Report No. 35239-TN Tunisia Agricultural Policy Review July 20, 2006 Water, Environment, Social and Rural Development Middle East and North Africa Region Document of the World Bank FNG Fonds National de Garantiernational GaranteeFund FODECAP Fonds de dkveloppement de la compktitivite dans les secteurs de l'agriculture et de la pkhe/Agriculture and Fisheries Competitiveness Development Fund FOPRODEX Fonds de Promotion des Exportations (Export Promotion Fund) FSA Financial Sector Assessment GDAP Groupement de De'veloppementde 1'Agricultureet de la P&he/ Agricultural and Fishery Development Group GFIC Groupement Forestier d 'Inte'rct CollectifBorestry Collective Interest Group GIC Groupement d 'Inte'ret Collectif/ Collective Interest Group GIP Groupement interprofessional/ Inter Professional Group GIS Geographical InformationSystems GOT Government of Tunisia IRESA Institut deRecherche et d 'EnseignementSupe'rieur/ The Agricultural Researchand Higher TrainingInstitute MARH MinistBre de 1'Agricultureet des RessourcesHydrauliques/ Ministry of Agriculture and Water Resources MDCI MinistBre du De'veloppementet de la Coopbration Internationalel Ministry of Development and International Cooperation MTEF Medium TermExpenditure Framework OEP OfJice de 1'Elevageet des PcituragesLivestock and Rangelands Agency ODESYPANO Office de De'veloppementSylvopastoral du Nord-Ouestmorth West Sylvopastoral Development Agency ONH Office National des HuiledNational Oils Agency PACFS Petite Agriculture a CaractBreFamilial et SocialBmall Scale Family and Social Agriculture PBB Performance Based Budgeting PISEAU Projet d'lnvestissement dans le Secteur de 1'EaulWater Sector Investment Project PMEACE Petites et Moyennes Exploitations Agricoles a CaractBre EconomiqueBmall and Mediun Scale Commercial Farms P S I A Poverty and Social Impact Analysis UTAP Union Tunisiennede 1'Agricultureet de la P&he/Fanners' and Fisherfolks' Union UTICA Union Tunisienne de 1Yndustrie, du Commerce et de 1'Artisanat/ TunisianUnion for Commerce, Industries and Handicrafts i Vice-president: Christiaan J. Poortman Country Director: Theodore 0.Ahlers Sector Director: Inger Andersen Sector Manager: LuisF.Constantino Dou las W. Lister This volume is a product o f the staff o f the International Bank for Reconstruction and Development/ The World Bank. The findings, interpretations, and conclusions expressed inthispaper do not necessarily reflect the views ofthe Executive Directors ofThe World Bank or the governments they represent. The World Bank does not guarantee the accuracy o f the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply anyjudgment on the part o f The World Bank concerning the legal status o f any territory or the endorsement or acceptance o f such boundaries. The material in this publication is copyrighted. Copying and/or transmitting portions or all o f this work without permission may be a violation o f applicable law. The International Bank for Reconstruction and Development/ The World Bank encourages dissemination o f its work andwill normally grant permission to reproduce portions o f the work promptly. For permission to photocopy or reprint any part o f this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, http://www.copyrigh t.com/. All other queries on rights and licenses, including subsidiaryrights, should be addressed to the Office o f the Publisher, The World Bank, 1818 H Street NW, Washington, D C 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. ... 111 Acknowledgements While the World Bank alone is responsible for the contents o f this report, many Tunisian organizations, too numerous to list, have contributed directly or indirectly. In particular, the report would not have been possible without the substantial support and coordination provided by the Studies and Agricultural Development Department o f the Tunisian Ministry o f Agriculture and Water Resources, and especially by its Director General, Badr Ben Ammar. This report is based on the findings o f a joint Bank/FAO-CP/AFD team which visited Tunisia in NovembedDecember 2005 consisting o f Douglas W. Lister (Task Team Leader), Alex Kremer (Senior Sector Economist), Jean-Philippe Tr6 (Agricultural Economist), Yaa Oppong (Social Development Specialist), John D. Nash (Lead Economist), Michel Debatisse (Consultant), Eric L e Brun (Consultant), Catherine Cormont-Tour6 (Consultant), Jean-Marc Bisson (FAOKP), Yves Rajat (AFD Tunis) and Jean-Franqois Richard (AFD Paris). Luis Constantino (Sector Manager - Rural Development & Social, MNSRE) joined the mission. Pierre Rondot (Senior Sector Economist), Nabil Chaherli (Senior Sector Economist) and Rory O'Sullivan (Consultant) provided support to the mission from Headquarters. The peer reviewers were: Stephen D. Mink (Lead Economist, EASRD), Hans Lofgren (Senior Economist, DECPG) and John D. Nash (Advisor, ARD). Marie- Franqoise How Yew Kin (Language Program Assistant, MNSRE) provided administrative and secretarial support. Nicole Wautiez de Blaye and HClbne Talon translated the report into French. iv TUNISIA .Agricultural Policy Review Table o f Contents Page 1................INTRODUCTION 1 A ............................................................................................ 1 B .. ....................................................................................................... Objectivesof the Review Approach ..................................................................................................................... 2 I1...............PASTPERFORMANCEOFTHE SECTOR ............................................................................................ ........................................................... 4 A . Five Year PlanPriorities 4 B Sector Performance ..................................................................................................... 5 C 6 D ... FoodTradeBalance .................................................................................................... ConsolidationofAgriculturalEmployment ............................................................. 7 E UnrealizedPotential .................................................................................................... 7 I11 CHALLENGESAND OPPORTUNITIES . ......................................................................... 14 A Policy andInstitutionalFramework ....................................................................... 14 B Professional Organizationsand Support Services ................................................. 38 C SustainableManagementofNaturalResources .................................................... 44 D 48 E Ruralland 58 F .. .... Resourcingthe Agricultural Sector .................................................................................................................. ......................................................................... Social PerspectivesonAgriculturalPolicy .............................................................. 68 IV OPTIONSFORTHE FUTURE . 78 A . ........................................................................................... Policy andInstitutionalFramework ....................................................................... 79 B ProfessionalOrganizationsand Support Services ..................................................... 81 C 82 D Resourcingthe AgriculturalSector ......................................................................... 82 E RuralLand ................................................................................................................. 83 F ... .. ManagementofWater .............................................................................................. Social Perspectives on AgriculturalLiberalization ................................................ 84 ANNEX 1 THE ECONOMICSOF LIBERALISATION . ........................................................ 88 ANNEX 2 THE "CARTEAGRICOLE" . ................................................................................... 91 ANNEX 3 TERRESDOMANIALES . ......................................................................................... 92 ANNEX 4 FORESTSAND RANGELANDS . ............................................................................ 93 V List o f tables Table 1.Productionlevels o fmain agricultural commodities (1000 T) ............................. 6 Table 2 .The contribution ofbeef andmilkto growth...................................................... 10 Table 3. Fruit and vegetables: annual growth rates 1980-2000 (%per year) ...................10 Table 4. Sub-sectors' contributions to growth.................................................................. 11 11 Table 6 Tariff Structure, 1994. 2004. and2005............................................................... Table 5.Performance i s inversely correlated with competitiveness ................................. 16 Table 8.Breakdowns of household expenditure ............................................................... Table 7 ..Importsunder tariff quotas. 2001-04. and customs dutyrates. 2005..................22 17 Table 10 Table 9.How the obenefits to consumers o fprice liberalisation wouldbe allocated........23 .Effects f agricultural trade liberalisation ........................................................ 23 Table 12.Area under cereals responds to simultaneous liberalisation o f all prices Table 11.Broad economic effects o f full agricultural liberalisation................................ .........24 27 28 Table 14.Recovery o foperations andmaintenance costs 1991-2000 .............................. Table 13.Winners and losers from cereals price liberalisation........................................ 46 Table 15. Arable landper person employed inagriculture. 2000..................................... 60 Table 17.Aggregate land availability............................................................................... Table 16.Labour and landproductivity (2000) ................................................................ . . . 61 61 Table 18.Evolution o f landholding size ........................................................................... Table 19.High-level agricultural strategy options for the Tunisian Government............62 78 List o f fimres Figure 1.Agriculture's share o fGDP................................................................................. . 58 Figure3.Agricultural employment 1967-2004 .................................................................. Figure2 Value-added perworker inagriculture 1992-2004............................................. 9 Figure4.Endingthe ONHmonopoly o folive oil exports in 1994.................................. 13 Figure 5.Tunisian wheat is more expensive thanimports................................................ Figure 6. Quantities collected bythe Cereals Board......................................................... 21 31 vi Executive Summary The sectoral context 1. Is agriculture making its full potential contribution to the nation? That is the question that the Tunisian Government is asking as it prepares for the llth 12th and Plan periods (2007-16). Tunisia's farm sector is entering new territory, with higher incomes, lower poverty rates, and new consumer tastes, market structures and trade agreements. Policymakers face challenges that simply did not exist a decade ago and are asking whether the solutions o f the past are still relevant. In2004, therefore, the Government asked the World Bank to review the entire farm sector for the first time inmore than20 years. This report is based on 11backgroundpapers prepared during 2005, with support from the FAO/CP and the Agence franqaise de dtveloppement. 2. Tunisia's agricultural policy has come a long way since the 1980s:from post- Independence dirigisme, through the emphasis on expenditure control inthe late `80s and productivity in the early `90s to the current preoccupation with international competitiveness. After nearly two decades of gradual liberalisation, many key prices are controlled andtariffs are high. 3. Agricultural production has performed remarkably well, with yields growing at 2.8% per year since 1989. Agriculture has kept pace with overall economic growth. It provided a quarter o f Tunisia's newjobs inthe 1990s, creating twice as manyjobs per unit o f GDP as the economy overall. 4. If one looks beneath the surface, however, there is plenty of evidence that agriculture is not makingits full potential contribution to the economy. A growing labour supply is disguising stagnant labour efficiency: farm labour went up 20% between 1993 and 2002 but there was no trend increase inits productivity. Agriculture's success imposes costs on the rest o f society: DT 180m per year in compensation subsidies, 4% extra on the cost o f living for consumers and taxpayers and 0.8% o f GDP in lost growth thanks to distorted prices. 0 Two-fifths o f agriculture's growth represents a loss to the economy because it comes from commodities which it would cost less to import - beef and milk. 0 Where Tunisia has the potential to be competitive, in h i t and vegetables, it is underperforming. Between 1980 and 2000 the value o f exports went downby 0.3% per year. EUimport quotas are underused. 5. The cases of citrus and olive oil illustrate how potential i s not being realized. The key issue is quality. Supply chains must be responsive to consumers' specific demands.But to get premium oranges and oil onto the market, farmers need first-rate technical and marketing support from research, extension and producers' vii organizations. And Government interventions, inthe form o f fixed retail margins and the Office des Huiles' control over quotas and testing, are interposed between the producer and the consumer. The result is that Tunisian oranges and olive oil command low prices and EUquota is left unused. Conversely, when Government has made partial reforms, such as the removal o f the ONH's olive oil export monopoly, the private sector has responded. 6. The next stage o f the analysis i s to ask how supply chains mightbecome more responsive. It looks at trade policies, domestic policies and institutions, producers' organizations, research and extension, natural resource management, the mobilization o f resources and land. Common themes begin to appear: 0 That the state's heavy presence in supply chains hampers their responsiveness. 0 That supply chains need first-rate public goods from Government: good research, extension andphytosanitary andhealth regulations. 0 That the prevailing logic is for Government to give top-down prescriptions e.g. for farmers organizations, credit packages, land tenure. Government could facilitate the private sector more effectively by seeking to understand and respond to its perceivedneeds. Institutions andpolicies 7. When it comes to making supply chains more responsive, one set o f reforms, removing price distortions, matters than others. Allowing markets to set prices freely is an "apex reform" because it makes other reforms more effective. For example, the more the price system rewards farmers for growing high-quality vegetables, the more vegetable-growers will take advantage o f improvements in water-management, land markets,farmer groupings, extension services and so on. 8. Trade policy is the starting point o f the analysis, because it i s the basis o f prices and profitability. Tunisia's farm tariffs are high by regional and world standards. Peak tariffs apply to 69% o f agricultural tariff lines and applied rates average 67%. According to the WTO, Tunisia's reforms o f the last two decades, "have not substantially liberalised trade." 9. This matters for consumers and taxpayers, whose cost o f living i s raised 4% by agricultural protection. And it also matters for the economy as a whole, which loses 0.8% o f GDP. In fact, the protection o f Tunisian cereals is particularly costly. Using trade protection to keep one person employed incereals costs the nation over 4 times national per capita income in terms o f lost output. This is because cereals- farminguses little labour. 10. But if the Government were to liberalise food trade, it would need to understand who would lose out. The two main sub-sectors, tree crops and small ruminants, would not lose out. Livestock farmers would even become better off, because tariff reductions would mean cheaper fodder. But 16% o f farms, covering Vlll ... 30% o f landarea, would be losers. They are generally cereals operations inthe north- andnorth-west o fthe country and include farms on state land.They include high-and low-income farms. The low-income losers are concentrated inK e f and BCja. Because cereals farming uses little labour, the impacts on income would be more pronounced thanthe impactson farmemployment. 11. Ontrade, therefore, the recommendation is that Government should gradually reduce cereals tariffs and the producer support price. Tariffs on barley and other feed inputs should be liberalised as soon as possible. A Poverty and Social Impact Analysis should identify vulnerable losers more precisely and work out mitigation measures. 12. Concerning domestic policies and institutions, the study focuses on three issues: cereals marketing, animal health and food safety and produce quality. A common conclusion emerges: the state should reposition itself. It should withdraw from direct market intervention. Instead it should concentrate on regulating markets, protecting health, natural resources and the environment and providing genuine public goods. 13. Government uses the Office des CCrCales' monopoly o f the cereals market to keep the producer price up and the price to millers down. But the cost o f its interventions is out o f proportion to the supposed benefits. At a budgetary cost o f DT 180 million per year, the Office's interventions reduce the price to millers by 2.9% at most. It costs $12 per consumer per year to deliver a benefit to the consumer o f $1.30. Moreover, the Office's monopoly has retarded the modernization o f quality control, transportation, storage, trading techniques, risk management etc.. 14. Government's reliance on the Office to ensure food security is understandable, in the light o f the food riots o f 1984. However, the causes o f food shortages in the past (foreign exchange shortages caused by macroeconomic imbalances) are no longer a threat. A small strategic reserve, with releases targeted at specific vulnerable groups during crises, would serve the purpose more cheaply than the Office des CCrCales' operations. 15. So the recommendation here i s that the state should transfer cereals importation and marketing to the private sector. The Office des CCrCales would concentrate on core public services: ensuring competition, maintaining a security stock and providing technical support to market players. A private inter-professional group should take the lead indeveloping trading instruments, such as forward pricing and standard contracts. 16. Food safety in Tunisia is inadequate. For example slaughtering arrangements are a health hazard and pesticide residue testing is incomplete. To protect the Tunisian consumer and export prospects, Tunisia should develop a comprehensive food safety system. It should be applied equally to products for domestic consumption and export. i x 17. A number of Government interventions in produce markets systematically send the message to producers: go for volume and cheapness rather than quality. Government controls retail margins, pressures wholesalers to keep prices down, sometimes imports vegetables when prices go up, pays inadequate quality bonuses for cereals and caps the prices o f processed foods. Meanwhile, supply chains need more o f the public services that would help improve quality: research, training, phyto- sanitary and food safety regulations, and up-to-date norms. Professional organizations and support services 18. In developed countries, farmers depend upon producers' organizations for technical and marketing services. In Tunisia there are 201 CoopCratives de Service Agricole, over 3,000 Groupements d'IntCrCt Collectif (mostly dealing with water), 171 Groupements de DCveloppement de 1'Agriculture et de la PCche, 7 inter- professional associations andUTAP, the farmers' and fisherfolk's union. 19. There are success stories. Some wine and dairy cooperatives have provided effective marketing services to their members. The transfer o f downstream irrigation management from the public sector to Groupements d'IntCrCt Collectif has also helped improve operatio ns and maintenance. However, there is great room for improvement. Cooperatives' coverage is low, many are inactive, and the turnover o f milk, wine and vegetable cooperatives is falling. Neither UTAP, nor the inter- professional groupings, nor the Central Cooperatives are seen as representative o f their interests. As 2004 legislation redefines GICs and Cooperatives as GDAPs, they will lose their ability to earn revenues. 20. Government can make professional groupings more effective by redefining their governance arrangements so that farmers are genuinely in control. This would improve ownership and accountability, and dispel the conception that they are arms o f Government. It could also encourage GDAPs to become commercial players. This would involve revising the 2004 statute as well as capacity-building. 21. Research and extension are core Government services. They also are not realizing their potential to support quality-driven supply chains. This came out strongly from the background studies. There are two broad problems. One is that extension and research are not responsive to farmers' perceived needs, for example for practical post-harvest advice. The other is that extension and research are under- funded. Government should reform the management o f research and extension, institutionalizing farmer consultation and showing a firm commitment to performance-based budgeting. There will be little incentive for improvement until research and extension are accountable for giving farmers what they want. Natural resources 22. Natural resources are the foundation o f the agricultural supply chain. As in other areas, Government could improve its effective by focusing less on quantitative investments andmore on systems and incentives. X 23. Tunisia has led the region in innovative irrigation management. The delegation o f downstream management o f 150,000 ha to GICs has improved operations. The cost-recovery rate for operations and maintenance costs overall is 115%. Government is currently experimenting with two-part pricing, so that farmers feel the marginal cost o f water. Thanks to Government subsidies, 75% o fthe irrigated areauses water-saving equipment. 24. Some anomalies need attention, however. The 50% tariff discount for cereals distorts farmers' choices. Also there is no real reason for Government to subsidise the replacement o fwater-saving equipment when it wears out. 25. When it comes to resource-allocation the emphasis i s still on massive water- mobilisation programmes. 55% o f the MARH's loth Plan Budget is for water investment, and 43% o f that is for large dams. The large dams allocation for 2005 was equivalent to 61% o f the M A R H ' s entire recurrent cost budget. Given the consensus that extension and research are under-funded, this is unfortunate. Moreover, with 90-95% o f available water already mobilized, diverting even more water cannot solve the key issues in the irrigation sector: dam siltation, groundwater depletion, pollution, salinization, and low water-use intensity. 26. The key tool for addressing this bias is economic cost-benefit analysis. Government could define its objective (e.g. volume or stability o f supply) and work out the least-cost combination o f measures for achieving it. This integrated water management approach would need to be underpinned by new skills in the MARH, especially inago- and socio-economics. Resourcing the agricultural sector 27. The public and private sectors' ability to respond to the market will depend upon their capacity to mobilize andmanage resources. 28. The MARH's management systems are biased towards the management o f hardinvestment. "Soft" activities, such as research and extension, are losing out. The Ministrycan therefore build on the approaches of the Agricultural Services Support Project (ASSP) and Performance Based Budgeting to improve its management o f "sofYYactivities. 29. Government's approach to rural finance has been to pursue quantitative investment targets with subsidies, interest rate caps and debt forgiveness. But on the ground people have not behaved as intended. Private investment and farm credit are contracting rapidly, andthe loan arrears rate i s around 50%. 30. If the current approach is not working, there are no easy alternatives. The Government should study decision-making in the credit market, to understand why bankers and farmers do not act as Government intends. On the basis o f this study, Government may see how to harness the behaviour o fbankers and farmers to increase investment. One option may be to allow micro-credit institutions greater commercial xi freedom. Another may be to make the BNA's objectives and accountability more transparent. 