Public Disclosure Authorized 64089 Environmentally and Socially Sustainable THE WORLD BANK Development Unit Europe and Central Asia Region Public Disclosure Authorized Public Disclosure Authorized The Dynamics of Vertical Coordination Public Disclosure Authorized in Agrifood Chains in Eastern Europe and Central Asia Implications for Policy and World Bank Operations The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Implications for Policy and World Bank Operations Environmentally and Socially Sustainable Development Unit Europe and Central Asia Region THE WORLD BANK This volume is a product of the work of the staff of the International Bank for Reconstruction and Development and The World Bank. The �ndings, interpretations, and conclusions expressed in this report do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The bound- aries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal sta- tus of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Copying and transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development and The World Bank en- courage dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 22 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, http://www.copyright.com/. All other queries on rights and licenses, including subsidiary rights, should be ad- dressed to the Of�ce of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. Contents The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Implications for Policy and World Bank Operations Acknowledgments vii Acronyms and Abbreviations ix Executive Summary xi Introduction and Background 1 Objective, Methodology, and Organization 3 Conceptual Framework 4 The Benchmark: Vertical Coordination in Other Countries 5 North America and Western Europe 5 Developing and Emerging Economies 7 There is More, and More Complex, Vertical Coordination in ECA 7 The Disruption and Reorganization of Supply Chains in Transition 9 iii Contents Multiple Models and Motivations 11 The Basic Model 12 Variations on the Basic Model 13 Vertical Coordination Varies by the Stage of Transition 14 Constraints on Vertical Coordination in ECA 17 Contract Enforcement Is Dif�cult but Essential 20 Sectoral Analyses 23 Dairy Sector 23 Sugar 25 Fruits and Vegetables (F & V) 26 Cotton 27 Grains 30 Impact 32 On Supplies and Productivity 32 On Quality 34 On Investment 35 Cross-Company Spillovers 36 Effects of Government Regulated Vertical Coordination 37 The Search for Quality is a Key Engine of VC 37 Competition Spreads Equity and Ef�ciency Bene�ts 40 Foreign Direct Investment is Engine of Change 41 Trade Protection and Vertical Coordination 42 Problems of Equity and Exclusion 43 Rent Extraction 44 The Problem of Exclusion: Small Farmers in the Chain Empirical Evidence 45 Why Contracting with Small Farmers? 46 The Farm Assistance Paradox 47 ECA is a “Supplier Market� 47 For Now . . . No Time for Complacency 48 Putting the Equity Issues into Perspective 48 Private vs. Public: Image and Reality 50 Small Farm Assistance 50 Farm Exploitation 50 iv Contents The Future of Vertical Coordination 51 Implications for Government Policy and World Bank Operations 52 Enabling and Stimulating Vertical Coordination 53 Improving Ef�ciency and Equity in VC 55 Implications for the Role of the Government in Policy Management and Public Goods and Services Provision 57 List of Background Studies and Related Studies 60 Bibliography 61 ANNEX 1 Importance of Contracting in ECA Agrifood Chains 71 ANNEX 2 Variations in Vertical Coordination Models 79 Various Companies Can Initiate Vertical Coordination 79 Complex Models 80 Multiple Stages 85 Full Vertical Integration: Ownership 86 Chain-Based Financial Instruments 86 Russian-Style Vertical Coordination 91 ANNEX 3 Vertical Coordination with Small Farmers 97 Empirical Evidence 98 Why Contracting with Small Farmers? 102 v Acknowledgments The main author of the report was Johan F. M. Swinnen. Background studies were prepared by Siemen van Berkum, John White, Matthew Gorton, Marc Sadler, Liesbeth Dries, E. Gataulina, V. Uzun, A. Petrikov, and Renate Yanbykh. The report made use of insights from other studies prepared by Tom Reardon, Hamish Gow, Daniele Giovanucci, Kees van der Meer, Steven Jaffee, Emmanuel Hidier, and Willi Meyers. William Sutton, Maria Amelina, and Maurizio Guadagni advised on the prepara- tion of the study. Mark Ludwick, Karin Rosskopf, and Gail Lee provided editorial assistance. Steven Jaffee, Richard Henry, Csaba Csaki, and Ali Mansoor reviewed the study proposal and the �nal report and provided very useful comments and suggestions. Rapeepun Jaisaard, Juergen Venema, Maria Amelina, Maurizio Guadagni, Iain Shuker, Kees van der Meer, Gotz Schreiber, Kym Anderson, and participants at seminars where the study �ndings were pre- sented contributed with helpful comments at various stages of the project. Finally, Benoit Blarel and Laura Tuck provided much appreciated encouragement and critical guidance throughout the duration of the study. vii Acronyms and Abbreviations ADB Asian Development Bank AFIRUM Agri-food Industry Restructuring in Ukraine and Moldova AIB Agroinvestbank BLG Bank Loan Guarantees CEE Central and Eastern European Countries CEO Chief Executive Of�cer CIS Commonwealth of Independent States CPR Cedula de Produto Rural EBRD European Bank for Reconstruction and Development ECA Europe and Central Asia EU European Union FAO Food and Agriculture Organization FCDF Friesland Coberco Dairy Foods FDI Foreign Direct Investment FFV Fresh Fruits and Vegetables FIG Financial-Industrial Groups F&V Fruits and Vegetables IFAD International Fund for Agricultural Development ISPA Instrument for Structural Policies for Pre-Accession NGO Non-governmental Organizations NRIF Non-registered individual farms OECD Organisation for Economic Cooperation and Development PFID Partnership for Food Industry Development PMO Producer Marketing Organization RF Reverse Factoring RIF Registered individual farms ix Acronyms and Abbreviations SAPARD Special Accession Programme for Agriculture and Rural Development SPS Sanitary and Phytosanitary SPV Special Purpose Vehicle USAID United States Agency for International Development USDA United States Department of Agriculture USDA/ERS United States Department of Agriculture/ Economic Research Service VC Vertical Coordination WB World Bank WBD Wimm Bill Dann WRS Warehouse Receipt Systems x Executive Summary Vertical coordination (VC) in agrifood supply chains is an important and growing phenomenon in the transition countries of Europe and Central Asia (ECA). VC is more important and more widespread than generally recognized. These changes have significant implications for the role of public policy and for the World Bank. There is a need to explicitly integrate these VC developments into policy thinking and program strategies. VC is more widespread in both scope and complexity in ECA than in Western economies. In the United States and Germany, around one- third of agricultural production is produced under contracts. However, VC in ECA differs signi�cantly from rich and poor market economies. First, there is signi�cant VC in sectors where we do not observe VC in other countries. Second, in sectors where VC exists in other countries, the forms of VC in transition countries are more extensive and complex. In the dairy and sugar sectors, extensive contracting arrangements have developed be- tween processors and farms, including the provision of credit, investment loans, feed, inputs, extension services, bank loan guarantees, etc. In the cotton sector, gins typically contract farms to supply seed cotton and pro- vide them with credit, seeds, fertilizer, etc. In fresh fruits and vegetables modern retail chains, quality and timeliness of delivery are required, along with developing supplier contracting with farm-assistance pro- grams. In grains sector, there is strong vertical integration in Russia and Kazakhstan, where huge agroholdings produce a large share of the grain crop in some regions. Private contractual initiatives have emerged to overcome disruptions of supply and poor public institutions for governing exchange. The pri- vatization and restructuring of the agrifood chain caused major disrup- tions. Widespread contracting problems were long payment delays or xi Executive Summary nonpayments for delivered products, causing drains on cash flow and con- straints in accessing inputs and selling products. At the same time, food processing companies have problems obtaining quality supplies. The problems are worsened by the lack of public institutions necessary to sup- port market-based transactions, such as for enforcing property rights and contracts. Traders, agribusinesses, and food companies contract with farms and provide inputs and assistance in return for guaranteed and quality sup- plies. Successful vertical contracting typically includes conditions for product delivery, prompt payments, and farm-assistance programs for suppliers. Farm assistance can include input supply programs, investment assistance, trade credit, bank loan guarantees, extension and management advisory services, etc. The search for quality is a key engine of VC. The shortage of quality supply, which is typical of transition countries, induces VC and spillover effects through farm support packages. The issue of quality has both ef�ciency and equity implications. Farms get a higher price for quality, but quality controls are not always transparent. Quality controls by indepen- dent institutions have both ef�ciency and equity bene�ts. Contracting requires access to finance. Initiators of contracting with supplier assistance include foreign investors who can access inter- national financial markets, companies who are investing profits from other sectors in the agrifood sector (e.g., financial-industrial groups [FIGs] in Russia), processors or traders who have liquidity by selling on international markets (e.g., grain traders in Kazakhstan), and processors who have contracts with international companies (e.g., cotton gins in Central Asia). Enforcement is an important problem. Enforcement is problematic where public enforcement institutions are absent. Trust is also often lacking as a base for business exchanges in many transition countries. Companies try to create self-enforcing contracts by designing the terms of the contracts such that nobody has an incentive to breach the contract. They also try to enforce contracts by interlinking markets. The enforcement of the credit trans- action (loan and repayment) occurs through the output market. However, there are many cases where enforcement failed. Even in successful cases it took considerable �ne-tuning of the contracts or adjustments as circum- stances changed. Creating the right conditions for successful and self- enforcing contracting requires extensive knowledge of the sector and of local conditions. Vertical coordination differs by the stage of transition. In early stages, VC focuses on securing supplies by overcoming basic supply problems such as input (feed, seeds, etc.) and credit (working capital) constraints. An xii Executive Summary important component of the early contracts is prompt payments. This is the case in some cotton supply chains in Central Asia and in emerging dairy and fruits and vegetables supply chains in countries such as Romania and the Caucasus. In more advanced situations, there is more emphasis on product quality. For this more sophisticated forms of VC are used, such as extension services and farm-level investments in technology and equip- ment, leasing, bank loan guarantees, investment assistance, etc. These pro- grams require more complex implementation and enforcement systems. Contract forms reflect different constraints faced by farms. For exam- ple, the dominant contract motivation for farms in Central Europe is guar- anteed access to markets. The motivation for Central Asian cotton farmers is access to �nance as credit constraints are most important. Successful vertical contracting has important positive effects, both di- rect and indirect. The direct impact is on increased output and productivity of the processing company that initiates vertical contracting. Indirectly, contract support measures have positive effects on farm productivity and product quality. Measures with the greatest impact on yields were specialist storage (cooling equipment in dairy), veterinary support, and physical inputs. Prompt payments, guaranteed prices, and market access also had large positive effects. Quality of output improved strongly in response to speci�c programs. Direct loans and loan guarantee programs stimulated farm investments. Programs that assist farms in accessing inputs (mainly feed) enhance investment indirectly by lowering input costs, or reducing transaction costs in accessing inputs, improving pro�tability. Horizontal spillovers occur as �rms compete for suppliers and have to offer similar contractual arrangements. This has resulted in contractual con- vergence. Contractual spillovers are not limited to the same sector. Firms in adjacent commodity sectors, competing for the same farm resources, are sometimes forced to offer similar contractual arrangements that have led to similar results. Not all examples of VC are successful. In particular, where govern- ments are heavily and actively involved in the management of the verti- cal integration, the effects are dubious at best. In cotton supply chains in Central Asia where the government has allowed private gins to develop and to compete, such as in Kyrgyzstan and Kazakhstan, farms have ben- e�ted from VC, with relatively high prices and strong cotton growth. In Tajikistan and Uzbekistan, where governments actively control input supplies, production, processing, and marketing of cotton, VC resulted in major rent extraction of cotton farms, with depressed prices and stag- nating cotton production. In Russia, the government-led re-creation of huge agroholdings has con- tributed not only to more inputs for farms and strong growth in output and xiii Executive Summary yields, but also to poor �nancial results and substantial debts. Pro�ts are worse than on nonintegrated farms. A key problem is the authorities’ interference with production plans and with decisions relating to which activities (and companies) should be maintained by a holding, sometimes imposing unpro�table activities and companies on the holding. Competition spreads equity and ef�ciency bene�ts. Competition is very important in supply chains for equity and ef�ciency. First, compe- tition induces VC spillover effects across the sector as other processors are forced to introduce similar supplier-assistance programs since sup- pliers may not want to deliver unless they get similar conditions. Second, competition prevents processing companies or input suppliers from ex- ercising monopoly power in setting contract conditions with farms. Competition among cotton gins in Kazakhstan allowed small suppliers to get better conditions by changing gins, induced investment by gins in local cotton seed collection centers reducing farm transport costs, and lead to better prices. Foreign direct investment (FDI) drives successful contracting and supplier-assistance programs. It is an initiator of change and institutional innovation. More sophisticated forms of vertical integration, with a greater emphasis on quality and standards, are often introduced by foreign com- panies because they tend to pay greater attention to quality standards. But we also �nd that spillover effects lead to convergence as domestic compa- nies start copying the management practices of foreign af�liates. A concern is that vertical coordination will exclude many farmers, in particular small farmers. First, transaction costs favor larger farms in sup- ply chains. Second, small farms are more constrained for making neces- sary investments. Third, small farms typically require more assistance per unit of output. Therefore, companies prefer working with relatively fewer, larger, and more modern suppliers. In reality, companies work with surprisingly large numbers of sup- pliers and of surprisingly small size. There are several reasons. Companies may have no choice if small farmers represent most of the supply base. Contract enforcement may be more problematic with larger farms. Farms’ willingness to learn and attitude are more important than size in farm- processor relationships. Small farms may have cost advantages in labor-in- tensive production activities. Processors may prefer a mix of suppliers. Cooperatives are more likely to work with small farms than corporate com- panies, either domestic or foreign. The farm assistance paradox. Small poor farms may be best off (in the perspective of “supply-chain-driven development�) if they are in an envi- ronment that is dominated by small poor farms. If small farmers must depend on farm assistance packages to make necessary upgrades, then it xiv Executive Summary will be a problem if suf�cient (quality) supplies are available because the processor is unlikely to come up with VC packages. ECA is a “supplier market,� for now. The collapse of farm output and livestock numbers created a gap between processing capacity and supply: hence there is excess demand based on processing capacity, especially for high quality. This makes it a “suppliers market� in most of ECA and this supports the farms’ bargaining position in the supply chain. However, an increase in competition among suppliers may lead to a consolidation of the supplier base. Supplier-assistance programs sometimes discriminate between farms with the focus of upgrading the better farms and ensuring a minimal supply base and quality from the rest as long as it is required. Hence, those who are concerned about the inclusion of small farms should not be complacent despite the observations of signi�cant contracting with small suppliers right now. Private vs. public: image and reality. The public policy debate (explic- itly or implicitly) frames the issue of vertical integration and small farms in terms of how public policy can prevent (small) farms from being exploited by large, sometimes multinational, agribusinesses in their contractual rela- tionships. However, reality suggests a much more nuanced picture. First, while profits are their primary concern for all agribusinesses, this does not seem to lead to exploitation of farms. In cases where suf�cient compe- tition exists there is more evidence that producers bene�t from VC than that they are exploited. Second, farm exploitation resulted from govern- ments either controlling input supplies and marketing or where authori- ties colluded with a private company, allowing and predating on the rent extraction by the private company in the cotton sectors of Tajikistan and Uzbekistan. Third, an important constraint on enterprise development in some countries is rent extraction by local governments, e.g., through tax- ation and ad hoc regulations. Only large corporations can withstand pressures from local authorities. This leads to a paradoxical situation that farms need to be large to withstand public pressures. Fourth, private sup- plier assistance schemes reach small farms that are left out of government programs. For them, the only source of credit and �nance is supplier credit. Implications for government policy and World Bank operations. The most important policy implication of this study is the recognition of the importance of the VC phenomena in ECA agrifood chains and the need to explicitly integrate these developments into policy thinking and program strategies. A government strategy to stimulate domestic growth in a supply-chain-driven development process while ensuring the inclusion of farms which face major constraints in this process, and an equitable distribution of rents in the chain, should include several policy components. xv Executive Summary Enabling and Stimulating Vertical Coordination Create the right conditions for stimulating investment. A poor policy environment has a negative effect on investments in the agrifood industry and on VC programs. As such it constrains the beneficial effects of VC. Ensure macroeconomic stability, a key condition for investments and for supplier-assistance programs or chain-based finance. Since VC is importantly a financial activity, instability may undermine contract enforcement. Refrain from direct intervention: bad policies are worse than bad weather. Direct government intervention in the supply chains may crowd out alternative financing systems or cause defaults. Companies are willing to incorporate temporary VC defaults due to unforeseen shocks such as the weather but not systemic risks due to government interventions. Improving Ef�ciency, Transparency, and Equity in Vertical Coordination Reduce transaction costs. The disadvantage of small suppliers is mostly due to transaction costs. Reducing transaction costs can be done in several ways: Lower transport costs through improvements of rural infrastructure. Rural infrastructure is a serious constraint on VC, and particularly for integrating small producers in remote areas. Reduce the number of transactions by investing in intermediary institu- tions. Intermediary institutions reduce the cost of exchange between farm and processor/input supplier. Invest in farm associations and collection points. Investment in farmers associations has several advantages, such as reducing transaction costs, enhancing suppliers bargaining posi- tion vis-à-vis suppliers and governments, and improving informa- tion distribution. Enforce competition. Competition in the supply chain is important for efficiency and equity. Competition induces more supplier-assistance programs and constrains rent extraction. Competition should be enforced through both domestic policies (e.g., competition policies, lower barriers of entry) as well as external policies (e.g., liberal trade policies). Stimulate and certify quality and safety standards and invest in projects, in- stitutions, and technical assistance stimulating higher quality. Modern sup- ply chains are based on quality. Preparing suppliers for quality-driven markets will make it easier for them to be integrated in the chains. xvi Executive Summary Empowering farmers is needed to strengthen their position in the chain and vis-à-vis governments in bargaining for better contract deals, better policies, etc. Policies include stimulating farmers asso- ciations, investing in quality control institutions, competition and trade policy, etc. Additional programs are to: invest in institutions to assist farms with contract negotiations and dispute settlements, invest in institutions for (independent) quality and safety control and certification, and encourage alternatives in input and output markets. Rethinking the Role of the Government and Policy-Making Policy analysis and information gathering. Policy analysis is complicated by the emergence of VC. Traditional instruments of information col- lection do not include information on VC. Rethinking traditional public investments. Traditional areas of public investment such as research and extension, market information sys- tems, veterinary services, and animal surveillance programs need to take into account the role that VC plays in these areas. Public-private partnerships: consider supply chains part of the solution, not the problem. Focus on collaborations between public authorities, non- governmental organizations, and private companies. Innovative �nance instruments. Chain-based �nancing instruments can be very successful. Focus on innovations that use the supply chain as a structural aspect of the �nancing problem, while being critical on which role international organizations and the government should play. Supply-chain development as part of a wider rural development strategy. Countries with many small farmers are typically characterized by overemployment in agriculture. Integration of the farms in modern supply chains cannot solve all structural problems. Supply-chain devel- opment models, even inclusive ones, can be only one part of a broader development strategy. xvii The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Implications for Policy and World Bank Operations INTRODUCTION AND BACKGROUND A major problem in the transition countries of Europe and Central Asia (ECA) during the transition was the breakdown of the relationships of farms with input suppliers and output markets. The simultaneous priva- tization and restructuring of the farms and of the up- and downstream com- panies in the agrifood chain has caused major disruptions. The result is that many farms and rural households face serious constraints in access- ing essential inputs (feed, fertilizer, seeds, capital, etc.) and in selling their products. The problems are worsened by the lack of public institutions necessary to support market-based transactions, such as for enforcing property rights and contractual agreements. In the absence of appropriate public institutions, private contractual ini- tiatives, often from large food and agribusiness companies, are emerging to overcome these obstacles. Large traders, agribusinesses, and food pro- cessing companies, often as part of their own restructuring or following foreign investment, start contracting with the farms and rural households and provide basic inputs in return for guaranteed and quality supplies. This process of interlinked contracting is growing rapidly in ECA agriculture and rural areas. These private contract initiatives can be quite substantial. Empirical evidence indicates that they include farm management assistance, extension services, quality controls, farm input assistance programs, 1 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia trade credit, and even bank loan guarantees. The programs generate im- portant improvements in the credit situation of the farms, as they con- tribute directly to improved access to �nance (e.g., through trade credit), and indirectly as they improve contracting farms’ access to loans from banks or external �nancial institutions (through loan guarantees, en- hanced farm pro�tability, and improved future cash flows). Arguably, the transition disruptions and contract enforcement problems have been even more severe in the rural credit markets than in other markets. In combina- tion, the direct and indirect effects of the farm-assistance programs create important bene�ts for the farms and households supplying to these com- panies: they lead to improved input access, productivity, product quality, and market access. There is growing evidence that these processes have been an engine of growth in the agrifood supply chains of the most advanced ECA countries. For example, almost the entire sugar sector in Central and Eastern Europe is based on supply contracts that include farm-assistance programs. Similarly, recent productivity growth and quality improvements in the dairy sector in many Central and Eastern European countries is driven by processing company investments and farm assistance packages. Yet, several important issues are unresolved in this process of verti- cal coordination (VC). Specifically, related to the efficiency implications, remaining questions include under which conditions such a process emerges spontaneously? Which are the key policy factors in this process? What is the role of foreign investment (FDI) in this process? What trig- gers beneficial spillover effects to other companies and suppliers? How general are these developments? Are they limited to certain subsectors? In which sectors is this process more likely to emerge? Does an optimal model of contracting exist? Are there ECA specific features of these VC programs? Concerning equity implications,1 key questions include whether this process of vertical contracting lead to the exclusion of small farms? Does the emergence of contracting with downstream companies lead to rent extraction of farmers by creating dependency? Finally, several policy issues need to be addressed. Under which con- ditions does such a process emerge spontaneously? In other words, is it suf- ficient for the government to create the right environment for private 1. An issue which is important both from an equity and ef�ciency perspective but which will not be addressed in this report is general equilibrium effects. The process of contract-driven market development and productivity growth may cause important positive general equilibrium effects for poor households, in particular in those coun- tries where agriculture makes up a signi�cant share of output and employment. 2 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia investments in the food industry and agribusiness in order to set this process in motion? If so, which are the key policy factors in this process? How can government policies or interventions contribute to the emergence of these institutions and to desirable ef�ciency and equity effects? OBJECTIVE, METHODOLOGY, AND ORGANIZATION The objective of the study is to analyze VC in agrifood supply chains in ECA and to identify options for improved policies, institutions, and investments which Governments could make, and which the World Bank could support, in order to improve links in the agricultural marketing and processing chain and increase access of farmers to input and output mar- kets. This is especially important in those countries where contractual arrangements are slow to develop. It is also important if farmers are to be lifted out of subsistence farming and into a modern agrifood economy. The report draws upon existing literature2 and a series of case studies and surveys implemented across several countries and (sub-)sectors in ECA agrifood supply chains in the framework of this regional study and other related studies.3 The combination of these case studies from a variety of sources, and the survey of already existing evidence, provides evidence across countries and commodities. The report starts with a conceptual framework and a discussion of the role of VC in other countries. Then it reviews the developments in ECA countries and provides an explanation for these developments. The sub- sequent sections discuss key factors in these developments, and the equity effects. The last part of the report discusses the policy implications. 2. Previous studies on supply chain restructuring and contracting in transition country food chains include Gow and Swinnen (1998, 2001), Dries and Swinnen (2004), Hobbs et al. (1997), Gorton et al. (2003). There is an extensive literature on experience in other parts of the world, in particular the United States, the European Union, Latin America, South Asia, and Africa (see, for example, Glover and Kusterer, 1990; Key and Runsten, 1999; Van der Vorst, 2000; Hobbs and Young, 2001). There is also a related, mostly theoretical, literature that focuses on optimal contracting and in- terlinked markets in developing countries (see, for example, Bardhan, 1989; Bardhan and Udry, 1999 for overviews). A more recent strand of literature studies the impact of new grades and standards imposed in agri-food markets. Recent studies analyzing the effects of standards on developing country farmers include those by Spencer Henson, Steven Jaffee, and a series of papers as part of a World Bank study on standards in agri- food chains; and studies on modern retail chains in developing and transition country farmers, by Tom Reardon and various collaborators. Finally, a series of studies analyzes supply chains from a �nance perspective, focusing on supplier credit as a source of �- nance for small farmers including recent studies by IFAD (2003) on East Africa and a series of studies organized for the World Bank by Renate Kloeppinger-Todd. 3. The studies are listed in annexes. 3 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia CONCEPTUAL FRAMEWORK Vertical coordination may occur at various stages in a supply chain. Two common examples are between a processor and a farm from which the processor purchases produce, or between a farm and an input supplying company. VC can take various forms, which can be thought of as institu- tional arrangements varying between the two extremes of spot market exchanges (0) and full ownership integration (1). Within this 0–1 interval, there is a large variety of different forms of coordination and an equally vast literature trying to classify these various forms, and to explain them.4 An often made distinction, which is useful for our purposes, is between marketing contracts and production contracts. Marketing contracts are (ver- bal or written) agreements between a contractor and a grower that spec- i�es some form of a price (system) and outlet ex ante. Production contracts are more extensive forms of coordination and include detailed production practices, inputs supplied by the contractor, quality and quantity of a com- modity, and a price (system). Key factors determining the use of various contracts or other forms of VC are the costs and uncertainties involved in the transactions, which themselves are affected by the economic and institutional environment, the need for transaction-speci�c investments, the frequency of interacting, and commodity characteristics such as its perishability and the costs of measuring characteristics. A key factor is asset speci�city. If buyers or sellers have to make ex ante investments that are speci�c to the transaction, they want to make sure that the transaction goes through. This encourages them to engage in con- tracting. Another motivation for contracting is perishability of commodi- ties: if a product is highly perishable, suppliers are in a weak bargaining position after harvest and want to make sure ex ante that there is an out- let for their products. An example is the extensive use of contracting in processed vegetables. A related factor here is the frequency of interacting (e.g., daily deliveries versus once a year harvests). Frequent deliveries allow for better exchange of information, the build up of trust, and lower costs of noncompliance. Uncertainty over product quality or reliability of supplies also tends to induce processors to contract suppliers. 