Presented on behalf of Research Team by Colin Poulton (CeDEP, SOAS)

1. Comparative analysis of organization and performance of African cotton sectors: Learning from reform experience Presented on behalf of Research Team byColin Poulton (CeDEP, SOAS) Other members of the team include: David Tschirley (MSU), Patrick Labaste, John Baffes and Julie Dana (World Bank), Gerald Estur and Nicolas Gergely (independent consultants)SOAS, September 19th 2008

2. Objectives of the Study • Comparative analysis of the lessons from cotton sector reforms implemented in SSA countries during the last 20 years • 9 countries cases : Tanzania, Uganda, Mozambique, Zimbabwe, Zambia, Mali, Burkina, Benin, Cameroon • Comparison across WCA and ESA • Understand • How a sector’s history and current structure influence the set of feasible reform paths • How the path chosen influences the types of challenges a sector might have most difficulty meeting (e.g. quality, productivity, competitive prices to farmers) • Through this, to provide a stronger analytical basis for public and private stakeholders to design their country’s reform path • Better understand to better advise (or decide)

3. Market Context • Price Trends • Decline of 55 percent between 1960-64 and 1999-2003 in real prices of cotton (similar to other major export commodities) • Driven by annual average yield gains of 1.8%, but stagnant per capita demand (competition from man-made fibres). • Subsidies to OECD producers • Full removal could raise world price by 10-15%, but progress slow • Exchange rate for WCA countries • Depreciation 1995-2001, but dramatic appreciation since 2002: in real terms, now back at 1995 levels • Demand Side: Increasing Importance of Lint Quality • Inherent fibre characteristics and hand picking should generate premium • Africa not capitalizing on its comparative advantage: contamination! • Valorization of by-products

4. Case Study Countries Mali B. Faso Cameroon U’nda Benin West & Central Africa East & Southern Africa Tanzania Malawi Zambia Mozambique Zim’we

5. Case Study Countries Mali B. Faso Cameroon U’nda Benin • West & Central Africa • Single channel systems • Many years of investment (rsch, AT) • Cotton as engine of rural development • Intensive input use, high yields • High total production • Huge economic and political weight • Limited reform • Recent very serious financial crises East & Southern Africa Tanzania Malawi Zambia Mozambique Zim’we

6. Case Study Countries Mali B. Faso Cameroon U’nda Benin • West & Central Africa • Single channel systems • Many years of investment (rsch, AT) • Cotton as engine of rural development • Intensive input use, high yields • High total production • Huge economic and political weight • Limited reform • Recent very serious financial crises • East & Southern Africa • - More diverse history • Generally much less investment • Not seen as engine of development • - Much lower input use, yields • - Lesser economic and political weight • Much more reform • No public financial crises Tanzania Malawi Zambia Mozambique Zim’we

7. Conceptual Framework 4 steps: • Typology of Cotton Systems (see Chart on next slide) • National Monopoly • Local Monopoly (Concession) • Concentrated • Competitive • Hybrid • Mapping SSA Cotton Sectors (see Chart on following slide) • Indicators of Performance (see Chart on following slide) • Process: Core activities and Service Delivery (prices paid to farmers, input credit delivery, quality, valorization of by-products, research) • Outcomes: yields and returns to farmers, value addition, efficiency and competitiveness, macro-impact • Linking Performance to Sector Types by analyzing sector performance vs the selected indicators

8. Is competition allowed for the purchase of seed cotton? Yes No Market-Based Regulated Are there many buying firms, or few? Is there more than 1 cotton buyer? Many Few No Yes Is each firm assigned an exclusive geographical area in which to buy seed cotton? Competitive Systems Concentrated Systems National Monopoly Yes No Local Monopoly Systems Hybrid Systems

11. Performance Indicators and Expected Performance

12. Data • Existing knowledge and secondary data summarised in 9 country case studies • Updated to 2007 through country visits • Supplemented by • Farmer focus groups (2-6 per country) to develop disaggregated production budgets by farm type • Interviews with ginners in ESA to piece together ginnery budgets; company accounts in WCA • Survey of lint quality for SSA lints undertaken by Gerald Estur

13. Core Activities and Service Delivery: Quality • Fibre characteristics of African cottons typically superior to Cotlook A Index • Also hand-picked (good for cleanliness) • But major problem of contamination (increasingly penalised on world market) • Analysis: • Compared average premium for top type of lint from each country in 2006/07 and mid-1990s • Estimated average export price differentials across the nine countries relative to the A Index, based on: • average premium of the quotation for the top type of each country in Cotton Outlook over the Cotlook A Index during 2006/07 • Usual world market price differences for grade compared to middling • Usual world market price differences for staple length, relative to 1-3/32” • Actual 2005/06 classing data for WCA countries and most recent available data or estimates for ESA countries • deduction of one cent per pound to reflect difference betweenseller’s offering price and actual negotiated contract price.