31. Rainfall variation bedevils rural investment. Government may therefore wish to promote rainfall-index insurance. However, Tunisian farmers have not beenwilling to pay for drought insurance inthe past. The reason may be that they expect drought- and debt-relief from Government. Rural land 32. If one compares Tunisia with developed Mediterranean countries like Italy and Spain and the new EU countries, one can see where increases in farm labour productivity come from. Much o f it derives, not from higher land productivity, but from higher land-labour ratios. Out-migration and land consolidation are part o f that process. 33. InTunisia, landfragmentation andabsenteeism are causing concern. Holdings o f under 5 ha have risen from 41% to 53% o f the total land area since 1976. In response, the Agence Foncihre Agricole (MA) has been active inreconsolidating and titling (mostly irrigated) land at the rate o f 12,000-14,000 ha per year. But is land fragmentation really a problem for competitiveness? And, if it is, is it a deep-rooted social phenomenon or can Government intervene in land markets to prevent it ? Unfortunately, despite all the quantitative and legal analysis o f land tenure, there is very little understanding o f farmers' decision-making. Moreover, we only have anecdotal evidence o f how landtenure affects farming practices. 34. So the first step towards defining a landtenure policy is to define the problem from the farmers' perspective, by studying their perceived priorities. Based on the findings o f this study, Government could consider policy responses such as one-stop shops for land transactions, administrative simplification, an integrated land code, fiscal penalties andbringinginemphyteotic leases. Social aspects of agriculturalpolicy 35. Farmingis a source o f livelihoods as well as a source o f growth. Those who depend upon agriculture are often the most poor and vulnerable elements o f Tunisian society: the elderly, the unskilled, landless labourers, inhabitants o f less-developed rural areas and females left behind by male migration. One must understand how policy change will affect them. 36. This report has suggested the re-engineering o f institutions and processes to make them more responsive to farmers' needs. This approach would apply to extension, research, farmers' organisations, budgetary management and credit and land tenure institutions. But Tunisian farmers are a diverse group. As Government tries to become more responsive to their needs, there is a risk that the dialogue will be captured by larger farmers and men, to the exclusion of smaller farmers, women and the landless. Government should therefore ensure that farmer-Government dialogue (e.g. consultations andresearch) embodies the diversity o f stakeholders. xii 37. This report has also suggested that cereals tariffs and guaranteed producer prices should fall. The social impacts o f this reform, via food and labour markets, will be complex and heterogeneous. At this stage, however, we can make the general prediction that small cereals farmers and cereal farm labourers inthe North andNorth West are vulnerable to cereals price liberalisation. It may be difficult for labour to find an alternative to cereals farming. 80% o f these cereal farmers are over 40 years old, and few (only 15% in L e Kef) have a secondary source o f income. In order to understand the dynamics more precisely, a Poverty and Social Impact Analysis (PSIA) should accompany any cereals price reforms. 38. When it comes to mitigating the impacts o f reform, the emphasis should be on improving the targeting o f Active Labour Market and other social programmes. At the moment they focus on urban males with secondary education, marginalising the rural poor. 39. Countries like Mexico and Turkey have de-linked rural income support fi-om farm production. Inthis way, they achieve the political objective o f a controlled and predictable income transfer to rural society - but without the market distortions and fluctuations intransfer values that go with production subsidies. 40. A keypoint, however, is that well-designed ruralsocialprogrammesshould not&t be seen as mitigating the impacts o freform. They are desirableper se. Subsidizing outputs, credit, loan arrears andwater is inefficient and socially regressive. Most o fthe benefits go to better-off farmers. So better-targeted social programmes will achieve Government's social objectives more effectively than the current market distortions. xiii 1. INTRODUCTION A. OBJECTIVES OFTHE REVIEW 1. It is impossible to consider economic development in Tunisia without considering agriculture. It generates about 13% o f GDP and employs about 16% o f the labour force, contributing 25% of new jobs during the gthPlan period (1997- 2001), and about a halfo f Tunisianhouseholds' consumption is on food. 2. The preparation o f the 1lthand 12thPlans, covering the period 2007-16, is a historic opportunity for the Tunisian Government to take a fresh look at agriculture. Higher incomes, lower poverty rates and a new international trading environment present policymakers with opportunities and challenges that simply did not exist a decade earlier. Onthe threshold o fthe new planning period, the Tunisian Government i s asking fundamental questions. What makes agriculture competitive? I s agriculture contributing its full potential contribution to the nation? How can agriculture take advantage o f trade agreements with the EUandother partners? 3. In2004, therefore, after more than 20 years without a comprehensive World Bank study, the Ministry of Development and International Co-operation requested an Agricultural PolicyReview. The review was executed by the World Bank, with the Ministry o f Agriculture and Water Resources as its Tunisian interlocutor. Inaddition to contributing to the preparation o f the 1lth and 12thPlans, the report is intended to provoke a lively debate on the role o f agriculture inTunisia's future. BackgroundPapers 4. This report is based on 11 background papers, on topics identified in consultation with the Ministry o f Development and International Cooperation (MDCI) and the Ministryo f Agriculture and Water Resources (MARH).Support was received from the Food and Agriculture Organization (FAOKP) and the French Development Agency (AFD), without which the sector review would have been impossible. The studies covered: - 1 - Tunisia's agricultural competitiveness; cereals; fruits andvegetables; livestock products; fisheries; soil andwater conservation, forestry and rangelands; irrigation water; rural land; agricultural organizations; agricultural finance; and social aspects, including gender. The eleven studies have provided a wealth o f data, which is now available to the Government'. The purpose o f the present report, therefore, is not to duplicate the background studies with another encyclopaedic description o f the agricultural sector. It is rather to answer Government's request that we explore the "main themes" and see inwhich direction they mightreorient the policies o fthe Tunisian Government. B. APPROACH 6. The starting point for the review is the Tunisian government's own policy objectives, as expressed inthe Plan and other official statements. This report briefly describes the sector's past achievements in terms o f these objectives, and then presents evidence that it could do even better. This raises the question: how could Government help thefarm sector realize its potential more completely ? To answer this question, the study looks in detail at the different building-blocks o f agricultural competitiveness: trade policy, domestic policies and institutions, producer groupings andso on. 7. As the study examines each facet o f the agricultural sector, some common themes begin to emerge clearly: - That agriculture's output growth has beenimpressive. - That it has unrealized potential. - That the challenges o f the past are not the challenges o f the future: production is nowjoined by issues of quality and global marketing. Unfortunately the findings o f the fisheries study were not available intime to be incorporated into thls report. It was also felt that soil conservation, forests and rangelands were satisfactorily covered as separate issues by the background study andthat their inclusion inthis synthesis report would have distracted from its main messages. - 2 - - That meeting the new challenges o f quality and global marketing is a matter o f responsiveness to the market. Supply chains must be agile and flexible to give the modern consumer what she wants. - That the state's heavy and distorting presence in supply chains hampers their responsiveness. Facilitating the private sector to serve the consumer could help agriculture achieve its unrealizedpotential. - For the state, effectively facilitating the private sector means becoming more responsive and less prescriptive. 8. However, the objectives o f agricultural policy are not just about maximizing growth potential. They relate to social stability and food security as well. When the Tunisian state intervenes in the agriculture sector, the ostensible reason is often social. So there are trade-offs between the various objectives o f agricultural policy: some policies intended to maintain social stability and food security have negative impacts upon the sector's competitiveness. It is not the intention o f this study to prioritize one policy objective above another; social and economic aspects of policymaking are treated as o f equal importance. However, the study will argue that the targeting o f agricultural policy instruments to social concerns has created a mismatch between instruments and objectives. The result is that neither social nor economic objectives are being addressed optimally. 9. The study's overall approach therefore is to recommend policies that will deliver agricultural growth, by empowering the private sector to respond to consumer demand. At the same time, and in parallel, the study will recommend the mobilization ofpolicy instruments that are better suitedto addressing the needs o f the poor andvulnerable. 10. The aim at this point is to advise on the overall strategic direction o f Tunisian agricultural policy. With such a broad remit and without a recent history o f analytical work, it would be imprudent to propose a step-by-step blueprint for implementation. As the Government identifies specific areas for reform, the transition path will need to bemappedinmore detail, drawing on the experience o f other countries. - 3 - PAST PERFORMANCE OFTHE SECTOR 11. This section examines the performance o f the agricultural sector over the last 20 years. Its purpose is to recognize its achievements, and also to identify ways in which it may not be realizing its potential. This analysis will lead into the analysis of strategic policy choices insubsequent sections. A. FIVEYEARPLANPRIORITIES 12. The Tunisian Government executes agricultural sector policy within the framework o f Five Year Development Plans. The Plans' aims are expressed as quantified volumes o f inputs and outputs. Although Chapter 111,D on "Resourcing the Agricultural Sector" will argue that a new approach to objective setting could improve Government's effectiveness, the current Plan framework will be used inthis section to assess the sector's past performance. 13. It is important to recognize that Tunisia's agricultural policy has come a long way since the late 1980s. During these last two decades, Tunisia has pursued a twofold agricultural development strategy oriented towards economic growth and social stability. The goals were sustainable productivity, access to foreign markets andimproved farmers' livelihoods. 14. Duringthe 7thEconomic and Social Development Plan2(1987-1991)' Tunisia embarked on structural reforms, supported through the Agricultural Sector Adjustment Program3(PASA). The reform program aimed at improving the balance o f payments, public finances and employment. It addressed the roles o f public and private entities, the investment climate and sustainable resource use. A number o f sector reforms were implemented, mostly inthe irrigation sector, inorder to improve resource-use. New water saving technologies and tariff systems were introduced in order to encourage efficiency andreduce water waste. 15. During the 8th Plan (1992-1996), further efforts were made to enhance productivity. This involved investments in applied agricultural research and extension, and the delivery o f farmer support services. Reforms in land tenure, agricultural credit and fiscal systems w ere carried out to encourage private entrepreneurship. * VIIemePlande Developpement Economique et Social (PDSE) 3Programme d'Ajustement du Secteur Agricole (PASA) - 4 - 16. In the late 90s, increased access to international markets brought new priorities. The gth Plan period (1997-2001) was dominated by a new Association Agreement with the European Union, WTO liberalization and the creation o f the Arab Free Trade Area. This kick-started domestic reforms in price and subsidy policies. Input subsidies were cut and most commodity prices were liberalized. However, prices remained fixed for some crops such as hard wheat and soft wheat, external tariffs remained high and the Ministry o f Commerce maintained formal and informal controls on consumer prices. Meanwhile, Tunisia pursued parts o f its previous agenda: (i) hrther mobilization o f water resources; and (ii)increased self- sufficiency in products such as olive oil, cereals, potatoes, tomatoes, milk, red meat andseaproducts. 17. The 10thPlan (2002-2006) has continued what the 9th Plan had begun, while emphasizing private investment and smallholder agriculture's roles in social and regional development. B. SECTOR PERFORMANCE 18. Agriculture has kept pace with Tunisia's economic growth overall. Within an economy growing at average rate o f 4.3%, the sector has shown average rates o f growth o f lo%, 2%, 2% and 5.6% during the 7th, 8th, 9thand lothplans respectively. So the contribution o f the agricultural sector to GDP has remained on a flat trend, averaging 13.4% over the past twenty years. This is quite unusual. Often agriculture's share o f GDP falls steadily when a middle-income country grows steadily. 19. However, agricultural growth has shown large fluctuations, mostly because o f rainfall variation; The growth rate o f agricultural value-added varied from 8.6% during 1987-1991 to 2.1% during 1991-1996, averaging 3.5% over the 9th and 10th Plans Figure I.Agriculture'sshare of GDP 100% 90% I 80% 70% BO% 50% 40% 30% 20% 10% 0% 1980 1986 1990 1995 2000 2005 Years Source: WDI. IBRD - 5 - 20. Some o f agriculture's growth has come from land productivity increases. Average yields have shown substantial growth, estimated at 2.8% for the period 1989-2003, through greater use o f chemical fertilizers, mechanization and improved seeds, more skilled farmers, and the expansion o f irrigated cropping. Table 1.Productionlevels o fmain adcultural commodities (1000 T) 7th Plan 8th Plan 9th Plan 10thPlan* Cereals 1408 1607 1394 1867 Olives 563 740 915 628 Citrus fruits 240 209 223 229 Dates 71 77 102 118 Potatoes 197 217 305 326 Tomatoes 487 507 759 905 Meat 124 238 325 401 Milk 387 526 806 905 Fisheries 95 86 93 104 Source:Tunisia7".,8".,9"a nd 10" PlansEvaluationRmorts 'Authors estimatesbased on 2002-2005 average outputvalues at 1990constantprices 21. Looking at the record o f different commodities, cereals crops declined during the gth Plan period with two consecutive years o f drought (Table 1). Preliminary production estimates for the lothPlanperiod show a recovery. Tree crops contribute a major share o f agricultural exports. They consist principally o f olives, dates and citrus Plan. Dates followed with 13% .The share o f citrus fruits has been only 2%, andtheir fruits. The share o f olive oil intotal agricultural exports averaged 34% during the loth contribution to total commodities exports is decreasing. 22. The livestock sector has grown rapidly across the board. Milk and poultry meat have exhibited especially highannual growth. Incentive policies have resulted in a strong increase in milk production and a large decrease in milk imports, despite rising national demand. Local milk supply currently meets more than 90% o f domestic demand. Fisheries outputs, which account for 18% o f exports, have practically stagnated during recent years. Nevertheless, their export value has shown considerable growth. C. FOOD TRADE BALANCE 23. The lothPlan (e.g. volume 2, para. 1) explicitly endorses food self-sufficiency as a national objective. Section 1II.A o f this report will argue that the pursuit o f self- sufficiency for its own sake cramps growth and is an inefficient way o f promoting - 6 - food security. However, is informative to assess progress against the self-sufficiency goal that Tunisiahas set for itself. 24. Tunisia's food trade balance has remained negative during the two last decades with the exception of the years 1999 and 2004. It i s chronically dependent on cereals imports, which represents about 40% o f total food imports. However, the overall coverage rate has risen from about 64.8% during the period 1987-1991, to 69.1% in 1992-1996,81.5% in 1997-2001 and 78.4% during2002-20064. The reason i s that Tunisian milk and sheep meat now cover domestic demand. D. CONSOLIDATIONOFAGRICULTURAL EMPLOYMENT 25. The sector's share in national employment has significantly decreased over the years. From 46% in 1960 and 23% in 1995-2000, the sector now contributes to only 16% o f national employment. During 1989-1994, net employment creation was negative inthe agricultural sector. In 1994-1997, however, its gross contribution was estimated at 23.5% o f total new employment creation and increased to 25% in 1997- 2001. In other words, over the last 12 years agriculture has created twice as many jobs per unit o f GDP as the economy as a whole. E UNREALIZED POTENTIAL 26. So one could fairly ask whether there i s an agricultural policy problem in Tunisia at all. With such a respectable growth and employment record, it seems at first sight as though the status quo i s perfectly satisfactory. But if one looks beneath the surface, one finds plenty of evidence that Tunisian agriculture is performing below its potential. 27. What does "potential" mean in this context ? One can think o f it as cost- competitiveness: the ability to produce a commodity more cheaply than competitors. Then "realizing potential" means actually increasing the value of competitive commodities purchasedat home and abroad. 28. To identi@ Tunisia's agricultural potential, therefore, the MARH's Carte Agricole data was used to compare economic production costs ("domestic resource cost") with border prices. Ifthe economic cost i s higher than the border price, this means that Tunisia would not be a competitive exporter and would save resources by importing this commodity. Ifthe border price is higher than the cost, this means that Tunisia i s potentiallv a competitive exporter - providing that quality and supply chain issues are addressed.The main findings are as follows: 4The lapse inthe food trade balance in2002-2006 is principally due to the 2002 drought, which resulted in a coverage rate o f only 48.7% inthat year. - 7 - * Trcecrops, ~ ~ andi ~e~etabfesarc ~otentia~fycompet~t~9e.The dorncstic ~ s resour~ecost of prod~cing~regetab~esand rnelons ranges from IO% to 5594 of the border price. For treecrops it r a n ~ from e ~ ~~~Q to 85%. I Dairy and beef p r o d u c ~ i ~are ~ n c o ~ ~ e ~because ~ofe ~ n i t i ~ high domestic on costs and 0 rt ~~bs~dies.Tti d ~ c ~ini o mi~ a ~ 125o/no f border prices, The mo which costs range from 175% e o ~ ~ c with ta domestic reso~rcecost of ~ i i ~ ~ ~ 65% ofborder prices. The ~ ~ m ~ e t ~ ~of~cereals ~ t e sest in the north and west where ~ ~ e s most fasourable d areas, Hard wheat is most Xi vc and soft wheat md barley least likely. The larger the f a n size, more conipccitive cereals ~ r o ~ u c t iisnlikely to be. o On this basis \vc whether Tunisia is realizing its g o are signs t is not. 29. Firstly, sonic of a ~ r i ~grot$l ~ ~simply~ contes from ~ has ~ ~ and more labor to the same area o f land, This is unusual for s1 cconomy with slow p o p u ~ n growth, and may be b e ~ a ~ sthe ind ~ e tanable to a ~ s o rlabour s ~ fy growth.$ A g r i c u l ~ ~e r ~~ ~ i ~ l ~ ~by~ 20% r n between 1993 and 3002, whiIe there is no trend increase in labour ~ r o d u ~ t i ~ ~ ~ y (Figure 2). Over the gthand IOth Plan p e ~ o ~d a~~ oprodL~cti~iry a ~ r ~ ~ u ~didu r e u r in t not grow, but it rose 4.4% and 4.7% per year in m a n u f a ~ t u r ~andgservices ~ ~ ~ e s p e c ~ i ~~~ e~l y oBank, ~d ~ ~ This ais w o. ~ ~ because it mems that labour r ~ 4 ~ g , supply growth is djsgu~silow cfficiency growth, n ~ Figure 2. V ~ l ~ ~ - ~irx d~ dg ~ ~~I~ ~ ~c~ ~ ~ ~~ ~ ~~ ~ ~~~~~~ ~ - r ~ ~ ?990 1992 1994 I996 1998 2000 2002 2064 year Source: UrBi CDId a ~ a b ~ s ~ 31. ~ h j ~ dmuch of` a ~ ~ ~ 1 t ugrowths has bccn driven by s u ~ s ~ ~ande s ~ y . r ~ ' ~ ~ro~ectionfor c o ~ ~ i o d ~intwhich Tunisia is not c o n ~ ~ e t ~As~Table .1 showed, i ~ s t ~ ~ e beef and dairy p ~ o d ~ ~ac ct ~~o~~for ~much of the Sector% ~ a ~ u e growth: ~ ~ d u i ~ ~ ~ d ~ ~ ~ c ugrew~ by 47016 over 1 9 ~ ~ -of2which ~ ~ u r e ~ ~ over, a third (IS%> was thanks to beef and m ~ lBut,Tunisia is grossly ~ n c Q ~ ~ i pinethemiboth. It costs 1.8 to 2.6 ~ t i ~ ~ ~ times as much to p r ~ d u beef in T ~ ~ as~itsdoes to import it ~ ~ d c a ~ o ~2s~u l~t ~5 a ~ c ~ ~ a and 1.2-2.1 times as muchto ~ ~ omilk.~Socgrowthofthe beef andmilk industry is d ~ ~~~~a~~~a net toss to the Tunis~~n ion and much of Tunisia's a ~ ~ ~sectoru ~ ~ u ~ l ati ~ ~ o \ ~recordis s t h ~ ~ ~ ~ o ~ ~ . Table 2. The co Source: ldeacons~lt~ ~ 0 ~ ~ a ~ 32. fourth^^?, where Tunisia & c o ~ ~ e t i t i v ein, fruits aid vegetables, it is not ~ a p i t a i i ~ i non its advantage. As Table 3 shows, a l t h o ~ ~exports of fnlit and g g ~ ~ ~ ~ ~ ~areap be r~f oe ~s i well in vo~umetenns, the valuc o f exports in current US t n ~ dollars is going dowi, slot up 33. Looking at Tunisia's major export market, the European Union, these are more signs that the p o t e ~ t ~cao~m ~ e ~ ~at ~dv~e F ~in~fruits and~ vegetables is not i ~ ~ e being fully t ~ a n s l ~ ~intodexports JCNEA, 2 0 0 ~ hT~ .t ~ ~ iis~ using only 55% o f its e i a citrus export quota, and there are no exports of t a ~ ~ e ~ and s ~ n eclemen~inesat all because ~ ~ ~ famis~cansonly sa~ i~~ thc~focal market. Tunisia's apricot exports n ~ p y peaked at 338 tomes in 1998, c o ~ i ~ with d ~ e an EU quota of 2,240 tonnes. Tunisia only ttscs 1,000 tonnes of its 4,000 tonne EU quota for ~ o ~ nc~ot~o ~ ~ ~ ~ r (Theaprice. t e of fresh tomato delivered to the factory is ~ ~ i ~than rin Italy, Spain, Portugal and h e ) Tunisia is agro ically suitedto getting onds onto the market I month ahead of EU iers, and yet i s rriiahle to usc its 3 120tonne EU quota, Even the quotafor olive o story, is 21% untiscd. across the sector cts are u n d e r - p e r f o ~ i ~ g while its u n ~ o r n p e t ~ t ~ ~ e ,horticult d fisheries, the mast corn 34% ofg despite stafliTlgaut 53% o f 34. Table 43. Overall, ~ o ~ - c o ~ ~ pp ~e o~d~u ~t arev c~ ~ n ~ ~ ~ b40%t iofngrowth, it s u ~ even t ~ ~they u gwere only 26% ofo ~ t ~during 1 ~ 8 6 -(Table 53. ~ ~ ~ u t 9 ~ - 10- % Of 1986-1990 % of growth between output 1986-90and 2001-5 Treecrops I 28% 19% Horticulture I 16% 16% Fisheries I 9% -1% .Other 3% 1% Source: MARH Products % of output in Shareofgrowth 1986-90 1986-90 to Competitive Hard wheat (50%), 73 2001-560 treecrops, horticulture (exc. Potatoes), small ruminants, fisheries, other products (50%) Non- Soft wheat, non-wheat 27 40 competitive cereals, potatoes, beef, milk, other products (50%) What prevents Tunisia from realizingits potential, and how can Government helpthe sector to performbetter? 36. Box 1 presents the cases o f citrus and olive oil. As it shows, much o f the answer lies in the fact that supply chains are not showing the agility and responsiveness to give consumers the quality produce that they want to pay for. This leads to the question o f how Government can best help agricultural supply chains to become responsive. In the cases of citrus and olive oil the answer i s the same: Government should focus on supplyingcore public goods, remove market distortions, and encourageprivatebusiness associationsto seek profit by developing technologies andmarkets. 