Deepening Poverty and http://www.outreach.psu.edu/programs/had- WTO’s Doha Agenda. Washington, DC: World Bank. Accelerating Climate Change.” Briefing Paper 114. ly-waters/files/2012_3pl_study.pdf. [AQ: Please http://go.worldbank.org/L8904MG5A0. Oxford, UK. update link] MEA (Millennium Ecosystem Assessment). 2005. Eco- PAM/IBGE. 2010. Pesquisa Agrícola Municipal. www.ibge. LCSSD Economics Unit. 2010. “Understanding the Impact systems and Human Well-Being: General Synthe- gov.br. of Transport and Logistics Costs on Food Prices in sis. Washington, DC: Island Press. www.millenni- Pardey, P.G., S. Wood, and R. Herford, eds. 2010. Research LAC: Supply Chain Analyses of Grain Imports into umassessment.org/documents/document.356. Futures: Projecting Agricultural R&D Potentials for Nicaragua, Honduras, and Ecuador.” World Bank, aspx.pdf. Latin America and the Caribbean. Washington, DC Sustainable Development Department, Latin Nelson, G., M. Rosegrant, A. Palazzo, I. Gray, C. Ingersoll, and St. Paul, MN: IFPRI, IDB, and InSTePP. America and the Caribbean. R. Robertson, S. Tokgoz, T. Zhu, T. Sulser, C. Ringler, Parry, M.L., C. Rosenzweig, A. Iglesias, M. Livermore, and Levchenko, A. 2004. “Institutional Quality and Interna- S. Msangi, and L. You. 2010. “Food Security, Farm- G. Fischer. 2004. “Effects of Climate Change on tional Trade.” IMF Working Paper 04/231. Interna- ing, and Climate Change to 2050: Scenarios, Re- Global Food Production under SRES Emissions and tional Monetary Fund, Washington, DC. sults, Policy Options.” IFPRI Research Monograph. Socioeconomic Scenarios.” Global Environmental Limao, N., and A.J. Venables. 2001. “Infrastructure, International Food Policy Research Institute, Change 14 (1): 53–67. Geographical Disadvantage, Transport Costs and Washington, DC. Perrotti, D.E., and R.J. Sánchez. 2011. La brecha de Trade.” World Bank Economic Review 15: 451–79. OECD (Organisation for Economic Co-operation and De- infrastructura en America Latina y el Caribe. Lin, J.Y. 2012. The Quest for Prosperity: How Developing velopment). 2011. Agricultural Policy Monitoring Santiago: CEPAL. www.iadb.org/intal/intalcdi/ Economies Can Take Off. Princeton, NJ: Princeton and Evaluation 2011. Paris. PE/2011/09214.pdf University Press. O’Çonnor, E. 2011. :The Agricultural Economy of Argen- Portugal-Perez, A., and J.S. Wilson. 2012. “Export Perfor- Mandel, B. 2012. “The Differentiation and Dynamics of tina: Recent Trends and Prospects.” FAO/CP Pro- mance and Trade Facilitation Reform: Hard and Latin American and Caribbean Agricultural Ex- gramme, TCI Division. Background paper for this Soft Infrastructure.” World Development (forth- ports.” Background paper for this report. report. coming). Manova, Kalina B. and Zhang, Zhiwei, Export Prices across Oliveira, A.L.R. 2005. Panorama mundial e nacional da Rabobank. 2010a. “Sustainability and Security of the Firms and Destinations (2012). Quarterly Journal soja, Safras 2004/05 e 2005/06. Informações Global Food Supply Chain.” Report prepared for the of Economics, 127 (2012), p.379-436. Available Econômicas 35 (11): 57–61. Duisenberg Lecture, October. at SSRN: http://ssrn.com/abstract=1326712 ———. 2007. Transporte de soja do estado do Mato Grosso para exportação: uma aplicação de pro- 156 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development Rabobank. 2010b. “Sustainability and security of the Sánchez, Ricardo J., and G. Wilmsmeier. 2009. “Maritime yond Booms and Busts? Washington, DC: World global food supply chain.” Economic Research De- Sector and Ports in the Caribbean: The Case of Bank. partment of Rabobank Nederland. CARICOM Countries.” www.eclac.org/publicacio- Tokgoz, Simla, Prapti Bhandary, and Mark Rosegrant. Ranjan, P. and J. Y. Lee. 2003. “Contract Enforcement nes/xml/6/36706/lcl3008i.pdf. 2012 “Forces shaping present and future agricul- and the Volume of International Trade in Different Schott, P. 2003. “One Size Fits All? Heckscher-Ohlin Spe- tural trade trends in Latin America and the Carib- Types of Goods.” Mimeo, University of California– cialization in Global Production.” American Eco- bean: Alternative Scenarios.” Background study Irvine. nomic Review 93: 686–708. prepared for this report. Regúnaga, M. 2010. “Implications of the Organization of Schott, P. 2004. “Across-Product versus Within-Product U.S. Department of Agriculture. 2011. “Production, Sup- the Commodity Production and Processing Indus- Specialization in International Trade.” Quarterly ply, and Distribution Online.” www.fas.usda.gov/ try: The Soybean Chain in Argentina.” Background Journal of Economics 119: 647–78. psdonline/psdHome.aspx. paper for Sinnott, Nash, and de la Torre, 2010. Schwartz, J., J. Guasch, and G. Wilmsmeier. 2009. “Lo- UN (United Nations). 