14. Estimated Average Premium over A Index (US cents/lb), 2005/06 Source: Gerald Estur quality survey

15. Change in Price Premia for Top Lint Types, mid-1990s to 2006/07 Source: Gerald Estur quality survey

16. Core Activities and Service Delivery: Quality II • Performance best in concentrated sectors • major reduction in contamination in Zambia • Zimbabwe has suffered since entry of new companies post-2001 • Variable in national and local monopolies • management culture and regulatory effectiveness are key • Poor in competitive sectors, especially Tanzania • Two conditions for a sector to produce high quality lint • ginners can control their supply chain • ginners have incentives to achieve high quality lint • Both fulfilled in concentrated sectors, but: • In monopoly systems, control over supply chain may be constrained by political requirement to buy all seed cotton at pre-set price • In competitive sectors unregulated competition undermines ability of ginners to control their supply chain, whilst limited vertical coordination between independent ginners and exporters weakens ginners’ incentives to produce high quality lint

17. Core Activities and Service Delivery: Quality III • Quality Improvement: • Major challenge for all sectors, across types • Many sectors not exploiting comparative advantage • Potential 10-20% price gain for some sectors through better quality management (given inherent fibre characteristics of African cottons) • Tanzania’s comparative advantage may lie more in low cost, modest quality cotton than high quality • Local auction system as way to improve quality?

18. Core Activities and Service Delivery: Prices to Farmers • WCA price setting: administered, panseasonal, panterritorial, announced before planting, purchase guaranteed at official price • Move in recent years to pricing systems providing more linkage to world price and more flexibility • Greater farmer voice in pricing decisions • ESA (excl MZ, UG post-2003): price leadership or competition • Indicator: share of FOT price paid to farmers 1995-2005 • Competitive systems (TZ, UG) best performers (68-70% of FOT) • Concentrated: good in post-reform years, but sharp drop since 2000s • WCA monopolies: very low in the 1990s, sharp rise in 2000s but in context of sector financial unsustainability

19. Farmer Share of FOT Lint Price, 1995-2005

20. Core Activities and Service Delivery: Input Credit and Extension Delivery • Well developed in WCA single-channel systems and has permitted intensification • But yield stagnation within these systems since mid-1980s • To the contrary, highly competitive post-reform structures in TZ and UG led to collapse of input and extension systems • Concentrated models (ZB, ZM) have performed better, but repayment adversely affected by new entry • Overall, monopoly and concentrated sectors best able to ensure provision of inputs on credit and also to provide some level of extension advice • Provision of these services is undermined by side selling in more competitive sectors

21. Core Activities and Service Delivery: Research • Critical for Africa’s cotton sectors to improve international competitiveness and contribute to poverty reduction • African sectors seem to be lagging behind many of their major global competitors in this critical area • At best, weak linkages with sector types • history exerts a strong influence over research performance (lags inherent in agricultural research) • governments slow to allow private stakeholders to contribute to research management, so predicted advantage of concentrated sectors in demanding and/or organizing effective research not seen in practice • Two new technologies potentially interesting for cotton growers in Africa over the near to medium term: • genetically modified (Bt) cotton • low-volume herbicides

22. Core Activities and Service Delivery: Valorization of by-products • Value of cottonseed oil and cake : 20 to 25% of the value of lint • Potentially growing markets with world market demand for vegetable oil and protein on the rise • Currently domestic markets more profitable (import substitution) • Whether country is landlocked or not has big influence on price paid for cotton seed • Cotton seed prices in Burkina and Mali should be higher • Landlocked, high demand for cake from livestock sectors (but also smuggled oils and difficulty of product differentiation) • High processing cost (high quality oils) • Underdeveloped markets for oil

23. Outcomes: Yields, Crop Budgets and Returns to Farmers • Results confirm that performance on yield levels is • Strongly related to performance on input provision and extension • Heavily influenced by past investments (WCA) • Yield performance in ESA is correlated with sector organization: • More concentrated systems (Zambia and Zimbabwe) achieve higher yields than more competitive models (Tanzania and Uganda) • Crop budgets: • Much higher proportion of cotton producing households found in the higher producing groups in WCA than in ESA • Weighted average returns to both family labor and to all labor are higher in WCA than in ESA • Between 25% and 75% of cotton producing households (depending on the country) would be better off hiring out their labor than applying it to their own cotton plots

24. Weighted Average Returns to Family Labor and All Labor in Study Countries (US$ / day) Source: project focus group exercises