37. Despite its difficulties with quality, the olive oil sector is proof that supply chains can respond to better Government policies. In 1994, when Government removed the ONH's monopoly on olive oil exports, the private sector responded. The ratio o folive oil export earningsto olive productionhit new levels after 1994 38. Figure 4). The value o f olive oil exports in 2004 was USD 569 million (FAOSTAT), or 6% o fthe country's total exports. - 11- Box 1. Unrealised potential incitrus and olive oil Unrealisedpotential: citrus The citrus supply chain illustrates how Tunisian supply chains could capitalize more on their cost-competitiveness. Citrus production has stagnated for more than a decade and growing domestic demand absorbs more than 90% of local production. Current citrus exports to the EU amount to 20,000 tons and represent only 50% of the country's preferential quota. To take advantage o fthis opportunity, Tunisia needsto increasethe quantity and quality of production. Tunisia citrus fruits are graded as being o f "average" quality. Many citrus orchards are old and unproductive. The conversion o f old orchards into younger and more productive farms i s slow. Yields are low and fruit are too small to get goodprices. Negligence at harvest is damaging fruit. Fruitsthat are tree-harvested and those collectedonthe ground are often mixedtogether. More efforts should therefore be made in applied research and extension service to develop appropriate harvest and post-harvest techniques that ensure high fruit quality for exports. These techniques must be developed for all stages of the supply chain and be easy to implement by citrus growers, fruit processing centers and traders. Despite price liberalization, retail margins are still regulatedby a 1988 decree: retail margins are set based on official purchase prices. Fruits growers and collectors are required to sell their produce to the official wholesale market while fruit retailers must purchase their goods at the same market. The permitted retail margin i s low, encouraging retailers to avoid the formal wholesale market and directly purchase fruit from local producers or collectors. This trend is coupled with quality-damagingpractices where fruits of all quality levels and sizes are mixedand sold as a whole regardless o f size and quality differentiation. What would improve the quality of Tunisian citrus ? o Government should focus on "public goods" like quality regulations, applied research and extension onpost-harvest handling. o Government should reduce its market-distorting interventions. Retail margins should be set by the market instead of being state-regulated. This would encourage private farmers and traders to go for quality. o Private sector associations should give producers and traders the technical and market information they need to supply what the consumer wants. Unrealised uotential: olive oil' Tunisia i s the world's fourth producer o f olive oil behind Spain, Italy and Greece and exports more than 70% o f its production. Olive oil prices depend on quality. While virgin olive oil, the highest olive oil quality, currently represents about 70% o f the international market, Tunisian oil exports mainly consist of the lowest grade o f such virgin oil, the lumpunte olive oil, which amounts to about 60% of its oil exports. Only 25% to 30% o f its oil i s high-quality compared to rates o f 70% to 86% observed in European competitor countries. This low rate i s the result o f various factors such as: (i) inappropriate harvest, storage and transport methods; (ii) long harvest and storage cycle; and a (iii) obsolete extraction equipment, which accounts for 42% o f capacity. Olive mills with modem 'Source: CEPI(2000) - 12- e y u i p ~ e ncurrently represent 31% compared to 70%~to 80% t n Italy, Greece and Spain. The ~ result is that about 97% of Tui~is~an olive oil ts still traded unbra~dedand in bulk. While $2 origin ~ a d e ~ a r cxist in Spam, 32 in Italy and 27 in Greece, Tunisia ol~zeoil prod~i~ers k s have deve~o~edorigin ~ a ~ e ~andrquatity labels.. no a k The cxisttng yuality testing labora~ories(three ~ a ~ o r a ~ ~ rTunis, i e s : Soussc and Sfax) ~ e ~ o togthe n s ~ a ~ e - oOffice~National des Huiles ~~~~~, ~ ~ n d which it makes available so pnvate compan~eson a ~ ~ e - ~basis. i EXowever, tt is said that in practice very few p r ~ ~ a ~ e ~ oliveooil ~ e d a y ~ ~ ~ ~ - ~ in%centres have accessto yu and control Facilities. tons ofexports to the EU. that a s not mads even if the ONH does not fix W ~ i ~a ~ ~iniproveuthe qualitydand p r o ~ t a b i lof~Tunisian olavr oil '? ~ ~ i o Again, ~ o ~ ~should~focus on ' ~~~ u~b goods" like applied research and e x ~ e n ~ono ~ ~ e ~ e ~ti c i a p p r ~ p r t~c~c~~en i ~ forethe harvest, storage and ~ a n of~olives.o ~ ~ s ~ o Again, ~ ~ ~ ~should~rcducc its dn ~ t ~ ~market ii~ n e~~ v e n ~Thenasl l.o ~ ~ t i ofnquota e ~ e o ~ ~ ~ o o shoutd be trans par en^^ A profitable export ~ o m r n o dshould~ not ~ross-~~bsidiseother ~ ~ activities. Quality analysts and control ~ a ~ i ~~~ t h i ebe~~ more~accessibledto the pnvate ~ sector. - x x Source: FAOSTAT - 13- 2002, but international experience shows that it could rise to 18% with the liberalisation o f foreign direct investment (Traill, 2006). Experience from other countries in the region shows that supermarkets will try to reduce uncertainty by centralising procurement and shifting from market- to contract-based purchasing (Codron et al., 2004). 40. The rewards for producers who can penetrate this market are high'. But "major changes in the global agro-food market are tilted toward those economic agents who have capital, superior organizational and marketing skills, and an acute understanding o f consumer requirements" (World Bank, 20040. So the challenge is to stay in the game, because there i s no place for second-best, the small wholesaler, the unreliable supplier, the consignment o f mixed quality or bureaucratic delay. As supermarkets expand in Tunisia and its export markets, the reward for excellence - andthe cost o fmediocrity - will increase. 111. CHALLENGESAND OPPORTUNITIES A. POLICY AND INSTITUTIONALFRAMEWORK 41. Chapter I1 concluded that Tunisia's agricultural sector, despite its growth record, is performing well below potential. This section will examine how the state intervenes in agricultural markets, and how reforms could boost agriculture's contribution to the national economy. It begins by looking at trade policy, which helps set the prices that farmers receive. It follows with an analysis o f how trade policy reforms would affect the national economy, consumers and different types o f farms. It then looks at government interventions in domestic agricultural markets. The section concludes that specific trade policy measures could boost growth whilst minimising employment impacts, and that government can improve agriculture's efficiency andcompetitiveness byrecastingits own role inthe sector. 42. This section focuses upon how trade policy and other Government interventions affect prices. This is because, when it comes to making supply chains more responsive, one set o f reforms, removing price distortions, matters than others. Allowing markets to set prices freely i s an "apex reform" because it makes other reforms more effective. For example, the more the price system rewards farmers for growing high-quality vegetables, the more vegetable-growers will take advantage o f improvements in water-management, land markets, farmer groupings, extension services and so on. 8For example, the producer price o f Gazan strawberries and cherry tomatoes destined for the EUcan be 4- 6 times hgher than the producer price for local sales. - 14- Trade Policy 43. For decades Tunisian trade policy has encouraged import substitution through high import barriers, with special export-promotion regimes and direct government involvement in commerce. Inthe case o f agriculture, this general approach has been overlaid with policies aiming at self-sufficiency, low prices and protection for certain farmers. Since the 1980s structural adjustment has brought macroeconomic stability. But reforms "have not substantially liberalized trade" and the trade regime remains "in needoffurtherliberalization" (WTO, 2005, p.vii). Import Policy Regime 44. In agriculture, Tunisia's applied tariffs are high by regional and world standards. Applied rates currently average 67% for agricultural products (compared to 31% for all products - see Table 6), with a maximum rate o f 150%. The current average rate i s about double the average at the time o f the WTO's last Trade Policy Review in 1994, although much o f the increase is associated with the "tariffication" o f non-tariff barriers since 1994. "Peak tariffs" (those over 15%, as per the WTO definition) account for about 69% o f agricultural tariff lines, compared to 59% for non-agricultural products. Among agricultural product categories, the highest overall tariffs (averaging around 100% in 2004 and 2005) are in the "exportable" fruits and vegetables sub-sector (Table 6). Protection o f these products in the local market not only reduces consumer choice', but also impedes the development o f a quality- oriented sector better able to compete inglobal markets. 45. Bound tariff rates (that is, the c e i h g rates above which Tunisia has committed in the WTO not to raise its applied rates) are much higher than applied rates, averaging close to 117% for the product categories agriculture and food. This creates considerable uncertainty as to whether applied rates might be raised in the future. The average bound rate for these two product categories is around twice that o f the thirdhighest category, clothing and textiles (WTO, p. 29). 46. Reliance on non-tariff barriers has been reduced over time: only 6.7% o f imported goods currently require a license, compared to around 13% in 1994, and most o f the remaining requirements are ostensibly imposed for technical reasons (WTO, p. 39). But imports o f about 1.6% o f all tariff lines - including a number o f important agricultural products - are under a tariff quota regime (WTO, p. 75), with above-quota tariffs averaging 98%. ( See Table 7). The quotas are allocated by an inter-ministerial committee in a procedure that is discretionary and less than fully transparent (WTO, p. 75). 9For example, Tunisian tomato paste manufacturers pay a higher price for raw tomatoes than their counterparts inItaly, Spain, Portugal and Turkey (CNEA, 2005b). - 15 - Table 6. Tariff Structure, 1994,2004, and2005 (Percentagesand US$million) PreferentialTariff for MFNTariff EU 1994 2004 2005 2005 Imports2004 Description Simple Duty Simple Duty Simple Duty Simple Duty (USSmillion) average Average average duty range duty range duty range averageduty range Total 30.7 0-43 32.7 0-200 31.7 0-150 18 0-150 12,727.10 Agriculture 35 10-43 69.3 0-200 66.8 0-150 63.6 0-150 1182.7 Live animals: animal products 38.5 10-43 97.4 20-180 91.8 20-150 91.8 20-150 40.5 Dairyproduce 23.5 1543 92.1 15-180 95.3 15-150 94.2 0-150 42.2 Coffee: tea, cocoa, sugar, etc. 38 15-43 73.1 0-200 72.1 0-150 67.7 0-150 178.5 Cut flowers andplants 27.3 20-43 37.3 0-180 36.9 0-150 34.7 0-150 10 Fruits& vegetables 42.1 15-43 104.3 0-200 96.8 0-150 94.6 0-150 61.7 Cereals 29 15-43 45.3 0-100 45.3 0-100 45.1 0-100 342.3 Source: World TradeOrganization, Trade PolicyReview Body WT/TF'RfS/152, September 7,2005 TradePolicy Review: Report by the Secretariat, Tunisia - 16- Table 7. Imports under tariff quotas, 2001-04, and customs duty rates, 2005 (Percentages,unlessotherwise indicated) Tariff quota 2001 2002 2003 2004 Descriptionof Product Average Average commitment Utilization Utilization Utilization Utilization quota rate Out-of-quota rate (tomes) rate rate rate rate ~~ ~~~ ~ Calves andbullocks 27 82 3,000 0 0 0 97 Bovinemeat 27 88 8,000 0 0 0 100 Sheepandgoat meat 21 125 380 0 0 0 100 Milkpowder 17 76 20,000 10 19 35 43 Butter 35 100 4,000 60 88 48 49 Cheese 27 139 1,500 100 100 100 100 Beans 25 60 1,300 0 0 0 0 Durumwheat 17 73 300,000 100 100 0 0 Soft wheat 17 13 600,000 100 100 85 100 Barley 17 73 200,000 100 100 100 100 Sugars 15 42 100,000 100 100 100 228 Shelledalmonds 43 60 1,335 70 100 0 0 Tomato concentrate 43 100 155 0 0 0 -100 Source: WorldTrade Organization,TradePolicyReviewBody WT/TFWSIl52, September 7,2005 Trade PolicyReview:Reportbythe Secretariat, Tunisia 47. InTunisia, as inmanyother countries, the highprotection givento agriculture i s intended to compensate for the protectionism o f other countries - especially industrialized countries. The argument i s that global liberalisation will eventually raise world prices, and until then temporary protection will avoid adjustment costs and irreparable damage to the production base. While this argument is dubious from an economic perspective, it has political appeal. But in any case, if this reasoning i s applied, the magnitude o f the tariffs applied should at least be grounded in solid analysis. A wide range o f modelling work indicates that for most commodities the price increases with global liberalization are expected to be o f the order o f 10% or less, although there is a fairly wide variance in these estimates*'. For some o f the most distorted markets - sugar and dairy are good examples -- the increases from full liberalization would be larger, but these are unlikely to be fully liberalized soon in any case. So tariffs should be fairly modest if they are meant "compensate for distorted world prices." Clearly, the high levels of protection in Tunisia cannot be justified on this basis. lo acommodity-by-commodity analysis For ofthe effects of liberalization inglobal agricultural markets, see Aksoy andBeghin (2004). - 17- Export Promotion Policies 48. Inthe 1990s, Tunisia liberalized its export regime and eliminated taxes on exports o f olive oil, fruits and vegetables, hides and skins, and cork (WTO, p. 55). But a number o fproducts canbe exportedonly with the permissiono fthe Ministryo f Trade. These include cereals and cereal preparations, forage and animal feed, molasses, and oilseed cake (WTO, p. 55). The aim is to keep down the domestic prices o f these goods, some o f which are subsidised, but in practice it can give Tunisian exporters a reputation for unreliability, thereby reducing their competitiveness. Export subsidies are widespread, generally as a 50% subsidy on transport costs (Ideaconsult, 2005b, p. 32), administered through the Fonds de Promotion des Exportations (FOPRODEX). FOPRODEX also provides other aids to exporters, including market surveys, staff training, and subsidies (for 3 years) for the salaries o fnew graduates who work for exporters. Institutional Framework 49. Several state trading enterprises play important roles inagricultural trade. The Office des Chkales has monopoly rights on the import o f durum wheat, soft wheat, and barley, and may if it chooses also import maize and soya cake. The Office du Commerce has a de facto monopoly on the importation o f sugar, and also imports several other food products, including tea, coffee, and potatoes. The Office National des Huiles imports edible oils and exports olive oil. It does not have sole legal rights to olive oil exports, but does control access to EUquotas, and uses this to maintain its virtual monopoly o f this market. The "rents" from controlling access to high EU prices cross-subsidize its other activities, and the lack o f transparency inthe way that quota is allocated betweenproducers may discourage investment. 50. Slow customs procedures have been identified in several studies as a major barrier to trade and to increasing Tunisia's international competitiveness' * .The World Bank (2004) notes that clearance for many products involves multiple institutions, with physical inspections and technical controls at several stages. Agricultural and food products -- which have the highest incidence by far o f technical import control requirements (WTO, p. 46) - are particularly susceptible to these problems. The length o f time to clear goods from their arrival in Tunisia until their release currently ranges from 7 to 20 days; the Bank's Export Development Project I1 aims to reduce this to 3- 7 days. l1WTO (2005) cites Cahiers de Z'IEQ,"Competitivitt de l'tconomie tunisienne", no. 18, Jan. 2004 ;and proceedingsof the Third Meeting o fthe EU-Tunisia Association Council, 29 Jan 2002, available at http://wwweuroparl.eu.in~meetdocs/delegations/~g~200203 13/012.pdf - 18- Box 2. The WastefulnessofFarmPriceProtection Diaz-Bonilla, Diao, and Robinson o fIFPRIreport the results o f an interesting simulation exercise inwhich a strategy o fprotection through trade policy is compared to a strategy o f agricultural investment as means to promote agricultural development and food security. They use a general equilibrium model with47 countries and regions and 37 sectors o fproduction. Inone scenario, they model the effects o f a 50% tariff on grain imports imposedby all developing countries and regions. This tariff stimulates productiono f those products, while acting as an implicit tax on consumption. Inthe second scenario, the implicit tax is transformed into an explicit tax, andthe proceeds usedto fund increased agriculture research. Using rates o freturn to such research inspecific countries and regions (which have beenestimated on the basis o f historical experience), they model the effects on overall welfare, GDP, food consumption, labor employment, and agricultural exports. By every one o fthese measures, the second strategy was hugely more successful inmeeting the objectives than the strategy o fprotection. While Tunisia was not broken out inthe model as a separate country, Morocco was, andby almost every measure, Morocco and the region o f "Rest o fNorth Africa" were among the biggest losers from protectionism. 51. To decide whether liberalisation makes economic and political sense it is usehlto understand the impacts androughly how bigthey will be. It is also important to know who will win and who will lose. That is what the next section o f this report will analyse. The cost o fprotectionism-the effects o f liberalisation 52. The result o f high rates o f protection is that food prices on Tunisian markets are significantly higher than those on world markets - even ifthe domestic market is apparently liberalised. We can take the example o f two commodities, wheat and tomatoes, one controlled by the government and one freely traded within Tunisia. - 19- Figure 5 shows movements since the early 1990sinthe wheat market. While the EU has droppedits producer prices to world prices, the Tunisian wheat producer price has remained well above the world price, being 43% higher in 2002 and 2003. For tomatoes, Tunisian processingplants are paying more than their counterparts inItaly, Spain, Portugal and Turkey (CNEA, 2005b), even though it is supposedly 55% cheaper to grow tomatoes in Tunisia than to import them. So strong tariff protection is enough to keep prices high, whether or not the product is competitive and government i s fixing prices. 53. Highproducer prices might appear beneficial in a country that is concerned about rural welfare. However, there are important costs attached to farm protection and these costs usually outweigh the benefits. Annex 1 explains in economic terms why this is so. 54. Annex 1 also describes how this study used economic models to estimate the effects o f protection. Readers who want a detailed technical explanation o f the models may refer to Ideaconsult (2005), which is the basis for the analysis in paragraphs 55 to 80. - 20 - Figure 5. Tunisian wheat is more expensivethan imports Wheat: buying pricefor domesticwheat, implicitimportprice, implicitexport prlceforselectedexporting countries. $250 $200 $150 $100 $50 $0 1998 I999 2Mx) 2001 2002 2W3 2004 2005 hplcnprceof imparts ($4) 5154,l $122.5 $116.1 $136.4 $145,8 $159,9 $ I S $ ` k h p i c t ,prce of exportsby $69.7 $129,9 $129.2 $81,7 $82.2 $89.5 113.1 - Maine (W) -& hplicRprce of exports by $142,7 $123,4 $1142 $115.3 $120,9 $141.3 $171.4 - a hpllcdprceofexportsbyLE -France (W) $137.5 $125.9 $121.5 $131.2 $149.8 $155.7 $164,0 Wtl -Buying price for domstk $208.0 $2208.0 $2208.0 5208,O $208.0 $208,0 216 c o m nw heat Sources: FA0 database and MARH Impact of Agricultural TradeLiberalization on Consumers. 55. High farm prices make food more expensive for consumers. As Table 8 indicates, foodstuffs remain by far the largest category o f expenditure for Tunisian households. They account for 38% o f household spending for the country as a whole, and44% for "non-communal" households (i.e. those that are not locatedindesignated municipalities - the difference is probably due to lower incomes and cheaper housing inrural areas.) 56. Food will be a big part o f household spending for a long time. The share o f foodstuffs in household expenditures has fallen very slightly over the last twenty years, from 41.7% in 1980 to 38% in 2000. Real per capita household expenditure rose by 42% over the same period. This i s inline with global estimates o f the income elasticity o f demand for food for middle-income countries, which are generally inthe range o f -0.5 to -0.7 (Seale et al. 2003). -21 - Table 8. Breakdown Source: National survey onhouseholdbudgets, consumption and living standards, INS (2000) 57. Lower tariffs would mean cheaper food -both Tunisian andimported. The net effect o f total liberalisation would be worth around 4% o f consumers' expenditure, o f the same order o f magnitude as a bonus year's growth. As Table 8 illustrates, moreover, food expenditures are higher than average proportion o f low-income households' expenditure, so the benefits o f price liberalisation would be greatest for the worst-off households. If consumers reallocate their spending in response to changes in relative prices, the benefits o f liberalisation are higher: 5.6% - 5.8% o f spending. 58. Part o f the benefit to consumers from price liberalisation derives from developed countries' farm subsidies, which depress world food prices. If the EU were to liberalise at the same time as Tunisia, the gain to the Tunisian consumer would average 2.6% o f expenditure instead o f 3.8%. 59. A few points should be borne inmind when interpreting the figures in Table 9. Firstly, the model assumes that consumers are actually affected by changes in producer prices. However, the consumer price o f wheat is controlled by the state, through the Office des CCrCales. So long as this policy stays in place, therefore, the benefits from a lower Tunisian producer price for wheat would go to the taxpayer, not to the consumer. Secondly, some domestic food prices, inparticular meat prices, are lower than world prices. This is not because o f trade policy but because Tunisian produce is o f lower quality than internationally-traded commodities and because o f a variety o f obstacles to exports. In assuming that farm price liberalisation would equalise domestic and international prices, the model assumes that these domestic food prices will actually rise. In reality this would not happen, because of quality constraints, so Table 9 underestimates the benefit o f liberalisationto the consumer. - 22 - Expenditure class % o fTunisian % o f Benefit fi-om (DTheadyear) population household price expenditure liberalisation allocated to (as % o f foodstuffs expenditure) 400 9.3 53.5 4.3 400-600 14.7 51.0 4.5 600-800 14.7 49.1 4.6 800-1200 22.8 45.1 4.4 1200-1600 13.9 41.2 4.3 1600-2400 13.4 36.6 3.8 2400-3600 6.8 31.5 3.4 >3600 4.4 31.3 3.5 Overall 100.0 38.0 3.8 Impacts of Agricultural TradeLiberalisation on the National Economy Another way o f capturingthe costs o fprotectionism is to look at its effects on economic growth. Protecting agricultural prices encourages the allocation o f labour and capital to farmingwhen it could be usedmore efficiently inother sectors. Bymakingfood more expensive, it also raises the cost o f labour, which discourages employment andmakes the economy less competitive inexport markets. The CGE model (see Annex 1) therefore estimates that farm trade liberalisation will raise GDP by 0.8%. 60. Table lo), shifting around 87,000 jobs out o f agriculture and producing total benefits o f 7.1 billion DT over 25 years. This leaves a positive net benefit o f over 6.1 billion DT over 25 years, when adjustment costs o f 1.0 billion DT12 have been factored in. Table 10. Effects o f amicultural trade liberalisation Agriculture contracts 1I I 1.4% o fGDP IRest o f economy expands 2.2% o f GDP Overall gain 0.8% o f GDP Source: Ideaconsult (2005) 61. The rise inGDP is 0.5% instead of 0.8% ifEUagricultural subsidies are lifted at the same time as Tunisian liberalisation. This i s because Tunisian consumers would no longer benefit from food imports subsidised by the European taxpayer. What this model suggests therefore, i s that other countries' protectionism i s not an ''Thsis the estimated cost of retraining labour and the upgrading [mise A niveau] of investments. -23 - economic argument for Tunisian protecti~nism'~.In any case, as the earlier trade policy analysis mentions, the removal o f global farm subsidies would only push world prices up by around 10% for many commodities. So Tunisian protection levels are far higher thanwhat is needed to countervail them. Variables andparameters Baseline Total liberalisation scenario scenario With EUagricultural Withremoval ofEU subsidies agricultural subsidies Economic growth (%p.a. in 5.7 6.5 6.2 year following liberalisation) Agriculture labour (% o f 20.2 17.4 18.3 employedpopulation) Agriculture labour movingto 87 67 other sectors ('000 jobs) Adjustment costs (million 984 874 DTto 2025) Adjustment gains (million 7107 4441 DTto 2025) 62. Half the benefits o f full liberalisation can be achieved by liberalising field crops (essentially cereals). Cereals liberalisation also has a very low cost interms o f farm jobs lost. Assuming that field crop liberalisation would raise production efficiency by 7.5%, the full liberalisation o f the field crop sub-sector would produce a one-off increase inGDP o f 0.4%14. This implies a gross benefit o f 3.6 billion DT over 25 years, or a net benefit o f 2.9 billion DT after the costs o f adjustment have been factored in. It would be achieved at a cost o f only 9,000 jobs, because field crops have a relatively low labour requirement. This means that the annual cost o f protecting onejob inthe cereals sector is over 4 times national income per head. 63. One reason that cereals liberalisation has such a small employment impact is that it encourages farmers to switch land into crops which use more labour. This has important implications for government policy. Preferential policy treatment o f cereals, such as raised prices, cut-price water and reserved access to state land, may beharmfulfor agricultural employment as well as detrimental to growth. 