4. The basic explanations draw often on the seminal work of Ronald Coase and Oliver Williamson. However, in two recent surveys of the literature (Hobbs and Young, 2001; and Rehber, 2000) no less than seven different strands of literature are identi�ed as being important to understand and explain those differences: transac- tion costs economics, agency theory, competency/capability models, strategic man- agement theory, convention theory, life-cycle theory, and contract economics. 4 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Another important factor is the costs of monitoring commodity character- istics. If commodity characteristics can easily be observed at the time of delivery (e.g., color), spot markets may work well. However, when superior characteristics cannot easily be observed or tested, this may induce VC. For example, recent developments in food safety regulations and biotechno- logical advances may require not just testing of the product, but also extensive monitoring of the production process to guarantee product characteristics. Suppliers may also prefer such vertical integration if they need to make extra efforts or investments to obtain quality standards that are hard to observe in order to receive adequate rewards for their efforts and investments. THE BENCHMARK: VERTICAL COORDINATION IN OTHER COUNTRIES While there is an extensive literature on the theory of VC and contract- ing and a series of case studies, there is relatively little systematic empirical evidence, even for developed countries, such as the United States and the European Union.5 What is available shows (a) that VC is moderately important and (b) that it varies widely between sectors and countries. North America and Western Europe In the United States and in Germany, around one-third of the total value of agricultural production was produced under various types of contracts in the 1990s.6 Contracts were used mostly by larger commercial farms. Only 11% of US farms used contracts in 2001 (see Table 1). However, more than 40% of commercial farms used contracts. In terms of output, 13% of small farm output was under contract, while the commercial farms contracted for more than 40% of their output. Marketing contracts were more widespread than production contracts: almost four times as much farms used marketing than production con- tracts (double as much for commercial farms). However, in terms of share of output, production contracts were almost as important as marketing contracts. 5. Interestingly, the need for better and more precise information on this issue is also felt in countries as the United States. The Canadian government (AgCanada) recently produced a report on vertical linkages in the agri-food supply chains in Canada and the United States (Hobbs and Young, 2001). The USDA/ERSAQ6 just published a major study on contracting in US agri-food chains (MacDonald et al., 2004). A set of US academic experts are currently working with USDA/ERS to develop a database on contracting in US agriculture. 6. Rehber (2000), based on Perry et al. (1996). 5 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia TABLE 1. Importance of Contracting in US Agriculture, 2001 Rural residence Intermediate Commercial All farms farms farms farms Percentage of Farms Production contracts 2.4 0.5 2.1 15.2 Marketing contracts 9.1 3.2 14.2 29.2 Either contract 11.0 3.6 16.0 41.7 Percentage of Output Value Production contracts 16.0 5.0 6.1 20.0 Marketing contracts 20.3 8.3 18.0 22.1 Either contract 36.4 13.3 24.2 42.2 Source: USDA. The main reason is that different commodities use different contracts. Production contracts are important in some of the livestock sectors and especially in hogs (54%) and poultry and egg production (81%). Marketing contracts are mostly used in crops: more than half of cotton and fruits were produced under marketing contracts. Other studies indi- cated that also in potatoes and sugar beets marketing contracts are very important. Contracting is not very important in grains, with the exception of malt- ing barley which is mostly under marketing contracts. The USDA data in Table 2 show that more than 52% of dairy production is under market- ing contracts. However, this likely includes contracts between farmers and their cooperatives as more than 80% of milk was sold to or bargained for by dairy cooperatives in the United States. TABLE 2. Importance of Contracting by Commodity in the United States, 2001 Either contract Marketing contract Production contract Corn and soybeans 11.0 10.9 0.1 Wheat 5.6 5.5 0.1 Barley* 19.3 — 19.3 Cotton 51.7 51.7 — Fruit 59.0 56.5 2.5 Vegetables 36.9 30.0 6.9 Cattle 20.9 3.2 17.7 Hogs 60.5 7.1 53.4 Poultry and eggs 88.1 6.8 81.3 Dairy products 53.1 52.2 0.9 All commodities 36.3 20.3 16.0 *1997. Source: USDA. 6 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Developing and Emerging Economies In developing and emerging economies, VC and contracting is different in terms of its nature and to some extent also in terms of the commodities where it is important. In Latin America, contracting is important in the pro- duction of, for example, sugar, fruits and vegetables (F&V), broilers, malt- ing barley, and in dairy production (Dirven, 1996). In Turkey, beet sugar processing and the commercial part of broiler production (about 40% of total) as well as most processed vegetables were based on contracting between farms and processors (Rehber, 2000). In these countries, contracting plays an important role in providing inputs to farms. Sometimes this is through interlinked contracts with landowners or by agribusiness contracting with farms. For example, a recent study by IFAD (2003) found that input credit provided by agribusiness companies with interlocking arrangements to buy the smallholders’ crops under farming contracts is an important institu- tional arrangement in several countries in East Africa. In many cases, input and credit assistance is limited to basic inputs such as fertilizer, seeds, or working capital. THERE IS MORE, AND MORE COMPLEX, VERTICAL COORDINATION IN ECA Empirical evidence (which is documented in more detail in Annex 1) shows that the pattern of VC in transition countries differs from these ob- servations in rich and poor market economies in several aspects. First, there is signi�cant VC in sectors where we do not observe VC in Western Europe and North America. Second, the nature of VC in ECA is more sim- ilar to that in developing and emerging countries; however, VC in several transition countries is more extensive and more complex.7 At the end of the 1990s, in the Czech Republic, Slovakia, and Hungary, 80% of the corporate farms, who dominated farm production in these countries, sold crops on contract, and 60–85% sold animal products on contract; numbers that are considerably higher than the shares of even commercial farms in the United States. A survey of agrifood processors in five CIS countries (Armenia, Georgia, Moldova, Ukraine, and Russia) found that food companies that used contracts with suppliers grew from slightly more than one- third in 1997 to almost three-fourths by 2003. There was also a strong growth in company ownership of farms. Enterprises directly engaged in 7. Only in some postliberalization emerging economies, such as in Latin America and South Asia, do studies suggest some similar developments. 7 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia farming increased from 6 to 26% of all interviewed firms—with most of this vertical ownership integration occurring recently. There is a sig- nificant growth of supplier support measures as part of the contracts and more farms are getting access to these. Monetary credit, prompt payments, transportation, physical inputs, and quality control are the most commonly offered forms of support. In 2003, over 40% of proces- sors in their sample offer credit to at least some of the farms that sup- ply them and 36% offered inputs. Key �ndings from analyses of various commodity sectors—of which more details are given in section “Sectoral analyses�—also show more extensive contracting in ECA than elsewhere: In the dairy sector, there is no production contracting in countries like the United States. We observe extensive production contracts between dairy processors and farms in ECA, including the provision of credit, investment loans, animal feed, extension services, bank loan guaran- tees, etc. In the sugar sector, we �nd, as in the developed economies, extensive marketing agreements, but the contracts are much more extensive in ECA, including also input provisions, investment loan assistance, etc. In both the dairy and sugar sectors, the extent of supplier assistance by processors also goes considerably beyond some of the trade credit and input assistance provided by agribusiness to farms in some developing countries. In the cotton sector, the standard model in the United States and Australia, two major cotton producers, is that the cotton (from seed to baled cotton) remains in ownership of the producer and the process- ing is paid for as a service. In ECA, the dominant player in the chain is the gin who typically contracts farms to supply seed cotton and pro- vides them with a variety of inputs. This model, which has developed in some of the poorer ECA countries in Central Asia, resembles that of the gin supply-chain structure in developing countries, such as in Africa. However, the extent of contracting and supplier assistance seems to be more extensive in ECA, with credit, seeds, irrigation, fertilizer, etc. being provided by the gins. In the fresh fruits and vegetables (FFV) sector, the rapid growth of mod- ern retail chains with high demands on quality and timeliness of delivery is changing the supply chains. New supplier contracting, which is developing rapidly as part of these retail investments, includes farm-assistance programs, which are more extensive than typically observed in Western markets. They resemble those in emerging economies, but appear more complex in several cases. 8 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia In the grains sector, there is extensive and full vertical integration in Russia and Kazakhstan, where large agroholdings and grain trading com- panies own several large grain farms in some of the best grain-producing regions. Interestingly, ECA seems to be lagging in those areas where VC is most developed in the West, such as intensive hog, poultry, and egg production. To understand the reasons for these differences, their likely develop- ments, and the implications, it is crucial to see these developments as an integral part of the process of transition, a process that involved a major change in the institutions governing exchange and enforcement of contracts. THE DISRUPTION AND REORGANIZATION OF SUPPLY CHAINS IN TRANSITION A major problem in the ECA agricultural sector and rural areas was the breakdown of the relationships of farms with input suppliers and output markets during transition. The simultaneous privatization and restruc- turing of the farms and of the up- and downstream companies in the agri- food chain has caused major disruptions. Widespread forms of contracting problems during transition were long payment delays or nonpayments for delivered products (see Box 1). Such payment delays caused major drains on much needed cash flow for sup- pliers. This was a major problem for all companies in the food chain. Also, food processing companies in Eastern Europe in the late 1990s considered late payments one of their most important obstacles to growth (Table 3). In a survey, companies ranked it as the most important obstacle to com- pany growth in Czech Republic and Slovenia, and the third most impor- tant (out of 12) obstacle in Hungary. Considering that these are some of the most advanced transition countries, one can imagine that this prob- lem was at least as important in others. Also farms breach contracts. Guaranteed supplies of quality raw ma- terials are crucial for processors. In transition countries, processors often have severe problems in obtaining suf�cient quality supplies. Suppliers may not deliver the quality or quantity of raw materials agreed to. The problems are worsened by the lack of public institutions necessary to support market-based transactions, such as for enforcing property rights and contracts. As a result of these and other exchange disruptions, companies lacked reliable supplies of quality deliveries while farms faced serious con- 9 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia TABLE 3. Importance of Various Barriers to Growth for Food Processors in 1998 Average survey scores (1–4*) Type of barrier Czech Republic Hungary Slovenia Exchange rate instability 2.16 2.52 2.04 Inflation 2.41 3.17 2.69 The level of interest rates 3.19 3.15 2.77 Access to credit 2.45 2.43 2.38 Activities of organized crime 1.58 2.03 1.73 and gangster Government price controls 1.74 1.57 2.67 Other government intervention 1.25 2.62 2.67 Your company having high 2.48 2.02 2.23 levels of debt Late payment by customers 3.44 2.70 3.54 Enforcement of bankruptcy laws 2.10 1.29 2.21 Activities of state monopolies 1.90 1.91 1.83 Problems with privatization 1.61 1.49 0.94 *1 describes “no problem,� 4 describes “a major barrier.� Source: Gorton et al. (2000). straints in accessing essential inputs (feed, fertilizer, seeds, capital, etc.) and in selling their products. In the absence of appropriate public institutions, private contractual initiatives, often from large food and agribusiness companies, have emerged to overcome these obstacles. Traders, agribusinesses, and food processing companies, often as part of their own restructuring, start con- tracting with the farms and rural households and provide basic inputs in return for guaranteed and quality supplies.8 The evidence in this report shows that this process of VC is growing. As a result of supply-chain restructuring and VC, these exchange and payment problems have often been importantly diminished in the most advanced ECA countries. However, in many countries problems of pay- ment delays continue until today, even in some of the European transi- tion countries. A survey of Polish pig farmers in 2004 found that problems with payments by contractors were their second most important problem (Wilkin et al., 2004). Another illustration is from the following quote: “Romanian farmers are holding back supplies of milk as they are experi- encing considerable delays in being paid by processors and other buyers. Many farmers have to wait more than two months to be paid for their 8. Interviews with agrifood companies suggest that one of the �rst actions new investors undertake as part of a company restructuring is to pay suppliers on time. 10 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia BOX 1 Contract Problems in Transition9 The widely observed phenomenon of farms not being paid, or much too late, for delivered products is an example of a contracting problem what is called a “hold- up� in the economics literature. In simple terms, it means that once a farm has made an investment for supplying to a processor, it weakens its bargaining position if the processor tries to renegotiate the contract. Consider the case of a farm sugar beet producer. Before the farm invests in seeds, etc. it will discuss with the sugar company the price and conditions of delivery (time, sugar content, etc.). The farm can then decide to produce sugar beet, or produce something else. However, once the farm starts with the beet production and invests in seeds, labor, fertilizer, pesticides, etc., the farm is in a weaker bargaining position. It cannot undo the investments already made and therefore may be forced to accept a worse deal if the company does not honor its commitments: a “hold-up�! The danger of a hold-up is more acute when there are less alternative options for the farm to sell its product. In this case it is said that the investment is “relationship specific�. This is more likely for products which are perishable or which require processing (such as sugar beet) than for commodities which are storable or for which alternative uses exist (such as grains). For the same reason, the availability of competition in the processing sector reduces the likelihood of hold-ups. When the farm realizes the likelihood of a hold-up, it will refrain from making such relationship-speci�c investments unless there are ways to protect itself. Such ways are, e.g., contracts which can be enforced by institutions, such as courts, or trust which has been built up over several years of contracting, or the reputation of a company. Notice that it is also possible that processing companies are held up by farms. This is, for example, the case when processors need a minimum amount of supply of a certain quality to make an investment pro�table. Once the company makes such investment, it is subject to hold-ups by suppliers who may refuse to deliver unless they get a better deal. milk. Some started bringing milk into towns themselves as they will get their money immediate� (AgraFood East Europe, March 2003, p. 23). MULTIPLE MODELS AND MOTIVATIONS Our �ndings suggest that empirically successful models are commodity specific, transition-stage specific, heterogeneous (varying from rather 9. Based on Gow and Swinnen (2001). For examples and economic analyses of contracting and hold-up problems, see Klein et al. (1978), Williamson (1985), Milgrom and Roberts (1992), Klein (1996), and Gow and Swinnen (1998). 11 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia simple to complex), and often “nontraditional�10 because successful mod- els address speci�c transition-related problems, some of which are not prevalent in a “normal market economy environment.� Part of these vari- ations are determined by the same factors that determine variations in contracting in developed market economies, such as transaction cost dif- ferences and commodity characteristics (see above). However, and more importantly in the framework of this study, there are also transition- speci�c factors that affect the VC developments. To understand how these factors affect the development of VC, let us start with a simple framework of a farm (supplier) and a processor and then later look at some more complex examples and institutional developments. The Basic Model After the initial transition-disrupted supply chains, processing companies face a lack of quality supplies. There are several reasons for this. First, farms may not be willing to supply their output to the processor because they fear not being paid once they deliver the product. Second, if farms want to supply, they may not be able to because they cannot access basic production factors. Third, if farms want to supply, they may only supply poor-quality supplies because (a) they lack the necessary inputs to improve the quality and (b) they lack expertise and know-how for producing high-quality goods. A strategy to address these problems typically involved some form of VC. Successful vertical contracting has taken many forms, but has typically included conditions for product delivery and payments as well as farm- assistance programs for suppliers. Typically, payment conditions imply immediate payment for delivered product. Farm assistance has taken many forms including, in some cases, input supply programs, investment assis- tance programs, trade credit, bank loan guarantee programs, extension and management advisory services, etc. Finance Once the company develops such programs, two conditions need to be ful- �lled. First, the processor needs suf�cient funds and cash flow to �nance the supplier contracting system, including immediate payments and the assis- tance programs. Therefore, initiators of contracting with supplier �nancing include: Foreign investors who have access to �nancial means because they have “deep pockets� or because they can access financial markets 10. In the sense that they are not normally observed in market economies. 12 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia internationally (e.g., foreign/multinational processing companies active in dairy, sugar, oilseeds, etc.). Companies who made money in other sectors and are interested in investing these funds in the food sector (e.g., FIGs in Russia). Domestic processors or traders who sell on the international market and have suf�cient turnaround to have �nancial liquidity (e.g., grain traders in Kazakhstan). Domestic processors who have links with the international �nance through contracts with international companies (e.g., cotton gins in Central Asia). Contract enforcement Second, the processor needs to enforce the new contracting system. Enforcement problems are an integral aspect of VC. Enforcement is crucial to make any of the contracts or supplier-assistance programs sustainable. Enforcement is especially problematic in environments where public enforcement institutions are absent. These problems have been overcome in some cases, but in several cases enforcement problems have continued to plague the contracting, sometimes leading to failure of the VC. In sec- tion “Enforcement is difficult but essential,� we will discuss contract enforcement problems and solutions in more detail. Variations on the Basic Model Vertical coordination, both in terms of its type and its extent, differs obvi- ously by commodity, as the commodity and process characteristics affect transaction costs in the exchange. They may also include different compa- nies than just the farm and processor. Annex 2 presents a variety of empir- ically observed models of VC. These include: Triangular structures where processors and retailers work with banks, e.g., via loan guarantee programs, to reduce �nancial constraints of suppliers. We found examples of this in the sugar sector in Slovakia, the retail sector in Croatia, and the dairy sector in many countries (see Boxes A2.1–A2.3 in Annex 2). Vertical coordination with multiple stages is the case, for example, with some brewery investments (where breweries vertically coordinate along the brewing-malting-farm-input supplier chain) or some cotton chains (Box A2.5 in Annex 2). Sometimes different models are developed because processors them- selves do not have access to �nance and other agents in the chain drive the VC. For example, we found several cases where input suppliers, 13 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia banks, local, and international traders vertically coordinated with the farms as processors did not have the �nancial means. In some cases, processing or trading companies take ownership over the farm and fully integrate it in their company. For example, in Russia and Kazakhstan, traders and processing companies own many farms. Large, vertically integrated grain companies are the dominant types of farming in the north of Kazakhstan (Box A2.6 in Annex 2). They use tens of thousands hectares of farming land. Vertical coordination in Russia has grown rapidly since 1998, but the Russian forms of VC are profoundly different from those discussed above as in several cases it was the state (or politicians) that were the driving force behind the VC (Gataulina et al., 2004). After the 1998 Russian �nancial crisis, local authorities encouraged Russian companies to invest in the agrifood system by offering privileges and guarantees. Large industrial holdings became large agricultural holdings as well. The vertical integration process is most active in the Belgorod and Orel regions where a large share of all agricultural enterprises are part of such integrated companies, often farming more than 100,000 hectares (see Box A2.7 in Annex 2). Another somewhat different motivation for vertical integration that is mentioned in some reports is tax incentives. However, in themselves, tax considerations seem to be a poor motivation for engaging in farm- ing by companies with little experience in farm management. Either these initiatives fair poorly or are only one of the motives.11 Vertical Coordination Varies by the Stage of Transition In the early stages most of the emphasis in VC is on securing supplies. Therefore, most emphasis goes to overcoming basic supply problems, such as input (feed, seeds, etc.) and credit (working capital) constraints. This is still the case in some of the cotton supply chains in Central Asia and in 11. This was apparently one of several motivations in Russia where some FIGs in- vested in agricultural production and where taxes were lower for companies having a minimal share of their output in agriculture. They responded to this (and other) in- centives by vertically integrating, and becoming active in farming itself to bene�t from these tax advantages. The same is reported for Ukraine, where vertical integra- tion in the oilseed crushing industry was stimulated by the favorable tax regime of the agricultural sector (EBRD, 2002, p. 23)—although important problems have con- strained, and even reduced, vertical coordination between processors and farms. For example, the EBRD report stipulates that tax incentives were only one reason and seed supply concerns were another, possibly more important, motive for the Ukraine oilseed industry to integrate with farms. Also in Russia, other motivations seem to have induced nonagricultural companies to get involved in farming. (See section on Russian Vertical Coordination in Annex for more details.) 14 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia emerging dairy and F&V supply chains in countries such as Romania and the Caucasus. An important component of the contracts is prompt payments. White and Gordon �nd that 90% of farms get prompt payments in the �rst year of contracting in their study across �ve CIS countries. Similarly, other studies on early stage VC, such as dairy in Romania in 2000 or sugar processors in Slovakia in the mid-1990s, all start with the introduction of prompt payments. In more advanced situations, as is the case in many sectors in Central Europe, there will be more emphasis on product quality. For this, more sophisticated forms of VC are needed, such as extension services and farm-level investments in technology and equipment, leasing, bank loan guarantees, investment assistance, etc. These latter assistance programs require much more sophistication and more complex implementation and enforcement systems. Interestingly, these advanced assistance programs are less widespread in CIS countries than in CEE countries, which more advanced in the process. These pro- grams are also apparently not found in developing country VC strategies. Studies on VC in other regions, such as the IFAD (2003) report on East Africa and Key and Runsten (1999) on Latin-America, do not mention such sophisticated measures. This may reflect the lower quality standards for supplies in these regions, or larger problems of enforcement for such programs, or both. The different contract forms in different regions and at different stages of transition may also reflect different constraints faced by farms, as reforms and their effects impact on the constraints in input and output markets. For example, Table 4 shows how the dominant motivation for farms in Central Europe (Hungary, Slovakia, and Czech Republic) at the end of the 1990s was guaranteed access to markets (52% of the farms listed this as their primary motive) and to a lesser extent guaranteed prices (21%). For TABLE 4. Contract Motivations for Farms in Central Europe Czech, Slovak, Hungary, Most important reason for contracting 1999 1999 1997 Average Contract price higher 9 8 10 9 Avoid price uncertainty 7 22 33 21 Guarantee product sales 64 50 43 52 (Part) prepayment 7 13 3 8 Easier to get credit 0 0 9 3 Contract—inputs and TA 7 6 2 5 Other 6 2 0 3 Source: Leuven ACE datasets. 15 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia a very few, access to credit or other inputs was the main motive. The farms in Table 4 are mostly large farms as they are the dominant contractors. However, the motivations for small cotton farmers in southern Kazakhstan to enter into contracts with gins are very different. For them credit constraints are by far the most important constraint still in 2003, as is clearly reflected in the survey results in Table 5. Financing of inputs also played an important role in the emergence of very large integrated grain companies in northern Kazakhstan. Grain-trading companies �rst introduced pre�nancing for farms in the northern Kazakh grain belt. However, with increasing defaults and bankruptcies, the grain companies ended up taking over the farms to reduce the risk of nonrepayment. An important additional benefit from vertical integration was the increased bargaining power of the integrated grain farms vis-à-vis local and regional authorities, who continue to intervene in farming opera- tions (Gray, 2000). These advantages were particularly important in the grain sector in Kazakhstan where most farms continued to depend on the local authorities for access to key inputs (such as seed and fuel) and for �nancing of these inputs, and where the authorities used this depen- dency to influence farm decision-making. TABLE 5. Contract Motivations for Cotton Farms in Kazakhstan, 2003 Reason for contracting (%) Yes No Most important reason Guaranteed product sales 9 91 8 Guaranteed price 4 96 3 Access to pre�nancing 81 19 75 Access to quality inputs 11 89 10 Access to technical assistance 0 100 0 Other 4 96 3 The progress from simple to more complex assistance programs reflects also the stage of development of contracting between processors and suppliers. Assistance programs start off simply and then gradually become more complex as (a) reforms progress in the countries and (b) as the processors learn from their experiences and the local situation. Not only assistance programs may change, but the organization of the supply chain may be entirely restructured. For example, in the case of modern retail investments, important changes in procurement systems occur step-by-step in the supplier-retailer relationship. These changes include: (a) a shift from local store-by-store procurement to nationally centralized big distribution centers; (b) an incipient shift to regionalization of procurement over countries; (c) a shift from the use of traditional bro- 16 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia kers to new specialized/dedicated wholesalers; (d) increasing local use of global multinational logistics �rms; (e) a shift to preferred supplier sys- tems; (f) a shift to high private standards of quality and safety. These changes dramatically change the contracting relationships between retailers and suppliers. CONSTRAINTS ON VERTICAL COORDINATION IN ECA Several factors may prevent the emergence of VC. First, if there are no problems with securing quality supplies, there is no reason to introduce VC. This is the main reason why there is no VC by supermarkets in, e.g., the Czech Republic and Slovakia, where fruit and vegetable suppliers are mostly large farms, better able to deliver good quality, and to �nance investments themselves. In contrast, in Croatia and Serbia, where food and vegetables suppliers are mostly small farms, supermarkets have intro- duced input assistance and supplier investment support (Reardon et al., 2003a; Dries et al., 2004). FIGURE 1. Impact of Economic Reforms on the Growth of the Modern Retail Sector in ECA* 60 Share of modern retail in total (%) 50 40 30 20 10 1998 2002 0 2.0 2.5 3.0 3.5 4.0 EBRD reform index * Correlation (R2) is 0.79. Data include Bulgaria, Croatia, Czech Republic, Hungary, Poland, Romania, Russia, Slovakia, Ukraine. Source: Dries, Reardon and Swinnen, 2004 The second reason why VC may not emerge is because of lack of invest- ments in food processor companies, which are typically a precondition for VC to emerge. The absence of foreign or domestic investments may be due to a variety of reasons.12 However, typically it reflects political or 12. See, e.g., the 2005 World Bank report on “Improving the Investment Climate.� 17 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia economic instability, insecure property rights, and the absence of key reforms. Several studies con�rm that poor government policies are a major constraint for investments.13 This is also illustrated by Figures 1 and 2, which show a very strong positive correlation between reform progress in transition countries and investments by multinational retailers (see Figure 1) and VC by dairy companies (see Figure 2). FIGURE 2. Impact of Economic Reforms on Farm-Assistance Programs in the ECA Dairy Sector* 80 70 (% interviewed companies) Farm assistance programs 60 50 40 30 20 10 0 2.0 2.5 3.0 3.5 4.0 EBRD reform index * Data include observations from Albania, Bulgaria, Poland, Slovakia Source: Dries and Swinnen (2005) Third, VC may not develop even if investments take place, due to prob- lems of contract enforcement (see section “Enforcement is dif�cult but essential�). For example, in the Ukrainian oilseed sector, pre�nancing from crushers to farmers has decreased markedly since the late 1990s, after crushers experimented with prepayment for seeds, but many had their �n- gers burned with signi�cant defaults (EBRD, 2002). This factor may also be captured by Figure 2, which illustrates the strong effect of reforms on VC as key reforms are needed to facilitate contract enforcement. Fourth, investments in the agrifood chain tend to come in “waves.