2011. World Population Prospects: Regúnaga, M. 2011 “Agricultural Policy Note for Argenti- gistics, Transport and Food Prices in LAC: Policy The 2010 Revision. New York. na.” Unpublished Report Prepared for World Bank, Guidance for Improving Efficiency and Reducing UNECLAC (UN Economic Commission for Latin America). LCSAR. Costs.” World Bank Policy Guidance Paper, LAC 2012. “International Transport Database (BTI).” Roberts, Peter, Shyam KC, and Cordula Rastogi. 2006. Ministerial Meetings, Chile. Santiago. “Rural Access Index: A Key Development Indica- Shearer, M., J.S. Almeida, and C.M. Gutierrez, Jr. 2009. UNCTAD (United Nations Conference on Trade and De- tor.” World Bank Transport Papers. World Bank, “The Treatment of Agriculture in Regional Trade velopment). 2011. “Review of Maritime Transport Washington DC. www.worldbank.org/transport/ Agreements in the Americas.” Washington, DC: In- 2011.” http://unctad.org/en/docs/rmt2011_en.pdf. transportresults/headline/rural-access/tp-10-fi- ter-American Development Bank (Vice Presidency Wilson, J.S., C. Mann, and T. Otsuki. 2003. “Trade Facilita- nal.pdf for Sectors and Knowledge, Integration and Trade tion and Economic Development: A New Approach Rosegrant, M.W., S. Msangi, C. Ringler, T.B. Sulser, T. Zhu, Sector) and International Food and Agricultural to Measuring the Impact.” World Bank Economic and S.A. Cline. 2008. International Model for Poli- Trade Policy Council. Review 17 (3): 367–89. cy Analysis of Agricultural Commodities and Trade Silveira, Borges, and E.M.M.Q. Buainain. 1983. Regulam- ———. 2005. “Assessing the Benefits of Trade Facilita- (IMPACT): Model description. Washington, DC: In- entação do mercado de leite e laticínios no Brasil. tion: A [AQ: Please complete] ternational Food Policy Research Institute. São Paulo: Universidade de São Paulo.. World Bank. 2008. Organization of Eastern Caribbean Runge, C.F., and B. Senauer. 2008. “How ethanol fu- Silveira, J.M.F.J., I.C., Borges, and A. M. Buainain. 2005. States: Increasing Linkages of Tourism els the food crisis.” Updated May 28, 2008, “Biotecnologia e agricultura: da ciência e tecnolo- with the Agriculture, Manufacturing and Service Sectors. Foreign Affairs. www.foreignaffairs.com/arti- gia aos impactos da inovação.” São Paulo em Per- Report No. 44060-LAC. cles/64909/c-ford-runge-and-benjamin-senauer/ spectiva 19 (2): 101–14. how-ethanol-fuels-the-food-crisis. Washington, D.C. Sinnott, E., J. Nash, and A. de la Torre. 2010. Natural Re- sources in Latin America and the Caribbean: Be- References 157 ———2009. World Development Indicators on CD-Rom. World Bank/Instituto Mexicano para la Competitividad. Washington, DC. 2007c. “United Mexican States: Integration of the ———.2010. Export-Led Growth v2.0. Economic Prem- North American Market for Sensitive Agricultural ise, Number 3. Washington, DC. Commodities/ Policy Notes Overview and Sugar Market: Implications for Mexican Producers and ———. 2011a. World Development Indicators. Washing- Consumers.” Washington, DC: Agriculture and Ru- ton, DC. ral Development Team, Sustainable Development ———. 2011b. Trade Competitiveness Diagnostic Toolkit. Department, Latin America and the Caribbean Vice International Trade Department. Washington, DC. Presidency. ———. 2012a. Global Monitoring Report 2012: Food WTO (World Trade Organization). 2011. World Trade Re- Prices, Nutrition, and the Millennium Development port 2011. Geneva. Goals. Washington, DC: World Bank. Xiang, C. 2007. “Diversification Cones, Trade Costs, and ———. 2012b. Inclusive Green Growth: The Pathway to Factor Market Linkages.” Journal of International Sustainable Development. Washington, DC. Economics 71: 448–66. World Bank/Instituto Mexicano para la Competitividad. 2007a. “United Mexican States: Integration of the North American Market for Sensitive Agricultural Commodities/ Policy Notes Overview and Corn Market: Implications for Mexican Producers and Consumers.” Washington, DC: Agriculture and Ru- ral Development Team, Sustainable Development Department, Latin America and the Caribbean Vice Presidency. World Bank/Instituto Mexicano para la Competitividad. 2007b. “United Mexican States: Integration of the North American Market for Sensitive Agricultural Commodities/ Policy Notes Overview and Bean Market: Implications for Mexican Producers and Consumers.” Washington, DC: Agriculture and Ru- ral Development Team, Sustainable Development Department, Latin America and the Caribbean Vice Presidency. 158 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development Chapter 4. 161 The World Bank 1818 H Street, NW, Washington, DC 20433, USA. www.worldbank.org 162 Agricultural Exports from Latin America and the Caribbean: Harnessing Trade to Feed the World and Promote Development