64. Infact, policyreforms that encourage the agricultural sector to shift towards higher-value h i t and vegetable crops can create off-farm employment by stimulating l3This is not to say that Tunisia should necessarily liberalise unilaterally under all conditions. Linking Tunisian liberalisation to other countries' concessions could be inprinciple an effective strategy for obtaining market access for products inwhich Tunisia has a competitive advantage. However, this consideration would only trump the benefits o f unilateral liberalization innegotiations inwhich Tunisia has substantial leverage on its own, which is probably not the case for the ongoing Doha Round. And, inany case, the bargaining chip o fthese multilateral negotiations is the bound tariff, so unilateral reduction o f applied rates does not diminishwhatever leverage a country may have. l4This figure includes an increase of 7.5% infield crop productivity, which works out at approximately 7.5% times the share o f cereals inagricultural production (12%) times the share o f agriculture inGDP (14%) i.e. 0.1%. - 24 - activity in the handling and agro-processing sectors. In Chile, for example, off-farm jobs related to agriculture contribute more than farming itself (Valdes and Foster, 2003). In contrast, the Tunisian cereals sector offers very little opportunity for additional post-harvest employment. 65. So far we have discussed the so-called "static" effects o f liberalisation, which derive from shifting factors o f production towards more productive activities. One would also expect "dynamic" effects: improvements in the competitiveness o f agriculture as farmers respond to the pressure o f foreign competition. Impacts of Agricultural TradeLiberalisation on the Tunisian Farm Sector. 66. But before policymakers will choose the benefits o f trade liberalisation, they need to know exactly where the changes will be felt. This will help them to understand the political and social consequences o f their decisions and to think about accompanying measures. So the linear programming models (see Annex 1) take the Carte Agricole data one step further, simulating the changes in cropping patterns and profitability, farm-type by farm-type, that would follow the equalisation o f domestic andworldproductprices. 67. Farms can be `divided into three categories: those benefiting from liberalisation, those whose profitability would be more or less unchanged and those which would be worse off. 68. The farm types classified as benefiting from liberalisation would have increases in gross margins [marge brute] ranging from 55% to 294%. They account for 41% o f Tunisia's farms by number and 30% by area. The predominant farm types in this group are those producing olive oil and sheep, off-season geothermal horticulture inthe Gabes region and citrus production inthe Nabeul area. 69. As noted above, however, the assumption that domestic farmgate prices will move towards world prices implies that some Tunisian prices will move upwards. This would probably not be the case inthe short runbecause it is quality rather than trade policy that keeps the price o f Tunisian olive oil, sheepmeat, fruit and vegetables below world prices. It would be safer to conclude, therefore, that these farms may be better o f f as a result o f liberalization. Inthe medium term, liberalization would be expected to leadto dynamic private andpublic sector investment inimproving quality o f export products. 70. The farm types classified as having their profitability more or less unchanged as a result o f liberalisation are modelled as having gross margin increases o f up to 47%. They account for 42% o f Tunisia's farms by number and 41% by area. They are typically mixed tree-crop / sheep farms o f the Centre and South and farms within irrigated perimeters. Infact, 60% o f farms with irrigation are inthis group. 71. The farms that lose out would experience falls ingross margins o f 1%to 79%. They account for 16% o f farms by number and 30% by area. However, only one farm type, accounting for 4% o f farm numbers and 5% o f area, would experience a - 25 - fall in gross margins o f over 34%. Farms in Category 3 are generally cereals operations in the north and north-west o f the country. This is the zone in which rainfall is most abundant. A highproportion o f government farming operations inthe state lands [terres domaniales] are cereals-based and in the fertile north and north- west, andwould therefore fall into this category. 72. This permits some conclusions about how the liberalisation impacts would be spread. 70% o f farming operations would at least not lose out from liberalisation, and may gain. These farms would be spread across the more arid centre and south o f the country andwould be producers o f sheep, olives, fruit and vegetables. The "winning" sub-sectors, livestock, treecrops and horticulture, account between them for around 60% o f agricultural labour use and are geographically dispersed. The farms that would lose out from liberalisation would tend to be cereal-producers in the better- watered north andnorth-west, including government farms on state land. 73. There is no correlation between a farm type's profitability and the impact o f liberalisation upon it. Farm types that would experience a fall in gross margin fall into all size categories; in this respect liberalisation is not specifically "anti-poor". (Before liberalisation the farms in the bottom quartile by farm gross margin would have a total gross margin o f around DT 100million. After liberalisation the farms in the bottom quartile would have a total gross margino f around DT 110million.) This is because the lowest-income farm types inthe Carte Agricole tend to produce olives, other treecrops, sheep andonly small amounts o f wheat andbarley. 74. The most vulnerable farm types would be those with low gross margins to begin with a fall in gross margin as a result o f liberalisation. The two "losing" farmtypes with gross margins o funder 7000 DT/year, are inKef andBkja. Together they represent 5.6% o f national farm numbers and 2.5% o f national farm area. It is also important to understand the form o f the adjustment that such "losing" farms and areas would face. A major shedding o f labour, leading to migration, is not expected: cereal production only employs 9% o f Tunisia's agricultural workdays, compared with over 40% for bothlivestock and fruit andvegetables, and 80% o fcereals farmers are aged over 40 years (CNEA 2005a). However, cereal-dependent areas such as K e f and Bkja would experience localised but significant falls in farm income, which would also affect demand for the local service sector. 75. Farm price liberalisation would result in an adjustment o f cropping patterns, as farmers respond to changes inrelative prices. Cereals land would spend more time as fallow, a change recommended by government agronomists (CNEA 2005a), and environmentally-fragile marginal land would be taken out o f cereals cultivation. Table 12 shows how the linear programming models predict that cereals production will respond to the simultaneous liberalisation o f f all farm prices; hard wheat would be least affected and soft wheat most affected. Soft wheat is ill-adapted to Tunisian ago- ecological conditions and hard wheat is a substitute for it in production, so the highlevel o fprotectionfor soft wheat isparticularlywasteful ineconomic terms. - 26 - 0 Sheep and fodder production would increase by 3% and 6% respectively. 0 There would not be any significant change in the overall area under horticultural crops, but within the horticultural sub-sector there would be a shift away from potatoes. 0 The profitability o ftreecrops will be maintained. Crop % change in % change in % change inarea price cropped area % change in I Hard wheat I -19 I -7 I .4 I Soft wheat -72 -50 .7 Barley -52 -27 .5 76. It is possible to derive from these findings a rough picture o f the winners and losers from cereals price liberalisation. The taxpayer would save DT 207 million per year on purchases o f cereals through the Office des CCrCales. 24% o f this saving would come from buying imported instead o f home-grown cereals. The other 76% would come from paying less for domestically-produced cereals. The cereal-growers, on the other hand, would see their profits fall by DT 182 million. So the net gain to the economy would be o f the order o f DT 25 million per year. This figure represents the cost savings from importing cereals that used to be producedmore expensively in Tunisia". 15A straight-line approximation o fthe cereals supply curve was used. The net welfare gain is therefore estimated as 0.5 x the change inquantity producedx the change inprice. Changes incropped area are assumedto be proportional to changes inproduction. - 27 - I Money savedto taxpayer (millionDT) Monev taxDaver I 207 I Profits lost to farmer (millionDT) 182 Net gains (million DT) pains 25 Source: Ideaconsult (2005) and World Bank staff calculations 77. The liberalisation of farm prices would produce an important benefit for livestock producers, namely a fall in the price o f feedstuffs. Livestock contributes 37.5% of agricultural GDP, and tends to be concentrated in the more disadvantaged arid areas o f the country. According to Carte Agricole data, fodder represents 50% of the cost o f raising a ewe. Hay accounts for around half o f fodder costs and barley grain for about a quarter. A 52% fall in barley prices following liberalisation would therefore reduce the cost of sheepmeat production by about one sixteenth. In addition, the linear programme models forecast that farm price liberalisation would lead to an increase in fallow land areas of 23%, and a 6% increase in forage cultivation, both of which would tend to lower the costs of grazing and feeding animals. The WayForward on agricultural trade 78. Reforming the trade regime would involve reducing tariffs and government controls andphasing out the commercial roles o f state economic enterprises. It would also be helpful to transform the tariff quotas into simple tariffs. This all could be done in a gradual process and without a high risk that import surges would severely damage local producers o f sensitive products. Tunisia has reserved the right under WTO rules to avail itself o f the special safeguard procedures for the products that are now under a tariff quota regime; this right could be invoked should the "tariffication" o f the tariff quotas be followed by the threat o f injury to producers. Dismantling the complex web o f state economic enterprises and subsidies would produce fiscal savings that could be channelled toward agricultural investments, and would also create more space for dynamic private sector development. 79. The Tunisian government can consider some farm price liberalisation options which offer a "soft landing": a good ratio o f economic gain to loss o f farm employment. "Soft landing" options are available because Tunisia is most competitive in products (treecrops, fruit and vegetables, sheep) which are labour intensive. A substantial phased reduction -- beginningimmediately and completed ina few years -- o f tariff protection for hard and soft wheat, accompanied by a matching reduction inthe guaranteed producer price (for as long as that instrument remains - 28 - inplace). More analysis would be needed of the best design and sequencing of tariff-reduction, price-reduction and privatisation o f the Office des Ckrkales' functions. The total liberalisation o f barley imports, and also o f livestock concentrate and its components. This would not only benefit livestock producers, but would also allow barley farmers to cultivate crops inwhich they are more competitive. 80. As explained above, it is expected that the employment impact o f these measures would be limited and localised. However, it is recommended (see section 1II.F) that the social impacts o f cereals liberalisation be verified by means o f a thorough Poverty and Social Impact Analysis (PSIA). Government should then assess the need for mitigatory measures on the basis o f the PSIA's findings. OECD countries have compensated farmers for reduction o f protection by direct, decoupled (and WTO compliant) area-based payments. Incountries where this is feasible, tariff reduction (and phase-out o f other subsidies) can be carried out quickly16. The experience o f Mexico and Turkey shows that this is pra~tical'~.However, inTunisia, affordability and institutional capacity (in particular, a land registration system) would need to be evaluated. Other policy reforms should be put in place to complement the trade policy reforms. However, some o f the most important reforms and investments may not be inthe agriculture sector itself, but rather would support rural non-agricultural employment, the integration o f rural-urban labour markets, and out-migration. 81. Finally, it should be emphasised that the "static" efficiency gains from f m trade liberalisation, like the employment costs, are modest compared with the size o f the Tunisian economy. The figures are measured intenths o f a per cent o f GDP. Far more important are the dynamic efficiency gains that could result from translating Tunisia's potential competitiveness intreecrops, sheepmeat, h i t and vegetables into actual growth and exports. This will involve making supply chains more responsive to consumer demand, particularly where quality is concerned. Supply chain responsiveness is a recurrent theme o f sections 1II.Bto 1II.Eo f this report). Domestic Policies andInstitutions 82. Chapter I1argued that Tunisia's agricultural performance is falling short o f its potential. This section will look at bottlenecks that are preventing Tunisia from responding to consumers' demands for quality and efficiency. The analysis will not examine exhaustively every aspect o f agricultural policy. It will focus on three areas l6 To the extent possible, this needs to be done with a credible commitment to keeptariffs low. Otherwise, it may set upperverse incentives for lobbyingto raise tariffs inthe fiture not only to collect the initial rents, but also to collect the compensation when the tariffs are later reduced. "Caremust be taken to ensure that these do notbecome permanent drains on the budget or distract from more productive forms o fpublic expenditure inthe sector. The mediumterm vision should be a phaseout o f such payments and integration o f poor f m e r s over time into a national targeted safety net system. Baffes and de Gorter (2004) examines lessonso f experience indirect income support, which canbe useful indesigning new schemes. - 29 - which are not only important in themselves, but also highlight the key issues for the sector as a whole: o Cereals markets o Animal healthand food safety o Produce quality Ineach case it is argued that the role o fthe state should be recast: a general strategic conclusionwhich the Tunisian government could then apply across the sector. Marketingof Cereals 83. The GOT'S cereal market interventions are meant to address two distinct objectives: a) a reduction in the level o f dependency upon imports through a guaranteed producer price; and b) a subsidised consumer price for wheat flour, bread andsemolina. 84. Farmgate prices for wheat andbarley remained unchanged for six years until 2005/06 at DT 290 per ton for durum wheat (US$ 232), DT 260 per ton for bread- makingcommon wheat (US$208) and DT 170per ton for barley (US$ 136). These prices are buying prices by agents o f the Cereals Board. They were adjusted in 2005/06 to DT 270 for common wheat (US$216), and TD 300 for durum wheat (US$ 240). This pricing system (i.e., pan-territorial buying price) for domestically produced wheat is directly administered by the Cereals Board either directly through its regional agencies (about 44% o f the total quantities collected) or through selected intermediaries, the "mandataires", mainly the "coopbratives centrales", with two o f them handling on average more than 42% o f the total quantities collected (CNEA, 2005a). Although it is illegal to market wheat through a trading institution other than the Cereals Board, about half o f the production remains outside the official channel (either kept for domestic use by the producers themselves, or marketed through informal or traditional trading channels, such as itinerant traders, local weekly markets, artisanal millers, etc.). - 30 - Figure 6. Quantities collected by the Cereals Board 12000 I4Oo0 B Barley Breadwheat Durum wheat Source: CNEA (2005a) 85. The Cereals Board also enjoys an import monopoly through its three seaport elevators at Bizerte, Tunis-La-Goulette and Gabbs. Tunisian private traderdmillers cannot directly tender for wheat intended to be consumed domestically. However, they can import wheat under toll millingandre-exportation arrangements. These arrangements ensure that Tunisian farmers earn much more than the world price for wheat (see Figure 5). 86. Since 2000 the Cereals Board has implemented payment according to technical criteria: "paiement Ci la qualite'". However, millers remain dissatisfied with the quality o f wheat. A substantial part o f the domestic wheat producers deliver low quality produce andthis does not seem to be satisfactorily compensated by the quality o f imported wheat. 87. The OC's costs are disproportionate to its turnover. One should keep inmind that the total quantities collected from Tunisian farmers by both the Board and its mandataires remains extremely small with an average over the past ten years o f 844,000 tons o f cereals (incl. 619,000 tons o f durum wheat, 159,000 tons o f common wheat). Even when adding the three port terminals and the 1.36 million tons ofwheat imported, this would notjustify an institution o f such size. 88. The fiscal cost o f wheat price compensation is also disproportional to the effects. This can be illustrated by splitting the fiscal cost (US$145m) between producer price support (US$39m) and consumer subsidies plus operating costs (US$106m). - 3 1 - 89. The US$39m producer subsidy was equivalent to US$20 per ton, or 12% o f the implicit price o f imports, in2004: down from 79% in2000. Large farmers inthe north have beenthe mainbeneficiaries. 90. Ifall ofthe remainingUS$106m subsidy benefited consumers, itwould work out at USDlO per person. This is worth around 2.5% o f average per capita food spending (see Table 8), or 1% o f mean per capita expenditure. Infact o f course, the benefits are even less because o fthe OC's operating costs. 91. Inother words, now that the wheat producerprice is closer to the landedprice o f foreign wheat, the whole superstructure o f the OC is becoming redundant. So Government can ask whether anyjustification remains for the current intricate system o f market intervention, the resulting absence o f market-based incentives to produce, process and use cereals, and the retarded modernisation o f marketing (both in hardware, e.g., quality control, transportation, storage) and in software (trading techniques, risk management, standard contracts, arbitrationmethods, etc.)? 92. Inshort, the administrative centralisationofthe marketing chain is harmful in many ways: (i) is costly for the budget, (ii) does not target the poor well, (iii) it it it discourages the private sector from restructuring (excess capacity is observed both at the milling and baking levels), (iv) it artificially favours the cultivation o f uncompetitive cereals at the expense o f other crops (see above) (v) it cramps the development o f competitive markets, and moreover (vi) it often has a detrimental impact on environment because it induces inefficient use o f scarce water resources. 93. The GOT has made modest reforms, opening the list o f mandataires to two small traders and to reducing the involvement o f the Cereals Board in the collection o fbarley. The total number o f employees o f the Board has been reduced from a total o f 2,800 to 1,500. It has mostly withdrawn from retail sales and indicates that it i s contemplating the privatizationo f a number o f its local facilities. 94. But there is no consistent movement towards reform. The monopoly o f sales to mills remains intact. Although 94% o f cereals are produced in the north, the Cereals Board is maintaining stores in all the regions o f Tunisia. The Board intends to keep its port facilities and is involved in the building o f a new port elevator in Zarziz to supply the Centre and South o f Tunisia. 95. Government's caution i s understandable, given the "bread riots" o f 1984. Nevertheless, the causes o f food shortages in the past (e.g., foreign exchange shortages caused by severe macroeconomic imbalances) are no longer a realistic threat. With reasonable macroeconomic policies and better systems to forecast drought and transport food, food insecurity now more commonly stems from insufficient purchasing power by poor segments o f the population than from any domestic shortages. - 32 - Animal Health and Food Safety 96. Severe weaknesses in the fields o f animal health, sanitary conditions and hygiene (in particular in the field o f animal slaughtering), and in quality control and food inspection needto be addressed. 97. For instance, informal large-scale slaughtering activities for poultry are maintained in the middle of cities. With one exception only, the Ministry o f Agriculture reports that no treatment o f waste and other effluents i s in place in the entire country for animal slaughtering activities. In order to address food poisoning outbreaks in tourist facilities, the Ministry o f Tourism has had to develop an ad-hoc task force to inspect hotels, restaurants, cafes, etc. intourist areas. 98. Another example o f defective food safety controls concerns pesticide residue testing. Government only tests for those pesticides which are approved for the crop inquestion. However, there is widespread use of unapprovedchemicals, such as the application o f systemic pesticides to vegetables (CNEA, 2005b), leading to ingestion bythe consumer. 99. It is dangerous and costly to have a dualistic approach to food safety: one dealing exclusively with foreign trade and the other with domestic markets. The prevalence o f unsafe food in consignments for domestic and foreign consumption can seriously limit Tunisia's access to international markets. Inpartnership with private sector organisations, therefore, the Government should develop a food safety and animal health assurance system, applied to products for export and domestic consumption inan integratedmanner. Produce Quality 100. As the Domestic Resource Cost analysis above has shown, Tunisia should already be cost-competitive in a wide range o f fruit, vegetable and small ruminant products. IfTunisia is not realising its export potential, it is because weaknesses on the quality and marketing side are limiting Tunisian agricultural exports, not crude cost (see CNEA, 2005b). 