� This is illustrated by Figure 3, which shows how investments in confec- 13. For example, Codron et al. (2004) show how legal constraints on foreign in- vestment limited FDI in the retail sector in Morocco until very recently. Similarly, for- eign investments in the food industry in Slovakia, a country with many characteristics similar to the Czech Republic, Poland, and Hungary, was much less than in the neigh- boring countries because political instability and lack of consistency in reform under the Meciar government in the 1990s drove investors away. After the change of gov- ernment in 1998, FDI in dairy (and other sectors) picked up rapidly. 18 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia tionary and breweries have preceded investments in dairy and edible oil processors in Russia. The latest wave is the retail investments. Hence, even if reforms are implemented, not all investments will come at the same time. Fifth, like the investment waves discussed here, there are also VC “waves.� Processors �rst focus on securing supplies and later start working on upgrading the quality of the supplies. Retail chains also go through sev- eral steps in developing and upgrading their supply chains, going from wholesale markets to preferred supplies schemes and moving to distrib- ution centre and cross-border supply systems. For example, one can dis- tinguish two phases in the transformation of the ECA retail system, with different types of vertical relationships with supplying farms: an (early) transition phase when privatization and market liberalization took place; and a later globalization phase when major investments in the retail sector by multinationals take place.14 FIGURE 3. Investment Waves in Russia Edible oil 99-04 Dairy 97-01 Yug Russi Beer 95-01 Wimm-Bill-Dann Sloboda Baltika Ehrmann Louis Dreyfus Packaging 99-07 EFES Campina Others? PLM SAB Danone Confectionery 91-99 Huhtamaki Sun-Interbrew Mars Nestle Cadbury Bolshevik Retail 98-09 Wrigleys Migros Dirol Metro Chupa Chups Others? 1990 2000 2010 Source: EBRD 14. Countries such as the Czech Republic, Hungary, and Poland are “�rst wave countries� in terms of retail transformation, starting the globalization period around 1996. Some Balkan countries such as Croatia, Romania, and Bulgaria are part of a sec- ond wave, where retail globalization started in the late 1990s. In a third wave of coun- tries, including Russia and Ukraine, retail globalization did not really start until 2002, but is growing very rapidly now. In 2003 and 2004, Russia was the top destination for global retail investments (Reardon and Swinnen, 2004). 19 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Sixth, growth of VC comes when VC programs spread to other com- panies, other suppliers, and even to other sectors. The main engine behind this is imitation and competition. Once one company introduces suc- cessful VC programs, either competitors copy these programs or suppli- ers shift to such processors. This competitive pressure for suppliers induces an expansion of VC programs even across sectors—as farms may shift crops to access VC assistance.15 In summary, the emergence of VC will depend on the level of reforms in a country, other characteristics which affect private sector in- vestment, the functioning of the rural factor markets, and sector speci�c characteristics. CONTRACT ENFORCEMENT IS DIFFICULT BUT ESSENTIAL Enforcement is crucial to make any of the contracts or supplier-assistance programs sustainable. Enforcement is especially problematic in environ- ments where public enforcement institutions are absent. Evidence sug- gests that court enforcement of contracts is generally not ef�cient; even approaches based on collateral are often flawed because either farms can- not provide the necessary collateral or collecting in on the collateral is problematic in many circumstances in transition. In such environments, the best one can do is create self-enforcing con- tracts by designing the terms of the contracts such that nobody has an incentive to breach the contract (see Box 2). This can be done by in- creasing the costs of breaching the contract or by introducing flexible terms that reduce the chance of breach in case conditions change unexpectedly. However, this is not a simple exercise. An illustration of the problems of setting up self-enforcing contracts is from an FAO project in Macedonia. The project attempted to create markets for vegetable farms by contracting between farms and processors/traders. A large number of farms and processors joined the project, but many contract breaches occurred, on both sides: “Quantities and quality of products delivered to the processors did not meet expectations, since contracted farmers sometimes preferred selling on (high-price) fresh markets and supplied processors with prod- ucts they were not able to sell on the fresh market. In some cases, this caused processors, who target quality product niche markets, to reject the delivered goods. [On the other hand], some processors did not honor 15. There is considerable evidence on this growth process in the reports on in- vestments in Croatian supermarkets (Reardon et al., 2003a), in the Polish dairy (Dries and Swinnen, 2004), in Bulgarian dairy (Noev et al., 2004), in the Kazak cot- ton sector (Sadler, 2004) and in the CEE sugar sector (Gow et al., 2000). 20 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia the signed contract because they failed in identifying export markets� (Martinovski, 2004, p. 6). There are many stories of enforcement failure. In some cases this caused the cancellation of the VC program. For example, Gow and Swinnen (2001) report cases of an international dairy company in Romania and an international brewing/malting investment project in Croatia, which ended up canceling their input prefinance program as farms continuously diverted the inputs for other uses. In other cases, foreign investors left after they failed to obtain sufficient quality of raw materials from their supplying farms, despite extension, training, and support programs, as suppliers regularly sold produce to other compa- nies or traders.16 Even in the successful cases it took considerable fine-tuning of the contracts over time to make the contracts self-enforcing. In addition, cir- cumstances change so rapidly in transition that contracts required con- tinuous adjustments as the self-enforcing range itself changes. Creating the right conditions for successful and self-enforcing contracting requires extensive knowledge of the sector and of local conditions and an ability to flexibly adjust the contract terms to circumstances that can change rapidly in transition. Institutional innovations to ensure supplies for processors or payments for input suppliers also help to enforce contracts. Effectively, what compa- nies do is interlinking markets. The enforcement of the credit transaction (loan and repayment) occurs through the output market. Yet, whether this is suf�cient as an enforcement mechanism depends on a variety of factors, and, as the evidence shows, it may not always be suf�cient. Ultimately, the best way of solving the exchange, contracting, and col- lateral problems in transition countries is to base exchanges and contract enforcement on trust. Unfortunately, due to traumatic experiences during both the communist and the transition periods, trust is generally lacking as a base for business exchanges in many transition countries. However, empirical evidence does suggest that once companies are able to success- fully instigate new contractual exchange forms, trust as a basis for business exchanges can develop relatively rapidly. An interesting example of such trust-based lending to suppliers is the ISPA/Promilch project in Romania where a dairy company collaborates with a farmers association to provide loans to small farmers without collateral (see Annex 2 for details). Hence, 16. This is not unique to transition countries. Enforcement of supplier contracts with assistance programs is a major problem in other regions as well, as, for exam- ple, illustrated by the IFAD (2003) report on supply chain �nance in East Africa. 21 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia BOX 2 Private Contract Enforcement and Self-Enforcing Contracts17 Enforcing contracts through courts is sometimes not viable due to a combination of litigation costs, ineffective contract law, poor third-party veri�ability, and the poten- tial loss of the only suitable trading partner for that commodity. This is especially true in transition economies. For example, in a 1999 survey of Polish hog farmers, more than 60% believed they could not use courts to enforce contracts with their most important customer (Beckmann and Boger, 2003). In this situation, contracts may be enforced without legal institutions by includ- ing flexible conditions to anticipate market changes and by including suf�ciently large private sanctions. Private sanctions include both the losses that result from termination or nonrenewal of the business relationship and from reputational losses, including increased costs of doing business in the future. To understand how this way can be effective, it is important to understand that typically, when there are no changes in factors that affect the contract conditions, there will be no contract breach—otherwise rational partners would not have agreed to the contract in the �rst place. However, if important changes occur in the market environment, it may become interesting for some partner to breach the con- tract. Consider the case when a farm and a processing company agree up front on a price to be paid by delivery of the commodity. Say, the contract price is set at the ex- pected market price. However, the actual market price may deviate from the con- tracted price. If the market price is higher than the contracted price, the contract provides unanticipated bene�ts to the processing company, but is costing the farm, since it could sell its products at a higher price on the market. The farm will compare the costs of staying with the contract (i.e., the losses it incurs by selling at a lower contract price than the market price) with the costs it would incur from breaching the contract. As long as the costs of contract breach are larger, the farm will con- tinue to supply. However, if market prices increase suf�ciently, it may become bene- �cial for the farm to breach the contract and sell its products to another company that pays the market price. Inversely, if the market price falls below the contracted price, the farm gets unexpected bene�ts from the contract, and the processing com- pany has to pay more than if it would buy the commodity at the market. Now the pro- cessing company considers whether it would honor the contract or not. Hence, as long as the market price varies within a certain range around the contracted price, the contract will be honored by both parties. This range is called the “self-enforcing range� of the contract. More generally, the self-enforcing range measures the extent to which market conditions can change without precipitating a hold-up by either party. As long as the relationship remains within the self-enforcing range where each transacting company’s bene�ts of a hold-up are less than the costs, a hold-up will not occur. 17. Based on Gow and Swinnen (2001). 22 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia the problem seems to be primarily to “get the thing going initially� and enforcement costs may decline over time. SECTORAL ANALYSES Dairy Sector Dairy companies played an important role in the restructuring of the dairy farms. Assistance programs that dairies offer to their suppliers help farmers gain access to capital, make investments, and upgrade the quality of their deliveries. Table 6 shows the forms of investment assistance being offered by dairy companies to their suppliers in each country in Poland, Slovakia, and Bulgaria, based on a series of interviews in the three countries with quite different farm structures and at different stages of transition. In Poland, each of the six interviewed dairies has an input supply pro- grams in which they provide access to inputs such as feed (or seeds and fertilizers for on-farm feed production). Five out of six companies assist farm investment through credit programs. Most of the companies also provide extension services to their suppliers. Five of the dairies provide guarantees on bank loans made to farmers, most of which include pref- erential interest rates. Most interviewed dairies also co-sign bank loans when farmers lack suf�cient collateral. In Slovakia, all of the six interviewed dairies assist farms through credit programs for dairy-specific investments. Three of the six interviewed companies assist their suppliers in accessing inputs. Most of the compa- nies also provide extension services. Three of the dairies provide guaran- tees on bank loans made to farmers. The respondents indicated that they offer these types of programs in order to upgrade milk quality and TABLE 6. Share of Interviewed Dairy Companies Having Assistance Programs (in %) Year Country Credit Inputs Extension Veterinary BLG Total Poland 50 67 50 0 50 43 1994 Slovakia 0 0 83 17 27 23 Bulgaria 9 18 9 0 0 7 Poland 83 100 83 17 83 73 1998 Slovakia 17 17 83 17 33 33 Bulgaria 45 64 18 18 18 33 Poland 83 100 83 17 83 73 2002 Slovakia 100 33 83 17 50 57 Bulgaria 82 91 73 18 36 60 BLG, bank loan guarantees. Source: Dries and Swinnen (2004). 23 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia to secure their supplier base against loss to other dairies who do offer these valuable services. Even in Bulgaria, the country least advanced in reforms and transition and which displays the lowest degree of VC, most of the 11 interviewed dairies offer assistance to their suppliers. Nine companies assist farms through credit programs for dairy-speci�c investments, with two of these indicating that they also offer credit for general investments. Ten of the selected dairies assist their suppliers in accessing inputs for on-farm feed production. The majority also provide extension services. Five out of the 11 companies offer bank loan guarantees. Securing the supply base is indicated as the main reason for offering these programs in almost all cases. The share of companies offering assistance has increased in all three countries. Dairy companies in Poland implemented assistance programs quickly. In Bulgaria, the number of dairy companies offering assistance increased gradually. In Slovakia, the increase in assistance accelerated after 1998, consequent with increased inflows of FDI.18 Programs initiated through the use of FDI may have forced local dairies to implement assis- tance programs in response to increased competition as well as providing examples for local dairies to emulate. Two countries that are even less advanced are Romania and Azerbaijan. Both are characterized by small-scale dairy farms and domination of street sales of milk.19 However, since 2000, there have been major changes in Romania, where growing FDI has induced important changes. VC and contracting has developed rapidly importantly with FDI. Interviews show that the large dairy companies contract with small and large farms and offer their farmers assistance programs. Improving milk quality and securing the milk supply base are the major reasons behind offering these assistance programs. Extension services include support to farmers from making feeding plans for their herd, how to increase milk quality, clean- ing practices, and also full business plans. Several dairies provide pre- �nanced inputs and medium-term investment credits. However, except for the dairy owned by a farmers association, other dairies offer these ser- vices mainly to larger farms. 18. In particular, Slovakia’s agrifood industry restructuring was held back by the lack of foreign investment due to poor government policy credibility until the late 1990s. Bulgaria was even further behind in the reform policies. 19. In Romania over 95% of all farms have one to two cows. Only 20–25% of the milk production is processed. Farm use and direct sales on street markets are the main outlets. The processing industry is very fragmented; there are around 550 dairies of which 250 have a capacity of less than 1000 ton/year. The six largest dairy companies account for 25% of dairy processing. In Azerbaijan, dairy farms are also small and less than 10% of the dairy products sold on the market are processed by the industry. Most of the products are sold directly from the farmer to the customer. 24 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia No such change has occurred in Azerbaijan due to the absence of FDI. Vertical linkages are almost nonexistent in the dairy markets, and the local industry remains weak. A particular problem of the Azeri produc- tion system is the lack of basic infrastructure such as reliable energy sup- plies. Regular electricity disruptions in rural areas prevent investments in basic mild cooling and processing equipment. In summary, these �ndings for the dairy sector support key hypotheses that (a) VC is growing in the ECA region; (b) that its emergence is strongly influenced by reform policies; and (c) that VC starts �rst with input sup- port, extension, and simple credit programs; later more sophisticated pro- grams, such as bank loan guarantees and investment loans, are developed; (d) that VC is very important in the most advanced ECA countries and that less advanced ECA countries will converge to this. Evidence from other countries and sectors largely con�rms this picture. Sugar Like in the dairy chains, VC is also prominent in the restructured sugar (beet) chains in ECA countries where foreign investments took place in the sugar sector. By the end of the 1990s, 80–90% of the East European sugar sector was taken over by foreign investors, often in anticipation of high rents with integration of these sugar sectors in the highly subsidized EU sugar markets (see Table 7). Foreign investors have introduced a vari- ety of farm assistance and contracting programs for sugar beet growers when they invested in sugar processing companies. Case studies by Gow et al. (2000) and others provide evidence of input and technical assistance programs, prompt payments policies, and �nance assistance programs, in- cluding bank loan guarantees. These VC forms have been copied by do- mestic producers. TABLE 7. Investment and Restructuring of Sugar Processing Plants in 1999 No. of processing plants Estimated # % of % of sustainable Processing Production Country 1989 1999* processing plants plants w/ FDI capacity w/ FDI Czech Republic 52 10 5 70* 90* Hungary 14 7 3 100 100 Lithuania 4 4 — 100 100 Poland 70 76 15 12* >50* Romania 33 17 10 50* >75* Slovakia 9 6 2 86 95* Slovenia 1 1 1 100 100 *Estimates. Source: Swinnen and Gow (1999). 25 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Fruits and Vegetables A study of �ve CIS countries agrifood sectors shows that VC is also im- portant in the F&V sectors in these countries (White and Gorton, 2004). In contrast, a similar study on Azerbaijan found very little VC in the F&V market. However, this seems to reflect mostly the dire status of the Azeri food industry and its poor record in attracting FDI in the food industry. There is only a very modest amount of foreign investment in two of the agroprocessing plants, from neighbors in the region, particularly Turkey and Iran (Giovannucci, 2004). In more developed ECA countries, FFV contracting is developing fast with modern retail investments. As we explained earlier, with supermar- kets becoming more important outlets, a restructuring of the supply base, including extensive contracting, is taking place in the more developed ECA countries and is likely to spread to other countries in the region. There are two motivations for contracting by supermarkets in FFV: se- curing high-quality supplies and securing year-round supplies. Examples of the �rst type of contract are those of a growing number of retailers in more advanced ECA countries (Dries et al., 2004). Retailers implement contracts with growers to procure higher quality produce. Codron et al. (2004) explain how contracts are designed to encourage the growers to make the speci�c investments required by such a transaction and to pro- tect them from any opportunistic decision of termination of the contract by the retailer. Producers are eager to enter into such a contractual arrange- ment with a retailer.20 A second reason for supplier contracting is to secure supplies through- out the year. An interesting illustration is from Turkey, where retailer Migros supplied the Antalya market with local contracts with tomato growers. The tomato packing unit is in Antalya where the tomato season is from September to June with a gap during the summer. Until recently, summer-harvested tomatoes were bought in northern Turkey. However, since no sorting facilities were available in that distant region, Migros contracted in 2002 with a whole village in the mountains nearby Antalya to grow 1000 tons of tomatoes during the summer period. Migros orga- nized the sorting and packing operations directly in the village, as well as the transport (Codron et al., 2004). 20. Supermarket contracting to stimulate others to invest in new technology to in- crease the product quality is not restricted to farms. Codron et al. (2004) document how a retail chain in Turkey contracted with truckers to upgrade their transportation vehicles. Migros encouraged truckers (with whom it had been working for a long time) to form a cooperative as a framework for a collective contract. Contracts offered employment guarantees to truckers who made speci�c investments in refrigeration equipment. 26 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Cotton21 A comparison of VC in the cotton supply chains in four Central Asian countries (Uzbekistan, Tajikistan, Kyrgyzstan, and Kazakhstan) shows that VC is widespread in each of the Central Asian countries. However, it takes on different forms due to the different reform approaches governments have taken; and the impact on farms has been very different. In Kyrgyzstan and Kazakhstan, farms have bene�ted from the reforms and the VC, with relatively high prices and strong cotton growth, while in Uzbekistan and Tajikistan VC resulted in major rent extraction of cotton farms, with depressed prices and stagnating cotton production (see Table 8). What is remarkable is that in countries where the government has allowed the private gins to develop and compete, cotton farms are doing much better. In Tajikistan and Uzbekistan, where governments actively control (directly or indirectly) input supplies, production, pro- cessing, and marketing in the cotton chain, farm prices are considerably lower than in Kyrgyzstan and Kazakhstan where the private sector has taken on these roles. However, not all is perfect in the latter countries. In Kyrgyzstan, the influx of illegal finance in the cotton chain caused contract breaches and disruption of prefinance agreements between gins and international traders, with major negative repercussions through- out the cotton chain. TABLE 8. Variations in Central Asian Cotton Production, 1992–2003* Measure Kazakhstan Kyrgyzstan Uzbekistan Tajikistan Annual growth rate Harvested area (Ha) 1993–1998 12.3 6.0 −1.7 3.7 1993–2003 5.8 7.6 −1.7 −0.1 Seed cotton production (1000 MTs) 1993–1998 26.7 11 −2.3 8.4 1993–2003 8.9 11.5 −2.8 0.1 Baled cotton production (1000 MTs) 1993–1998 12.6 20.4 −2.7 0.4 1993–2003 5.4 25.9 −2.6 −3.5 Seed cotton Price per MT, 2003 $550.00 $450.00 $200.00 $165.00 Source: Based on Sadler (2004). *There are signi�cant differences in seed cotton production and baled cotton production. The most important reason for these differences is probably smuggling of seed cotton from Uzbekistan and Tajikistan to Kazakhstan and Kyrgyzstan, although there are no data to quantify the amounts of smuggled seed cotton. 21. See Sadler (2004) for more details. 27 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Uzbekistan Cotton exports are a major source of government revenue in Uzbekistan and the state has continued to impose strict controls on the cotton chain. Market reform has been slow. The government reinstituted a state monop- oly on the purchase of cotton in 1995, with prices �xed at amounts based on estimated production costs. Of�cially, producers are required to sell 50% of production. In practice, they often sell their entire crop to the state because they have no �nancing options except to use government �nancing for cotton. Vertical coordination is also controlled by the state. Nearly all gins remain under government ownership, and even the privately owned gins are subject to government control. Cotton producer �nancing is available through a single form of contracting offered by the state through the two main state banks. These banks provide loans in amounts dictated by the central government. Funds are automatically transferred out of a pro- ducer’s account to repay these loans as soon as payments are received by the producer. Inputs are provided through a centralized system of state controlled enterprises. These enterprises give priority to large farms. Small private farmers often obtain access to equipment through unofficial arrangements with equipment operators in these farms. With cotton prices set low by the state, households tend to sell their inputs for cash, or use them on their own private plots. Tajikistan Also in Tajikistan, the government continues to be heavily involved in the cotton chain, although less transparently. Cotton gins are jointly owned by the government and the so-called “investors,� �nancial institutions with (informal) links to the government. There is no competition between the gins. They operate as monopolists in clearly delineated areas, and pre- vent farms from delivering to other gins. The vast majority of cotton is produced under VC �nance package schemes, controlled by the “investors.� They provide crop �nance and sales contracts to the farms, and control also the processing of the cotton. Their �nance comes through pre�nance from the government and a sin- gle international cotton trader. The vast majority of Tajik cotton exports is controlled by one international cotton trader, Reinhart. This company negotiated a “financing package� in 1997 with the government of Tajikistan. Since then, the company, through a local bank (AIB), provides 77% of all agricultural credit, 90% of which is for cotton, of Tajikistan (ADB, 2002a). Reinhart/AIB provides loans to gins that use these loans to further provide �nance to producers through the provision of physical deliveries of fuel, seed, and fertilizer. 28 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia This monopolized system leads to rent extraction from farmers with low seed cotton prices and inflated input prices (ADB, 2002b). With strict control of cash accounts held by producers and the lack of com- petition among gins, cotton producers have no alternatives. The situa- tion is worsened by government involvement in farm production plans and farm debts. Producers have no choice in production decisions because they cannot get financing for the production of anything but cotton. Moreover, accumulated debts on their imposed cotton produc- tion leave them no choice then to follow local authorities’ (who guar- antee the debts) production plans. Kazakhstan The situation is entirely different in Kazakhstan, where VC is also wide- spread in cotton production, but where both producers and processors have been freed from government control since a few years and where competition among gins has produced much better conditions for cotton farms. Kazakhstan initially also took a slow approach to the privatization of farms and gins, and much has changed since 1998. Gins were fully pri- vatized by 1998 and, since then, many new gins have been established. Most gins began purchasing or hiring seed cotton delivery points to trans- port seed cotton from places outside their immediate area. The resulting competition and reduced transport costs have bene�ted (small) farms. Cotton producers are generally too small to attract commercial credits directly, as they lack suf�cient collateral and present a high default risk. They are mainly financed by gins. Eighty-nine per cent of producer respondents said that they received �nancing from a ginner. Gins provide crop �nance, as well as supplying inputs and some agricultural services (see Table 9). TABLE 9. Gin Assistance to Cotton Farms in Kazakhstan, 2003 Kazakh cotton farms Assistance to suppliers (%) Yes No Finance 89 11 Water 73 27 Seeds 65 35 Fertilizer 40 60 Fuel 20 80 Agrochemical inputs 4 96 Agroconsulting 4 96 29 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Gins obtain funds for these �nancing operations from three sources— trader �nancing, facilities with domestic banks, and from their own cash reserves. Trader �nancing takes the form of forward sales of cotton, against which the gins receive a percentage of the value of the cotton that is due to be delivered under the contract. Ginners and traders have established good trading relations through this system over the past 10 years. Kyrgyzstan The situation in Kyrgyzstan is more complex. Privatization, removal of government control, and competition seems to have induced a rapid expansion of the Kyrgyz cotton sector, albeit from a very small base, with similar effects as in Kazakhstan for farms. However, a poor supporting infrastructure and contract breaches with international traders a few years ago have negatively affected the growth of the cotton chain and VC. That said, cotton production and processing continues to expand strongly, partly based on smuggled Uzbek cotton, induced by the large price gap for seed cotton between Uzbekistan and Kyrgyzstan. Many new gins have been constructed in recent years, often with investments by Russian and Turkish textile companies. VC is not needed for the processing of smuggled Uzbek seed cotton which is bought on a cash basis. However, locally produced cotton is based on prefinance contracts by the gins. This system was functioning well until a few years ago as gins were themselves financed under pre- finance contracts with international traders. However, ownership and management of several gins changed around 2000 as the cotton sector was a target of money laundering strategies, and contracts were breached and prefinance from international traders has largely ceased. This has strongly affected contracting with farms as gins themselves now have problems accessing funds to finance the contracts. As a result, ginners have to provide financing out of their own cash reserves and this ham- pers their ability to finance large amounts of seed cotton. Privatized gin- ners supply finance under seed cotton “forward� contracts, with the producer contracting to deliver a predefined quantity of seed cotton to the gin and the ginner agreeing to supply the producer with local cur- rency and inputs at certain times of the season to cover the cost of inputs and labor. Grains VC in the grain sector has taken a somewhat different road as that in the other sectors discussed so far. In fact, several forms of VC have emerged in the grain sector, both public and private. Public interventions have played an important role, and in different ways: 30 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia First, in several East European countries, warehouse receipt systems (WRS) have been introduced, e.g., in Slovakia and Bulgaria, to help farmers overcome working capital shortages. Second, quasi-WRS have developed in Russia and are widely used to collateralize inventories (e.g., of grains and oilseeds). However, there are limitations on this system due to the absence of strong reg- ulatory and enforcement systems or independent inspectors (see Annex 2). Third, authorities in three large grain-producing countries (Russia, Ukraine, and Kazakhstan) have actively intervened in the grain chain. In all three countries, authorities used their control over input sup- ply channels (in particular seeds) to influence farms’ grain produc- tion and marketing decisions. In all countries, this had a detrimental impact on the development of private input supply industries (see OECD, 2004 for Ukraine; Csaki et al., 2002 for Russia; and Gray 2000 for Kazakhstan). More recently, strong vertical integration took place in Russia and Kazakhstan in grain production. In both countries VC was strongly affected by economic events in 1998. This induced the emergence of huge agroholdings in Russia. For example, huge vertically integrated holdings use 80% of the land in Belgorod region and 41% of all arable land in Orel region, two important grain-producing regions in Russia (see Annex 2 for details). The changes in Kazakhstan were different.22 VC was already important in Kazakhstan, and driven by the private sector, before 1998. Until 1998, contract farming was the main organization form, with grain-trading companies entering into annual contracts to provide inputs against deliv- ery of an agreed level or proportion of output of grain. Three important events occurred in 1998, thoroughly affecting the structure of Kazakh agriculture. First, the strong devaluation increased the competitive position of Kazakh grain producers, while improved rev- enues with the increase in international energy prices contributed to more liquidity in the Kazakh economy. Second, the Government initi- ated effective bankruptcy proceedings to remove the burden of farm debts and to change ownership and management of farms. Third, a seri- ous drought in the 1997–1998 season left many farms with increased debts and, importantly, hurt many grain traders who had contracted with grain farms and re�nanced some inputs and were unable to recover their 22. Based on Gray (2000). 