25. Production Cost of Seed Cotton, incl Labour (US$/kg) Source: project focus group exercises

26. Outcomes:Cost Efficiency, Overall Competitiveness, and Macro Impact • Ginning costs: • Sharply lower in market based systems (Zambia, Zimbabwe, Tanzania) than in monopoly or hybrid systems (WCA, Mozambique, and Uganda). • Total net cost from farmgate to FOT: • Lower in market based systems, be they competitive or concentrated (due to lower ginning costs, overheads and financial costs; higher sales value of seeds) • WCA monopolies perform especially poorly in terms of company efficiency • WCA sectors are the least competitive: barely breaking even (Cameroon) or generating large deficits (Mali and Burkina Faso). • ESA sectors appear to be competitive in world markets • No budget support in ESA, periodic bailouts of companies in WCA

27. Estimated Average Ginning Costs at 2006 Capacity Utilization Rates in Study Countries Source: company accounts (WCA), project interviews (ESA)

28. Outcomes:Cost Efficiency, Overall Competitiveness, and Macro Impact II • Higher value added per ha in WCA than ESA • Farm-level value added per kg of lint as high or higher in ESA than WCA • Higher soil fertility • Higher value added at ginning level in ESA than WCA (low or negative) • Total value added (farm and ginning levels) per capita of total population reflects size of sector • Burkina scores most highly

29. Total Value Added per capita by Cotton Sector, 2006/07

30. Realized Performance by Sector Type

31. Conclusions: Lessons from Reform • The analysis has revealed strengths and weaknesses in various systems, particularly when one looks at them over a long time period. • no one model has proven superior to all others in all respects over time, and • none of the systems under review offers a fully satisfactory and sustainable response to the challenges of future competition in the world cotton market • Structure matters, even if other factors may count as well: good predictive capacity of the typology • Clearly“reform” does not imply a movement from one stable set of rules of the game to another stable set • These objectives are important, notwithstanding factors that are beyond the direct control of SSA governments and stakeholders: by-products, such as: • the evolution of the euro/$ exchange rate, and • the slow progress in reducing market distortions due to OECD subsidies

32. Conclusions: Opportunities and Challenges for African Cotton • Some challenges are common to all SSA sectors • Three major challenges: • Achieving greater value through improved quality, marketing, and valorization of by-products, • Bridging performance and competitiveness gaps through farm-level productivity and ginning efficiency, and • Improving the sector’s sustainability through institutional development and capacity-building of stakeholders, as well as strengthening of governance structures and management systems

33. Conclusions: Summary of findings for particular sector types • National monopoly model • has generated strong returns to very large numbers of farmers, • but poor cost efficiency has undermined these sectors’ competitiveness • Competitive sectors • are cost efficient and pay attractive prices to farmers, • but their inability to provide input credit and extension, or to raise quality makes them unlikely to make substantial contributions to poverty reduction • Concentrated sectors • have performed well in quality and service delivery (input and extension), have been more efficient than the monopolies, and have also generated attractive value added per capita while making the highest contributions to state budgets, • but their performance on seed cotton pricing has been disappointing and they may be inherently unstable .

34. Conclusions: Ways forward for particular sector types • Monopoly systems • Cost reduction from farm gate to FOT needs to be a top priority: greater role for private companies to achieve this • Price setting rules must continue to be reformed • Inter-professional committees and farmer organizations need to continue to be developed, with special emphasis on the operational abilities of the latter • Clear rules for evaluating and re-tendering concession areas need to be developed • Reforms in research organizations to make them more responsive to inter-professional committees • Allow investment in the oil sector to create more competition

35. Conclusions: Ways forward for particular sector types (II) • Concentrated sectors • Key challenge is to develop appropriate regulatory regime that understands strengths and weaknesses of the concentrated model • Concentrated sectors need barriers to entry: licensing rules that specify capabilities and conduct of firms wishing to participate in the sector • Given problems of relying entirely on the threat of entry to discipline incumbent firms, develop formalized price setting mechanisms to replace price leadership?

36. Conclusions: Ways forward for particular sector types (III) • Competitive sectors • Key state role in sector coordination, but need to strengthen the accountability of regulatory bodies towards both ginners and farmers • Regulatory bodies and/or ginners’ associations work with other actors (e.g. local government or donors) to develop long-term programs to enhance soil fertility or promote animal traction

37. Conclusions: Final Reflections on Sector Types and Looking Ahead • Some degree of convergence in the forms of cotton sector organization in Africa likely to happen over the next decade • Increase in the number of local monopoly systems in the short/medium term? • Transition to concentrated systems desirable, if regulatory challenges can be overcome • More competitive systems are probably part of the long-term future in most countries • Need more effective rural financial markets and farmers’ organisations first