101. A number o fGovernment interventions andpolicies systematically discourage Tunisian farmers from producing high-quality commodities for the domestic consumer. Inthis way they encourage a dualistic market, inwhich highquality goods are produced only for export. This makes it difficult for ago-businesses to progress via the top end o f the domestic market towards developing exportable production. For example: 0 Formal retailers have to purchase from official wholesale markets. Formal wholesale markets have to pay a 12.5% tax, and observe controlled and inadequate retail margins - unchanged since 1993. The result is that a large proportion o f production (70% for citrus) avoids the official marketing channel, and formal retailers prefer to sell low-quality produce. This inturn discourages the emergence o f domestic demand for high-quality, well presented products, - 33 - which in turn would make it harder for Tunisian producers to progress towards export-grade quality. Ministry o f Commerce inspectors monitor pricing in wholesale markets and take measures to discourage highprices, e.g. by ordering traders to take a commodity out o f the formal market when the price is high to avoid its inclusion in official statistics. This creates uncertainty for farmers and traders, which i s particularly marked for the potato marketing chain (CNEA, 2005b). Such measures are unnecessary; because there i s patently a highlevel o f competition between sellers in wholesale markets, but discourage the trading o f high-quality goods which would command a higher price. Tax concessions for 100% exporters discourage companies from selling into local and foreign markets at the same time, which prevents the high-quality end o f the local market from becoming a stepping-stone towards export-production. This is an issue for dates, (CNEA, 2005b) and the same issue may arise for other commodities. Government support for quality (e.g. research, training, coordinated approach by private sector and state agencies) is under-funded by comparison with infrastructure. Top-down budgetary procedures prevent organizations from being flexible and responsive, for example by concentrating more research on post harvest issues. Quality norms are underdeveloped. For example, the single norm for citrus h i t dates from 1985; the tomato norm dates from 1983; Tunisian tomato concentrate norms do not correspond to those in the Codex Alimentarius (CNEA, 2005b). Again this would encourage a dichotomy between production for the domestic and foreign markets, which discourages agro-processors from making the transition to exporting. The Office des Ckrkales quality bonuses are insufficient, with the result that many consumers prefer to buy more expensive but better-tastingblack market couscous through inefficient small traders. The Ministry o f Commerce's supervision o f retail prices for processed goods (e.g. tomato concentrate and milk) feeds back to depressed "reference prices" for raw products (e.g. tomatoes and raw milk), and discourages the payment o f higher prices for better-quality produce. Government support for exports focuses on reducing the cost-to-market for Tunisian exporters. For example, FOWRODEX subsidises freight costs and COTUNACE subsidises exporting costs. 102. Taken together, such measures send a clear message to the producer: go for volume andcheapness, but not quality. - 34 - The Way Forward: Redefining the Role of State vis-&vis the Private Sector 103. The preceding sections have analysed how government interventions have confined and distorted the agricultural sector's growth and competitiveness. By planning, steering and sometimes supplanting the activities o f private farmers and traders, the state i s cramping agriculture's and ago-industry's expansion. The challenge now, therefore, i s for the state to reposition itself. Experience in other countries suggeststhree mainroles for the State: to design and enforce the legal framework to ensure an efficient finctioning o f competitive markets for products, services and factors (finance, land, labour); to protect the population's health, natural resources andthe environment; and 0 to address issues with large externalities Le., supplying products or services with a public goods nature. 104. For example, in the cereals sector it is now appropriate for government to transfer importation and marketing to the private sector. The Cereals Board would then concentrate on providing core public services that the private sector cannot deliver, ensuring that: no collusion nor market disruptions are observed; modem methods o f trading, transporting, storing, inspecting grain are developed byprivate entrepreneurs; 0 security stocks are properly maintained, separately from commercial flows and under government supervision. 105. Modest strategic reserves o f "produits de premiBre ne`cessitd", against catastrophes (such as natural disasters, wars and social unrest) would be governed by strict operating rules and be small relative to the market so as not to displace private storage. Their release would be targeted at groups specifically affected by the breakdown o f markets duringacute crises. 106. An analysis o f social transfer schemes that could succeed the public wheat monopoly is beyond the scope o f this note. It is now important to identify alternative targeted social transfer options that would allow government to lift current restrictions upon the development o f an efficient wheat market. 107. A privately-governed interprofessional organization for cereals and cereal products would be governed by representatives o f the various professions operating in the whole marketingchain for cereals. It would ensure: Support for the physical exchange o f products by the private sector in order to reduce transaction costs and improve transparency (i.e. the provision o f price information as well as production, trade, storage, millingstatistics); The creation o f a system o f forward pricing; - 35 - 0 The development o f risk management instruments (for price and counterpart risks) 0 the development o f contracting practices and an efficient system o f private arbitration. This would result in simple written rules for the exchange o f agricultural products, including for delivery (on spot or later, Le., forward), for sorting and grading (when possible), for payment at or after delivery, for product inspection(ifnecessary), and for the designationo f recognised private arbitrators. 108. Looking at the issue o f food hygiene and safety, a coordinated approach between i)the concerned Ministries and State agencies; and ii)the private sector (processing, retailers, wholesalers) in the promotion o f an efficient and modern system to protect human and animal health is needed. The cost would be trivial comparedwith the current costs o fprice support. 109. As for the issue o f quality, measures would include: 0 Removing controls on retail margins and other interventions o f the Ministry o f Commerce to limit prices. Competition in food markets will ensure reasonable margins and the Ministry should focus its interventions on identifying and attacking collusive behaviour. 0 Developing appropriate norms and standards in conjunction with (privately- governed) inter-professional organisations. Work has already begun in this area, under the Agricultural Services Support Project (ASSP). Redefining the role o f the Office des CCrCales (see above). 0 Ensuring that the objectives defined for the Ministry o f Agriculture under Performance-Based Budgeting(see section 1II.D) emphasise quality. 110. The basic lesson from these three examples is that the state should concentrate on delivering core public goods, genuinely private business groupings should be encouraged to develop their marketing chains, and government should remove distortions from the marketing chain. The same principles could and should be applied across the sector. - 36 - Box 3. Realisingpotential value-added bv improving qualitv An action plan for improving the value-added of agricultural and agro-industrial production should cover the areas detailedbelow. Many o f them are the role o f the private sector, others the role of the public sector, and a thirdcategory would depend onjoint public-private collaboration.. a. Supportingprimary and secondary processing industries, inthe areas of: 1) quality management and HACCP certification ii) the development of company laboratories, and the strengthening of public laboratories specialised inagro-food quality control; iii) themodernisationofagro-foodprocessing,'presentation,packaging and logistics; iv) the treatment o f industrial waste and effluent (particularly from slaughterhousesand the meat and milkprocessingsectors); b. Training and support for the private sector ingrading and quality-based pricingfor farm products; setting up arrangements for traceability and quality-based pricing. C. Public investment in supporting infrastructure (access roads, waste treatment, water supply, electricity and telephone connections), and the development of private fruit and vegetable packing houses in rural areas, in irrigated areas and date and almond- producing areas; d. Pilot traceability initiatives for plant and animal supply chains; e. Analysis of health and environmental risks, leading to the improvement of the regulations and institutionsresponsible for managing them; f. The co-ordination and strengthening of institutions involved in quality assurance training, sorting, packaging and quality control, for stores, factories, collection and packing centres, and warehousing and transport services. g. The immediate and complete abolition o f export bans and controls, except in genuine crises, such as natural disasters. This is particularly important for oils, tomatoes and tomato products, milkand almonds; h. Immediate and complete access to duty drawback arrangements for unprocessed and semi-processed agncultural products, to enable agro-processors to operate modem processes at an optimal scale; 1. A gradual reduction in import tariffs on agncultural and ago-industrial products, in order to encourage efficiency, modernisation and quality in the domestic sector. In parallel, the impact of value-added tax at each level o f the marketing chain should be studied; j. The gradual removal ofthe Ministryof Commerce's resale margin controls; k. The reform of the inter-professional associations, to make them independent of Government (see below). - 37 - B. ProfessionalOrganizations and Support ServicesI8 Producer organisations 111. Chapter I1 homed in on the importance o f improving the sector's responsiveness to consumer demand, especially where quality is concerned. In all developed countries, private producers' organisations take the lead in providing marketing services, market information and technology to their members. [add Pierre's para. on other countries...I Turning to Tunisia, this leads to the question o f how the private sector organises itself, and how government promotes or discourages private organisations. 112. This section will describe and analyse the performance o f Tunisian producer organisations, before making recommendations on new strategic directions. The broad conclusion is that agricultural professional organizations in general are under the control o f state-run institutions, with very limited financial independence. They are perceived as another layer o f the administration, which limits their technical and economic responsiveness to the market. The report therefore makes a small number o f strategic recommendations which would have the effect o f changing the incentives structure under which producer groupings operate, making them more responsive to the perceived and expressed needs o f their clientele. 113. Cooperatives for delivery of economic services (Cooperative de services Agricoles CSA). These number 201 with 90,000 members, i.e. 20 YOo f total Tunisian farmers. Their turnover is 5.5 % o f total agricultural production value. Most o f them started by collecting milk and providing services to livestock owners. They then diversified their activities into agricultural input supply and sometime crop marketing. 80% o f cooperative business is inthe coastal Governorates from Nabeul to Sfax. 114. The cooperative sector is stagnant. Although a few are doing very well, mostly the ones engaged inmilk collection or wine-production, the majority are non- performing or inactive. The gthPlan objective o f covering 50% o f producers was not reached. The turnover o f milk, fruit and vegetable cooperatives is falling. Only 5 new cooperatives were created during the last five years whereas 100 new cooperatives were created between 1990 and 2000. 115. The cooperative sector is facing several challenges. 0 There is little sense o f farmer ownership. Cooperatives were set up as Government institutions, and farmers and government employees still see them as such, not as farmer-controlled bodies. Managers are unqualified and neither members nor board members participate incooperative management. '*See Ideaconsult-IRAM-FERT-AMCIDA ~ (2006) - 38 - Central Government disengagement from cooperatives has been replacedby local government intervention. 0 Because o f their bureaucratic management practices, cooperatives have difficulty competing against genuinely private operations. There i s no bottom-up regional andnational cooperative union or federation. 116. Where the Tunisian cooperative sector has had some successes, in milk and wine, there are lessons to be learnt. A successful cooperative supplies services to members as requested by members, with pricing and payment arrangements suited to them, andhas professionalmanagers, staff andmanagement systems. 117. Collective Interest Groups (Groupements d'Inte`re^t Collect$ GIG`) number more than 3,000. Half deal with drinking water supply (for 219,000 families) and a third with irrigation management (110,000 families on 150,000 ha). They are operating successfully for the most part. However, the law o f March 15, 2004 requires them to adopt the status o f "Agricultural and Fishery Development Groups" byMarch2007. 118. Agricultural and Fishery Development Groups (Groupements de De`veloppementde I'Agricultureet de la PEche: GDAP) are non-profit associations of farmers and fisherfolk. Their objective is to manage collectively the natural resources o f a clearly identified zone and to contribute to its development. This status was created by the law o f 10 May 1994. 171 GDAPs existed at end-2005, often where there is no cooperative. 119. The law o f 15 March 2004 redefined GDAPs. They must no longer have any commercial or economic role, and must and engage instead in (i) extension (ii) promoting cooperation, (iii)protecting and managing natural resources (iv) supporting local investment andbasic infrastructure (v) addressing landtenure issues. They will neither receive Government fundingnor have access to loans or revenues. 120. There are seven Central cooperatives, mainly playing a role in the administered cereals market. The total average annual turnover o f the 7 cooperatives was 190 M TND over the last three years. The Cereal Office delegates 54% o f its cereals collection monopoly to three Central Cooperatives (CCGC, COCEBLE and CCSP). The Central Cooperatives also sell cereals, and supply agricultural inputs and equipment. 121. Members and users believe that these cooperatives are arms o f Government, since they are not managed as cooperatives. They have no member cooperatives, and election to their governing bodies i s not easy. Management is centralized, without analytical accounting systems. They have financial deficits and are highly indebted towards the Office des Cbrbales andthe banking sector. 122. TheFarmers and Fisherfolk's Union :I'Union Tunisienne de I'Agriculture et de la Pe^che (UTAP). UTAP is the only farmers union permitted by law, and - 39 - represents a quarter o f Tunisian agricultural producers. It is active on the boards o f cooperatives, inter-professional groups and technical agricultural centres. UTAP leaders are large farmers, with political ties, and are usually former civil servants (the current UTAP leader is a former Minister o f Agriculture). UTAP is therefore not perceived as representing ordinary farmers' interests, which negatively affects the functioning o f the many farmers' organizations which it i s involved in. Interprofessional organisations and commodity-specijk technical support centres 123. The Inter-professional Groups (Groupement Inter Professionel - GIP) are intended to cover all the actors in a supply chain. They number five: milk, fish production, animal production, fruits and vegetables. A quarter o f their board- members are from Government and the others are from UTAP and its equivalent for industry, UTICA. Financing comes from para-fiscal taxes collected by the FODECAP: 1% o f agricultural products sold on wholesale markets. GIPs are managed by a General Director, a civil servant appointed by the Ministry o f Agriculture. The Department o f Investment o f the Ministry o f Agriculture also fixes the allocation o fthe FODECAPto each GIP. GIPs are therefore effectively anarm o f the administration. 124. The Technical Agricultural Centers (Centres Techniques Agricoles: CTA) were created less than ten years ago, at UTAP's initiative, with the objective o f linking research and extension to farmers. There are three Centers, covering cereals, potatoes andbiological agriculture. They are financed by FODECAP, and inpractice underthe authority o fthe Ministryo fAgriculture which nominatestheir Directors. Key issues 125. The main issue, therefore, is that the rigid framework imposed on producer organisations makes it very difficult for them to become farmer-owned organisations. The laws defining their status are simplistic and inflexible: farmers are either members o f a cooperative or o f an Agricultural Development Group (Groupement de diveloppement agricole, GDA), or they are not affiliated at all. Inshort, a genuinely independent, profit-making cooperative would be illegal inTunisia. The shift from cooperatives to agricultural service organizations (Sociitis mutuelles de service agricole, SMSA) will place them under the direct control o f the local governorate. GDAs will combine the functions o f Collective Interest Groups (Association d 'intiret collect$ AIC) and o f collective interest organizations (Groupement d'intiret collect$ GIC). This will harm their performance: an association managing drinking water collection points has little in common with an association o ftomato producers. The latter needs a statute that will allow them to earn and distribute commercial profits -which a GDA is not allowed to do. - 40 - Inter-professional groups are currently managed by the administration and their job is to implement government policy: market and currency regulations, security stocks and so on. It i s not surprising that they have no membership base and cannot respond to producers' demands for information, services and market linkages. The wayforward 126. The associative sector has the potential to become an engine o f supply chain improvement if Government empowers it to do so. It is therefore necessary to recast its role, by: (a) Redefining the governance arrangements for UTAP and other associations (GIPs, GDA, GIC etc..) to encourage the perception that they are owned by farmers and not arms o f government. (b) Create a diversity o f legal frameworks for farmer organisations. Whilst maintaining the new SMSA concept, develop a legal basis for fanner-governed, independent, profit-making cooperatives, updating cooperative legislation and regulations to be inlinewith best practice defined bythe international cooperative movement. (c) Redefining Government's role as providing a facilitating framework rather than becoming directly involved in management. A semi-public agency with majority farmer ownership should promote farmers' groupings and agricultural extension. This would primarily involve capacity-building, linking different types o f organisation andhelpingcooperatives form unions and apex organizations. (d) Using actual Government services, rather than the associative sector, to perform core government functions like regulating phytosanitary standards and food safety. 127. Under this reoriented strategy, Government would adopt a revised approach to GDAPs: i)helping them to serve as intermediaries between Government, banks and farmers; ii)allowing them to receive revenues; iii)encouraging contractual partnerships between GDAP and cooperatives; iv) supporting new GDAPs in new irrigated areas and help old ones to diversify; v) developing a communication planto improve farmers' understanding o f GDAPs' role; andvi) capacity-building. 128. There is a role for apex cooperatives to play, but the future o f the Central Cooperatives would depend upon the willingness o f Government to liberalise cereals trading and regularise their financial situation. 129. For the Inter-professionalGroups (GIP), the new approach would involve: i)freeingthemtomanagetheFODECAPbasedonpreviouslyagreeduponrules,and encouraging a shift towards private funding; ii)breaking up responsibilities for heterogeneous supply chains (regrouping some chains was a mistake); iii)capacity- building in management, especially marketing and financial management; iv) ensuring Boardrepresentation o f ordinary farmers and cooperatives. -41 - Public agricultural support institutions 130. Chapter I1o f this report argued that Government should re-centre its activities on the provision o f high-quality public goods, such as research and extension. This section will examine briefly the agricultural support services provided by Government, before recommending a way forward based on increased responsiveness to producers' expressed needs. 131. The Agricultural Research and Higher Training Institute (Institut de Recherche et d'Enseimements SupCrieur: IRESA) was created in 1990, and regrouped 4 research institutes, 9 higher training centers, 5 regional research and development poles and their experimentation networks. IRESA includes 590 researchers and professors, including 270 full time researchers. IRESA has created national programming and evaluation commissions to program and budget agricultural research in 10priority areas. 132. The common theme incritiques o f IRESA (Ideaconsult et al., 2006) is that its research should become more receptive to the needs o f stakeholders and the market. Links with extension and fanners are weak and feedback from farmers plays little role, although this i s changing. Research priorities and research planning are neither results-oriented nor participatory. There is limited input from regional research poles anduse o fresearch results (this is also changing). Varietal improvement andcropping technology are prioritised to the detriment o f applied agronomy, socio-economic and post-harvest issues. The top-down mono-sectoral philosophy precludes an integrated farming systems approach. Finally, human resources are lacking, especially in the social sciences, and there is no obvious rationale intheir allocation. 133. The Agricultural Extension and Training Agency (Agence de vulgarisation et de formation agricole : AVFA) was created in 1990 and is responsible for all extension. Its central services are in Tunis. At regional level, under the Regional Center for Agricultural Development (CRDA), are the Territorial Extension Cells (Cellule territoriale de vulgarisation: CTV) at governorate level and the Agriculture Sector Cells (Cellules de Rayonnement Agricole; CR4) at sector level. Some Offices, some Inter-professional Groups (GIP) and some technical Centers also have their own specialized extension services. To complete the picture, there is an association o f private farm advisors, composed mainly o f retirees from the Ministry o f Agriculture. 134. The need to strengthen extension services came up inmany o f the sub-sectoral studies commissioned during the preparation o f this reportlg. Responsiveness to farmer priorities is again the key issue. The top-down training and visit approach fails to adapt extension messages to farmers' needs. Extension is perceived as part o f the government apparatus, helping farmers access Government subsidies, and not transferring any updated technology. Extension programmes are fixed by the AVFA with limitedconsultationwith producers andresearch. ~ l 9e.g. CNEA (2005a), CNEA (2005b), Ideaconsult-IRAM-FERT-AMCIDA(2005) - 42 - 135. At the same time, the extension service is under-resourced relative to its mandate. Bothinterms o f the quantity and quality o f services available, it i s unable to meet increasing demand. The lack o f resources is particularly acute at the CRDA level. The wayforward 136. The first key recommendation, therefore, is that Government should ensure that farmers' perceived and expressed priorities determine the planningo fresearch and extension services. Given the fact that farmers' organisations are not representative, farmers' perceived priorities should be elicited through quantitative and qualitative surveys o f representative samples, and by adopting a focus group approach2'. It would be essential to ensure that female farmers are properly represented, as they are increasingly acting as de facto farm managers (see chapter 1II.F). It would also be essential to ensure that socio-economists experienced in interpreting farmer priorities play an increasing role in determining the overall direction o f research and extension. One approach tested successfblly in Chile, Costa Rica and Nicaragua has been to give farmers extension vouchers that they can use to pay for public or private extension advice (Feder et al. 2000). Inthe longer term, as farmer groupings become more genuinely representative, it would be advisable to institutionalise their role inthe governance o fresearch and extension. 