31 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia costs. In combination, these factors induced major changes in the supply chains and, in particular, the growth of large-scale corporate farming in the northern wheat-growing region. While some grain companies continued contracting farms despite poor results during the 1998 drought, several companies acquired farms. Within the northern oblasts, grain and food companies bought several entire former state and collective farms, resulting in huge farming com- panies that now produce a large share of grain output. These large inte- grated farm complexes are located in the most favorable areas with the best growing conditions. Even within the same oblast the drier areas are attracting no such investment and the situation of the remaining farms is desperate. For example, in Kostanai oblast no large investor has acquired any entity in the drier southern regions. These forces lead to the survival of only the best and most pro�table areas. The remaining areas, which constitute a numerical majority of farms, lack working capital to operate and continue to be dependent on the local authorities to provide guaran- tees for delivery of fuel and seed inputs. IMPACT The impact of these contract innovations is dif�cult to quantify as sev- eral other factors affect output simultaneously and as company-level information is dif�cult to obtain. Still, the evidence we collected from a series of case studies suggests that successful private contract enforce- ment with vertical contracting has important positive effects, both direct and indirect. On Supplies and Productivity An important direct impact is on the output and productivity of the pro- cessing company that initiates vertical contracting and its suppliers. Supplying enterprises have experienced bene�cial effects on output, pro- ductivity, and product quality through better access to inputs, timely pay- ments, and improved productivity with new investments. Case studies indicate that the programs can lead to double-digit annual growth in out- put and productivity. A case study of the Slovakian sugar sector shows (a) how contracting with timely payments and technical assistance had immediate effects on productivity and output, (b) how the introduction of �nancial assistance packages was the main engine behind the growth in the supplier base, (c) how the bene�ts spilled over to the entire sector as other companies started copying the contracts and farm assistance packages (Gow et al., 2000). Table 10 shows how Juhocukor, the largest sugar processor in 32 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Slovakia, had a decline in sugar beet deliveries and contracted hectares between 20 and 30% before 1993. The company was taken over by for- eign investors, restructured, and introduced new contracts and farm- assistance programs in 1993. After that, its output grew by an average of 33% annually and its contracts by 25% annually between 1993 and 1997. Average yields increased from 33 tons per hectare with 13% sugar content in 1993 to an estimated 45 tons per hectare with 16% sugar con- tent for the 1997 season.23 While output increased immediately after the introduction of input and technical assistance programs in 1993, there was little impact on the number of contracted hectares until 1995. In 1995, a �nance assistance package for suppliers was introduced. After this pack- age was launched, contracting grew very rapidly: by almost 50% annu- ally between 1995 and 1997. TABLE 10. Impacts of contract innovations by Juhocukor* 1993–1997 Indicator 1990–1993 Total 1993–1995 1995–1997 Sugar production −8.8 33.4 30.1 36.7 sugar content 1.2 3.6 3.0 4.2 Contract hectares −6.6 25.3 1.7 48.8 *Average annual changes, in %. Source: Based on Gow et al. (2000). Other studies confirm that relatively small changes in the industry’s practices can already have a major impact at the farm level. For exam- ple, Leat and Van Berkum (2003) indicate that dairy farmers, willing to learn, can achieve better performance even when they have access to only modest farm assistance. Another example from Friesland dairy investments in Romania is illustrative. In 2001, the Friesland company bought a Romanian dairy, which utilized less than 50% of its capacity and had a bad reputation with respect to paying its farmers. Without changing anything but paying-in-time, Friesland succeeded in taking in 20–30% more milk within a time period of 3 months. If farmers are convinced that a processor is reliable in making its milk payments, producers are generally prepared to deliver (more of) their milk (Van Berkum, 2004). 23. For comparison: EU average yields are 50–55 tons sugar beet per hectare and 17.5% sugar content. 33 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia In their survey of ECA agribusiness enterprise executives, White and Gorton (2004) concluded that various contract support measures had caused an average increase in yields of 9.6%. The measures with the greatest impact on yields were specialist storage (especially cooling equipment in the dairy sector), veterinary support, and physical inputs. Specialist storage in the form of on-farm cooling tanks has been partic- ularly important in raising yields and quality in the dairy sector, an effect also found in other countries (Dries, 2004). Market measures such as prompt payments, guaranteed prices, and market access also had significant positive effects. On Quality Quality of output also improved due to these measures, as evidenced by increases in the percentage of output reaching higher or basic quality standards in response to speci�c programs. The programs with the great- est impact on quality were quality control support, veterinary support, physical inputs, market access, and prompt payment programs. Firms that saw an increase in product quality procured a signi�cantly greater proportion of agricultural raw materials using contracts and also employed a signi�cantly greater number of contract support measures. FIGURE 4. Share of Extra Class Milk in Total Deliveries in Poland* 100 90 Warmia Dairy Share of extra class milk in total (%) 80 Kurpie 70 60 Mlekpol ICC Paslek 50 40 Mazowsze 30 Mleczarnia 20 10 0 1996 1998 2001 * Dairy companies in the North East of Poland. Source: Dries & Swinnen, 2004. Figures 4 and 5 show how milk quality rose rapidly following contract innovations by dairy processors were introduced in Poland in the mid- 34 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia 1990s and in Russia in 2000. In Poland, the share of the market held by highest quality milk increased from less than 30% on average in 1996 to around 80% on average in 2001 (see Figure 4). In the Russian Campina factory, the share of the highest quality milk increased from 6% in 2000 to 55% in 2004, while the lowest quality fell from 37% to less than 10% over the same period (see Figure 5). FIGURE 5. Quality standards of milk supplied to the Campina factory in Stupino, Russia, 2000–2004 (% of quality class in total milk supply)* 70 Campina 60 First Second Share of total milk supply (%) 50 40 30 20 10 0 2000 2004 * The Campina quality class includes milk complying with Campina’s international quality standards as well as Premium milk (the top quality class in the Russian classification system). Source: EBRD/FAO (2004); Dries and Reardon (2004). On Investment Direct loans and loan guarantee programs contributed strongly to farm investments in small and medium dairy farms in North Poland (Dries and Swinnen, 2004). More than three-fourths (76%) of all farms made investments in the past years after VC was implemented, includ- ing many small farmers. Of those who invested, 58% used loans for investments in enlarging and upgrading the livestock herd (30%) and cooling tanks (56%). Also, programs that assist farms in accessing inputs (mainly feed) enhance investment indirectly by lowering input costs, or reducing transaction costs in accessing inputs, and consequently, through improved profitability. Figure 6 illustrates the strong impact on investment in equipment, in particular cooling tanks, for small dairy suppliers. 35 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia FIGURE 6. Growth in Dairy Farm Investments in Poland, 1995–2003 (% of suppliers with own cooling tanks for Northern Polish dairies) 100 Mlekpol 90 80 Share of suppliers with own c.t. 70 60 50 Lowicze Kurpie 40 Mazowsze 30 20 10 0 1995 1998 2001 2003 Source: Dries (2004) Cross-Company Spillovers Indirect effects,24 in particular cross-company spillovers, occur as �rms competing for the same suppliers, and their �xed inputs, are forced to offer similar contractual arrangements. For example, in the case of the Slovak sugar sector, competition induced other sugar processors, none of which were taken over by a foreign com- pany until 1998, to imitate Juhocukor’s contractual arrangements. With a 1 or 2 year delay, this resulted in increases in output and productivity in the other sugar companies and their suppliers. On aggregate, with spillover effects starting to be reflected in output after 1995, aggregate sugar output in Slovakia increased from 140,000 tons to around 250,000 tons between 1995 and 1998. Other studies con�rm the importance of this competition 24. The literature on VC in developing countries identi�es two additional sets of spillover effects. First, regional spillover effects may result as increased investments by farmers induce other input supply companies to invest in the region and to in- crease the quantity, quality, and variety of input supplies (Govereh and Jayne, 2003). A second effect is household-level spillovers. For example, Govereh and Tayne (2003) �nd important spillover bene�ts from vertical coordination in contracted cotton pro- duction on increased productivity of noncontracted activities resulting from man- agement advice, and better input use in Zimbabwe. A study on contracted vegetable production in Uganda by Spencer Henson shows that bene�ts from VC for rural households come also from reduced risk (receiving guaranteed prices for contracted crops) in the absence of insurance markets, and improved access to credit (cash for contracted crops) in the presence of imperfect capital markets. 36 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia effect. Noev et al. (2004) and Dries et al. (2004) �nd that, respectively, in the case of the Bulgarian dairy sector and in contracting by modern retail companies in Croatia, competition for suppliers forces other companies to replicate farm-assistance programs in order to secure supplies. Effects of Government Regulated Vertical Coordination Not all examples of VC have equally impressive results. In particular, those cases where the government is heavily and actively involved in the creation and management of the vertically coordinated organizations the effects are dubious at best. In the discussion of the cotton chains (see section “Sectoral analyses�), we already pointed out the negative effects of government intervention in the price setting and product planning and exchange for the farm incomes and ef�ciencies. Also in Russia, the impacts of the government-led creation of huge agro- holdings are mixed (Gataulina et al., 2004; see Annex 2 for details). Vertical integration has contributed to a better supply of inputs to farms and growth in output and productivity, and also to poor �nancial results. Grain productivity has increased strongly since their creation in the late 1990s, also helped by strongly improved prices. Despite this, the pro�tability of grain, milk, and pork for these vertically integrated farms decreased over the same period. Moreover, in contrast to the cases of private-sector-driven VC discussed above, the vertically integrated farms in Russia performed consistently below the regional average. As a result, they have accumulated substantial debts since their creation. Hence, while the vertical integration process seems to have contributed to sometimes spectacular growth in out- put and yields, the pro�ts of the vertically integrated farms seem to have been worse or not better than those of nonintegrated farms. A key prob- lem is the authorities’ interference with production plans and with deci- sions relating to which activities (and companies) should be maintained by a holding, sometimes imposing unpro�table activities and companies on the holding. THE SEARCH FOR QUALITY IS A KEY ENGINE OF VC The shortage of quality supply, which is typical of transition countries, induces VC and spillover effects through farm support packages. Quality demands and SPS standards are likely to increase both for ex- ports and for domestic consumption. They will reinforce the tendency to- ward VC. On the public side, SPS regulations and other requirements will continue to increase for exports (Roberts, 2004). These will have spillover effects on quality requirements and standards for domestic consumption. 37 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia On the private side, modern processors and retail chains impose their private high standards even in countries where consumers may not de- mand such standards. There is extensive evidence on this for supermarkets (e.g., Dries et al., 2004). In a survey of foreign investors in the ECA food industry, all food companies identi�ed quality as an important require- ment for suppliers (see Table 11). There are several reasons for this. They use quality as a strategic tool and as an instrument to differentiate their products from competition. Another reason is that consistent quality stan- dards reduce transaction costs in cross-border supply chains. Private stan- dards also act as a substitute for missing public standards, infrastructure, and institutions. The relationship between public and private standards is nuanced. Modern processors and retailers often have higher private standards than the existing public standards. However, in several cases the standards are also complementary. For example, in the Czech Republic, public standards require FFV producers to keep records of fertilizer and pesticide applica- tions. The implementation of safety (residue analysis) and traceability measures are private initiatives (EBRD/FAO, 2005). Also, Van Berkum and Bijman (2004) find in their survey of food multinationals that these ccompanies demand that their suppliers comply with both public- and company-speci�c quality standards (see Table 11). In order to reach the speci�c quality requirement levels, most of these companies have devel- oped programs to assist farmers to improve their production methods. There is an important issue of what the role of the public sector is in this process, and the issue is quite complex and nuanced. There is a bias toward private institutions in rich countries and toward public institu- tions in poor countries for standards and control (see Table 12). Yet, at the same time, public institutions in rich countries are typically better equipped and reliable to implement these controls. In many poorer countries, private companies are often the only organizations that have both the capacity and the incentives to enforce high product standards. The issue of quality control has both ef�ciency and equity implica- tions. In all the quality-induced VCs, farms get a higher price for higher quality. Since price is related to quality, control of quality is important. Often, quality controls are done by organizations that have an incentive to bias the assessments, both in private and public sectors. Table 11 indi- cates that quality control is always taking place at the factory gate, and sometimes also at the farm gate. Quality controls are implemented by the processors/retailers. The involvement of an independent third party to evaluate company control systems and production methods is still lim- ited. Van Berkum and Bijman (2004) conclude that only in dairy in some ECA countries public inspection services play a substantial role. 38 TABLE 11. Quality Issues in Contracts Between Foreign Investors and Local Suppliers in the Agrifood industry Friesland Campina Oerlemans Farm Frites d a ns SVZ Frites McCain Cehave Explicit contractual agreement on quality Yes Yes Yes Yes Yes Yes Yes Public versus company quality standards Company speci�c Company speci�c Both Both Company Company Both Quality control by Company and Company (each Company Company Company Company Company third party deliv.) and third party (3 monthly) Quality control point At farm level, at At farm level and At factory At factory At factory At factory At factory collection at factory gate gate gate gate gate gate point, and at factory gate Quality determined by Processor and Processor and Client Processor Processor Clients Clients client client and client and client Quality determined by product or by Product Product and Both Both Both Product Both production process production process Involvement of public inspection Control at dairy Control at dairy Very limited Only with Only in food Only in food Sets level, not at and farm level role disputes safety safety issues standards farm level issues Source: Van Berkum and Bijman (2004). 39 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia TABLE 12. Public versus Private Standards Institutions in 2000 Countries Public sector Private sector Public-private Industrialized countries 27% 52% 21% Developing countries 88% 6% 6% Total 73% 17% 10% Source: Henson and Spencer (2004). This causes problems when there is no transparency or trust. For example, prices for cotton seed for farmers in Kazakhstan and Uzbekistan are based on quality. In Kazakhstan, the quality control is done by the private ginneries in their labs. A World Bank survey of small cotton farm- ers in Kazakhstan found that only 8% of the farmers always trusted the evaluation by the ginneries; 60% of farmers say they do not always trust the evaluation, and 32% never trust the evaluations. In Uzbekistan, qual- ity controls are implemented by a government-controlled agency. Yet producers also complain that they are unfairly treated and that their cot- ton deliveries are downgraded. Farms, anticipating this, mix good- and bad-quality seed cotton and attempt to pass it off as good-quality cotton. Improving the quality controls, e.g., by introducing an independent con- trol institution, or by letting farm representatives participate in the eval- uation would have both ef�ciency and equity bene�ts. Improving quality controls, e.g., by introducing independent control institutions, would therefore have both ef�ciency and equity bene�ts. COMPETITION SPREADS EQUITY AND EFFICIENCY BENEFITS Competition plays a very important role in supply-chain effects in induc- ing both bene�cial equity and ef�ciency effects: First, competition induces VC spillover effects across the sector as other processors are forced to introduce similar supplier-assistance pro- grams since suppliers may not want to deliver unless they get simi- lar conditions. This finding of our study is a specific case of more general conclusions that competition is a key factor for encouraging innovation and productivity and that technological development is primarily encouraged through the presence of competition. These are key conclusions in the World Bank’s forthcoming World Dev- elopment Report on Improving the Investment Climate for Growth and Poverty Reduction. 40 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Second, competition prevents farmers from being “exploited.� Competition prevents processing companies or input suppliers from exercising monopoly power in the setting of the contract conditions with farms. A comparative analysis of VC in the cotton sector in Central Asia con- �rms the importance of competition as an important factor to protect small farms against rent extraction by large processors. The only places where we �nd clear evidence that farmers are consistently exploited are in government-controlled monopolized systems, such as the cotton system in Uzbekistan, Tajikistan (and Turkmenistan). In contrast, in Kazakhstan, the cotton chain is characterized by strong competition among private gins buying cotton seeds from small farms for processing. Competition among gins for supplies has induced several developments in Kazakhstan: Ability of small suppliers to get better conditions by changing gins. Almost all (92%) of the interviewed farmers said that they had changed ginnery over the past years if they got better prices or conditions, indicating independence and competition. All farmers (97%) said that they would be able to change gins if they wanted to.25 Investment by gins in local cotton seed collection centers. This means that a farmer now has the option of delivering his seed cotton locally, to a ginner who is not from that area. The cost of the transportation of the seed cotton now falls on the ginner. The provision of basic inputs to small farmers on reasonable conditions. Better prices: as Table 9 illustrates, prices for Kazakh cotton farmers are two to three times higher than those in Uzbekistan or Tajikistan, where competition does not exist. While there remain important problems in the Kazakh cotton system, compared to the situation in Uzbekistan and Tajikistan, the Kazakh situ- ation seems to be considerably more favorable in terms of both equity and ef�ciency. FOREIGN DIRECT INVESTMENT IS ENGINE OF CHANGE An important issue is the role of foreign investment, both as an engine of change and as a potential source of exclusion. 25. There appear to be regular contract break-ups by farmers during the season when another gin offers a higher price. There is a system that if you get pre�nance and inputs from one gin that you can repay the pre�nance and the inputs to the gin, plus a penalty for nondelivery, and then deliver your cotton seeds to another gin. 41 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia The first issue is whether FDI is the engine of VC in supply chains. Conceptually there is no need for FDI to implement successful contract- ing and supplier-assistance programs. However, empirically we observe that FDI has been the most important driving factor behind these pro- grams. Foreign investment plays an important role as an initiator of change and institutional innovation. The introduction of basic forms of vertical integration requires access to outside �nancial sources, which foreign investors have, but which other investors also can have. However, more sophisticated forms of ver- tical integration, with a greater emphasis on quality and standards, are often introduced by foreign companies because they tend to pay greater attention to quality standards. But we also �nd that spillover effects lead to convergence. For example, in the Polish dairy sector, in the mid-1990s, there was a signi�cant differ- ence between foreign-owned processors and local processors. However, by 2001 this gap had disappeared largely as domestic companies started copying the management practices of foreign af�liates. The second issue is whether foreign investors are more likely to exclude small suppliers from the supply base. Foreign agribusiness companies in interviews expressed a preference to source their supplies from larger suppliers (Van Berkum, 2004). However, many domestic companies have similar pref- erences. Empirical evidence based on surveys does not �nd that foreign- owned companies are more likely to cut off small farmers from their supply base than domestic companies. In fact, our surveys in Poland and CIS indicate that FDI �rms source at least as much from small farms as domestic companies. FDI can play an important positive role in the sur- vival and growth of small farms directly and indirectly, by initializing farm-assistance programs and institutional innovations and providing an example of how such innovations can work. What seems to be a key difference in small farm inclusion is the own- ership structure of the companies. Cooperatives are more likely to work with small farms than corporate companies, either domestic or foreign. Our studies from the Polish and Romanian dairy sector indicate this. TRADE PROTECTION AND VERTICAL COORDINATION The argument made is that (some) protection simulates VC. Obviously, there is a relationship between investment and protection—either because of the direct effects of trade policy, or indirectly because it stimulates FDI. The first is trivial, and the second is widely analyzed, as part of many studies on the FDI-trade relationship. Hence, the key issue here is whether protection stimulates VC beyond these well-known effects. 42 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia The answer is not clear. The only cases where trade protection would seem to stimulate VC with domestic suppliers is where the initiators of VC are processors who use as input supplies not the basic commodity, but an intermediate (processed) product that is easily tradable, such as skimmed milk powder for secondary dairy processing or malt for brew- eries. However, a considerable share of VC in ECA is with suppliers of bulky, perishable products (milk, sugar beet, vegetables and fruits, etc.), where it is not evident to source supplies internationally. For example, the CEO of a large retail chain in the Balkans explained that, “for every- thing which can be made from milk powder we use highly subsidized milk powder from the EU, but for the rest we continue to contract with thousands of small local dairy farms,� where contracting involved in importing heifers, giving loans to dairy farms, etc. Another case is cotton in Central Asia. Some gins in the Kyrgyz Republic do not provide VC because they use imported cotton seeds, but this is only smuggled cotton seeds from Uzbekistan. However, as soon as Uzbekistan would liberalize its trade and cotton policy, this would (a) reduce smug- gling because domestic supplier prices would double in Uzbekistan and (b) induce signi�cant investments in the Uzbek cotton gins, most likely with VC. Hence, liberalization would increase VC on both sides of the border. Even in cases where products can be sourced internationally, it is not clear that trade protection would lead to more VC or less. Consider super- markets and F&V supplies. Generally, the arrival of supermarkets and their development of preferred supplier lists, with cross-border sourcing, ini- tially tend to increase VC with local suppliers, and later with investment in logistics to increase cross-border sourcing. This process causes at least as much international sourcing than it reduces. Here again the impact of pro- tection on VC is much less clear than suggested. Finally, open trade can play an important role in providing competi- tion in supply chains. For example, in Moldova the main oilseed crush- ing company wants to ensure its monopoly on sourcing from farmers through export restrictions on oilseeds. An open trade regime would pro- vide farms with alternative outlets, improving their situation—similar to cotton trade in Central Asia where government export restrictions hurt farmers. PROBLEMS OF EQUITY AND EXCLUSION There are two potential equity problems with the VC process. The first is the possibility of rent extraction by the dominant company in the chain. The second is exclusion from the VC process. 43 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Rent Extraction By introducing an interlinked contract, farms can access credit, inputs, etc., which was unavailable before and processing companies can have access to higher quality and timely supplies. Productivity and income increase for the chain as a whole. However, a key question is who bene�ts from this increase in ef�ciency and total income? If both the supplier and the proces- sor bene�t from the institutional innovation, everybody is better off. However, the very rationale for the emergence of these VC institutions may at the same time act as a barrier to entry for other companies and may give the dominant partner in a transaction additional leverage. If the processing �rm can set the terms of the contract such that it captures most or all of the rents, the productivity growth may not bene�t the farms. Moreover, if the interlinking of transactions bestows additional monop- oly power upon the processing company, the farm’s income could even be lower after the contract innovations, despite the fact that total income has improved.26 Hence, an important—and outstanding—issue is how to combine ef�ciency gains with an equitable distribution of the bene�ts in the chain. As we have explained in a previous section, competition can play an important role in this process. Competition prevents companies from ex- ercising monopoly power in the design of the contract conditions and makes it more likely that the farms share in the bene�ts. Interestingly, there may be, at least initially, a problem of sustainabil- ity of the new contracts in a competitive environment. For example, with pre�nanced feed by dairy companies, or seed and fertilizer by crop pro- cessing companies, farms sold their output to competing processors who could offer higher prices since they did not have to incorporate the costs of the assistance programs. In other words, while competition is impor- tant to induce a desirable distribution of the gains, competition may undermine the ability to obtain the gains. Empirical evidence on these issues is limited and one should be care- ful drawing conclusions. Preliminary evidence suggests that so far in ECA (a) both farms and processors bene�t from private-sector-induced VC, especially when suf�cient competition is present, (b) that competition can cause enforcement problems, complicating the VC, and (c) that rent extraction seems most problematic where governments are directly or indirectly involved in restricting competition, as, e.g., in some Central Asian cotton chains (see section “Sectoral analyses�). 26. See Bardhan (1989) and Bell (1989) for an analysis of these equity effects in developing countries in the framework of landlord-tenant transactions. 44 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia The Problem of Exclusion: Small Farmers in the Chain A key concern is that the process of VC will exclude a large share of farm- ers, and in particular small farmers. There are three important reasons for this. First, transaction costs favor larger farms in supply chains, since it is easier for companies to contract with a few large farms than with many small ones. Second, when some amount of investment is needed in order to contract with or supply to the company, small farms are often more constrained in their �nancial means for making necessary investments. Third, small farms typically require more assistance from the company per unit of output. Our studies and interviews with companies generally confirm the main hypotheses coming out of global observations (see Annex 3 for more details): Transaction costs and investment constraints are a serious consideration. Companies express a preference for working with relatively fewer, larger, and modern suppliers. But empirical observations show a very mixed picture of actual con- tracting, with much more small farms being contracted than predicted based on the arguments above. Empirical Evidence Empirical evidence from transition countries and from emerging and developing countries shows a largely consistent picture. Our studies and interviews with companies generally con�rm the main hypotheses that transaction costs and investment constraints are a serious consideration and that companies express a preference for working with relatively fewer, larger, and modern suppliers. However, our empirical observations also show a very mixed picture of actual contracting, with much more small farms being contracted than predicted based on the arguments above. In fact, our surveys in Poland, Romania, and the CIS �nd no evidence that small farmers have been excluded over the past 6 years in developing supply chains. In the CIS, the vast majority of companies have the same or smaller suppliers in 2003 than in 1997. In terms of the supplier assistance, better and more assistance seems to go to larger farms, although there is signi�cant variation with the type of assistance. Other studies con�rm that VC does not exclude small farmers from the supply chains and that all major companies contract with small farmers, but that more sophisticated supplier-assistance programs tend to be more available for larger farms. Often, supplier programs differ to address the characteristics of these varying farms. For example, in case studies of dairy processors investment support for larger farms include leasing 45 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia arrangements for on-farm equipment, while assistance programs for smaller dairy farms include investments in collection units with micro- refrigeration units. Hence, despite the apparent disadvantages noted earlier, the empirical evidence suggests that VC with small farmers is widespread. Furthermore, our empirical evidence indicates that companies in reality work with sur- prisingly large numbers of suppliers and of surprisingly small size. Why Contracting with Small Farmers? There are several reasons: First, the most straightforward reason is that companies have no choice. In some cases, small farmers represent the vast majority of the potential supply base. This is, for example, the case in the dairy sector in Poland and Romania, and in many other sectors in ECA. Second, our case studies suggest also that company preferences for con- tracting with large farms are not as obvious as one may think. While processors may prefer to deal with large farms because of lower trans- action costs in, e.g., collection and administration, contract enforce- ment may be more problematic, and hence costly, with larger farms. Processors repeatedly emphasized that farms’ “willingness to learn, take on board advise, and a professional attitude were more important than size in establishing fruitful farm-processor relationships.� Third, in some cases small farms may have substantive cost advantages. This is particularly the case in labor-intensive, high-maintenance, pro- duction activities with relatively small economies of scale.27 Fourth, processors may prefer a mix of suppliers in order not to become too dependent on a few large suppliers. Finally, processing companies also differ in their willingness to work with small farms. Some processing companies continue to work with small local suppliers even when others do not. These companies have been able to design and enforce contracts which both the small �rms and the companies �nd bene�cial. This suggests that small-scale farm- ers may have future perspectives when effectively organised. 27. For example, Key and Runsten (1999) present evidence that small farmers’ production costs in Mexican vegetable contract production were 45% lower than that of specialized farms owned by the processing companies. Small farmers had signi�- cantly lower labor costs because of access to unremunerated family labor for which markets are missing, and much lower costs of supervising, transporting, and recruit- ing labor input; and because they did not pay any government bene�ts. And also pest control costs were lower due to better crop monitoring and thereby lower chemical use. Further, small farmers yields in vegetable production were 20% higher than on the �rm’s own farms. 46 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia The Farm-Assistance Paradox The evidence presented so far suggests an interesting paradox. Small farm- ers in ECA may not be able to make the necessary upgrades to satisfy the demand of modern supply chains without support packages by processors or agribusiness. If there are sufficient (quality) supplies available for processors, they have no interest in introducing such VC support packages. If there are not suf�cient supplies, VC will be forthcoming. Hence, we have the paradoxical situation that small poor farms may be best off (in the per- spective of “supply-chain-driven development�) if they are in an environ- ment which is dominated by small �nancially constrained farms. There is some empirical evidence for this hypothesis. Companies seem to be most likely to reach out to small farms when they face a supplier base that is dominated by small farmers not able to supply the commodities they want, and least likely when there is a heterogeneous farm structure with some farms able to deliver the desired supplies. For example, some inter- national dairy companies and foreign investors target larger farms as their preferred suppliers and only reach out to smaller suppliers if they need them to secure supplies. For example, in the case of dairy processors, the same processors have different minimum size thresholds for different countries, reflecting the structure of the farm sector in these countries. An international comparison suggests that in Hungary, with many medium and large farms, suppliers have to have more than 20 cows; in Poland, with many small farms, more than �ve cows, and in Romania, where almost the entire herd is in semisubsistence farms, processors work with farms with two cows. ECA is a “Supplier Market� The collapse of farm output and livestock numbers created a gap between processing capacity and supply: hence there is excess demand based on pro- cessing capacity. There is even more excess demand for high-/better-quality supplies because quality is low due to (a) a history of poor quality in the system and (b) reduced access to inputs and finance affect the qual- ity as well. This makes it a “suppliers market� in most of ECA and this, in turn, sup- ports the farms’ bargaining position vis-à-vis the processing sector in the dis- tribution of supply chain rents.28 Moreover, in cases where quality supplies 28. Codron et al. (2004) argue that also in Morocco and Turkey, where there is also a shortage of preferred suppliers to retail chains, these suppliers have strong bargaining power (“most of them have more bargaining power than the retailer�) because, first, as modern retailing of FFV is still in its infancy, there are only a few modern and large grower-shippers, and, second, with a high price gap between ex- ported produce and local produce, the cost of contract termination for the suppli- ers is small since they can get the high export price for their high-quality products. 47 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia are scarce and nontrivial investment is required for quality upgrading, the bargaining power of quality suppliers may increase substantially (post- investment) vis-à-vis the processor or trader.29 These arguments are important both for the issue of exclusion and for the rent distribution in the chain because it suggests that the “power rela- tionship� (and the rent distribution) is endogenous in the development of the supply chain integration. For Now . . . No Time for Complacency What will happen when the market turns? If competition among suppliers increases, or if demand falls, pressure on processors may lead to a consol- idation of the supplier base. At this point there is no systematic evidence that this is happening in ECA. However, there is some ad hoc information from the most advanced countries which suggest that this may be the next phase of VC. Studies from other parts of the world, in particular emerging markets in Latin America, suggest that these pressures may be real and important (Reardon and Berdegué, 2002; Berdegué et al., 2003). Moreover, we �nd that even companies willing to invest in upgrading small farms only go so far, and tend to have a strategy in the long run to upgrade part of their supply to larger, more ef�cient, and fewer suppliers. In many cases supplier-assistance programs explicitly discriminate between larger and smaller farms with the focus of upgrading the better farms and ensuring a minimal supply base and quality from the rest as long as it is required. Hence, in combination, these factors indicate that those who are con- cerned about the inclusion of small farms in these supply chains should not be complacent despite the observations of signi�cant contracting with small suppliers taking place right now. Putting the Equity Issues into Perspective Clearly the equity issues are important challenges. However, several fac- tors suggest that the impact of VC in ECA supply chains will be nuanced, also for small farms. First, the impact of VC is likely to differ signi�cantly between countries and sectors. In some ECA countries, farm output is mostly produced by larger corporate farms. For example, in Slovakia and the Czech Republic, the vast majority of output is produced by corporate farms. In other countries, the importance of farm organizations often 29. Studies on international FFV supply chains in East Africa also �nd that with increasing demand for traceability, the “dependency relationship� between suppliers and processors changes; as processors/traders are now more dependent on their sup- pliers. By working with fewer suppliers, but with higher quality and traceability con- tracts, the suppliers become more “powerful�—and tend to get higher prices. 48 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia differs significantly among subsectors (e.g., grains versus vegetables), reflecting economies of scale. Second, the impact of VC is likely to be a continuation of important agri- food chain restructuring that started 15 years ago. ECA farms have under- gone a dramatic restructuring over the past 15 years. Besides privatization, this included a massive outflow of labor in the most advanced ECA coun- tries. In countries such as Estonia, Hungary, the Czech Republic, and Slovakia, more than 50% of (of�cially registered) workers left agriculture early on in transition.30 This process continued as investments in the food industry and the need to enhance the international competitiveness of the domestic farms have continued pressure for restructuring. In other coun- tries, this adjustment process has been delayed by a variety of problems, but a signi�cant reduction in agricultural employment will be necessary with economic growth, with or without VC. Third, the VC processes have positive effects by addressing major weak- nesses of the ECA farm sector. The ECA farm sector is most in need of finance for investments, technology, and quality improvements, and access to high-value markets. All these factors weaken the competitiveness of ECA food-supply chains with negative effects on their trade balances. Investments by modern processing companies and VC with suppliers can play a signi�cant role in addressing these weaknesses and improving the global competitiveness of the ECA supply chains. Fourth, modern agribusiness and food company investments will not only affect farms, but will have a wider impact on rural development. This includes improved access to better quality and a wider variety of foods and other products for rural households, and the creation of off-farm employ- ment, directly or indirectly, in the supply chain. Investments in packaging, quality control, extension services, etc., are likely to create new jobs in the rural areas; while at the same time the competition from the new chains will cause traditional shops and processors to close. Modern agribusiness and food companies, as motors of market development, will also generate opportunities for differentiation of products and value added. In summary, these arguments suggest that VC in modern supply chains have the potential for important positive implications for rural households in ECA, despite the challenges that they pose. These investments may bring very signi�cant bene�ts to the region, but could also pose signi�- cant threats where inef�cient or undercapitalized farmers cannot “make the grade.� It is important for policy to focus on the most effective and 30. The average annual decline in agricultural employment in the �rst wave coun- tries Hungary, Czech Republic, and Poland was -12% in 1989–1992, -6% in 1993– 1997, and -4% in 1998–2001 (Swinnen et al., 2005). 49 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia appropriate methods for developing “win-win� solutions for companies and farmers. PRIVATE VS. PUBLIC: IMAGE AND REALITY Small Farm Assistance The public policy debate (explicitly or implicitly) often frames the issue of vertical integration and small farms in terms of how public policy can assist small farms who are excluded by the private sector from their supply base. However, the reality suggests a much more nuanced picture. In fact, in a number of cases, instead of excluding small farms, their supplier- assistance schemes seem to reach many small farms that are left out of gov- ernment programs. For example, our study on the Lithuanian dairy sector shows that small farms do not get access to government programs such as SAPARD. For them, the only source of credit and �nance is supplier credit. Furthermore, public policies may even worsen the situation for small farms. Private processing and trading companies only implement their assistance policies out of necessity to enhance their supply base, and seem to do so with some but relatively little discrimination toward small farms. However, if government policies allow medium and large farms to upgrade their technology and farm infrastructure directly or to get access to formal bank loans, they may induce processors to drop their general supplier-assistance packages and start working with the medium and larger suppliers with minimal assistance only. In this way, some gov- ernment rural credit and investment policies may have both direct and indirect anti-small farm biases. Farm Exploitation The public policy debate (explicitly or implicitly) frames the issue of verti- cal integration and small farms in terms of how public policy can prevent (small) farms from being exploited by large, sometimes multinational, agribusinesses in their contractual relationships. However, again, the real- ity suggests a much more nuanced picture. First, while for all agribusinesses pro�ts are their primary concern, this does not seem to lead to exploitation of farms. In cases where suf�cient competition exists there is more evidence that suggests that producers bene�t from VC than that suggests they are being exploited. Second, and equally important, the main cases of farm exploitations that we found is those cases where governments had a heavy hand involvement in either the control of input supplies and product processing/ marketing themselves or where the authorities colluded with a single private company, allowing and predating on the rent extraction of farms 50 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia by the private company. This is the case in the cotton sectors of Tajikistan and Uzbekistan.31 Third, evidence from Ukraine, Russia, and Kazakhstan indicates that an important constraint on enterprise development is rent-extraction by local governments, e.g., through taxation and ad hoc regulations. A major bene�t of large (often vertically integrated) farming corporations in these countries is their ability to withstand pressures from local authorities. This leads to a paradoxical situation that instead of public policy assist- ing small farms to grow in a market environment dominated by large companies, farms need to be large to withstand public pressures. THE FUTURE OF VERTICAL COORDINATION The underlying reason for VC is that these institutional designs address (transition-speci�c) problems that traditional �nancial instruments do not address. This holds also for farm-assistance programs, leasing, WRS, pre�nancing in vertical contracts, etc. Hence, when markets start working better, there is less need for VC. An intriguing question is therefore whether the process as described here represents a transition-speci�c phenomenon or not. The transition from a centrally planned system to a market economy in most of the countries coincided with the break-up of the old state system of strong vertical inte- gration into independent units as illustrated in Figure 7. However, the transition disruptions of the exchanges in product and factor markets caused independent private companies to take initiatives to vertically inte- grate to enforce contracts and improve coordination within the supply chain. In other words, will vertical integration in the supply chain be rein- forced (path A), or will it retreat once public institutions are suf�ciently strong to enforce contracts, with the development of new public institu- tions and market actors, and once factor markets work better (path C)? Most likely a hybrid path will develop in the medium term (path B): for some aspects, vertical integration will remain important. Some forms of VC will remain important, as they are in developed market economies (described earlier). However, in other aspects, which are more closely aligned with transition conditions, VC may retreat. For example, recent information suggests that some of the multinational companies, where possible, return to their core business and leave farming to farms, with as little involvement of the company as possible to keep the quality and reliability of supplies up to a desirable level. 31. Interestingly, the IFAD (2003) also �nds that the main case of rent extraction in the East African farm contracting chains is in Mozambique where government- controlled monopolies dominate. 51 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia FIGURE 7. Vertical Integration in the Agrifood Chain During Transition A Vertical Integration B C Transition PRE EARLY LATE AFTER In general, processing companies have vertically integrated out of necessity rather than intrinsic interest. These companies want to get out of VC if they can, because it is not their core business and because they do not want to carry the risk. These companies prefer no longer to be involved in several aspects of this VC. They would prefer working with “normal� partners—such as �nancial institutions for farm lending. The change of company VC strategies at various stages of transition is illus- trated by the case of Farm Frites, one of the largest potato processors in Europe, which started investing in Poland in 1992: “In the early years of its presence in Poland, Farm Frites found too few Polish farms had the knowledge and means to supply Farm Frites with good quality potatoes. Therefore, the company leased land and cultivated potatoes by itself to sup- ply the factory with good quality potatoes. Next, Farm Frites contracted mainly foreign potato growers already settled in Poland, which were used to West-European quality standards. Step by step Farm Frites has been able to increase the local supply by Polish farmers by assisting them in improving their performances both in yields and the quality of their pro- duce. Nowadays, the company has reduced most of its own potato growing activities, while Polish farmers supply almost 50% of the com- pany’s total purchase of potatoes� (Van Berkum and Bijman, 2004, p. 11). IMPLICATIONS FOR GOVERNMENT POLICY AND WORLD BANK OPERATIONS This report has touched upon a wide variety of issues where policy plays an important role. The issues include the role of privatization, investment 52 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia policy, macroeconomic reforms, land reform and farm restructuring, cap- ital market liberalization, competition policy, etc. It would go too far here to address all these issues and the policy implications related to each of them. This section concentrates on a series of key messages and implica- tions that come directly out of our report and are most relevant.32 A government strategy to stimulate domestic growth in a supply-chain- driven development process while ensuring the inclusion of farms which face major constraints in this process, and an equitable distribution of rents in the chain, should include several policy components, encompassing changes in the regulatory environment and public investments. The policy issues and components can be classi�ed, roughly, in to three groups— although some of the policies could �t in more than one of these groups: (a) the enabling environment for the emergence of VC, (b) policy and programs for addressing equity and efficiency concerns of VC, and (c) implications of VC for public interventions and role in agriculture and agribusiness development. Before discussing these in detail, it is important to emphasize more gen- eral policy issues, which are arguably the most important policy implications of this study. The �rst is the importance of the VC phenomena in ECA agri- food chains and, therefore, the need to explicitly integrate these develop- ments into the policy thinking and program strategies. One of the key �ndings of this study is that VC is much more important and more wide- spread than generally recognized. Policies, in general, have not integrated this structural development so far. The second is that there is signi�cant variation across countries and sectors. The implication is that, as there is no one-size-�ts-all VC but instead several models of VC, reflecting com- modity characteristics, stages of transition, and development, there is no one-size-�ts-all policy. Instead, optimal policies and policy components will also need to differ and change to reflect these differences. In this perspective, the focus of the discussion of policy implications in this report is primarily on less advanced ECA countries, where the WB will be primarily active in the coming years. Enabling and Stimulating Vertical Coordination Create right conditions for stimulating investment Two general conclusions made in the World Bank’s forthcoming World Development Report on Improving the Investment Climate for Growth and 32. Several of these issues and the policy implications are also discussed in related World Bank studies and reports, such as the 2004 Agricultural Investment Sourcebook, “Challenges, strategies and cost of compliance in agrifood standards� by Jaffee et al. (2004); “The Use of Grants to Address Market Failures� by Van der Meer and Noordam (2004); and the 2005 World Development Report on “Improving the Investment Climate for Growth and Poverty Reduction.� 53 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Poverty Reduction are that a good investment climate is the driving force behind economic growth and poverty reduction, and that policy uncer- tainty is the primary concern of �rms in developing countries. In our study, we found ample evidence that a poor policy environ- ment has a negative effect on investments in the agrifood industry and on VC programs. As such it constrains the beneficial effects of VC, which have been shown to be important. For example, the CIS supply- chain study identified problems related to “market governance� as the most important constraints cited by processors. They identify two such problems in particular: the continued operation of “shadow� produc- ers who undercut their competitors through smuggling, counterfeiting, and the avoidance of tax and social security obligations and corrupt national administrations that impose burdensome regulations and demand frequent bribes. Ensure macroeconomic stability Macroeconomic stability is a key condition not only for the investments but, even more so, for supplier-assistance programs or other forms of chain-based �nance. Since VC is importantly a �nancial activity, signi�cant instability may cause such changes in the contract conditions that self- enforcement is no longer possible. This was the case, for example, with the development of quasi-WRS in Russia during the 1998 crisis, or input pre- �nance programs in Kazakhstan during the same period. Hence, macro- economic stability is not only necessary for more traditional �nancing systems but also for nontraditional �nancial instruments. Refrain from direct intervention: Bad policies are worse than bad weather Direct government intervention in the supply chains may undermine the emergence of nontraditional VC �nance systems in several ways; for example, direct interventions in the commodity or input supply markets may crowd out alternative �nancing systems or cause defaults. Large-scale defaults in VC have been caused by both bad weather and bad policies, as they lead to contract enforcement failures. While both can lead to contract breaches, one may be worse than the other in the long run by undermining the private sector development of input provi- sion, both with and without VC. Governments continued to play a very heavy-handed role in the allo- cation of fertilizer and seeds in several countries. For example, in Ukraine, government intervention in input supplies in the 1990s caused private input supply contracting to stop. All private companies moved to input supplies and output purchasing on a cash basis. Private input supplies to the sector declined by about 80%. 54 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Large-scale defaults due to weather effects induced adjustments of con- tract systems and temporary disruptions, but the long run effects were much less negative. Farms defaulted on their repayments after bad harvests due to poor weather. While processors suffered signi�cantly, they were willing to re-engage in VC, with some contract adjustments to reflect risk- reducing strategies. This suggests that processors and input supplies are willing to incorporate temporary defaults due to unforeseen shocks such as the weather but not systemic risks due to government interventions. Improving Ef�ciency and Equity in VC Reduce transaction costs The disadvantage of small suppliers is mostly due to transaction costs.33 Therefore, focus on reducing transaction costs, i.e., the costs per trans- action. This can be done in several ways: Lower transport costs through improvements of rural infrastructure. Rural infrastructure is identi�ed as a serious constraint on VC, and particular on integrating smaller producers in more remote areas. For example, bad roads make it dif�cult to collect farm produce in several countries; regular electricity interruptions in Azerbaijan constrain investments in processing or storage facilities which require electricity, e.g., cooling equipment. Public investments in such infrastructure would stimulate (a) agribusiness investment, (b) VC with suppliers, and (c) inclusion of small farmers in these regions. Reduce the number of transactions by investing in intermediary institutions. Intermediary institutions reduce the cost of exchange between farm and processor/input supplier. Investments in this include both the cre- ation of farm associations and investment in collection points. Collection centers are used by processors and retailers to source from many small suppliers many small farmers, such as Poland, Bulgaria, and Romania, and in Latin America.34 33. Another factor is scale economies. There is a large literature on scale economies in agriculture. Scale economies are commodity speci�c (e.g., wheat vs. dairy), but in general they are rather limited in agriculture, and can often be captured by (larger) family farms. In combination with signi�cant transaction costs in monitoring labor, that is one of the reasons why the family farm is a dominating farming organization in much of western market economies. 34. In some cases we found that retail companies and processors asked farms to organize themselves into associations (see, for example, in Annex 2 the case of the PMOs in the Czech F&V sector for supplying to supermarkets). This suggests that for processors and retailers reducing transaction costs is more important than increased bargaining power of suppliers with farm associations. 55 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Investment in farmers associations has several advantages, such as reduc- ing transaction costs, enhancing suppliers bargaining position vis-à-vis suppliers and vis-à-vis governments, and as an instrument for commu- nication and information distribution. Stimulating farmer associations is an often mentioned policy. In fact, it is hard to �nd a policy document that does not mention it as an important policy. However, the forma- tion of associations in transition countries is hampered by farmers’ neg- ative communist experience with imposed collectivization and farm cooperation.35 Therefore, innovative approaches are important here. Enforce competition The report has documented the great importance of competition in the supply chain, both for ef�ciency and equity. Competition induces proces- sors, retailers, and input suppliers to provide more supplier-assistance programs and it constrains rent extraction of suppliers by up- or down- stream companies. Given these strong bene�ts of competition for farms in the chain, ensuring competition is an important role for the govern- ment. Competition should be enforced through both domestic policies (e.g., competition policies, lower barriers of entry) as well as external poli- cies (e.g., liberal trade policies). The importance of competition does not only apply to private com- panies, but holds also for the case when the government is directly or indirectly imposing a monopoly system and thereby extracting rents from farms. Competition is also important on the input side. The existence of alter- native channels of credit or inputs will constrain rent extraction in the sup- ply chains. Therefore, investments in alternative sources of farm �nance, such as cooperative credit associations, microcredit institutions, etc., should be supported and continued. Stimulate and certify quality and safety standards Invest in projects, institutions, and technical assistance stimulating higher quality Modern supply chains are based on quality, both domestically and inter- nationally. Therefore, preparing suppliers for quality-driven markets will make it easier for them to be integrated in the chains. Technical assistance to strengthen public sector quality testing and certi�cation schemes are use- ful, both for participating in domestic chains and international trade. 35. This is not only the case in transition countries. Key and Runsten (1999) also document how farmers in Mexico are reluctant to work cooperatively because many collective and cooperative rural organizations, including the ejido system, have a his- tory of being used for political ends and subject to manipulation by corrupt leaders. 56 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Furthermore, with prices positively related to quality, it will increase incentives for quality production and, ultimately, incomes. This can also include systems for accreditation and certi�cation, pro- moting better farm and postharvest practices, developing better record- keeping/traceability systems. Enhance bargaining power of (small) farmers Empowering farmers is needed to strengthen their position in the chain and vis-à-vis governments in bargaining for better contract deals, better policies, etc. Several of the policies mentioned earlier will contribute to this objective, such as (among others): Stimulating farmers associations Investing in quality control institutions Competition policies Invest in institutions to assist farms with contract negotiations and dispute settlements. Measures to increase the transparency of contracts, provide for dispute settling arrangements, provide market benchmarks for price negotiations, training farmers in their rights/obligations as contractors, etc. are all important to increase the transparency of the contracting system, competition among contracts, and thereby the bargaining position of farms. As it is generally either not possible or too costly to resolve dis- putes in courts, alternative dispute settlement institutions can play an important role. Invest in institutions for (independent) quality and safety control and certi�- cation. Investing in quality control centers has additional advantages of enhancing the bargaining power of suppliers and ensuring correct payments for quality in the chain. This will lead to better investment incentives and more equal distribution of rents. Improving quality controls, e.g., by intro- ducing an independent control institution, or by letting farm representa- tives participate in the evaluation has both ef�ciency and equity bene�ts. Encourage alternatives in input and output markets. Empowering farmers will also come importantly from alternative options in accessing inputs and selling their products. Competition and liberalization of export regimes will also enhance the farms’ situation. Here also investments in projects and institutions supporting higher quality will contribute to this goal. Implications for the Role of the Government in Policy Management and Public Goods and Services Provision The development of supply chains and VC requires a fundamental rethink- ing of the role of the government and policy-making. Large companies develop 57 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia their own standards, their own extension services, their own supply chan- nels, and wholesale exchange institutions, quality testing, etc. Some of these activities are in areas where traditionally governments were consid- ered to play an important role. Hence, there are fundamental and dif�cult questions on the role of the government in such a changed environment. Policy analysis and information gathering Policy analysis is complicated by the emergence of VC for a number of reasons. One reason is that basic models of supply reactions to policy changes may have to be adjusted for the more complex organization of the supply system. Another reason is that traditional instruments of information collection on which policies are typically based do not usu- ally include information on VC. Hence, information collection (e.g., sur- vey instruments) may need to be explicitly designed or adjusted to account for this.36 Rethinking traditional public investments Traditional areas of public investment such as research and extension, market information systems, veterinary services, and animal surveillance programs require rethinking to take into account the role which VC plays in this area. Optimal government policies in a VC environment will be based on public-private partnerships. Public-private partnerships: Supply chains as part of the solution, not the problem Since private investments and strategies play a crucial role in the supply- chain process, collaborations between public authorities, nongovern- mental organizations, and private companies should be at center stage. There are some successful examples of where such partnerships have con- tributed to positive developments from each perspective. For example, a recent collaborative project, �nanced by USAID, between the Michigan State University-based Partnership for Food Industry Development (PFID), South African retail chains, and local NGOs has lead to the creation of a framework approach where small farmers access to seeds, services, finance, and output markets were integrated—much like VC in private- sector-driven models—and which has led to upgrading of small farmer supplies and integration of small farmer groups in South African supply 36. This is not just a problem in ECA countries, but almost everywhere. For ex- ample, a recent study by USDA/ERS points at important gaps in the understanding of markets in the United States as their information base does not capture important contractual developments (MacDonald et al., 2004). 