137. The second key recommendation is that the MARHshould introduce internal management reforms: firstly, to ensure that research and extension are properly resourced; and secondly to make them more accountable for results. As chapter 1II.D o f this report will argue, the MARHcould achieve both these objectives by showing a firm resolve to implement Performance Based Budgeting. (chapter 1II.Dwill make the case that the MARH's management systems are inherently biased against "soft" activities like research and extension, and that Performance Based Budgeting could be a powerfbl instrument for revitalising them.) One way o f making research institutions more responsive is to offer competitive research grants. Longthe norm in the U S A and EU, competitive grant schemes also operate in countries such as Romania, Vietnam, Brazil and Turkey, andtend to stimulate research efficiency when they replace a public sector research monopoly (Gill and Carney, 1999, Echeverria et al., 2002), InEcuador, for example, as many as 45 organisations, public and private, are executing competitively-awarded research assignrnents2l. 138. So far as research is concerned, this new approach would create a favourable strategic framework for updating priorities, involving farmers in research planning, measuring the results o f research, promoting an integrated supply chain approach, interdisciplinary research and dissemination. For extension, the new approach would involve reinforcing the role o f farmers in managing extension services and defining extension programs, better targeting, and management procedures that allow front- line agents to respond to client demand. ''IntenseBank 2o consultations with a small representativepanel of farmers. World Agricultural Investment Sourcebook, inpreparation - 43 - 139. But it is important to be realistic. Until and unless those responsible for research and extension feel themselves to be accountable to farmers and to their employer for the impact o ftheir work, there will be little incentive for change. c. SUSTAINABLE MANAGEMENTOFNATURAL REiSOURCES 140. Natural resources are the beginning o f the agricultural supply chain and the basis for agricultural livelihoods. 5.4 million ha o f cultivable and 400,000 ha o f irrigable land are the foundation o f the cropping sectors. Forests and rangelands cover another third o f the territory or 5.5 million ha, and are home to 10% o f the Tunisian population. Finally, the 4.3 million ha o f rangelands provide up to half o f the national herd's food requirements. This section will look at water alone, given its importance as a factor o f production, but an analysis o f forests and rangelands i s presented inAnnex 4. The implications for long-term direction o f government policy will be highlighted. Water mobilisation 141. Tunisia's well-developed water infrastructure can mobilize 90-95% o f the 4.8 billion m3 o f surface and groundwater available for use. Nearly half is mobilized by twenty-seven existing large dams, 5% from hill dams and lakes and the rest from groundwater. Water management inTunisia focuses on storing large volumes for use inthe increasingly frequent years o f shortage22. The situation is complicated by the fact that water resources are mainly in the north and inland whereas users are concentrated on the coasts, therefore requiring extensive and costly transportation. 142. The 2002-11water mobilization strategy aims at: (i) developing conventional infrastructure (mainly dams and groundwater) as well as non-conventional sources (recycling o f used and saline waters), (ii) protecting existing infrastructures (see also soiywater conservation), and (iii) promoting water savings (particularly in irrigation schemes). 11 new dams have been completed or are under construction under the 10th Plan, and 11 others (mainly for water storage) are planned under the 1lth and 12th Plans. Other water-mobilisation investment includes the recharging (under the 10thPlan) o f a number o f groundwater sources and the construction o fhilldams and deep wells. 143. Water infrastructure investment consumes a massive share o f the MARH's budget. 55% o f the loth Plan budget is allocated to water investment, and 43% o f water investment i s for large dams alone. The large dams' allocation in 2005 was 22 Water consumption averaged around 2.6 billion m3in 2004, or two thirds of the volume mobilized. Irrigated agriculture accounts for about 80% of water consumption; another 12% consists o f drinkable water, and the remaining 8% goes to industry and tourism. - 44 - equivalent to 61% o f the M M ' s entire recurrent cost budget for that year. Given the evidence above that Tunisia's agricultural competitiveness depends on quality, which depends inturnupon first-rate research and extension services, it is unfortunate that these massive infrastructure investments are accompanied by a real decline inthe MARH's recurrent budget. This theme ispursued inchapter 1II.D. 144. Moreover, buildingnew water-mobilisation infrastructure does not address the main constraints to water mobilization: dam siltation, overexploitation o f groundwater, water pollution and salinization. Siltation losses in storage capacity are estimated at around 17 million m3per year, or 0.7% o f storage capacity. The main cause is the degradation o f the watersheds upstream o f the dams; large-scale soil conservation works would be necessary to reverse this situation. So protecting the 90-95% o f surface water that is currently mobilised from siltation might be more economical thanmobilising the last 5-10%0. 145. Inaddition, the level of use of costly public irrigation infrastructure remains lower than design assumptions in most schemes. This is particularly so in the high- rainfall north o f the country where farmers have less incentive to irrigate. The intensity o f use is often less than 1 (i.e. less than one full irrigated crop per year), which is usually some 20% less than original design objectives. Weaknesses in extension services have been identified as an obstacle to raising the utilization rate o f existing infrastructure. 146. In short, dams are expensive, so it is important to match construction to utilisation. Currently on average only two thirds o f the water mobilized i s consumed. New dams will have to be built on less convenient sites, increasing their cost. It may be optimal to delay the planned construction o f new dams beyond the 11th and 12th Plans and thus liberate funds for other priority needs, such as rehabilitation, de- siltation, extension and research. Farm-level water management 147. Tunisia has made substantial progress in delegating the management o f downstream irrigation systems to user groups. Operating with the legal status o f Groupements d'InttrCt Collectif (GIC), they serve over 100,000 households and 150,000 ha under irrigation. They have changed the way the irrigation department i s working in Tunisia, improving the financial sustainability and the management efficiency o f irrigation water delivery. 148. There are two policy issues concerning GICs facing the Government. Firstly, they will be converted to the status o f GDAPs (see paragraph 119) as redefinedby the Law o f 15 March 2004. It remains to be seen whether this vision o f multi-purpose farmer groupings, managing water, production, purchasing and marketing, will live up to reality. It may be that the specific needs o f individual farmers or groups cannot be coveredby the standardized catch-all model prescribed by Government. It is likely that different sub-sets o f farmers would be interested in the water-management and commercial aspects o f group membership. Secondly, their large number - 1,100 so - 45 - far -i s beginning to stretch the capacity o f MARHto support start-ups andto provide advice. 149. Tunisia leads the region in the use o f water fees to promote financial sustainability and more efficient water-use. Water charges in public irrigation schemes cover 115% o f all operation and maintenance costs, up from 70% in 1991 (see Table 14). They are reportedly applied consistently except in some central and southern governorates with highconcentrations o f low-income irrigators. Region Recovery Rate 1991 RecoveryRate 2000 North 76% 119% Sahel 56% 81% Centre 44% '107% South 60% 83% Total 70% 115% 1Source: Bazza andAhmad (2002) I 150. Tunisia is experimenting with a two-part pricing system [tarification binbme] in 12 pilot public schemes over 10 governorates. The approach is to charge a fixed initial fee plus a volumetric fee. It is intended to ensure that operating and maintenance costs are fully-covered without discouraging efficient water use. The experience so far has beenmixed, but the consensus i s that the shortcomings -mainly a lack o f communication and enforcement - do not bring the two-part pricing approach into question. A major problem, however, is that Government decreed a 50% tariff discount for cereals after the two-part tariffs had been fixed, which is undermining the financial viability o fthe pilot schemes. 151. Concerning on-farm water-use efficency, the Government's objective is to have irrigation schemes filly covered by water-saving techniques23by 2009. About 305,000 ha are covered so far, close to 75% o f the irrigated area. The Government pays between 40% and 60% o f the total costs o f upgrading to water-saving equipment, and 20-30% o f the cost o f replacing it. The subsidies granted for water saving equipment to irrigators are expected to be around 40 millionTunisian dinars in 2006. It is estimated that field-level water use efficiency has improved from 50% to 75% over the last 10years, which is an excellent achievement byworld standards.24 The Way Forward 152. A theme running through the observations above is that there is a bias in favour o f new dam investment at the expense o f anti-siltation, rehabilitation and improving the utilization rate o f existing public irrigation schemes - let alone 23e.g. sprinklers, dnp irrigationor improved surface irrigation. 24Contrary to expectations, agricultural production has increased rather than water consumption decreasing. - 46 - recurrent spending on supporting farmers' quality-competitiveness. The key tool for addressing this bias is economic cost-benefit analysis. Government should define explicitly the objectives o f its dam-building policy. Ifthe objective i s to stabilize farm incomes, there may be lower-cost ways o f doing this (see section 1II.D). If,however, the goal is stabilizing the supply o f water in drought years, Government should use economic analysis to determine the least-cost approach to attaining this 0bjective.e. Alternative approaches would include anti-siltation work, dam rehabilitation and the transfer o f water from agricultural to non-agricultural uses. Cost-benefit analysis should be carried out for each proposed dam to determine whether it is the least-cost means o f achieving the objective compared with, for example, anti-siltationmeasures. 153. MAHRmust therefore shift its emphasis fi-om water mobilisationto integrated management. The strategy would be to improve the effectiveness o f existing infrastructure, with activities such as: (i) the protection o f existing water resources against pollution; (ii)research into dam protection and desiltation; (iii)better meteorological and hydrological analysis. On the demand side, activities would include: (i) the refinement o f pricing systems; (ii)continued promotion o f improved irrigation techniques, and (iii) continued increases inmanagement by water-users (iv) helpingfarmers to increase the value o ftheir production. 154. For use-values and demand-management to play a greater role in water resource planning, it would be useful to strengthen the capacity o f the central services the Ministry o f Agriculture and Water Resources. This would build on recent progress achieved by the creation o f the Bureau de la planiJcation et des kquilibres hydrauliques. It will be particularly important to improve skills in the fields o f agro- economy and socio-economy. 155. So far as farm-level water management policies are concerned, Tunisia should build on the reforms o f the last decade. The general approach should be to extend successful existing models andto eliminate the remaining anomalies. 156. Looking at the extension o f successful models, it i s recommended to extend the two-part tariff experiment to more schemes countrywide, in accordance with the results o f the evaluation workshop held in 2004. It would also be useful to consider the introductiono freplacement costs inthe calculation o f the fees. 157. Government should continue to support the creation o f water-user groups. However, it should monitor the impact o f the conversion to GDAP status. As recommended in paragraph 125, Government should be ready to adapt the GDAP model to the dynamics o f groupings on the ground, rather than expect conformity. As water-user groups become more established and co-ordinated decision making between different groupings becomes more important, regional water councils - comprising all types o f water users (including associations) -could be set up. 158. However, it will be difficult to improve extension services to water-user groups so long as the number o f GICs goes up and the MARH's recurrent budget - 47 - moves in the opposite direction. Sub-contracting advisory services to private organizations will not solve the underlying budget issue. 159. There is no economic or social justification for public subsidies for the replacement o f water-saving equipment, certainly when it involves the substituting depreciated equipment with new equipment o f the same type. The start-up subsidy could perhaps be justified on the grounds that the beneficiary does not yet know the technology. But if farmers are paying the fillcost o f water still are not motivated to buy water-saving equipment after usingit for a few seasons, then it is probably not profitable for them. The subsidy is also distributionally regressive, because access to itisproportional to ownership o firrigated land. 160. Government should also remove the preferential water tariff for wheat. It is cumbersome to assess how much o f a farmers' water use should be at the fillrate and how much at the discounted rate. It distorts farmers' cropping patterns, encouraging them to grow crops in which they are less competitive (see chapter 1II.A above). It undermines the financial viability o f the two-part tariff approach. And it is distributionally regressive, because access to the subsidy is again proportional to ownership o f irrigated land. D. RESOURCING THE AGRICULWL 161. We have seen that realising Tunisia's agricultural potential means pursuing quality and responding to the demands o f the market. The state's role will increasingly be to provide core public goods and a favourable business environment, so that producers can be in touch with and adapt to market demand. The capacity o f the public and private sectors to play their respective roles will depend in part upon their ability to marshal resources. This section will therefore comment on the performance o f the institutions that mobilise andmanage resources for agriculture. 162. A common theme emerges for boththe public andprivate sectors: a pervasive philosophy o f top-down quantitative planning i s making it difficult for the government budget and the financial sector to respond flexibly to the needs o f the farmer. Government has erected a complex superstructure o f programmes and prescriptions, which responds imperfectly to the perceived priorities o f farmers. The report therefore makes a small number o f strategic recommendations to address the incentives framework under which decision-makers inbanks and the MARHoperate. The intention is to make them more aware o f andresponsive to their clientele. Management o fthe Ministry o f Agriculture's Humanand FinancialResources 163. Whatever new policy direction the Tunisian government decides to take, it will only be effective to the extent that the ministry has the management tools to reorient its activities and to allocate staff and money to its new priorities. Currently, 25Based on CNEA (2006), a study financed by the Agence Francaise de Developpement - 48 - however, the ministry lacks some o f the critical management systems for translating policy priorities into action on the ground. 164. A key observation is that the MARH's systems of planning, budgeting and human resource management are better suited to the management o f physical investment (e.g. irrigation infrastructure, plantations and fishing ports) than to managing so-called "soft" activities. The phrase "soft activities" here means the management o f knowledge, information, systems, procedures, institutions and behaviours. It covers critical functions o f the MARH, such as research, extension, improving product quality, the establishment o f maintenance systems and the involvement o f communities inthe management o fnatural resources. 165. The background studies provided strong evidence that the MARH's efforts are focussing on the creation o f new physical infrastructure at the expense o f "software". The assessment o f the irrigation sector is that ambitious plans for new storage dams are accompanied by the poor maintenance, high siltation and relatively low rate o f utilisation o f existing dams. Similarly, the MARH's programmes o f support for the fisheries sector emphasise the provision o f additional ports, whereas it is more important, given the over-exploitation o f limited fish stocks, to ensure that the potential value-added from the existing catch is fully realised. Meanwhile, analyses o f the fruits and vegetables, livestock, olive oil and cereals marketing chains all emphasise the lack o f effective research, extension and quality assurance systems. 166. The management systems which create a bias infavour o fphysical investment at the expense o f "soft" activities include: 0 The emphasis on quantified physical targets in the Plan process, in the presentation o f the Budget Economique and in sectoral analyses, to the exclusion o f more qualitative indicators. The definition o f targets and objectives in terms o f volumes (e.g. quantity o f tomato puree produced) rather than values (e.g. U S D received from exports o f tomato puree), which implies that crude levels o f output matter more than marketingor quality. 0 The institutional separation o f Titles I (recurrent operations) and I1 (investment) o f the budget inthe MARH- even though they are managed by the same Directorate General inmost other ministries. 0 The apparent lack o f an objective procedure for allocating resources between Titles Iand 11. 0 The different degrees of strategic management applied to Titles Iand I1o f the Budget. The Title IBudget, especially where personnel costs are concerned, is the product o f incremental adjustments to the previous year's allocations within very narrow margins. The Title I1budget, on the other hand, benefits from discretionary resources, planning, analysis and heavy donor - 49 - involvement. Inpractice, Title Ihas to match Title I1with staff and running costs as well as it can within its much tighter constraints. 167. These imbalances are visible in the MARH's resource allocations. Out o f a total ministerial budget o f DT 722 million for 2005, DT 523 million are classified as "investment expenses", o f which DT 218 million are funded by external grants and loans. Ofthe remaining DT 199 million available for recurrent expenditures, DT 175 million are allocated to staff costs, leaving only DT 24 million for non-staff recurrent expenditures, which are often crucial for the delivery o f high-quality "soft" activities like research and extension. Slightly over a half o f lothPlan allocations, DT 1206 million out o f DT 2200 million, are allocated to irrigation activities. This compares with DT 56 million for research, studies and extension. Inother words, the savings from reducing water investment by 5% could double spending on research, studies and extension. The total agricultural Plan size increased by DT 105 million between the gthand lothplans, whereas expenditure on water resources alone increased by DT 134 million. 168. The Government could ask itself some hard questions about the implications o f this bias for the development o f the agricultural sector. So far this report has stressed that Government's role should be to provide core public goods to support high-quality production: research, extension, phytosanitary controls and food safety regulations. But does the MARHhave effective tools for managing the performance o f such "soft" activities? For example, how would it mobilize its front-line staff to put more emphasis on quality andefficiency as opposed to productivism? Is the staff training programme planned and funded so as to support the reorientation o f the Ministry's priorities? How can a CRDA planits staff use when its personnel (funded byTitle I) are pulledin different directions by special programmes and donor-funded projects (funded by Title II)? Will the MARHsee a progressive deterioration inthe ratio o f the stock o f infrastructure to the resources available to fund its maintenance andto advise the farmers who use it? 169. The ongoing Agricultural Services Support Project (ASSP) is achieving positive results in a number o f "soft" areas such as research, extension and quality. This is due in part to the fiamework o f objectives provided by the project cycle management approach. The ASSP demonstrates that it i s possible to define measurable objectives for "soft" activities, and to transmit them through the organizational hierarchy. The lesson from the ASSP i s that relatively small injections o f non-salary recurrent expenditure can have a large transformational impact, provided that they are backed up by a robust framework o f objectives for performance management. The WayForward 170. The Ministry o f Finance is currently preparing the government's transition towards Performance-Based Budgeting (PBB) [budgbtisation par objectifs]. The intention i s that PBB should be practiced across the entire government by 2010. The Ministry o f Agriculture is a pilot ministry, and within the MARH the Forestry -50- Directorate General, the Water Sector InvestmentProject (PISEAU) and the Regional Agricultural Development Commissariat (CRDA) o f Ariana will be pilot departments. 171. The main elements o f the PBB approach are: (a) A Medium Term Expenditure Framework (MTEF), which is the medium-term development strategy translated into financial flows, and implies by definition the integration o f budgetary planningfor "development" and "recurrent" expenditure (i.e. Titles Iand11). (b) Theorganization of the budget by objectives, so that the budget format maps onto the ministry's objectives and organigramme. This reform helps ensure that human and financial resources follow development priorities and that staff responsibilities for translating money into results are clearly defined. (c) The setting of performance objectives which cascade responsibility for the ministry's performance from top to bottom, while respecting the congruence betweenresponsibilities, budget lines andthe organigramme. (d) A modern approach tofinancial control and audit, which shouldbe concerned not only with procedural correctness but also with the effectiveness o f expenditures. 172. PBBwill therefore provide the MARHwith a tool that is currently lacking for the management o f "sofi" activities. It will help integrate the two titles o f the budget, to transmit the ministry's new priorities to the front line and to monitor and control staff performance across the range o f the MARH's activities. However, the MARH will only benefit from PBB to the extent that it is prepared for and committed to its introduction. This will involve: creating a PBB management cell within the MARH, bringing together the functions o f strategic planning, administration, finance and investment; establishing an implementation plan and mobilizing the required technical assistance; launching a communication and training campaign for MARH staff at all levels; and developing a long-term staffing and human resource development plan, taking account o f the MARH's new role and the high rate o f retirements expected incoming years26. 173. According to the Economic Budget, 55% o f public agricultural investments were externally funded in 2004 (DT 214 million out o f DT 387 million). There are indications that the modalities o f external funding may risk weakening the Ministry's management and planning systems. The existence o f many regional or thematic projects andprogrammes, each with its own budget, management unit, objectives and procedures, is in direct contradiction to the move towards integrated strategic planning described above. It is reported that projects' demands upon staff for "special duties" makes it difficult to plan and manage CRDA activities. Their 26Annual departures are currently 400-500 per year (around 2.5%), which is expected to rise to 900by 2010. - 5 1 - reliance upon contractual staff funded via Title I1o f the budget may also compromise their long-term sustainability. 