58 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia chains. Retail chains are interested in working with USAID in Africa to replicate this system.37 Innovative �nance instruments: Finance the chain A key conclusion from this report is that the most successful VC approaches have been nontraditional ones that address transition-specific con- straints, are flexible, and allow adjustments to reflect changes in transition. Innovative instruments using chain-based �nancing have been devel- oped in the private and the public sectors, and some have been very suc- cessful. In the annex to this report, we review several of these innovations, including: lending to farmers through associations, reverse factoring, quasi-WRP, leasing, contract-based lending without collateral, etc. Several of these are private initiatives and there is only a limited role for the gov- ernment. In other cases (e.g., WRPs) there is a more important role in, e.g., the regulatory and legal system that is required for these instruments to function; or there may be a role in co�nancing seed money to start up some of these innovations. The key conclusion here seems to be one of being open to innovations which explicitly take into account the supply chain as a structural aspect of the �nancing problem, while being critical on which role international organizations and the government should play. Supply-chain development as part of a wider rural development strategy Countries where small farmers make up a large share of the agricultural sector, and thus the supplier base, are typically characterized by signi�cant overemployment in agriculture from a long-run development perspective. Signi�cant productivity increases and growth can come from integration of the farm sector in modern supply chains and the associated inflows of inputs, technology, capital, and management. However, these bene�cial developments are unlikely to solve all structural problems in the rural areas. Therefore, it is unrealistic to assume that in such countries all house- holds currently employed or relying on agriculture will be able to be included in such a development. For a broader pro-poor development process, ultimately a broader process of rural development is needed with the creation of many off-farm employment opportunities in rural areas, or at least accessible for rural households. Supply-chain development models, even inclusive ones, can be only one part of such strategy. 37. This success contrasts with some other marketing projects where a similar approach was tried but with a strong focus on the supply side. The lack of attention to the demand side ultimately negatively affected the project as output markets were missing. 59 List of Background Studies and Related Studies “Dynamics in Vertical Coordination in the Romanian Dairy Sector,� by S. van Berkum. “Vertical Coordination and Foreign Direct Investments in the Dairy Sector in Central Eastern Europe,� by L. Dries. “Integration Processes in Agriculture-Industrial Complex: Agro-Firms and Agro-Holdings in Russia,� by E. Gataulina et al. “Vertical Coordination in CIS Agrifood Chains,� by J. White and M. Gorton. “Cotton Supply Chain Report,� by M. Sadler. “Study of Dairy Chain in Azerbaijan,� by E. Hidier. “Study of the Fruit and Vegetable Chain in Azerbaijan,� by D. Giovanucci. “The Impact of Foreign Investment on Supply Chains in ECA: Lessons from a Survey of Dutch Agri-Food Multinational Companies,� by S. van Berkum and J. Bijman. “The Rapid Rise of Supermarkets in Central and Eastern Europe: Implications for the Agrifood Sector and Rural Development,� by L. Dries, T. Reardon, and J. Swinnen. “Study on Market Linkages in Slovak Agriculture,� by G. Blaas, M. Bozik, J. Pokrivcak, and S. van Berkum. “Study on Rural Financial Constraints in the Framework of EU Accession of Lithuania,� by W. Meyers, R. Naujokiene, and I. Krisciukaitiene. 60 Bibliography Asian Development Bank, 2002a, Report and Recommendation of the President on a Proposed Loan for the Tajikistan Agriculture Rehabilitation Project. Asian Development Bank, 2002b, Cotton in Central Asia, A Review of Policy and Technology. Badiane, O., D. Ghura, L. Goreux, and P. Masson, 2002, Cotton Sector Strategies in West and Central Africa, IMF Working Paper, October. Bardhan, P. (ed.), 1989, The Economic Theory of Agrarian Institutions, Clarendon Press, Oxford. Bardhan, P. K. and C. Udry, 1999, Development Microeconomics, Clarendon Press, Oxford. Beckmann, V. and S. Boger, 2003, “Courts and Contract Enforcement in Transition Agriculture: Theory and Evidence from Poland,� paper pre- sented at the IAAE Conference in Durban, South Africa, 16–22 August. Bell, C., 1989, “A Comparison of Principal-Agent and Bargaining Solutions: the Case of Tenancy Contracts,� in Bardhan, P. (ed.) The Economic Theory of Agrarian Institutions, Clarendon Press, Oxford. Bende-Nabende, A. and J. L. Ford, 1998, “FDI, Policy Adjustment and Endogenous Growth: Multiplier Effects from a Small Dynamic Model for Taiwan, 1959–1995,� World Development 226(7), 1315–1330. Berdegué, J., F. Balsevich, L. Flores, and T. Reardon, 2003, Supermarkets and Private Standards for Produce Quality and Safety in Central America: Development Implications, Report to USAID under the RAISE/SPS Project, Michigan State University and RIMISP, July. Bivings, L. and D. Runsten, 1992, Potential Competitiveness of the Mexican Processed Vegetable and Strawberry Industries, Report for the Ministry of Agriculture, Fisheries and Food, British Columbia, July. Blaas, G., M. Bozik, and S. Van Berkum, 2004, “Study on market linkages in Slovakian Agriculture,� paper, World Bank, Washington DC, June. 61 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Blanchard, O., 1997, The Economics of Post-Communist Transition, Clarendon Press, Oxford. Blanchard, O. and M. Kremer, 1997, “Disorganization,� Quarterly Journal of Economics 112(4), 1091–1126. Blomstrom, M. and A. Kokko, 1997, “How Foreign Investment Affects Host Countries,� Policy Research Working Paper 1745, International Economics Department, World Bank, Washington DC. Boger, S., J. Hobbs, and W. Kerr, 2001, “Supply Chain Relationships in the Polish Pork Sector,� Supply Chain Management 6(2), 74–82. Borensztein, E., J. De Gregorio, and J-W. Lee, 1998, “How Does Foreign Direct Investment Affect Economic Growth?� Journal of International Economics 45, 115–135. Brenton, P. and F. di Mauro, 1999, “The Potential Magnitude and Impact of FDI Flows to CEECs,� Journal of Economic Integration 14(1): 59–74. Bruszt, L., 2000, “Constituting Markets: The Case of Russia and Czech Republic,� in Dobry, M. (ed.) Democratic and Capitalist Transition in Eastern Europe: Lessons for Social Sciences, Kluwer, Dordrecht. Bull, C., 1987, “The Existence of Self-Enforcing Implicit Contracts,� Quarterly Journal of Economics 102(1): 147–159. Codron, J.-M., B. Zouhair, F. Fort, E. Coudel, and A. Puech, 2004, “Supermarkets in Low-Income Mediterranean Countries: Impacts on Horticulture Systems,� Development Policy Review, forthcoming. Csaki, C., 2004, “Regoverning Markets in Food and Agricultura,� Unpublished report, Corvinius University of Budapest. Csaki, C., J. Nash, V. Matusevich, and H. Kray, 2002, Food and Agricultural Policy in Russia: Progress to Date and the Road Forward. World Bank Technical Paper No. 523, Washington DC. Dirven, M., 1996, “Agroindustria y pequeña agricultura. Síntesis com- parativa de distintas experiencias (LC/R.1663),� Santiago de Chile, CEPAL. Dries, L., 2004, “Vertical Coordination and Foreign Direct Investments in the Dairy Sector in Central Eastern Europe.� Report for the World Bank Study on “Dynamics of Vertical Coordination in ECA Agrifood Chains: Implications for Policy and Bank Operations,� World Bank, Washington DC. Dries, L. and T. Reardon, 2004, “Impact of Retail Transformation and Procurement System Change on the Russian Dairy Supply Chain,� mimeo. Dries, L., T. Reardon, and J. Swinnen, 2004, “The Rapid Rise of Super- markets in Central and Eastern Europe: Implications for the Agrifood Sector and Rural Development,� Development Policy Review 22(5): 525–556. 62 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Dries, L. and J. Swinnen, 2004, “Foreign Direct Investment, Vertical Integration and Local Suppliers: Evidence from the Polish Dairy Sector,� World Development 32(9), 1525–1544. Dries, L. and J. Swinnen, 2005, “The Growth of Vertical Coordination in Transition: Survey Evidence from Albania, Bulgaria, Poland and Slovakia,� Working paper, LICOS Center for Transition Economics, KU Leuven. Dunning, J., 1993, Multinational Enterprises and the Global Economy, Addison- Wesley, Reading, MA. EBRD, 2002, Agribusiness Operations Policy, paper, London. EBRD/FAO, 2002, “Ukraine: Review of the Sunflower Oil Sector,� FAO Investment Centre/EBRD Cooperation Programme, Report. EBRD/FAO, 2005, “Central and Eastern Europe: Impact of Food Retail Investments on the Food Chain,� FAO Investment Centre/EBRD Cooperation Programme, Report Series No. 6, February. Eggertsson, T., 1990, Economic Behaviour and Institutions, Cambridge University Press, Cambridge, UK. Foster, C., 1999, “Agribusiness Problems and Development in the for- mer Soviet Union,� in Agricultural Finance and Credit Infrastructure in Transition Economies, OECD Center for Co-operation with Non- Members, OECD Publications, Paris. Gataulina, E. A., V.Ya. Uzun, A. V. Petrikov, and R. G. Yanbykh, 2004, “Integration Processes in Agriculture-Industrial Complex: Agro-Firms and Agro-Holdings in Russia, Moscow,� The Foundation for Support of Agrarian Reform and Rural Development (RosAgroFond). Germenji, E., 2004, “Vertical Coordination in Albanian Agriculture,� Unpublished report, LICOS, KU Leuven. Giovannucci, D., 2004, “Azerbaijan Agricultural Markets Study. Current Constraints and Future Potential in Fruits & Vegetables,� paper, World Bank, Washington DC. Glover, D. and K. Kusterer, 1990, Small Farmers, Big Business: Contract Farming and Rural Development, St. Martin’s Press, New York. Gorton, M., A. Buckwell, and S. Davidova, 2000, “Transfers and Distortions along CEEC Food Supply Chains,� in Tangermann, S. and M. Banse (eds) Central and Eastern European Agriculture in an Expanding European Union, CABI Publishing, Wallingford, Oxon, pp. 89–112. Gorton, M., M. Dumitrashko, and J. White, 2003, “Overcoming Supply Chain Failure: a Moldovan Case Study of Agri-Food Industry Recovery,� Report for INTAS Research Project on Agri-Food Industry Restructuring in Ukraine and Moldova (AFIRUM). Govereh, J. and T. S. Jayne, 2003, “Cash Cropping and Food Crop Productivity: Synergies or Trade-offs?� Agricultural Economics 28: 39–50. 63 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Gow, H. and J. Swinnen, 1998, “Agribusiness Restructuring, Foreign Direct Investment, and Hold-Up Problems in Agricultural Transition,� European Review of Agricultural Economics 25(4): 331–350. Gow, H. and J. Swinnen, 1999, “Agricultural Credit Problems and Policies during the Transition to a Market Economy in Central and Eastern Europe,� Food Policy 24: 21–47. Gow, H. R. and J. Swinnen, 2001, “Private Enforcement Capital and Contract Enforcement in Transition Economies,� American Journal of Agricultural Economics 3(83): 686–690. Gow, H. R. and J. Swinnen, 2004, “How Private Contract Enforcement Mechanisms Can Succeed Where Public Institutions Fail: Contracting in Transition,� Working paper, LICOS Center for Transition Economics, KU Leuven. Gow, H., D. H. Streeter, and J. Swinnen, 2000, “How Private Contract Enforcement Mechanisms Can Succeed Where Public Institutions Fail: The Case of JUHOCUKOR a.s.,� Agricultural Economics 3(23): 253–265. Gray J., 2000, Kazakhstan. A Review of Farm Restructuring, World Bank Technical Paper No. 458, World Bank, Washington DC. Greif, A., 1997, “Contracting, Enforcement, and Ef�ciency: Economics beyond the Law,� in Annual World Bank Conference on Development Economics 1996, The World Bank, Washington DC, pp. 239–265. Greif, A. and E. Kandel, 1995, “Contract Enforcement Institutions: Historical Perspectives and Current Status in Russia,� Chapter 8, in Lazear, E. P. (ed.) Transition in Eastern Europe and Russia: Realities of Reform, Hoover Institute Press, Stanford, CA, pp. 291–321. Hart, O., 1995, Firms, Contracts and Financial Structure, Clarendon Press, Oxford. Henson and Spencer, 2004, “National laws, Regulations and Institu- tional Capabilities for Standards Development,� Paper presented at the World Bank Conference on Food Regulations and Standards: International Dimensions and Trade, Washington DC, January 27, 2004. Hobbs, J. and L. Young, 2001, Vertical Linkages in Agri-Food Supply Chains in Canada and the United States, Report, Agriculture and Agri-Food Canada. Hobbs, J. E., W. A. Kerr, and J. D. Gaisford, 1997, The Transformation of the Agrifood System in Central and Eastern Europe and the New Independent States, CAB International, New York. Hu, D., T. Reardon, S. Rozelle, P. Timmer, and H. Wang, 2004, “The Emergence of Supermarkets with Chinese Characteristics: Challenges and Opportunities for China’s Agricultural Development,� Development Policy Review, forthcoming. Humphrey, J. and H. Schmitz, 1998, “Trust and Inter-Firm Relations in Developing and Transition Economies,� Journal of Development Studies 34(4): 32–61. 64 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia IFAD, 2003, “Agricultural Marketing Companies as Sources of Smallholder Credit in Eastern and Southern Africa. Experiences, Insights and Potential Donor Role,� Rome, December 2003. Jacobsen, H. C., 2004, “Financing Agribusiness in Eastern Europe. The EBRD Experience,� EBRD paper. Jaffee, S., de Haan and K. Van der Meer, 2004, “Challenges, Strategies and Cost of Compliance in Agrifood Standards,� Report, World Bank, Washington DC. Johnson, S., J. McMillan, and C. Woodruff, 1999, “Contract Enforcement in Transition,� EBRD Working Paper No. 45. Kazlauskiene, N., 2004, Lessons Learned from Pre-accession Experience in Rural Development and Structural Programs in Lithuania, presentation at the World Bank seminar on Challenges of Effective Implementation of EU Rural Development and Structural Policies in Latvia, May 20–21. Key, N. and D. Runsten, 1999, “Contract Farming, Smallholders, and Rural Development in Latin America: The Organization of Agroprocessing Firms and the Scale of Outgrower Production,� World Development 27(2): 381–401. Klein, B., 1996, “Why Hold-Ups Occur: The Self-Enforcing Range of Contractual Relationships,� Economic Inquiry 36(July): 444–463. Klein, B., R. G. Crawford, and A. A. Alchian, 1978, “Vertical Integration, Appropriable Rents, and the Competitive Contracting Process,� Journal of Law and Economics 24: 297–326. Konings, J. and P. Walsh, 1999, “Disorganization in the Process of Transition: Firm-Level Evidence from Ukraine,� Economics-of-Transition 7(1): 29–46. Lall, S., 1980, “Vertical Inter�rm Linkages in LDCs: An Empirical Study,� Oxford Bulletin of Economics and Statistics 42: 203–226. Leat, P. and S. Van Berkum, 2003, “Dairy Sector Analysis,� in Van Berkum, S., S. Davies and S. Popov (eds) The Romanian Agrifood Chain: On the Road to Accession, Ministry of Agriculture, Forests, Water and Environment, Bucharest. Lerman, Z. and K. Brooks, 2001, Turkmenistan: An Assessment of Leasehold- Based Farm Restructuring, World Bank Technical Paper No. 500, The World Bank, Washington DC. Liefert, W. M., B. Gardner, and E. Serova, 2002, “Allocative Ef�ciency in Russian Agriculture: The Case of Fertilizer and Grain,� American Journal of Agricultural Economics 85(5): 1228–1233. MacDonald, J. M., M. C. Ahearn, and D. Banker, 2004, “Organizational Economics in Agricultural Policy Analysis,� American Journal of Agricultural Economics, May 2004. 65 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Macours, K. and J. Swinnen, 2000, “Causes of output decline during tran- sition: The Case of Agriculture in Central and Eastern Europe,� Journal of Comparative Economics 28(March):172–206. Martinovski, Z., 2004, Unpublished FAO Report on Contract Farming in Macedonia, FAO. Masten, S. E., 1996, “Introduction,� in Masten, S. E. (ed.) Case Studies in Contracting and Organization, Oxford University Press, Oxford. McMillan, J. and Ch. Woodruff, 1999, “Interfirm Relationships and Informal Credit in Vietnam,� Quarterly Journal of Economics 114: 1285–1320. Melyukhina, O. and I. Khramova, 2000, “New Players in the Food Marketing Chain in Russia,� in Wehrheim, P. and J. Von Braun (eds) Russia’s Agri-Food Sector: Towards Truly Functioning Markets, Kluwer Academic Publishers, Dordrecht, pp. 383–404. Meyers, W., 2004, Study on Rural Financial Constraints in the Framework of EU Accession of Lithuania, Interim report, Lithuania Technical Assistance Project, The World Bank. Milgrom, P. and J. Roberts, 1992, Economics, Organizations and Management, Prentice Hall, Englewood Cliffs, NJ. Moore, M., 1994, “How Difficult is it to Construct Market Relations? A Commentary on Platteau,� Journal for Development Studies 33(3): 287–263. Noev, N., L. Dries, and J. Swinnen, 2004, “Foreign Investment and the Restructuring of the Dairy Supply Chains in Bulgaria,� Working paper, LICOS-Centre for Transition Economics, KU Leuven. OECD, 1997, Review of Agricultural Polices: Slovak Republic, Centre for Co- operation with Economies in Transition, Organization for Economic Co-operation and Development, Paris. OECD, 1998, Agricultural Policies Markets and Trade in Transitional Economies: Monitoring and Evaluation 1998, Centre for Co-operation with the Economies in Transition, Organization for Economic Co- operation and Development, Paris. OECD, 2004, Achieving Ukraine’s Agricultural Potential: Stimulating Agricultural Growth and Improving Rural Life, Organization for Economic Co-operation and Development, Paris, September. Perry, J. D., M. Morehart, and J. Johnson, 1996, “Farmer’s Use of Marketing and Production Contracts,� USDA, Farm Business Economic Branch, ERS, AER 747. Pirrong, S. C., 1993, “Contracting Practices in Bulk Shipping Markets: A Transaction Cost Explanation,� Journal of Law and Economics 36(October): 937–976. 66 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Platteau, J.-P., 1994a, “Behind the Market Stage Where Real Societies Exist—Part I: The Role of Moral Norms,� Journal of Development Studies 30(3): 533–577. Platteau, J.-P., 1994b, “Behind the Market Stage Where Real Societies Exist—Part II: The Role of Moral Norms,� Journal of Development Studies 30(4): 757–817. Pomfret, R., 2000, “Agrarian Reform in Uzbekistan: Why has the Chinese Model Failed to Deliver?� Economic Development and Cultural Change 48(2): 269–284. Pomfret, R., 2002, “Resource Abundance, Governance and Economic Performance in Turkmenistan and Uzbekistan,� Paper prepared for the project “Energy and Welfare in the Caspian Region,� September, University of Adelaide. Reardon, T. and J. A. Berdegué, 2002, “The Rapid Rise of Supermarkets in Latin America: Challenges and Opportunities for Development,� Development Policy Review 20(4): 317–334. Reardon, T. and J. F. M. Swinnen, 2004, “Agrifood Sector Liberalization and the Rise of Supermarkets in Former State-Controlled Economies: Comparison with Other Developing Countries,� Development Policy Review 22(4): 515–524. Reardon, T., C. P. Timmer, C. B. Barrett, and J. Berdegué, 2003a, “The Rise of Supermarkets in Africa, Asia, and Latin America,� American Journal of Agricultural Economics 85(5): 1140–1146. Reardon, T., G. Vrabec, D. Karakas, and C. Fritsch, 2003b, The Rapid Rise of Supermarkets in Croatia: Implications for Farm Sector Development and Agribusiness Competitiveness Programs, Report for USAID, DAI and Michigan State University, September. Rehber, E., 2000, “Vertical Coordination in the Agro-Food Industry and Contract Farming: A Comparative Study of Turkey and the USA,� Food Marketing Policy Center Research Report No. 25, University of Connecticut, February. Roberts, D., 2004, “The Multilateral Governance Framework for Sanitary and Phytosanitary Regulations: Challenges and Prospects,� Paper presented at the World Bank Conference on Food Regulations and Standards: International Dimensions and Trade, Washington DC, January 27. Roland, G. and T. Verdier, 1997, “Transition and the Output Fall,� CEPR Discussion Paper No. 1636. Rozelle, S. and J. Swinnen, 2004, “Success and Failure of Reforms: Insights from Transition Agriculture,� Journal of Economic Literature, forthcoming. 67 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Rylko, D., 2002, “New Agricultural Operators, Input Markets, and Vertical Sector Coordination in Russia and the NIS,� Paper presented at FAO/EBRD Workshop on Investment in Agribusiness, Budapest, April. Rylko, D., S. Matthies, B. Zimbler, and E. Zharikova, 2000, “Using Warehouse Receipts as a Financing Instrument for the Russian Agribusiness Sector,� EBRD, FAO, USAID, Moscow. Sadler, M., 2004, “Comparative Analysis of Cotton Supply Chains in Central Asia,� mimeo, The World Bank. Swinnen J. and H. Gow, 1999, “Agricultural Credit Problems and Policies During the Transition to a Market Economy in Central and Eastern Europe,� Food Policy 21(1): 21–47. Swinnen, J., H. Gow, and I. Maviglia, 2002, “Modest Changes in the West, Radical Reforms in the East, and Government Intervention Everywhere: European Sugar Markets at the Outset of the 21st Century,� in Schmitz, A., T. Spreen, and W. Messina (eds) Sweetener Markets in the 21st Century, Kluwer Academic Publishers, Dordrecht. Swinnen, J. F. M., L. Dries, and K. Macours, 2005, “Transition and Agricultural Labour,� Agricultural Economics 32(1): 15–34. Telser, L. G., 1980, “A Theory of Self-Enforcing Agreements,� Journal of Business 53(1): 27–44. Top Agrar, 2004, “Russische Milchviehhalter bekommen neue Melktechnik,� Top Agrar, May 6. Van Berkum, S., 2004, “Dynamics in Vertical Coordination in the Romanian Dairy Sector,� Report for the World Bank Study on “Dynamics of Vertical Coordination in ECA Agrifood Chains: Implications for Policy and Bank Operations,� LEI-WUR, The Hague/ Wageningen. Van Berkum, S. and J. Bijman, 2004, “The Impact of Foreign Investment on Supply Chains in ECA: Lessons from a Survey of Dutch Agri-food Multinational Companies,� Report for the World Bank Study on “Dynamics of Vertical Coordination in ECA Agrifood Chains: Implications for Policy and Bank Operations,� LEI-WUR, The Hague/ Wageningen. Van der Meer, K. and M. Noordam, 2004, “The Use of Grants to Address Market Failures. A Review of World Bank Rural Development Projects,� Agriculture & Rural Development Discussion Paper, The World Bank, Washington DC. Van der Vorst, J. G. A. J., 2000, Effective Food Supply Chains. Generating, Modelling and Evaluating Supply Chain Scenarios, Wageningen. Von Braun, J., D. Hotchkiss, and M. Immink, 1989, Nontraditional Export Crops in Guatemala: Effects on Production, Income, and Nutrition, Research Report 73. IFPRI, Washington DC. 68 The Dynamics of Vertical Coordination in Agrifood Chains in Eastern Europe and Central Asia Wandel, P., 1999, “Vertical integration in the Russian Agro-Food Sector,� in Wehrheim, P. and J. Von Braun (eds) Russia’s Food Economy: Towards Truly Functioning Markets, forthcoming. Watanabe, S., 1983. “Technological Linkages through Subcontracting in Mexican Industries,� in Watanabe, S. (ed.) Technology Marketing and Industrialization: Linkages between Small and Large Enterprises, Macmillan, New Delhi. White, J. and M. Gorton, 2004, Vertical Coordination in ECA Agrifood Chains as an Engine of Private Sector Development: Implications for Policy and Bank Operations, World Bank, Washington DC. Wilkin, J., M. Juchniewicz, and D. Milczarek, 2004, Case Study: Poland. Regoverning Markets Project, Unpublished manuscript, Warsaw Uni- versity. Williamson, O. E., 1985, The Economic Institutions of Capitalism. Free Press, New York. World Bank, 2002, World Development Report 2002: Building Institutions for Markets, Washington DC. World Bank, 2004, Agricultural Investment Sourcebook, Agriculture and Rural Development, Washington DC. World Bank, 2005, World Development Report on “Improving the Investment Climate for Growth and Poverty Reduction,� Washington DC, forth- coming. 69 ANNEX 1 Importance of Contracting in ECA Agrifood Chains There is no systematic evidence on the emergence of various forms of VC. But various sources do provide a set of information that allows drawing some conclusions. We �rst review evidence from cross-sectoral studies from several countries, both in CEE and CIS. Afterwards we look more carefully into specific sectors, such as dairy, cotton, grain, and others. The sector-speci�c approaches are important because the extent and nature of VC will differ by sector, reflecting commodity—and chain—speci�c characteristics. Tables A1.1 and A1.2 present data on contracting on farms in the Czech Republic, Slovakia, Hungary, Bulgaria, and Romania based on rep- resentative farm surveys in these countries in the second half of the 1990s (1997 for Hungary and Bulgaria, and 1999 for Czech Republic and Slovakia).38 The data show that Contracting was widespread among farming companies in all coun- tries. Virtually all corporate farms were selling farm output on contract in Hungary, Czech Republic, and Slovakia. Notice that in these coun- tries corporate farms produce a large share of the output. Contracting was much less widespread among individual farms, although there is a difference among individual farms. In statistics and surveys individual farms typically include both subsistence farms and more commercially oriented family farms. The Czech data allow to 38. Boger et al. (2001) find that in a survey of pork farms in two regions in Poland in 1999 about 13% of the farms had a formal contract and 70% had an oral contract; only 16% had no contract. Two-thirds of the oral contracts only speci�ed time, quantity, and means of delivery, while one-third included quality and price premium speci�cations. 71 Annex 1 Importance of Contracting in ECA Agrifood Chains disaggregate the “individual farms� in “registered� (more commercially oriented) and “nonregistered� (more subsistence oriented) farms. The data show that contracting is much more widespread among registered family farms (46%) than among nonregistered farms. Contracting of farming companies was less developed in Bulgaria in 1997 than in the Czech Republic, Slovakia (1999), or Hungary (1997). This probably reflects progress in the restructuring and privatization of the agrifood industry, which had advanced the fastest in Hungary, and the slowest in Bulgaria in that period. TABLE A1.1. Share of Farms Selling on Contract (as % of total) Czech Type of contract NRIF RIF Slovak Hungary Bulgaria Individual farms Selling crop products on contract 4 37 29 8 5 Selling livestock products on contract 1 13 4 10 3 Selling animals on contract 2 7 6 na na Selling on contract 5 46 35 17 7 Corporate farms Selling crop products on contract 79 82 86 42 Selling livestock products on contract 73 83 59 23 Selling animals on contract 49 77 na na Selling on contract 96 98 94 43 RIF, registered individual farms; NRIF, nonregistered individual farms. Source: Leuven ACE datasets. TABLE A1.2. Share of Czech Farms with Inputs Provided as Part of a Sale Contract (%)* Individual farms Type of input Non-reg. Registered Corporate farms Seeds 0 22 62 Feed grain 0 0 13 Hay or other fodder 0 0 1 Industrial fertilizer 0 21 66 Chemicals 0 17 63 Concentrated feed 0 17 72 Fuel for machinery 0 0 44 Irrigated water 0 25 Other inputs 15 44 *Share (%) of farms using that input in 1999. Source: Leuven ACE datasets. 72 Annex 1 Importance of Contracting in ECA Agrifood Chains Only for the Czech Republic do we have data on input provision for as part of the contract. The data show that a majority of corporate farms received seeds, fertilizer, chemicals, and feed as part of the contracts. The share among registered individual farms was only around 20%. Nonregistered farms did not receive inputs. Another set of evidence comes from a study by White and Gorton (2004), who conducted surveys of 53 agribusiness enterprise executives from Armenia, Georgia, Moldova, Russia, and Ukraine that have made recent capital investments in the agrifood sector. The results of this survey, shown in Table A1.3, show that contracting has been increasing over the period 1997–2003. More speci�cally: Contracting between processors and farmers increased signi�cantly between 1997 and 2003. Companies that used contracts with suppli- ers grew from slightly more than one-third in 1997 to almost three- fourths by 2003. Contracting is signi�cantly higher for larger farms (around 75% of the companies contract with large farms) but contracting with small farms grew as well over the 1997–2003 period, and around 50% of them is now contracting with small farms. The use of spot markets is stagnating around 50% while contracting and also the use of other agents, such as intermediaries and traders, is becoming more prevalent. The study also indicates a strong growth in extreme versions of VC such as full ownership. In fact, the largest observed increase during this period was in the number of enterprises directly engaged in farming, TABLE A1.3. Supply Relationships in Sourcing Raw Materials 1997–2003 (% of companies) Relationship 1997 1999 2001 2003 Spot markets With all farmers 27.2 43.5 47.1 50.0 With small farmers 25.0 41.3 44.2 47.2 With larger farmers 15.6 25.5 25.5 23.1 Contracts With all farmers 41.3 61.7 73.1 77.4 With small farmers 36.2 43.8 46.2 49.1 With larger farmers 37.0 58.3 69.2 73.6 Own farms 6.4 8.3 17.8 26.4 Other agents 16.7 28.6 46.2 49.1 Source: White and Gorton (2004). 73 Annex 1 Importance of Contracting in ECA Agrifood Chains which increased from 3 (6%) to 14 (26%) of the �rms—with most of this vertical ownership integration occurring recently. The study also shows that the use of supplier support measures was introduced mainly after 1999, but has since been spreading to a larger number of companies and farms. Table A1.4 shows the different types of contract support measures being offered by the surveyed agribusinesses to their suppliers. These measures are listed in descending order of the fre- quency with which they are offered. Some conclusions are given below: Monetary credit, prompt payments, transportation, physical inputs, and quality control are the most commonly offered forms of support. Investment loans from processors to farmers, harvesting, and veteri- nary support are the most infrequently offered measures. When they are offered, virtually all farms bene�t from prompt payments and guaranteed prices. A great majority of farms also take advantage of quality control, agronomic support, and market access support when it is available. Over 40% of processors in their sample offer credit to at least some of the farms that supply them; and 36% offer inputs, as per 2003 data. TABLE A1.4. Importance of Supplier Support Measures Farms Firms imposing receiving this Farms receiving minimum farm % of support in this support in size for this Support measure Sample �rst year (%) current year (%) measure (%) Credit 43.4 39.8 50.9 60.8 Prompt payments 41.5 90.0 87.3 0.0 Transportation 39.6 67.5 76.2 45.0 Physical inputs 36.0 50.2 52.7 61.1 Quality control 34.0 75.5 78.3 16.7 Guaranteed prices 24.5 86.7 91.7 14.3 Agronomic Support 20.8 78.4 81.4 10.0 Farm loan guarantees 20.8 7.0 15.1 27.3 Machinery 16.9 10.4 22.8 66.6 Specialist storage 13.2 34.7 27.3 28.6 Business/�nancial management 11.3 45.8 47.5 50.0 Market access 11.3 68.3 69.7 0.0 Veterinary support 9.4 58.0 66.0 40.0 Harvest/handling 9.4 37.2 13.1 60.0 Investment loans 5.7 4.0 0.3 66.7 Average 57.0 60.7 34.0 Source: White and Gorton (2004). 74 Annex 1 Importance of Contracting in ECA Agrifood Chains The percentage of farms that receive a speci�c form of support has tended to increase after the �rst year the support measure is offered. This is true for almost every form of offered support. Finally, Tables A1.5 and A1.6 summarize a series of case studies dis- cussed in greater detail in Foster (1999), Gow and Swinnen (2001), Dries (2004), and Van Berkum and Bijman (2004). The tables indicate that the introduction of farm-assistance programs by agribusiness companies is a common phenomenon across countries, and increasingly important in sectors, such as the dairy sector (see also the discussion of the dairy sec- tor �ndings in the main report). 