174. Many ministries o f agriculture that are similarly dependent upon external fundingseek to incorporate donor fbndingwithin apredetermined sector strategy and resource mobilization plan, to manage it within existing ministerial structures instead o f project-specific management units, and to channel fbnding as far as possible through the departmental budget. The MARH should consider adopting such a "Sector Wide Approach." Agricultural Finance 175. There are signs that private investment in agriculture i s far below its potential level. Over 2002-4, private investment was 51% o f total investment, up from 40% in 1987-91, but a long way short o f the middle-income norm o f around 75% (CNEA, 200%). The share o f farmers investing inagriculture has fallen significantly over the last decade, from 36% in 1990-4 to 26% in2000-4, with the greatest drop (from 50% to 33%) occurring among the medium-sized farms o f 10-50 ha. The proportion o f farmers who invested fell in every region, and the drop is most pronounced in the north. This raises the question: to what extent has limited access to finance constrained private investment? 176. Under the 1993 investment code, Government encourages agricultural investment with an extensive and complex system o f direct subsidies (((encouragements de l'Etat D), administered by the Agence de Promotion des Investissements Agricoles (APIA). The schedule o f subsidies is intended to encourage exports, the development o f backward regions, modem technologies (e.g. organic farming) and protection o fthe environment. Subsidies represent 20.4% o fthe cost o f investments over DT 40,000, and 38.4% of the cost o f investments under DT 40,000. It is difficult to assess the impact o f these subsidies in terms o f growth and jobs. Ifthe farmer would have invested anyway without a subsidy, the subsidy only represents a transfer from the government to the farmer. 177. The Banque Nationale Agricole (BNA) is the main agricultural lender, with DT 0.7 billion o f loans outstanding. Although around 1,600 farmers borrow from other commercial banks, these are people with other businesses, for whom agriculture i s a secondary part o f their banking relationship. Meanwhile, the Banque Tunisienne de Solidaritk (BTS) has a portfolio o f individual smallholder loans and credits to microfinance institutions. Around 40% o f BTS-funded microcredits are classified as agricultural. 178. Bank credit accounts for around 20% o f the financing o f agricultural investment. The remainder i s divided between self-financing (61%) and subsidies. The relative importance o f bank loans in agricultural finance has fallen significantly over recent years. The ratio o f bank loans for agriculture and fishing to the sector's value added fell from 3.8% in 2000/01 to 2.1% in 2003/04. For seasonal loans [crkdits de campagne] the fall was even more marked, from 2.2% to 0.9%. - 52 - 179. The share of farmers expressing a demandfor credit is low, at 7 -8% (CNEA, 2 0 0 5 ~ )and is on a slight downward trend. When surveyed, around a quarter o f ~ farmers say that they do not demand credit for fear o f indebtedness, and another quarter because o f difficulties in accessing credit. Most of the remaining farmers gave a wide range o f reasons too diverse to capture statistically. At the same time, the share o f farmers reporting that their demand for credit was satisfied has fallen sharplyover the last decade, from 54% in 1990-4 to 36% in2000-4. Only the larger farmers, cultivatingover 50ha, have not seen their satisfaction rate fall. 180. The main factors underlying this contraction o f agricultural lending are well known. Thepoor repayment rate. The repayment rate o f the Banque Nationale Agricole (BNA) is around 50%. The repayment rate ofthe BanqueTunisienne de SolidaritC (BTS) for loans to individual farmers ("microprojets agricoles") i s also around 50%, and around 80% for on-lending via NGOs (MARH estimates). The fiscal cost o f these arrears is not transparent, because Tunisia's fiscal regime discourages provisioning and the central bank does not permit a loan asset to be written off if it can be shown that it i s backed by security (IMF, 2004). For the banking sector as a whole, therefore, provisions only cover 46.2% o f bad debts (IMF, 2004). Forthe BNA, provisions are only 5% ofthe DT 3.3 billion of loans outstanding. Unprovisioned bad loans stay on the creditor's balance sheet as an asset, backedbythe implicit guarantee o fthe state shareholder. The tightening of the market. The number and value of agricultural loans has been falling for several years, with 3,907 seasonal credits in 1998/9 becoming 1,446 in 2003/4, compared with an estimated 59,000 farms with some degree of commercial potential27. This drop can be attributed to (i) indebtedness, clients' (ii) BNA'sadoptionsince1999ofapolicyoflimitingitscostsbyfocusingon the larger clients (World Bank, 1999)28,(iii)real cost o fcredit o faround 15%29,and a (iv) increased farmer savings - the demand for credit always falls after a good season - and access to informal credit fiom friends and family. N o n - f m earningsare often the main sourceo f capital for farm investments. Suitability of credit instrumentsfor smallfarmers. 1.8% of small farmers receive short term credit (Guellouz, 2004), compared with 4,140 out o f 471,000 farmers, or 0.9%, in the sector as a whole in 2003/4. However, it is possible that classic bank credit, with its administrative formalities, earmarking o f funds and requirement for immoveable security, is ill-adapted to the needs o f the 225,000 farming operations categorized as "of weak viability" or worse. These rural 2' i.e. classified incategories B and C. 28 The sharpest fall inthe rate o f credit demand satisfaction has been at the bottomend ofthe BNA's clientele (CNEA, 2005c), i.e. among farmers too small to be the BNA's best borrowers but too large for BTS credit. 29 Interest rate 11-12%, contribution to the Fonds National de Garantie 2%, compulsory insurance 2 - 3%, variable commissions and charges, inflation 5%. - 53 - households, which often engage in multiple informal economic activities, may be better suited to microfinance products. 181. These phenomena are typical o f a rural finance sector that is tightening up after a period o f "easy credit" and bad debts. From a purely prudential point o f view, however, the situation appears stable, as Government has the fiscal resources to continue subsidising the BNA, the FNG and the BTS. This, along with the prominence o f high-profile co-operatives in the bad debts o f the BNA, helps explain the lack o f a sense o f urgency concerning the reform o f the rural finance sector. However, the status quo does have hidden costs for Tunisia. Firstly, there is the fiscal cost o f the government's implicit guarantees for unprovisioned bad debts. Secondly, there is the cost o f investments being foregone because bankable projects are not receiving finance. Thirdly, there is the cost o f the allocation o f savings to unproductive investment. 182. The supply o f credit to agriculture therefore depends critically upon the lending practices o f the BNA. These inturn depend upon the incentives under which its management operates. Key features o f this incentives fi-amework are that: The BNA has an obligation to maintain an expensive rural branch network and to offer agricultural finance products; 0 The BNA is encouraged to operate in non-agricultural markets, which are perceived as less riskyandmore lucrative; The BNA is under pressure to improve its operatingprofit; The BNA is under little pressure to collect bad debts, as these are not provisioned (see above)30. 183. Taken together, these factors could easily encourage a bank's management to view its agricultural portfolio as an imposition: a high-cost business line which must not be allowed to affect the bank's competitiveness in more lucrative markets. The result would be conservative lending practices, high charges and little effort actively to develop new agricultural business. 184. The lending and arrears figures quoted above show that the finance sector is failing agriculture. One reason is that Government policies have removed the incentive for financial institutions to lend sustainably. The conflicting social and commercial objectives o f the BNA and the BTS make it impossible for their performance to be measured, let alone rewarded or sanctioned. State subsidies compensate for poor financial results and discourage the entry o f private banks. Central Bank policies on provisions discourage proper accounting for bad debts. Publicly-funded repayments o f arrears discourage prudent lending and collection procedures. 30Of around 100,000 loans inarrears only around 7-8,000 are beingpursuedas "contentieux". There are no cases ofland security beingrepossessed. - 54 - 185. Ifgovernment were to seek to remobilize the agricultural finance sector, an essential first step would be a study o f the decision-making in the credit market. It would not only look at farmers' investment decision-making, demand for credit, and perceptions o f the financial system, but bankers' decision-making and banks' organizational behaviour. There is currently no hard data available upon whether a bankable clientele is currently excluded from the credit market and, if so, what are its characteristics and needs, andhow it is segmented into differentiated sub-markets. 186. This recommendation for a field level study is not simply an evasion, a cop- out, an excuse for not making firm recommendations at this stage. It is part o f an important paradigm shift, in which the perspective o f decision-makers moves away from prescribing schemes that may or may not suit the sector's needs towards understanding the behaviour o f key stakeholders and responding to their perceived priorities. 187. Depending on the results o f this study, two broad policy options would suggest themselves: Adopting international best practice on the environment for microcredit. The World Bank /IMF 2001 Financial Sector Assessment (FSA) concluded that Tunisia's microcredit environment is far from ideal. For example, Tunisia's 170 Microcredit Associations (AMCs) are not allowed to charge positive interest rates and the recent decision to limit banks' interest rates to 5% for f m s with incomes o f under DT 3,500 / year make it impossible to cover expenses. AMCs depend for their income on a DT 20 subsidy per loan dossier from the BTS, funded from BTS's DT 18 billion / year state subsidy. Best practice policies, on the other hand, would include: allowing micro- finance institutions to set rates that cover their costs, the use o f subsidies for start-up costs only, the dismantling o f competing "social" credit schemes and legislation facilitating the creation and growth o f micro-finance institutions. The reform of the BNA. The purpose o f reform would be to strengthen the incentives for bank management to manage credit risk effectively, and to develop improved lending practices. This could involve recasting the BNA as an autonomous commercial bank, having transferred its irrecoverable bad loans to a separate financial entity. If the Government wished to maintain credit subsidies, these could be administered more efficiently and transparently via competitive tendering between commercial banks, which would have the added advantage o f increasing competition in rural banking. Separating the agricultural and non-agricultural accounts o f the BNA would enable government to identify clearly the economic and fiscal cost o f its agricultural credit policies. Reform o f the BNA would require improved policies on the accounting treatment o f bad debts. It would also have serious implications for debtors with major arrears, including a number o f important co-operatives. It was beyond the scope o f this study to examine this option in detail. - 55 - Agricultural Income Stabilisation 188. Tunisian rainfed agriculture is drought-prone. For example, wheat yields between 1961 and 2005 were on average 18% above or below the trend line (FAOSTAT data). In seven o f these last 45 years they were at least 25% below the trend line. CNEA (CNEA, 2005a) data paint a more dramatic picture: over a period o f 19 years, 5 high-rainfall years produced at least 18 million quintals o f wheat and 9 low-rainfall years produced less than 5.1 million. Exposure to climatic risk can reduce farm incomes significantly (World Bank, 2005a), as farmers stick to low-risk, low-return activities. In addition, a part o f the BNA's bad debts are attributed to drought losses. 189. Meteorological data show that temperatures have been rising gradually during last century. Intheory this could increase the frequency o f extreme droughts, even if statistical analysis does not yet show this to be happening (MART3/ GTZ, 2005). 190. A number of instruments for stabilizing farm incomes operate in Tunisia. Taken together, these instruments leave many farmers exposed to drought risk, since they have limited coverage andthe farmer cannot predict how they will operate : (a) The provision o fpublic irrigation services, which were discussed earlier. (b) The Fonds National de Garantie (FNG), which pays interest costs when loans are rescheduled (over 5 years) after droughts. Of course, the FNG's coverage is limited to the minority o f farmers that borrower and the benefit i s proportional to a farmer's borrowing capacity i.e. the benefit is focused on richer farmers. In practice, moreover, rescheduled loans are often not serviced and the bank is left carrying the remaining debt. It is estimated, for example, that 51,000 farmers benefited from rescheduling in 1999, o f which 18-20,000 are keeping up with repayments . (c) Commercial Crop insurance, which is seen only as a precondition for obtaining a bank loan. It only covers non-systemic losses such as hail, fire and damage to greenhouses. The Caisse Tunisienne des Assurances Mutuelles Agricoles (CTAMA) has 80% o fthe market. The C T A M A covers its costs, and is gradually losing market share to other insurance companies, which suggest that this market is competitive. The market is shrinking inparallel with the number o fbank loans; in spite of the CTAMA's marketing campaigns there is little demand for insurance except when it is linked to bank credit. (d) The "Fonds des CalamitCs" and other ad hoc interventions. In drought years Government allocates a part o f its budgetary reserve to disaster mitigation as the Fonds des CalamitCs. Plans for the use o f these resources are proposed by the regions, consolidated by the centre and implemented through sectoral ministries' development (Title 11) budgets. Typical MARH activities might include fodder and transport subsidies and the provision o f water-tankers. The structure o f the - 56 - government budget makes it difficult to identify Fonds de CalamitC spending. Other ad hoc interventions include the decision in 1993/5 to write off the debts o f 7,557 farmers, and in 1999 to write off all arrears below DT 1,000 and to reschedule arrears betweenDT 1,000 andDT 4,000 over 7 years. 191. The MARH is interested in exploring agricultural drought insurance. Traditional drought insurance involving individual crop-loss assessments, however, is unlikely to succeed in Tunisia. The subjectivity o f crop loss assessments and the problem o f "moral hazard" (once insured, farmers may take insufficient care to reduce the likelihood and severity o f losses) will make it difficult to bring in reinsurance, which is crucial because drought risk is systemic i.e. many claims will be made at the same time. Also the cost and delay o f administering many individual field loss assessmentswould be prohibitive. 192. A possible alternative option would be rainfall index insurance. This would involve the creation o f a set o f rainfall indices, inwhich rainfall at different times o f the year i s weighted to maximize correlation with the yields o f different crops in various locations. The insurer would pay out an amount proportional to the rainfall index's shortfall below a predetermined threshold. Such index-based insurance is offered in Canada and has been tested since 1998 in Nicaragua, India, Ukraine, Ethiopia and Malawi. The advantages o f index insurance are that reinsurers can trust the objectivity o fmeteorological data andthat there is no need to assess farmers' field losses. 193. Global experience o f index insurance provides two relevant lessons (World Bank, 2005). Firstly, policymakers must be clear whether the objective i s to reduce poverty and vulnerability on the one hand or to provide a commercial insurance product on the other; mixingthe two can lead to suboptimal outcomes. Secondly, an index insurance product is unlikely to succeed in a high-risk environment where government intervenes with ad hoc and unpredictable compensation, for example in the form o f loan write-offs. This is because farmers will not pay the large premia requiredwhen they hope that government might provide compensation fi-ee o fcharge. That was the experience o f the ((Fonds de mutualit6 pour l'indemnisation des dommages agricoles aux calamitks naturelles )) inMorocco (1986-88). Inspite ofa history o f drought, a 30% government subsidy and a 20% reduction in premiums in year 2, the insurer sold very few policies because farmers assumed that government would cover their losses and bad debts. Another initiative in the late 1 9 9 0 ~ ~ supported by the International Finance Corporation and the Italian Government, was abandoned when the value o f premia was fully understood. (Indeed, when the Tunisian government considered launching drought insurance in 1987 and 1997, premiums o f 6% o fthe insured amount were considered too expensive). 194. However, there is scope for Government to conceive more systematic social drought insurance programmes targeted on the most vulnerable. This would involve moving away from input subsidies (credit, write-offs, fodder, and transport) which mostly benefit larger farmers, towards targeted direct transfers triggered by drought events. Ifnecessary, Government may consider stabilizing the fiscal impact o f such - 5 7 - programmes by purchasing drought insurance from international reinsurers, or by entering into conditional borrowing agreements with international financial institutions. E. RURALLAND 195. Land i s the foundation o f agricultural production and livelihoods. It is also an area in which the MARH has actively intervened. This section explores the issues surrounding rural land tenure in order to identify possible new orientations for Government. 196. On one level, the analysis is inconclusive. The data available do not say much about how land tenure affects farmers' behaviour and competitiveness. On another level, however, a firm conclusion does emerge: before offering prescriptions, we need to understandwhat is driving farmers' choices on the ground. 197. Tunisianlaw recognizes four forms o flandtenure: e Private land. [terres privCes]. This legal form was introduced in 1885. 4.7 million hectares are held as private property, a figure increasing at the rate o f around 70,000 haper year. e Collective land [terres collectives]. This legal form, defined in 1901, integrated the concept o f tribal land tenure within the framework o f colonial French law. The Tunisian law o f January 4, 1964, incorporated many o f the provisions o f earlier legislation on collective land, giving tribal collectivities legal personality and exclusive land rights. It also facilitated the conversion o f collective landinto private or forest tenure [terre soumise au regime forestier]. e Habous land [terres Habous]. Habous landtenure was historically created by religious endowments. Habous land incomes financed religious and social activities implementedby the Djemai'a des Habous. e The state estate [terres domaniales]. This was the dominant form o f land tenure during the colonial era. 820,000 in 1964, it now covers around 500,000 ha o f highly fertile land. It is farmed by pilot farms, "agro-combinats" and co- operatives farming under the Office des Terres Domaniales. By law the state can not alienate such land. But Government can delegate it [affectation] to public institutions, lease it, or grant the usufruct to young farmers and agricultural graduates. The study was not able to look into Terres Domaniales, but some questions for hture reflection are inAnnex 3. 198. Before Independence, titled tenure were the preserve o f the colonial state and foreigners. A key theme o f land tenure policy since 1956 has been the promotion o f registered private title for the Tunisian citizen. Habous land was abolished in 1956 - 58 - andtaken into the state estate3'. The policy o f nationalisation and collectivisation o f state and private land was reversed by the law o f September 22, 1969. Since Independence 3 million hectares o f collective land have been converted either into private property or put under forest tenure. Although it i s not compulsory to register titles, several laws enacted since the late 1990s require that private property be registered in the cadastre [livre foncier] in order to create property rights. This, however, leaves some ambiguity around the implications o f the law o f 10 July 1974, which recognises that unchallenged occupation o f land for five years can lead to issuance o f a Certificate o f Possession, which can be mortgaged and converted into title after a further five years. 199. A second central theme o f land policy has been the consolidation o f land in public irrigation schemes in accordance with fixed minimumand maximum levels o f ownership. The maximum levels are to widen ownership, while the minimum levels are for viability and technical efficiency. Since 1977 the Agence de la Refonne Agraire des PCrimbtres Publics IrriguCs, and then the Agence Foncikre Agricole, have been reconfiguring land ownership in public irrigation schemes. They do it by expropriating private land, transferring it into the state estate and then reissuing it to private owners. New legislation in2000 widened the AFA's mandate. Ifrequested by 75% o f the owners, it can reconsolidate landholdings in privately irrigated or productive rainfed areas. Since 2000, the AFA has processed 12,000 - 14,000 hectares per year, out o f a publicly-irrigated area o f 200,000 ha. 200. The rest o f this section will examine land policy in the light of the government's stated policy objectives. It will look at the role o f land productivity increases in generating higher farm household incomes, at the possible implications o f landmarket performance for agricultural growth and at landlaw as an incentive for investment and effective natural resource management. As we will see, the legal and institutional framework is muchbetter understoodthan farmers' decision-making. Profile o f LandAvailability 201. One o f the Tunisian government's 21 stated policy objectives is "un revenu meilleur pour l'agriculteur." In this context, it is important to appreciate that increased land productivity alone is unlikely to provide most o f the desired increase inper capita farm incomes. This point can be illustrated by comparisons with high- income and transition countries. 202. Around 4.5 million hectares o f land are classified as agricultural in Tunisia, equivalent to 1.3 hectares per rural inhabitant and 4.7 hectares per person employed in agriculture. 2.9 million hectares are classified as arable, which works out at 3.0 hectares per person employed in agriculture. To put this in perspective, Tunisia's arable land availability per.person employed, although low, is typical for the southern 31Although 100,000 ha remainunder haboustenure, it has negligibleproductivepotential. - 5 9 - Mediterranean region (see Table 15), and is o f the same order o f magnitude as the countries acceding to the EUin2004: 4.8 h d p e r s ~ n ~ ~ . ITable15. Arable landper personemployedinagriculture, 2000 Countrv I ha/worker Libya 16.8 Israel 4.8 Lebanon 4.0 Tunisia 3.O Syria 3.0 Algeria 3.0 Morocco 2.1 I Jordan 1.3* Egypt I 0.3* Source: World Bank / GDIdatabase * largely irrigated 203. A comparison with developed agricultural economies (Table 16) suggests that increasing per capita incomes from agriculture in the long run will depend largely upon farm labour shiftinginto sectors, and by the consolidation ofrural landholdings. A comparison between the Tunisian and Spanish farm sectors, for example, shows that, for Spain's land productivity i s only 70% higher than Tunisia's, some o f which can be explained by climatic and market access advantages, whereas Spain's land to labour ratios are over three times higher than inTunisia. This means that only a third o f the difference between Spain andTunisia's agricultural labour productivities is due to higher land productivity; the rest is due the absorption o f Spanish rural labour by the industrial and service sectors. The difference between Australia and Tunisia's agricultural labour productivities is entirely due to Australia's higher land to labour ratio. Inthe case o f Italy, which has greater natural advantages than Spain, Australia or Tunisia, still as much as 35% o f the difference in labour productivities i s due to Italy's higher landto labour ratio. 204. A comparison with EU pre-accession countries is also revealing. Although Tunisia's value added per worker in agriculture, at USD 2557/year, is o f the same order o f magnitude as that o f countries such as Hungary (USD 339Uyear) and Poland (USD 1182/year), the number of people employed in agriculture fell in Hungary and Poland by 29% and 15% respectively duringthe same period (Table 16). 