75 Annex 1 76 TABLE A1.5. Vertical Contracting and Support to Producers by Agribusinesses in ECA and CIS Case number Foreign investor Foreign partner Local �rm or partners Activity Countries of investment Producer type Importance of Contracting in ECA Agrifood Chains 1 UK/Fr sugar processor – Local sugar processor Sugar production Hungary, Czech, Slovakia Sugar beets 2 UK sugar processor – local sugar processor Sugar production Poland Sugar beets 3 Belgian beer producers – Local brewery Malting Romania, Hungary Barley 4 Belgian malter – Local malting plant Malting Croatia Barley 5 European seed merchant – Foreign food co. Seed merchandising Central and Eastern Europe Crops 6 – – Oilseed producer Vegetable oil production Slovakia Rape and sunflower 7 US combine producer Ag. equipm. dealer – Equip. leasing and sale Ukraine Grain and oilseed 8 Swiss confectionary co. dairy plant – Confectionary Russia Dairy producers 9 US dairy processor Dairy plants – Dairy products Ukraine, Moldova, Kazakh Dairy producers 10 US food catering �rm Food catering – Fast food service Russia, Ukraine Vegetable producers 11 US nut processor Facilities owners – Walnut processing Moldova Walnut growers 12 US agri-chemical mnf. Ag. input distribut. – Fertiliz. and pesticide sales Ukraine, Russia Crop producers Case number Types of support 1 2 3 4 5 6 7 8 9 10 11 12 Training Y Y Y Y Y Y Y Y Y Y Y Y Credit access programs Y Y Y Y Y Y Y N N N Y Y Input provision Y Y Y Y Y Y Y Y Y Y Y Y Machinery procurement Y Y N N Y N N N Y Y Y Y Agronomical support Y Y Y Y Y Y Y Y Y Y Y Y Veterinary support N/A N/A N/A N/A N/A N/A N/A Y Y N/A N/A N/A Harvest and handling support Y Y Y Y Y Y Y Y Y Y Y Y Quality control Y Y Y Y Y Y Y Y Y Y Y Y Transportation Y Y Y Y Y Y Y Y Y Y Y Y Specialized storage Y Y Y Y Y Y Y Y Y Y Y Y Bus. and �n. mgmt support Y Y Y Y Y Y Y Y Y Y Y Y Market access Y Y Y Y Y Y Y Y Y Y Y Y Annex 1 Timely payments Y Y Y Y Y Y Y Y Y Y Y Y Source: Based on Foster (1999) and Gow and Swinnen (2000). 77 Importance of Contracting in ECA Agrifood Chains Annex 1 Importance of Contracting in ECA Agrifood Chains TABLE A1.6. Farm-Assistance Programs Offered by Dairy Companies Company Credit— Credit— Input Extension Veterinary Bank loan name* speci�c general supply service service guarantee Poland** Mlekpol Y Y Y N Y Mleczarnia N Y N N Y Kurpie Y Y Y N Y Mazowsze Y Y Y N N ICC Paslek Y Y Y N Y Warmia Dairy Y Y Y Y Y Bulgaria Merone Y(2000) N Y(????) Y(1992) N N Fama Y(1994) N Y(1994) N N Y(once) Mlekimex Y(1997) Y(1998) Y(1997) Y(1999) Y(1997) Y(1998) Danone Y(1997) N Y(1998) Y(2000) Y(1995) Y(1999) Iotovi N N Y(1995) N N Y(1995) Milky World Y(1999) Y(2000) Y(1999) Y(1999) N Y(1999) Markelli Y(1999) N Y(1998) N N N Mandra Obnova Y(1998) N Y(2000) Y(2000) N N Meggle Y(2001) N Y(2001) Y(2001) N N PRL N N N Y(2002) N N Serdika 90 Y(1997) N Y(1997) Y(1997) N N Slovakia Liptovska Y(2000) N N Y(1994) N N Mliekospol Y(1999) N N Y(1992) Y(1992) Y(1992) Rajo Y(2001) N Y/N Y(1992) N N Levicka Y(1998) N Y(1998) Y(0000) N Y(1998) Tatranska Y(2001) N Y(2000) Y(0000) N N Nutricia Dairy Y(2000) N N N N Y(2000) Romania Danone Y Y Y Y Friesland Y Y Y Y Promilch Y Y Y Y Rabaul N Y Y N *Either the company provides inputs and the farmer pays back later, or the company offers forward credit, which the farmer uses to buy inputs. *In Poland, no distinction is made between credit for dairy-speci�c investments and general investments. Farm-level evidence shows that the dairy companies mainly support dairy-speci�c investments. Source: Dries and Swinnen (2004) and Van Berkum (2004). 78 ANNEX 2 Variations in Vertical Coordination Models Vertical coordination, both in terms of its type and extent, may differ by commodity, country, company strategy, etc. Here we review a series of structural differences in observed forms of VC which are important to understand. Various Companies Can Initiate Vertical Coordination The discussion in the report focused mostly on companies purchasing farm output (processors, traders, supermarkets) as initiators of VC. However, like food processing companies, agribusiness companies trying to sell farm inputs, such as seed, machinery, fertilizer, etc. face important problems. They are confronted with farms that cannot pay for their inputs because they are not paid in time, or who do not have output markets for their products. They ask “how can we ensure that we can sell our products and get paid for them?� The answer is: by assisting our customers (the farms) in �nding market outlets for their products. To ensure payments by farms, input suppliers have engaged in a variety of, sometimes quite unconventional, forms of contracting. For example, in one case in Ukraine, described by Foster (1999), a US farm equipment manufacturer partnered with local farm equipment distributors to sell combines and tractors to farms. The US company provided a large inven- tory of equipment for the distributor to sell on credit and service. The farms received the equipment for a 25% down payment (in cash or kind) and after three additional payments received ownership. The equipment dealer, to ensure payment, was given the rights to certain agreed upon grain area as repayment; in addition, the equipment dealer was given the rights to harvest, transport, store, and sell the grain himself. Further, as part of the arrangement, the equipment dealer and the transactions themselves pro- vided the farms with training and skill development. 79 Annex 2 Variations in Vertical Coordination Models Sometimes different models develop because processors themselves do not have access to �nance. In some cases this induced VC driven by �nan- cial groups, such as by Rabobank international in Central Europe, or by financial groups in Russia. Another example is from the Ukraine oilseed sector in the 1990s. There, farmers preferred to sell seed to trading �rms through barter contracts against inputs, such as agricultural machinery and fuel oil, rather than to crushers. Because processors had poor access to credit, traders, equipment suppliers, and even banks procured seeds for the oilseed crushing factories. Many farmers also retained ownership of their product, leaving the crushing plants in their role of subcontractors, who charged a tolling fee for processing seeds. In 1999, around 80% of the crushers’ throughput of sunflower seeds was based on a tolling basis. Under the tolling system, crushers received 13–20% of the oilseeds deliv- ered to them as their toll payment for crushing. The oil obtained from the rest was returned to the owners (equipment suppliers, farmers, or traders), who sold the oil either in the domestic market (competing with the crush- ers) or exported it (EBRD, 2002). Alternatively, if domestic sources of �nance are lacking, with tradable commodities foreign traders may provide the necessary �nance for the whole chain. For example, in the Ukrainian oilseed sector, multinational traders purchased oil seeds locally from Ukrainian farmers and elevators and then had the seeds crushed by local crushers, which purchased part of the oil from the multinational. A similar development occurred in the Kazak cotton chain, where contracting between domestic private cotton gins and international cotton traders provides the gins with �nancial means to pre�nance the farms’ inputs (see Sadler, 2004).39 Complex Models In several cases, we �nd VC with several companies involved in the con- tract. For example, VC sometimes includes both input supplying compa- nies and processing or trading companies. VC then implies contracting “around� the suppliers. Examples are agricultural pesticide companies in Bulgaria who, in order to ensure payments for inputs delivered to farms, collaborate with local grain trading companies to market the grains of the farms to which they delivered inputs. First, triangular structures are used by processors and retailers to bring in �nancial institutions in the contract. Examples of this are processor or 39. Interestingly, the resulting ownership structure here is the opposite to that in the United States or Australia, as the Central Asian farms, mostly small farms that have limited access to �nance, sell the cotton to gins while in the United States and Australia farms maintain ownership of the cotton throughout the chain, and gins are paid as service providers. 80 Annex 2 Variations in Vertical Coordination Models BOX A2.1 Triangular Structures Retail/processing co. Farm Bank Example: Retailer/processor provides loan guarantees for bank loans to suppliers retailers who provide loan guarantees to �nancial institutions for loans to their suppliers. The underwriting is for specific loans, related to the contract, and restricted for contracting suppliers. Box A2.1 illustrates such triangular structures that are fairly common. They are imple- mented by sugar processors in Slovakia (see Gow et al., 2000), by retailers in Croatia for F&V supplier investments in greenhouses and irrigation (see Reardon et al., 2003b), and by dairy processors in several countries (see Dries and Swinnen, 2004). Second, an even more complex example of vertical integration “around� the farms, where both input suppliers and processors are included, is the use of so-called special purpose vehicles (SPVs). An SPV is a stand-alone com- pany jointly owned by the processor, input providers, and a project �nanced by the bank. The contract between the SPV and the farms includes all provisions on output, inputs, and credit (Box A2.2). One example of this was implemented by an international �nancial institution specialized in agribusiness and food supply chain �nancing, in Hungary, in collabo- ration with local agribusiness partners. An important advantage of such an institution is that the partners in the SPV now share the risk of contract breach. When a processing com- pany by itself implements input and investment facilitation programs, the processor carries the entire risk of farms’ breaching contracts, although both the input suppliers and the �nancial institutions bene�t from these contract innovations. Institutions such as SPVs allow sharing 81 Annex 2 Variations in Vertical Coordination Models BOX A2.2 Rabobank’s Special Purpose Vehicle Processing co. SPV Farm Bank Input supplier Example: Foreign investor is financial institution and the special purpose vehicle (SPV) is used to distribute risk equally among partners. Example: Group of small farmers formed cooperative to participate in SPV. of the risk between various agents, and, hence, will stimulate investments by companies who otherwise may be deterred by the risk. In some cases, such structures have developed with farmer participa- tion. For example, Gow and Swinnen (2001) report that in Eastern Hungary a group of sheep farmers set up a producers’ cooperative through which they participated in an SPV-like joint company. This gave them more bargaining power vis-à-vis the other partners, much in the same line as a marketing or input purchasing co-operative does in a normal environment. Another variation with farmer participation is from the Czech Republic, where modern retailer investments and the associated increase in standards and supplier requirements have induced the creation of pro- ducer marketing organizations (PMOs). The main reason for their estab- lishment was to gather suf�cient quantity and product variety to satisfy the requirements of large supermarkets chains (EBRD/FAO, 2004). However, in addition to reducing transaction costs and increasing bar- gaining power, the PMO provided important services to their member farms, such as extension services, storage, sorting and packaging, and information. Furthermore, PMOs enhanced access to inputs and to bank loans for their members by providing guarantees to input suppliers and banks (Dries and Van Kerckhove, 2004). 82 Annex 2 Variations in Vertical Coordination Models Another example of a triangular structure with a specially designed institution is the collaboration between the Russian dairy processor Wimm Bill Dann (WBD) and the Swedish dairy equipment seller De Laval in the region of Nischnyj-Nowgorod for the modernization project “Milkrivers.� This project plans to modernize the milking equipment on the farms through leasing contracts. Practically all dairy farmers in the area have to modernize and upgrade their equipment and facilities, but only a few have the �nancial resources to do so (see Box A2.3). The pro- gram allows dairy farms to lease milking equipment. They have to cover about 20–30% of the costs themselves and receive the equipment based on a 3–5 year leasing basis. The principal balance can be paid off by the farmers through delivering the raw milk to one of the dairy processors owned by WBD. The main condition in order to take part is the compli- ance with WBD quality standards. The equipment is being delivered by De Laval. The project costs are shared by WBD and De Laval. BOX A2.3 Leasing of Dairy Equipment by Joint Leasing Project (Wimm Bill Dann (WBD) and De Laval) Processor Project Farm Equipment seller Source: Top Agrar, 2004 Another interesting VC model is discussed in Van Berkum’s report on the developments in the Romanian dairy sector (see Box A2.4). He reports the case of Danone, an international food company, which has developed an expensive �nance scheme for farms, including a triangular structure with input suppliers. But Danone goes further than most other 83 Annex 2 Variations in Vertical Coordination Models BOX A2.4 How Food Processors Become Financial Institutions: Loans and Collateral in Contracting with Danone in Romania Danone has made prefinancing inputs a corner stone of its farmer-development program in Romania. The company supports farmers who are aiming at improving their business through investments in inputs like (spare parts of) field machinery and milk installations, and through purchases of feed com- pounds (concentrates), milk powder (as cattle feed), and detergents (of milking equipment). A farmer may apply for these inputs only after he has delivered good quality milk to Danone for at least 6 months, and when he has a certain minimum size. Furthermore, together with the requests for inputs one needs to submit a business development plan. If Danone accepts the plan, the com- pany and the farmer make a contract. Dependent on the investments to be made, Danone agrees to buy the farmer’s estimated milk production 1, 2, or a maximum of 3 years in advance and pays the farmer a sum of money that he is supposed to invest in the inputs agreed on. Danone normally takes the farm house and the land as guarantee for nondeliverance of milk or breach of con- tract. Danone then provides a security to input supplying companies, who are then willing the deliver the inputs to the farmers. The contract is signed in a notary’s deed. Danone offers this assistance only to medium and larger farms, with a minimum of 20 cows. In the Romanian context, which is dominated by two-cow farms, this severely limits the impact of the program. Processing Farm Bank Input supplier Source: Van Berkum (2004). 84 Annex 2 Variations in Vertical Coordination Models companies as it takes collateral itself from farms for medium-term invest- ments for which it provides loans. Multiple Stages Another example of complex VC is where the VC includes multiple stages, sometimes more than three, which complicates the enforcement prob- lems. This is the case, for example, with brewery investments or some cot- ton chains (see Box A2.5). Interbrew, a multinational brewery holding that made extensive invest- ments in ECA countries, has introduced contracts with barley farms via malting companies and international consultants. In this strategy, it col- laborated with foreign malting companies to ensure high-quality malt, to be arranged from imported malt or barley if possible and otherwise from domestic barley producers. For this, Interbrew put up an amount of capi- tal to pre�nance import and breeding of high-quality seeds, training of bar- ley specialists to supervise the barley growing via international consultants, etc. In Russia, this VC strategy included international experts managing the actual barley farms. BOX A2.5 Four-Stage Models: Brewery Investments Brewing co. s Core business = brewing s Forced to vertically integrate all the way to seed supply to ensure quality malt- Malting co. barley-seed s General strategy applied differently in countries b/c of local conditions Barley farm s Bring in co-foreign investors to assist with non-core activities and set up farm assistance programs Seed supply s Programs interlink markets An alternative version of this model is where the malting company is the driving investor, often in collaboration with brewing companies. For example, Soufflet, a large international malting company, signed a framework agreement with EBRD for investments in several ECA countries where it contracts with barley farms to provide high-quality 85 Annex 2 Variations in Vertical Coordination Models barley. These strategies are linked to EBRD financial investments by international brewers, such as EFES, who use Soufflet’s malt, e.g., in Russia. As should be clear by now, VC is not a black or white issue. It is, instead, a choice of a set of institutional arrangements that vary between a spot market arrangement and full ownership of the supplier by the processor (or vice versa). The examples discussed so far are of an intermediate form whereby there is more coordination than a spot market, but less integra- tion than full ownership. In most cases it implies some form of output contracting in one direc- tion and technology, input, and marketing assistance the other direction. However, in some cases the vertical integration has gone much further. Full Vertical Integration: Ownership The most extreme version is when processing or trading companies actu- ally take ownership over the farm and fully integrate it in their company. This is what happened in some cases in some of the CIS countries. For example, both in Russia and in Kazakhstan VC took extreme forms as traders and processing companies took over ownership of farms. In the northern wheat belt of Kazakhstan, many bankrupt farming cooperatives have been taken over by grain trading and investment companies after 1998, when the government imposed bankruptcy pro- ceedings. Box A2.6 presents several of these cases. Grain companies are now one of the dominant types of farming in the north of the coun- try. These are companies that have been acquired by specialized grain production and trading companies with substantial financial resour- ces, usually accompanied by vertical interests in the grain marketing process. Examples of this organization mode are shown in Case nos. 2 (Bisco-Trade) and 3 (Zhambul farm). Food companies are also invest- ing in farming: Case no. 1 (Arai JSC) is an example of this category. In several cases the farms that are managed by these integrated companies are very large; sometimes more than 100,000 hectares of farm land (see, e.g., Case no. 2). Chain-Based Financial Instruments Several of these institutional innovations that we have reviewed so far can be considered as nontraditional methods for providing �nancing to farms, which have developed to overcome speci�c constraints in transi- tional economies. Contracts and incentive structures have been designed to overcome working and investment capital constraints, a weak collat- eral base, information and enforcement problems, etc. 86 Annex 2 Variations in Vertical Coordination Models BOX A2.6 Integrated Grain Companies in Kazakhstan Case No. 1: Perelyevsky Cooperative Farm incurred debts since 1992 to the budget (tax, tax arrears penalties, and pension contributions) and to the Arai Joint Stock Company (JSC) food complex, which had contracted the entity to supply wheat. The livestock population fell from 7800 in 1986 to 1700 in 1996. After bankruptcy, the farm was auctioned in 1997 and Arai JSC acquired all assets. The farm now forms a branch of Arai JSC, an integrated complex of companies comprising grain, meat, and milk production and the production of alcohol, soft drinks, and beer. Land shares have been formally transferred to Arai JSC by all former owners in return for a commitment for regular wage income for all employees. Conditions improved radically after the takeover with an influx of working capital and investment. The full former labor force has been retained on the farm. The cattle herd has been rebuilt to 4200, and the com- pany produces wheat on 4000 hectares of land.40 Case No. 2: Bisco-Trade is an investment grain company which has taken over several bankrupt farms. The acquisitions followed poor experiences with contract farming as contracts could not be enforced through the courts. The company could chose from a large number of bankrupt farms which the company could choose among as most were in desperate conditions. It currently owns 11 farms covering 220,000 hectares. In 1999, the company cultivated 120,000 hectares of wheat and barley, with the remainder left fallow as part of the rotation. Case No. 3: Ivolga ISC, a large grain trading company owns 23 farms, most of them acquired after farm bankruptcies. The company’s farms produce grain, veg- etables, and potatoes. The company owns 12 grain terminals and elevators. Case No. 4: Agrocenter was one of 20–30 companies contracting farms in the wheat growing belt, which together contracted for one-half of all the cultivated wheat area in Kazakhstan. In 1998, the company contracted 11 former state farms totaling about 400,000 hectares. The company provided spare parts, chemicals, and seeds and met the wage bill for an agreed employment level. In 1999, these mecha- nisms led to large outstanding debts, when farms had accepted inputs on a barter basis but were hit by the drought and unable to deliver to pay off their input loans. The dif�culties experienced with contract and debt management following the 1999 drought caused Agrocenter to withdraw from contract farming in 1999. Source: Gray (2000). 40. The company has discontinued financing of the basic social services that are legally the responsibility of the raion authorities (notably teachers’ salaries). However, it uses its leverage as a major taxpayer to the raion to obtain assurances that raion obligations are in fact met (unlike in most neighboring entities). The company has undertaken a major investment is gasi�cation of the entity, including housing and heating for the school and other facilities. 87 Annex 2 Variations in Vertical Coordination Models In this last part of this section, we present three additional ways of how processors, and in one case the government, have tried to overcome these problems in practice.41 Lending to farmers without collateral through associations In Romania, the farmers association ISPA created a joint venture with a pri- vate milk processor ProMilch in Romania, and became majority share- holder. ISPA provides loans to small farmers with the assistance of a Dutch fund—received from Rabobank on very attractive terms. It provides small loans to farmers who want to invest in animals, (re-)construction of sta- bles and/or equipment. Farmers qualify for a loan through an interview in which they have to indicate their business plan. An average loan is around Euro 400, with a maximum of Euro 2000. ISPA loans are to be repaid after a 6–18 month grace period for animals, and 4-year grace period for con- struction investments. Farmers do not have to provide any collateral; the milk delivered is con- sidered the “collateral.� Eligibility criteria for loans include several ele- ments. First, the farmer needs to have a durable relation with ISPA. In practice, ISPA requires a delivery period of at least 6 months but preferably 1 year. Important is that a farmer uses an appropriate fodder base at his farm and agrees upon a commitment for further expanding the farm. ISPA personnel, who generally have a close contact with all individual members, need to con�rm the assessment on eligibility. The requirements are, how- ever, not too strict and subject to ISPA staff assessments. Trust and relia- bility are important. ISPA deals with the default risk by having a solidary liability of both the loan bene�ciary and the milk collection centre staff who guarantee for the reliability of the loanee (Van Berkum, 2004). 41. An interesting new instrument with potential importance for transition coun- tries has been developed in Brazil. The Cédula de Produto Rural (CPR) is a tradable Production Certi�cate offered by the Banco do Brasil to Brazilian farm producers, cooperatives, and cattle ranchers. Producers can sell the CPRs they have been issued to commodity buyers, by trading them in domestic commodity exchanges or by sell- ing them through e-commerce with other Banco do Brasil customers. For producers who choose to trade their CPRs, they act like forward contracts. The producer grant- ing the CPR receives cash immediately from the Banco do Brasil, in an amount deter- mined by prices on futures markets or by price averaging, in exchange for a promise to pay the amount, through the delivery of physical produce or money, to the buyer of the CPR at a �xed date (when the contract “matures�) sometime within 1 year from the date of issuance, at the end of the current harvesting or processing cycle. The value of the CPRs depends on the type of commodity, the phase of production, and indices of historical production and technical capacity. Today, the Banco do Brasil offers a majority of all credit �nancing in Brazil. In 2004, 40% of the credit it issued was granted to agribusiness companies. A signi�cant and fast-growing portion of this credit—3 billion Reals ($965 million) in 2004—was issued as CPRs. 88 Annex 2 Variations in Vertical Coordination Models Leasing Leasing can be seen as a specific kind of financial contracting between input suppliers and farms. The system was developed in the ECA region to overcome collateral problems. Often leasing was introduced by agri- business companies supplying equipment and farm machinery to sell their products to farms that could not come up with the necessary col- lateral for loans. Even when the companies selling the equipment had sufficient financial means to introduce the system, problems with enforcement may prevent the system from developing. If the company cannot reclaim the equipment in case of nonpayment by the farm it will not be willing to provide the leasing contract. In many ECA countries this is a serious constraint. There is little information on how important private sector leasing has become in the agrifood chains, since there are no formal statistics on this. Ad hoc discussions and interviews with agribusiness companies suggest that it is being used in many ECA countries, and it appears as an instru- ment with large potential in transition environments, as was illustrated by the Russian dairy case we discussed earlier (see Box A2.3).42 Reverse factoring Innovations are taking place in �nancing supply chains in Western coun- tries, with potentially important spillovers for transition countries. For example, Rabobank International has developed a financial instrument (reverse factoring, RF), which is designed to finance the supply chain (supplier-retailer/processor) and to make use of the better credit ratio of 42. Also in leasing, the Russian government has taken an active role. Rosagrosnab is a large government-designed leasing program for agricultural equipment, which had major problems and was revised later. Government funds were allocated and managed by Rosagrosnab, which entered into leasing agreements with the oblast- level suborganizations, which in turn entered into subleasing agreements with local (rayon) level agrosnabs. The latter leased equipment to farm producers. Farms were asked for upfront cash payments of 30% of the value of the leased equipment, fol- lowed by 3–5 years of additional payments, ending in a transfer of ownership. However, the heavy handed government involvement was problematic. The large number of intermediate organizations resulted in high prices for farmers. The cen- tralization of purchasing and allocation decisions also meant that farmers had very limited choices as to the type or manufacturer of the equipment they could lease. Lack of competition meant that machine producers and leasers had no incentives to im- prove their ef�ciency or quality standards. Signi�cant problems also existed in the ability of the leasing companies to repossess equipment in cases of nonpayment. Often farmers did not realize that they would not assume ownership until the end of the lease, and in other cases where local of�cials were able to block attempts by leas- ing �rms to repossess their equipment. As a consequence, the leasing program was modi�ed in 2001 (Csaki et al., 2002). 89 Annex 2 Variations in Vertical Coordination Models the retailer/processor to address the working capital needs of suppliers. For suppliers under contract, or short listed by the retailer or processor, the bank prepays the supplier to address their working capital constraints, quasiguaranteed by promissory notes provided by the retailer/processor based on delivery contracts. This system is thought to be particularly useful with new requirements being imposed on the supply chains, related to traceability/transparency, fair trade, biosecurity, etc., to address constraints imposed on suppliers, which they may not be able to address by themselves. In addition, the instrument could be seen as a risk-mitigating instrument as the risks are being distributed between more agents in the chain. These issues of over- coming �nancial constraints and distributing risk among partners in the chain are very similar to the issues we discussed in the “complex contracts� section in this report. While Rabobank International is just introducing this system in Western Europe to selected companies, the principle of the financing scheme seems particularly interesting for transition countries, where processors and retailers are often in a much better position to get access to bank loans than suppliers are, for a number of reasons including collateral base, loan history, lower transaction costs, transparency, etc. Currently, a study is being implemented by the World Bank to look into this and related schemes for financing farms and small suppliers in developing and transition countries. Warehouse receipt systems Warehouse receipt systems (WHR) have proven to be a successful instru- ment in providing �nance in the supply chains for transition countries, in particular for storable commodities such as grains. Such systems are now actively and successfully operating in several ECA countries. The bene�ts can be large. For example, a study by Rylko et al. (2000) estimated that the introduction of a WHR system in Russia could increase the liquidity in the system by approximately 2 billion US$, which was approximately the same size as the largest federal credit program in Russia. Interestingly, informal WHR systems have developed without waiting for the necessary government regulations. For example, in Ukraine in 2002, banks were making contracts with elevators who held grain and oilseeds as collateral for loans taken by farmers although at that time there was not yet a legal basis for this as a WHR system in place, which limited the ability of farmers and crushers to use stored seed as collateral (EBRD/FAO, 2002). Similarly in Russia, a “quasi-WHR system� has developed and is widely used to collateralize inventories (e.g., of grains and oilseeds) in a number 90 Annex 2 Variations in Vertical Coordination Models of transactions (Csaki et al., 2002).43 These quasi-WHRs were commonly used in preexport or preimport transactions and in domestic agribusiness transactions. These transactions would involve inspection by a bank or international organization that issued credits and assumed responsibil- ity for losses until the commodity was transferred. Domestically, the use of stored commodities as collateral in credit agreements was somewhat less sophisticated. Loans were often extended based on political or per- sonal relationships. In the absence of a regulatory system, there are important enforcement problems. In both Ukraine and Russia, there were problems in enforcing informal storage contracts with elevators or warehouses. In Ukraine, farms claimed they were cheated by elevators in both quantity and qual- ity of their deliveries and oilseed crushers had problems retrieving their seeds from the elevators (EBRD, 2002). In Russia, banks use their own collateral inspectors to physically inspect the warehouses and may post their own security guards at the warehouses. In times of high uncertainty, such as during and immediately after the 1998 crisis, the banks insisted on higher collateral requirements and combining warehouse receipts with other collateral, unless loans were guaranteed by the government. Often the transactions were based on warehouse acceptance receipts from the Soviet times (“form 13�), which was subject to manipulation and fraud. Yet, the fact that such systems developed and spread widely despite these imperfections, indicates the huge gains in ef�ciency from institu- tions supporting exchange and the large potential for better regulated sys- tems (Csaki et al., 2002). Russian-Style Vertical Coordination Agroholdings 44 After the 1998 Russian �nancial crisis, several factors coincided to cause profound changes in the vertical organization of the Russian agrifood system. First, the dramatic devaluation of the ruble profoundly changed the competitive position of domestic production compared to imports. Second, trading and processing companies faced an acute shortage of raw materials as imports became very expensive, while local supplies were either not available or of low quality. Third, growing social tensions put pressure on local and regional authorities to look for a solution to the 43. The devaluation of the ruble following the 1998 Russian �nancial crisis re- sulted in signi�cant losses to participants who held quantities of stored commodities through these programs, temporarily limiting them. However, the programs began to re-emerge in 1999–2000 with even greater domestic bank participation. 44. See Gataulina et al. (2004) and Rylko (2002) for more details. 91 Annex 2 Variations in Vertical Coordination Models continued crisis in agriculture and rural areas. Fourth, due to changed macroeconomic incentives and exchange rate realignments, a number of Russian nonagricultural companies became interested in investing in the food industry. In a number of regions, local authorities encouraged Russian compa- nies to invest in the agrifood system by offering privileges and guarantees. A number of large companies, including large industrial conglomerates such as the Stoylensk Ore Mining and Processing Plant, “Norilsk Nickel� and Gazprom, invested in the food industry and agriculture through ver- tically integrated business models. As a consequence of this process, the largest industrial holding, Gazprom, became the largest agricultural holding as well, with 91 agri- cultural enterprises in its structure controlled by 25 Gazprom daughter holdings. Some of these daughters were unpro�table because they con- trolled mostly bankrupt agricultural enterprises that held large debts. The vertical integration process is most active in the Belgorod and Orel regions. In Belgorod, more than two-thirds of all agricultural enterprises are part of such integrated companies. By 2002, these companies used almost 80% of the land while employing 25% of the farm workers.45 In the Orel region, the three largest holdings used 41% of all arable land and employed 25% of all people employed in agriculture. The total share of vertically integrated holdings in resource use and regional output is around 50%. To illustrate the size and scope of the holdings, Box A2.7 lists several cases. Benefits offered by local and regional authorities included budget �nancing and cheap credit in the Orel region and land assets, acquired by forcing landowners to cede land shares as capital in new companies, in the Belgorod region. The �rst holding in Belgorod was created in December 2000 after the Governor of the region “asked� the head of the Stoylensk Ore Mining and Processing Plant to invest in agriculture. Land owners, through the exchange of land shares that they used as capital,46 owned 49% of the company and investors owned the remaining 51%. In Orel, the regional governor, an influential politician, also played a key role in launching the process of vertical integration. His influence allowed companies to secure �nancial resources and external investment against securities of the local authorities. 45. Compared to an average of less than 25% in other regions of the Central Black Earth region. 46. Thus, the new company de facto owns all the land. 92 Annex 2 Variations in Vertical Coordination Models BOX A2.7 Agroholdings in Russia Case 1: Orel Niva holding Orel Niva controls 337,000 hectares of land and employs 16,000 workers. It processes 200–300,000 tons of wheat. Its activities include 102 large farms, 28 processing plants, 100 trade organizations, 32 service enterprises, etc. (see Box A2.8 for a scheme of the structure of the company). Case 2: Pshenitsa-2000 Orel holding This holding received a signi�cant (40 million US$) capital injection from Germany including new machinery, etc. The company controls almost 100,000 hectares of land and employs more than 3000 employees. Despite massive investments prof- itability is poor and debts have accumulated over recent years. Case 3: Orel Agro holding Orel Agro employs 12,000 people and controls around 200,000 hectares of land. Its activities include, except for many grain farms, dairy and pig production, animal and dairy processing companies, grain elevators, etc. BOX A2.8 Structure of “Orel Niva Public Company� “Orel Niva� public company 5 Servicing 26 Agro-�rms 15 Processing 5 Trade enterprises enterprises enterprises 4 Enterprises directly reporting to the head company 102 Agricultural 13 Processing enterprises plants 96 Trade 27 Servicing organizations enterprises Source: Gataulina et al (2004) 93 Annex 2 Variations in Vertical Coordination Models The leading motive of the regional authorities in Orel for assisting in the creation of agro�rms and large regional holdings was “to actively employ methods of state control to preserve and develop the agroindus- trial sector of the region, to rehabilitate bankrupt agricultural enterprises, to provide the enterprises with �xed and current assets, and to restore the broken integration links.� The Orel region was to become a model for developing a regional agricultural sector under market conditions but with a strong controlling role of the state. Impact of VC on Russian agroholdings The creation of huge agroholdings in Russia has signi�cantly affected the performance of the farms. Interestingly, the results appear to be quite mixed. Gataulina et al.’s (2004) study �nds are that vertical integration has contributed to: A better supply of inputs to farms. Growth in output and productivity. Poor �nancial results. Important heterogeneity among the farms. Some of the integrated farms with good management seem to have performed very well. The three holdings in the Orel region, which are described in Box A7, all operate between 100,000 and 300,000 hectares of land. Their main agricultural activity is grain production. Grain productivity in the Orel region has increased strongly since their creation in the late 1990s, increasing by 62% from 1999 to 2002 (an average of 20% per year), helped also by strong increases in real grain prices. However, the pro�tability of grain, milk, and pork for these vertically integrated farms has been decreasing during the same period (see Table A2.1). The holdings performed consistently below the regional average (see Table A2.2). Moreover, they have accumulated substantial debts since their creation. TABLE A2.1. Performance Indicators for the Agro�rms of the “Orel Agro-Complex� Indicators 1999 2000 2001 2002 Pro�tability of agricultural activity, % 25.5 25.3 20.3 4.7 Pro�tability of crop production, % 74 49 32 12 Grain yield (kg/hectare) 1180 1640 1710 2160 Grain yield, index 100 139 145 183 Pro�tability of milk sales, % 27 26 3 −18 Pro�tability of meat sales, % 47 18 21 12 Pro�tability of pork sales, % −8 −27 −21 −32 Source: Gataulina et al. (2004). 94 Annex 2 Variations in Vertical Coordination Models TABLE A2.2. Performance Indicators of Agricultural Enterprises of the Three Largest Holdings in the Orel Region Compared to Regional Averages of all Farms (2002) Orel Agro- Orel Complex Wheat— Region Indicators Niva (12 agro-�rms) 2000 average* Grain pro�tability, % 8 13 −45 19 Pro�tability of sold milk, % −6 −11 −4 2 Pro�tability of sold pork, % −39 −23 −38 −25 *Average values for large and medium enterprises in the region (per Goskomstat database for 2002). Source: Gataulina et al. (2004). Hence, while the vertical integration process seems to have contributed to sometimes spectacular growth in output and yields, the pro�ts of the vertically integrated farms seem to have been worse or not better than those of nonintegrated farms. In both the Belgorod and Orel regions, farm debts accumulated over the 2000–2002 period, particularly on the verti- cally integrated farms. For example, in 2002, among the 149 agricultural enterprises vertically integrated in 19 holdings in Belgorod only 39 were pro�table, down from 84 in 2001. The main problem is the same problem that plagued the Soviet-style agroholdings: poor allocation of resources, bad management, and the imposition of social and political objectives on economic institutions. A key problem with management is the authorities’ interference with production plans and with decisions relating to which activities (and companies) should be maintained by a holding, sometimes imposing unpro�table activities and companies on the holding. Through direct and indirect state ownership, there is signi�cant involve- ment of the state in the management of the agroholdings and agro�rms in both the Orel and Belgorod regions. Some of the holdings receive(ed) signi�cant budgetary support from the regional governments. However, even the management of “private� holding companies “need to take into account that any reorganization of agricultural business cannot avoid coordination with local authorities,� which have their own objec- tives and preferences. Holdings of all types “have to accept certain social obligations,� which can negatively affect their pro�ts knowing that “in return, authorities grant agroholdings certain privileges.� For example, Orel Agro holding receives interest-free loans from the Orel regional government in return for supplying grain (175,000 tons in 2002) to the regional food reserve fund at prices set by the regional authori- ties. The regional administration also determines minimum prices and marketing margins for milk. It also “requests a mandatory submission of production-financial plans on a form provided by the Ministry of 95 Annex 2 Variations in Vertical Coordination Models Agriculture of the Russian Federation. They insist on their own �gures and values, most often arable lands and gross charges, cattle population, volumes supplied to processing factories; and in doing so they show lit- tle concern [whether] the �rms’ resources match the values under dis- cussion. Ignoring the demands expressed by the [local authorities] can lead to displeasing results, therefore most often the decisions are passed on a consensus basis.� Another example is the case of “Soskovo� public company, an agro�rm47 created on the foundation of the district dairy plant in 1999, the manage- ment of the company planned to incorporate only the best dairy farms from the region to create a good raw material (supply) base. However, fol- lowing a “strong request from the authorities,� three agricultural organi- zations with a “very weak �nancial position� (essentially bankrupt farms48) were also integrated into the �rm. Yet, Soskovo is a very interesting case study. It shows that in some cases, where good management is present, dramatic improvements in performances of companies can take place within these vertically inte- grated structures. The management improvements have yielded spectac- ular results. Between 1999 and 2003, labor ef�ciencies and farm yields increased dramatically on each of the Soskovo farms despite signi�cant reductions in numbers of employees. Yields increased for every product type in each of the farms, with an average increase in yields across all farms and products of 176%, i.e., 44% per year. Due to these increases, the pro�tability of the farms also increased. However, the management of Soskovo, which explained the success of this case, appeared more exceptional than typical. This explains why the overall holding story is less successful. There are problems of sustainabil- ity and replicability of this model as the management position is unclear and since the personality of management seems to be a key factor. The overall conclusion is that the re-creation of vertical integration in these Russian regions by local authorities has coincided with a strong increase in yields while at the same time not increasing profits proba- bly because the misallocations are also reintroduced. However, allocating causality requires further analysis since the yield increase is not higher than that of nonintegrated farms in the region. Hence, it is not clear what the impacts have been. 47. An agro�rm is a more integrated structure and is typically a unit within an agroholding. 48. According to the study, based on evidence of the public company executives, in these enterprises no wages have been paid for 3 years, land was not cultivated, only one tractor was operational, land remained unfertilized, in several enterprises no turnover of cattle took place during several year’s time. The discipline of workers was extremely low. 96 ANNEX 3 Vertical Coordination with Small Farmers In most transition countries, farms vary widely in size and organization, from small household plots, over family farms, to large cooperatives or farming companies. For example, in Slovakia, almost all land is used by large farming corporations, while in countries such as Albania, Azerbaijan, etc., almost all land is used by small individual farms. Other countries have large differentiations internally by region or commodity. For exam- ple, in Kazakhstan, the northern regions are dominated by vast grain- producing farming corporations, while the cotton areas in the south are dominated by very small household farms. In Russia, the vast majority of the land is used by large farms, but around 60% of agricultural output is produced by household plots. A key concern is that this process of VC will exclude a large share of farmers, and in particular small farmers. There are three important rea- sons for this. First, transaction costs favor larger farms in supply chains. There is an important �xed transaction cost component in costs of exchanges between farms and companies, making it more costly for companies to deal with many small farmers than with a few larger suppliers. Second, another reason that may reinforce the �rst factor is that when some amount of investment is needed in order to contract with or supply to the company, small farms are often more constrained in their �nancial means for making necessary investments, either because they do not have suf�cient own resources or because they have problems accessing external funds in imperfect rural �nancial markets. Third, an additional reason is that small farms typically require more assistance from the company per unit of output, because they are more likely to lack essential management capacity or because they are less likely to have at least some of the investments themselves. For example, before the 97 Annex 3 Vertical Coordination with Small Farmers vertical integration process started, large dairy farms in Slovakia had cool- ing tanks and dairy specialists, while small Polish dairy farms had neither. The concern of the exclusion of small farmers is voiced often and raised in many studies, in particular also in the emerging literature on the impact of the growth of modern supply chains, which emphasize the shift to larger preferred suppliers and the exclusion of small farms (e.g., Reardon and Berdegué, 2002). However, what empirical evidence do we have on the exclusion of small farms from vertically integrating supply chains? We looked at empirical evidence from transition countries and from emerging and develop- ing countries. Interestingly, the two sets of empirical evidence show a largely consistent picture. Our studies and interviews with companies gen- erally con�rm the main hypotheses coming out of global observations: 1. Transaction costs and investment constraints are a serious consideration. 2. Companies express a preference for working with relatively fewer, larger, and modern suppliers. 3. Empirical observations show a very mixed picture of actual contracting, with much more small farms being contracted than predicted based on the arguments above. Empirical Evidence The CIS study by White and Gorton looked at both the share of small farms in contracting relationships with suppliers and at the terms and conditions of the contracts. They �nd no evidence that small farmers have been excluded over the past 6 years in developing supply chains. In the vast majority of cases companies have the same or more small farm- ers in 2003 than in 1997. In fact, 57% of the processors had more small suppliers in 2003 than in 1997. Moreover, the processing companies indicate, on average, that they are not likely to cut suppliers in the future. Interestingly, for those that expect to deal with less suppliers in the future, this expectation is based mostly on choices made by farmers themselves who may move out of small-scale agriculture and into more rewarding activities either inside or outside agriculture as the economy is improving. In terms of the supplier assistance that small farmers receive, White and Gorton �nd evidence that better and more assistance seems to go to larger farms, although there is signi�cant variation with the type of assistance.49 For example, there is little difference in the provision of quality control, 49. They de�ne “small farm� as having less than �ve cows (in dairy) or operating less than 1 hectare. 98 Annex 3 Vertical Coordination with Small Farmers guaranteed prices, agronomic support, prompt payments, or even farm loan guarantees between small and large suppliers. However, the major- ity of companies operate a minimum supplier size for providing credit, physical inputs, machinery, etc. In his study on the Romanian dairy sector, van Berkum also �nds that VC does not exclude small farmers from the supply chains and that all major companies contract with small farmers, but that more sophisticated supplier-assistance programs tend to be more available for larger farms. In other countries, we �nd similar conclusions. For example, most dairy companies in Slovakia place conditions on farms to gain access to these offered programs, meaning that not all suppliers have equal access. Three out of six interviewed companies said that farms need to have a minimum size to qualify for investment support; one indicated that only the bigger and better quality suppliers were allowed to use the (forward) credit pro- gram; one only offers investment programs to “�nancially healthy� farms. In Bulgaria, two domestic dairies indicate that they set a minimum size for farms to qualify for their programs. A third claims that small producers are not interested in investments and do not apply for support. Danone explicitly limits assistance programs to contracted suppliers. Meggle’s pro- grams are limited to large suppliers by default since only farmers with large milk quantities can deliver to Meggle. In Poland, we did not �nd such differential treatments and found that VC and increased quality requirements had positively affected the survival and growth of small dairy farms. Two hundred and eighty-three house- holds in the sample delivered milk to dairy processing companies in 1995. Of these, only 36 (13%) stopped delivering milk between 1995 and 2000. Ten of them (4%) stopped producing altogether while the rest kept some cows for home consumption. Hence, 87% continued delivering to dairies despite radical restructuring of the dairies and tightened quality demands. (Moreover, some of those who stopped delivering might have stopped anyhow: the average age of those who stopped producing is 56 years, compared to 45 years for the entire sample.) Often, supplier programs differ to address the characteristics of these varying farms. For example, in case studies of dairy processors in Moldova, Poland, and Romania, we �nd that investment support for larger farms include leasing arrangements for on-farm equipment, while assistance programs for smaller dairy farms include investments in collection units with microrefrigeration units. In the crop sector, Gow and Swinnen (2001) report how oilseed proces- sors and brewing/malting companies in Central Europe (Hungary and Slovakia) used several different contract types to deal with different farms and regions. The contracts differed, reflecting differences in enforcement 99 Annex 3 Vertical Coordination with Small Farmers problems and transaction costs. For example, several companies con- tracted directly with larger farms and with so-called integrators, who in turn contracted with smaller farms. Studies from other regions show a similar mixed picture. While many reports indicate the general preference of companies to work with fewer and larger suppliers, there are many cases where companies actually con- tract with a mixture of large and small farms or even primarily or exclu- sively with a large group of small farms (e.g., Bivings and Runsten, 1992; Glover and Kusterer, 1990; Von Braun et al., 1989). Key and Runsten (1999) document how, even within the same industry, a variety of con- tractual models coexist with some companies contracting with large sup- pliers only, and others with both large and small suppliers, and yet others where the contracting has changed over time.50 Furthermore, empirical evidence indicates that companies in reality work with surprisingly large numbers of suppliers and of surprisingly small size. Even studies pointing at challenges facing small farmers in evolving modern supply chains �nd that small and medium farmers can be suc- cessfully integrated in the chains, with processing and trading companies actively investing in institutions and infrastructure to reduce transaction costs, such as collection centers. Some examples are given here. In a recent report on the impacts of the growth of modern retailing in Central America, Berdegué et al. (2003) document that 70% of the sup- pliers for Hortifruti, a highly developed specialized retail FFV supplier in Costa Rica, are small farmers (who passed the selection process and ben- e�ted from extended supplier assistance), working via intermediate pack- ers who aggregate supplies, and that this suppliers group has been very stable over the past 5 years. The same thing occurred in Nicaragua, an even poorer country, where the supply chain invested in a system of collection centers in rural areas to collect small and poor farmers’ products. Friesland Romania, a subsidiary of the Dutch FCDF group, entered the Romanian market in 2000, and 3 years later processed approximately 200–250 tons of milk per day, in �ve factories. The company has simple contracts with approximately 40,000 small farmers from whom it pur- chases milk through 1050 collecting points and from some 600 larger farms (Van Berkum, 2004). Friesland owns the collection points and has upgraded them by investing in cooling and inspection facilities. 50. The evidence in Key and Runsten (1999) on contracting in the Mexican veg- etable industry indicates that most processing companies were contracting with small and large suppliers when the demand was high compared to the supply, and thus prices high, but that small suppliers were cut �rst when the market became less prof- itable. This suggests that as the processing companies come under pressure they tend to shift to larger growers and cutting off the small suppliers in a strategy to econo- mize on transaction costs. 100 Annex 3 Vertical Coordination with Small Farmers A recent study on VC in Chinese modern retail supply chains (Hu et al., 2004) also yields evidence on contracting with large numbers of small sup- pliers. They give the example of Beijing SanLu Vegetable Co. started con- tracting with farmers in 1999 for its Beijing supermarkets and for exports. It started with 300 farms in 1999, with average farm size of one-sixth of a hectare. By 2003, it was contracting with 4500 farms with an average size of one-third of a hectare. All the farms receive seeds, technical assistance, information on prices and market developments, and guaranteed prices for speci�c qualities. In addition, SanLu invested in collection centers in the main production sectors to reduce transaction costs. Another example from this study is Xincheng, a food processing enter- prise that produces packaged fresh vegetables for the Shanghai urban wet- markets. In 1997, Xincheng began to supply supermarkets and by 2003, it was supplying 500 supermarkets owned by domestic and foreign chains. Until 1998, Xincheng relied on the wholesale markets for raw materials, im- plying paying wholesale margins, inconsistent quality, and risk of food- safety problems. As of 1998, Xincheng shifted toward contracting supplies directly. It has contracts with 4200 individual farmers in the rural area near Shanghai. It supplies the farmers with seed, fertilizer, pesticide, and techni- cal assistance. All their output, produced according to Xincheng quality and safety standards, is sold to Xincheng, and the cost of the inputs is deducted from the output price. Since 2000, Xincheng also rented 1000 hectares of land, built greenhouses on it, and hired 200–300 farmer migrants from poor areas inland to operate the farm, under similar contracts as the farm- ers, except for land.51 At this moment, Xincheng sources 15–20% of its veg- etables from its own greenhouses and hired workers, and 50% from the 4200 outgrowers. The balance, 30%, is bought from wholesale markets. Finally, our Kazak cotton study also yields evidence of processor con- tracting with thousands of small farms, many of whom receive pre�nance and other inputs from the processors. To reduce the transaction costs, the gins have themselves invested in collection centers at various localities throughout the cotton-producing areas. An alternative mechanism of overcoming transaction costs is using ex- isting cooperatives as an instrument for reducing transaction costs. For example, in Poland, Danone has its own suppliers but also purchases milk from a Polish dairy cooperative, with many small suppliers, for fur- ther processing the milk into higher value products. Here, Danone makes use of the cooperative to reduce transaction costs in its dealings with small suppliers. 51. Xincheng has recently started a similar operation 500 km to the south, in order to have counter-season vegetables in Shanghai’s winter. 101 Annex 3 Vertical Coordination with Small Farmers Why Contracting with Small Farmers? Hence, despite the apparent disadvantages noted earlier, the empirical evidence suggests that VC with small farmers is widespread. The question is why? There are several reasons. First, the most straightforward reason is that companies have no choice. In some cases, small farmers represent the vast majority of the potential supply base. For example, over 95% of Romanian dairy “farms� have one to two cows. These farms have 83% of all milking cows in Romania. Only 0.25% of all farms with cows have 10 cows or more, with only 5% of the total herd of milking cows on these farms. Similarly, in countries such as Poland, much of the dairy supply is in very small farms. Hence, any dairy processor needs to deal with small farms by necessity, focusing, for example, on investments in collection points, etc. rather than on-farm equipment. Van Berkum (2004) writes: Although the larger farmers have some privileges in assistance programmes with respect to investment funds eligibility, there are no signs that the present vertical coordination arrangements in the Romanian dairy supply chain exclude small farmers. Despite high transaction costs dairies are willing to collect the milk from small plots, largely through collection points. The two foreign dairies interviewed in the survey explicitly indicated to like to reduce the number of small-scale sup- pliers and work with larger suppliers. Yet, the problem is that there are only very few dairy farms with more than 5 cows in Romania. For the moment, dairies have to accept this situation until restructuring and consolidation in the sector starts off. In the meantime, the larger dairies are keen to assist their supplying farmers in improving their conditions for producing higher quality milk by providing advice, improving access to inputs including investment means and enhancing access to output markets. Farmers who are willing to learn and develop get chances to further develop their business. Second, our case studies suggest also that company preferences for con- tracting with large farms are not as obvious as one may think. While processors may prefer to deal with large farms because of lower transaction costs in, e.g., collection and administration, contract enforcement may be more problematic, and hence costly, with larger farms. In several interviews company managers indicated that (smaller) family farms were less likely to breach contracts or to divert company investments than large coopera- tives or farming companies. White and Gorton (2004) also conclude in their study that processors repeated emphasized that farms’ “willingness to learn, take on board advise, and a professional attitude were more impor- tant than size in establishing fruitful farm-processor relationships.� Third, in some cases small farms may have substantive cost advan- tages. This is particularly the case in labor-intensive, high-maintenance, production activities with relatively small economies of scale. For exam- 102 Annex 3 Vertical Coordination with Small Farmers ple, Key and Runsten (1999) present evidence that small farmers’ pro- duction costs in Mexican vegetable contract production were 45% lower than that of specialized farms owned by the processing companies. Costs were lower primarily because of imperfections in labor and land markets. Small farmers had significantly lower labor costs because of access to unremunerated family labor for which markets are missing, and much lower costs of supervising, transporting, and recruiting labor input; and because they did not pay any government bene�ts. Land rents were lower because of restrictions in land markets constraining formal renting by large farms. And also pest control costs were lower due to better crop mon- itoring and thereby lower chemical use. Further, small farmers’ yields in vegetable production were 20% higher than on the �rm’s own farms. Fourth, processors may prefer a mix of suppliers in order not to become too dependent on a few large suppliers. For example, Caudron et al. describe the dif�culties of retailer Metro in Morocco in contracting with modern large shippers. Metro caters for hotels in tourist areas and other high-standard food outlets and requires high product quality and pack- aging. Metro has contracted informally with “Domaine Douiet,� one of the very few Moroccan FFV exporters. Douiet has committed to provide Metro with �xed volumes of high-standard FFV (such as cherry tomatoes sold in a transparent plastic box). However, Metro has not been success- ful in enforcing full compliance with the terms of contract and faced sev- eral shortages of produce, in particular in periods of high export prices when the supplier found it more pro�table to export. As a consequence, these retailers have started to contract with groups of smaller growers and in some cases, when those groups did not exist, to stimulate their creation. Fifth, processing companies differ in their willingness to work with small farms. Evidence indicates that some processing companies work with small suppliers even if others do not—this may reflect the process- ing companies’ own roots as cooperative organizations. For example, in the Romanian dairy sector, Friesland, and especially ISPA and Raraul seem to be more inclined to work with small farms than Danone, reflected in the way they assist small suppliers and invest in overcoming transaction costs (see Table A3.1). All of them contract many small-scale holdings and most of them also contract large-scale farmers. Friesland and Danone have outsourced the transportation of milk to independent conveyors but Friesland has invested signi�cantly in collec- tion centers. Raraul and Promilch/ISPA take care of the collection and transport themselves. ISPA, a dairy association, takes further initiatives to provide its mem- bers with basic farm-level support on matters of key importance (feed- ing, milk quality, and hygiene) and a secure market outlet. Furthermore, 103 Annex 3 Vertical Coordination with Small Farmers TABLE A3.1. Contracts partners and arrangements for collection and transport Arrangement Danone Friesland Promilch/ISPA Raraul Contracting small farmers X X X X Contracting large farmers X X X Owning collection centres X X X Arranging transport to farm- collection centre X X Arranging transport to collection centre—dairy X X Note: X means “yes� or “applicable to.� Source: van Berkum. by investing in further processing, this association adds further value to raw milk and strengthens the market position of its members. This example shows that small-scale farmers may have future perspectives when effectively organised. Key and Runsten (1999) also find that small farm inclusion may depend on company strategies in Latin America. They also document how some processing companies in the Mexican vegetable industry con- tinue to work with small local suppliers even when others do not (any- more). These companies continue to work with small farms because they have been able to design and enforce contracts that both the small �rms and the companies �nd bene�cial. The �rm bene�ts from the low-cost production on small farms, signi�cantly lower than on large farms, while providing inputs, assistance, and credit in a way that minimizes the �rms’ transaction costs. The latter is done by providing credit without collat- eral; by restricting the number and the location of the farms (farms have to be located along roads where they are easily accessible for company agronomists) and by demanding that farms come to the company to pick up seeds and fertilizer and to deliver the harvest. That said, even companies willing to invest in upgrading small farms only go so far, and tend to have a strategy in the long run to upgrade part of their supply here to larger, more ef�cient, and fewer suppliers. Yet, in countries like Poland, Romania, and many CIS countries dominated by household dairy production, “large� is a relative concept. As Van Berkum (2004) puts it: “In Romania, large farms are farms with more than �ve cows.� 104 THE WORLD BANK Washington, D.C.