32Weighted average, excludingCyprus, Malta andthe Slovak Republic. - 60 - Table 16.Labour and land productivi? (2000) Agricultural Arable land Value- Agricultural hdworker Value-added value added (ha) addedha labour force /worker (USDbn) (USD) (USD) Tunisia 2.4 2,864,000 838 941,000 3.0 2,550 Spain 18.5 13,400,000 1,381 1,333,000 10.1 13,878 Italy 27.5 8,479,000 3,244 1,352,000 6.3 20,347 Australia 13.3 50.304.000 , , 264 447,000 112.5 29.754 Tunisian agricultural workforce 4.0 2,864,000 1,381 941,000 3.0 4,202 with Spanishlandproductivity Tunisian agricultural workforce 9.3 2,864,000 3,244 941,000 3.0 9,874 withItalianlandproductivity Tunisian agricultural workforce 0.8 2,864,000 264 941,000 3.0 805 with Australian land productivity Source: WorldBank/ GDIdatabase 205. Moreover, Tunisia's agricultural labour force is not yet on a downwardtrend. Itrose by20%between1993 and2002, while labourproductivitystagnated (see para. 29). Table 17. Aggregate land availability. Source:FAO, LandandWater DevelopmentDivision, WorldBank/ GDIdatabase 206. Water availabilityis also a constraint upon strategies for raising agricultural incomes based upon improvements in land productivity. Only 2.3% of Tunisia's arable land is situated in humid or sub-humid zones with at least 600 mm rain per year. 78.6% is locatedin arid or desert areas with rainfall lower than 300 mm per year, andit i s estimatedthat 88% o firrigable landis already irrigated(Mtimet, 2004). 207. It is therefore in agriculture's long-term interest that industrial and service- sector growth should attract labour out of farming. This process will involve many elements, includingmacro-economicstability, removal o f the pro-agriculturebias in trade policies (see para. 43 onwards), and a favourable business environment. One important component, however, will be the ability of the farm sector to avoid land fragmentationand absenteeism by consolidatinglandholdingsunderthe ownership o f activefarrners. This requiresan effectivelandtenure system andmarket. -61 - Performance of the Land Market 208. Land-ownership trends in Tunisia are causing concern. These are fragmentation, absenteeism, and low involvement o f younger farmers. They are not specific to any sub-sector. A study commissioned by the Direction GCnCrale de Production Agricole in 2002/3 identified absenteeism as land fragmentation as constraints upon cereals production (CNEA, 2005a). 90% o f farmers growing early vegetables [primeurs] are cultivating fewer than 2 ha. Over 50% o f citrus farmers and cereals farmers are absentees. And as far back as 1994, over 62% o f tomato producers were over 50 years old (CNEA, 2005b). 209. Land ownership data (Table 18) shows that land fragmentation is indeed advancing; 73% o f landholdings are currently less than 10 ha in size, up from 64% in 1976. 210. The assumption is that small farmers are uncompetitive. Formal sector wholesalers and retailers inTunisia and abroad will prefer the greater predictability o f supply, low transactions costs and standardisation o f quality offered by larger farmers: for example, 10 citrus growers represent 80% o f export sales, and 4 date exporters have 36% o f the market, and it is estimated that it takes 10 ha o f tomatoes tojustify the cost o f a refrigerated truck. 211. It i s therefore often proposed that Government should actively intervene to prevent fragmentation and to encourage consolidation; for example the World Bank proposed in the 1980s that the Agricultural Sector Adjustment Loan action programme should involve legislation on minimum rainfed farm sizes, with government actively consolidating undersized farms, and that Islamic inheritance laws be amended to prevent subdivision below specified levels. * = estimate, basedon 1980 average farm sizes within each size range. Source: World Bank (1976, 1980 data), MARH(2005 data) 212. Land registration is also seen as a policy challenge. 25% o f titles were registered inthe 1960s, and for about a seventh o f registered titles the registrationwas out o f date, a condition known as a "frozen title" or "titre gelC". By 2001 around half the land stock was registered, and about half the remainder was considered not worth - 62 - registering (Rochegude, 2005). However, reliable up-to-date data is scarce: one estimate (Gharbi, 1998) i s that around 34% of farms have neither title nor a certificate of possession. The suspicion is that frozen titles are also widespread, encouraged by registration costs o f USD30-50 per hectare and land registration taxes which were as highas 15% inthe 1 9 8 0 ~ ~ ~ . 213. However, one should not automatically assume that consolidation and titling are a major policy issue. 0 Much landremains inlarge holdings: 60% inholdings o f 20 ha or more. This is not just because o f large, arid rangeland holdings inthe south. As o f 1980, landholdings o f 20 ha and over were over-represented in the higher-rainfall north-east and north-west zones. Since 1976 the share o f total land in holdings under 10 ha nationwide has only increased by two percentage points, hardly a major inroad into the nation's landresources. Inother words, plenty o f landis still available for large-scale farmingandmarketing34. Secondly, Government (MDCI) classifies 24% o f farms as either "social" or "familial". The practices o f such farmers would probably remain non- commercial even ifthey could consolidate their landholdings. 0 Thirdly, there is no empirical evidence that the land tenure regime making farms inefficient. A recent informal survey estimated that 5.7% o f land was "under-used" nationwide, falling to 1.9% and 1.8% in the more fertile north and central zones. Although the sampling methodology and definitions are not clear, this does not ci priori indicate an important market failure. Fourthly, we do not know whether landowners actually wish to consolidate, but cannot because o f laws and institutions. Ifthe desire to consolidate is not there, policy interventions cannot make it happen. 214. So, is it the land market or the landowner that is stopping consolidation ? Rochegude (2005) found that the Tunisian legal framework is favourable to the titling and registration o f land, although the accretion o f overlapping legislation over the years has made for a complex body o f law. Property transfer and registration taxes are probably no longer a issue. Real estate transfer taxes are now 5%, plus 1% for registration, down fi-om 15% in the 1 9 8 0 ~and ~ falling to 2.5% for successions to children or spouse. 215. If there is a major problem, it may be that the institutional set-up is inaccessible to the ordinary farmer. Landregistration is handled exclusively inTunis, and involves three separate ministries. The Ministbre des Domaines de 1'Etat et des Affaires Foncibres maintains the agricultural land registry. The Ministbre de 1'Equipement is responsible for land delimitation and mapping. The Ministbre de la Justice i s responsible for the Tribunal Immobilier, which takes decisions on land registration. It was estimated in the 1980s that the services responsible for registration had the capacity to handle 22,000 ha per year. N o w that the AFA is 33Real estate transfer taxes are now 5%, plus 1%for registration. 34A localised exception mightbe the landdeveloped for geothermal cultivation, where the lack oflarge holdings limits economies o f scale invegetable marketing (CNEA, 2005b). - 63 - inserting 12,000 - 14,000 ha annually into the system35and the urbanlandmarkethas grown exponentially, registration services may be a bottleneck for the private rural citizen. This would tally with the observation that slow, cumbersome land registration procedures are an obstacle to industrial investors (World Bank, 2000). If user surveys confirm that inaccessible institutions do indeed impede the landmarket, one solution may be to create local "one-stop shops" [guichet unique]. Another may be to promote, through legislation and communication campaigns, the use o f the Certificat de Possession, which is currently only recognised as a "transitory and exceptional form o f tenure" (Rochegude, 2005). 216. But land is a psychological asset as well as an economic one. Even though 64% o fthe population is now classified as urban, rural landownership is perceived as part of one's identity, which encourages absentee ownership. Inheritance is determined by Muslim law and custom. A surviving spouse receives an eighth o f the estate, and surviving sons and daughters the remainder on a 2 to 1 basis. This practice may produce fragmentation and joint ownership, but it makes sense from a social and religious point o f view. All these factors combine to encourage informal tenurial arrangements. Often the owner retains formal title while delegating de facto possession to a relation or neighbour. And it would still happen with a perfectly efficient landmarket. Foreign investment 217. Foreign direct investment in agriculture is politically contentious, but it does promote competitiveness and exports. The receiving country gets not only capital, but technology and a ready-made marketing chain. InChina, for example, foreign direct investment runs at around USD 1 bn per year, primarily in horticulture and floriculture for export (Sellami, 2005). Investors in Chinese agriculture get the same tax advantages as investors in remote and deprived areas. Vietnam passed legislation permitting foreign investment in farmland in 2001 and 15 agricultural projects received foreign investment in 2002 (Vietnam Economic Times Jan 8, 2003). In Chile there is no limit on percentage or duration o f foreign ownership o f joint ventures, local enterprises, buildings, or land(Decree Law 600 o f 1974). 218. InTunisia, foreignownershipofagricultural landhasbeenillegalsince 1964. According to 1989 legislation, moreover, only companies with 100% Tunisian shareholdings can own agricultural land. Many countries allow emphyteotic leases, whereby the rent is inversely related to the value o f investment by the lessee. The aim i s to encourage foreign inward investment without putting land in foreign hands. In Tunisia, however, emphyteotic leases are explicitly banned by the Code des Droits 35The law o f 11August 1976on landexpropriation inthe public interest, which covers suchoperations, laid downextremely onerous procedures. Delays oftwo years between expropriation and compensation were normal, rising to seven years inexceptional cases (e.g. n"8O- 72 du 28 mai 1980). The law of 14 April 2003 may speedthe process up. - 64 - R b e l ~ ~Furthermore, the law o f January 13, 1995 only permits long-term land leases ~ . for SociCtCs de Mise en Valeur Agricole - which are defined as 100% Tunisian- owned 219. Limitations on foreign land ownership are not unique to Tunisia; in Switzerland both rural and urban land ownership is restricted to the Swiss, and in Turkey legislationpermitting foreign ownership o f agricultural landwas struck down by the Supreme Court in its judgement o f 14 March 2005. However, in Tunisia a relaxation o f the 100% local shareholding requirement and the introduction o f emphyteotic and other long-term leases might at least keep the door open for would- be investors without any risk o f surrenderingnational control. 220. On the face o f it, it seems as though Tunisian land law could be an obstacle for foreign investment. But once again, the legal situation is better understood than the economics: we do not know whether potential agro-business investors see farmland tenure as an obstacle. LandTenure andNaturalResources 221. Land tenure is crucial to the sustainability o f renewable natural resources, such as rangelands and forests. Analysis o f an extensive sample o f case studies suggests that there is no single "correct" form o f tenure: the arrangement that controls use effectively ina specific context is the appropriate one (Gibson et al. 2004). 222. The legal instrument for the government's forest and rangeland conservation objectives is the Code Forestier. The Code applies to 2.8 million ha, up from 0.7 million in 1987, and cuts across several forms o f landtenure: 1million ha inthe state estate, 1.7 million ha o f collective rangelands [parcours collectifs] and 47,000 ha o f private property. 223. The government's performance (16,000 ha) against 2002-5 Plan targets (12,000 ha) for plantation forests in the state estate cover suggests that state ownership is not a constraint on the growth o f plantation forest cover. However, the highest rate o f achievement o f Plan targets for both forest and rangeland plantations took place inprivately-owned land, where the execution rates were 163% and 118% respectively for 2002-6. Land tenure has been identified as an obstacle to afforestationalongside oueds androads, inurbanareas andinirrigation schemes. 224. Once again, there is an issue about we represent what is happening in the forest sector. In the logic o f quantitative planning, the focus is on the number o f hectares planted and the number o f GFICs . This is what drives the Plan, which is what drives data-collection, Part I1budget allocations and the work programmes o f Ministry staff. But one can fairly ask whether we really understand the forces that 36Art 191 Code des Droits r6els :d o n t interdits, b compter de la date d'entrkeen vigueur duprksent code, toute constitution d'emphytkose, tout renouvellement des baux emphytkotiguesen cows et toute constitution d'un droit de superjcie, d'enzel ou de kirdar D. - 65 - drive local people to conserve natural resources - or to over-exploit them. And without understanding them, can we manage them? 225. Indeed the degradation o f non-plantation state forests is a concern. The code forestier allows the population limited use rights: the collection o f dead wood and brush, grazing, the collection o f specific forest products and the cultivation o f specified plots o f land. These rights are free o f charge, but restricted to production for domestic consumption. Inreality, however, government cannot stop illegal and excessive use o f forest resources; the government's limited resources and enforcement capacity mean that state ownership is no guarantee against land degradation (Rouchiche and Abid, 2002). Although land inthe state estate cannot be alienated, the government's response has been pragmatic: Article 75 o f the Code Forestier was revised in 2005 to permit Groupements Forestiers d'IntkrCt Collectif (GFIC) to enter into concessions o f up to 30 years for such land. The effect will be to introduce elements o f common property tenure for state forest lands. Land UsePlanning 226. Some policymakers are worried that urban expansion is eroding the stock o f productive agricultural land. It i s important, however, to put this perceived problem inperspective. Firstly, at an estimated rate of 4000 haper year (Mtimet, 2004) only one per cent o f the stock o f agricultural land is transferred to urban use every 12 years. This is less than a fifth o f the rate o f loss to erosion. Secondly, the conversion o f land from agricultural use usually indicates that it has a higher value as residential or commercial property, which is a net gain to the economy. Thirdly, global experience suggests that it is very hardto enforce controls on the conversion o f fertile farmland unless an ample stock o f alternative construction sites is available. Inshort, the conversion o f a small fraction o f Tunisia's agricultural land to other uses is an inevitable and positive feature of Tunisia's development. 227. Having said that, there may be scope to improve land-use planning. The Code de 1'AmCnagement du Tenitoire et de 1'Urbanisme o f 28 November 1994 provides a legal basis, among other things, for ((une rkpartition rationnelle entre les zones rurales et urbaines ))(Article 1). Article 5 o f the Code provides for the preparation o f G schkmas directeurs d'amhagement D, as an instrument to balance different land uses. This article makes specific reference to agriculture. However, whereas the code specifies detailed mechanisms for enforcing urban development plans, there appear to be no legal instruments to ensure that land identified as agricultural in the "schema directeur" is used as such. With land-use planning passing to the regions in 1989, the question is how the "schema directeur" applies to the decisions o f local government. - 66 - The Way Forward 228. Insummary, the first step towards defining a landmarket strategy is to define the problem. The problem o f farmers not aligning themselves with government policy by consolidating andregistering landholdings may not be a problem at all from the farmers' perspective. It is therefore essential to study the reality o f landtenure on the ground, based upon rigorous quantitative andperceptional data. This study would examine : The actual land tenure situation, not only land's formal legal status but also de facto ownership and management arrangements; Whether it is possible to demonstrate a causal correlation from land tenure status to farmingpractices; How the under-utilisation o f landcan be meaningfully defined andmeasured; Which legal, financial, institutional and cultural factors determine people's participation in the rural land market, and how they affect such phenomena as fragmentation, frozen titles and absenteeism; What are the costs and capacity constraints involved inlandtransactions; Whether the ban on land ownership discourages foreign investment in agro- business; H o w different land tenure regimes affect the sustainability o f forests and rangelands; Whether farmers are ever unable to make a rational decision about converting land to non-agricultural use e.g. because they are coerced by powerful local interests. 229. To reiterate a point made earlier, the recommendation for further field-level study is not a pretext for a lack o f firmer prescriptions. Field-level research into farmers' perceptions and attitudes is a very important component o f the proposed reorientation o f the MARH's role. As in other domains, the role o f Government in landtenure would become less one o f laying down administratively logical solutions. Instead, Government would increasingly seek to understand what stakeholders require from a landtenure system and respond to that perceived and expressed demand. 230. Based upon the findings o f this study, a number o f policy measures may be considered, including: Creating local one-stop shops for landtransactions; Simplification o f administrative procedures; Updating o f landlaws into an integrated code; 0 Legislation and communication campaigns to promote recognition o f the Certificat de Possession; 0 Fiscalpenalties for landfragmentati~n~~; The introduction o f long-term or emphyteotic leases for foreigners; 0 New mechanisms for zoning land for agricultural use. The Carte Agricole (see Annex 2) provides much o fthe information required. 37 However, landtaxes are also an incentive for non-registration. - 67 - 0 A final policy option: recognising that the current situation reflects farmers' cultural and economic priorities in a time o f rapid social transition, and maintaining the status quo. F. SOCIAL PERSPECTIVESON AGRICULTURAL POLICY Introduction 231. Agriculture is not only a source o f growth. It is the basis o f livelihoods, so reform has to be analysed through a social as well as an economic lens. The objective o f this section, therefore, is to i)describe agriculture's social context; ii)outline the potential social implications of the changes suggested in earlier sections and iii) tentatively offer some potential mitigation measures, to be investigatedby a PSIA3*. 232. With a longstanding Government commitment to social development, the lothPlan i s that all social groups should participate in the benefits o f economic Tunisia's social and gender equity indicators are among the region's best. A goal o f and social development. Since the 9thPlan, Government has made efforts to promote beneficiary participation in design and implementation o f agriculture projects. Particular attentionhas also been given to the inclusion o f rural women. 233. Tunisia enjoys one o f the lowest poverty rates in MENA. The headcount poverty rate fell from 8% to 4% between 1995 and 2000. Nevertheless, important rural-urban differences exist. 8.3% o f the rural population i s poor compared with 1.61% o f the urban.Pockets o f severe poverty are found inremote rural areas. 234. In1995 povertyrates were highestinthe North-West, followed bythe Centre- West and the South. Together these three zones accounted for about 80% of the 2000 data revealed that poverty levels were falling more steadily inthe North- West than the Centre-West and South, perhaps because o f better rainfall and access to remittances. By 2000, the Centre-West included 40% o f the poor as against only 13% o f the non-poor. Also by 2000, the Southhad as manypoor as the North-West (World Bank,2003). Thesocial significance of farming 235. Poverty rates are not particularly highinthe agriculture sector as a whole. But in2000, households headedbylandless agriculturalworkers were amongthe poorest- second only to the unemployed (World Bank, 2003). Poverty was also associated with large household size. 38A Poverty and Social Impact Analysis (PSIA) evaluates the distributional impacts o f policy reforms on the well-being o f different stakeholder groups, with particular focus on the poor and vulnerable. It also addresses issues o f sustainability o f reforms and the risks to successful implementation arising from the social impacts o fpolicy changes. 39Republic o f Tunisia: Social Conditions Update. Volume 1(2000). - 68 - 236. Demographic change i s transforming rural areas. The rural population fell from 45% o f the total population in 1986 to 34% in 2001. 67% o f Tunisians now reside in cities and towns. The agricultural labor force represented 16% o f the total labor force inTunisia in 2004, down from 24% in 2000. 78% o f farmers live inrural areas. 237. Farm labour is essentially family-based and accounts for approximately one million persons, most o f whom are perhaps female. Only a minority o f famil workers work full-time infarming: the equivalent o f 275,000 permanent employees4K . There are only 190,000 salaried agricultural workers. 238. The average age o f farmers is increasing rapidly as young people migrate to the city. In 1995, the date for which we have the most recent data, the average age o f farmers was 53, with 37% over the age o f 60, compared with 20% inthe 1960's. The level o f education also remains low among farmers; 88% have not gone beyond primaryeducation. 239. 56% o f farmers depend principally on agriculture for their livelihoods (1995 data), for 35% agriculture is an important secondary source o f income, and for 6% agricultural production makes a limitedcontribution. A larger farmer is more likely to depend principally on agriculture than a smaller one. There are also regional differences. For example, half of Medenine farmers have an important secondary activity, compared with 15% inL e Kef. 240. Across the region, women are not leaving agriculture as fast as men and a growing share o f women's jobs are in agriculture (World Bank 2004a). Tunisian women have always played a critical - but often unrecognized - role as non- remunerated family labour, hired labourers and, a minority, as farm managers. A recent survey noted that 96% o f female members o f farming households said that they were unpaid family agricultural workers. Agriculture is the third biggest employer o f Tunisianwomen. Most female salaried agriculturalworkers work seasonally andtheir numbers are increasing, especially in larger tree-farming and market-gardening enterprises. 241. Moreover, woman's role in agriculture is becoming qualitatively more important. The ageing rural population, male out-migration and diversification into livestock all place new responsibilities upon women (World Bank, 200%). However, extension services target men more than women, even where women are the de facto farmmanagers. 242. This is in a context where rural women from poor families still have fewer educational opportunities than other social groups. Illiteracy among older rural women remains a particular problem. In 1999 half were illiterate compared with a quartero furbanwomen (World Bank2005~). 40Presentationby M. Badr B e nArmnar, Directeur GBnCral Etudes et DCveloppement Agricole, Ministtre de I'Agriculture et des Resources Hydrauliques, Tunis, December 2005. - 69 - 243. Farm size i s extremely skewed. 53% of.farmers use less than 5 hectares while a quarter o f the country's agricultural land i s cultivated by the 1% o f farmers with over 100hectares. 244. Government classifies farms into three categories on the basis of farm size and incomes: - Large Farms (les Grandes Exploitations); - Small and Medium Sized Farms which are economically viable (Les Petite et Moyennes Exploitations Agricoles a Caractere Economique - PMEACE) - Small Family-Based Farms (La Petite Agriculture a Caractere Familial et Social - PACFS4'). The Ministry o f Development and International Cooperation (MDCI) classifies 24% o f farms as either `social' or `familial.'42. 245. To understand poverty and vulnerability in Tunisian farming, we will look at the PACFS in detail. At the same time one should bear inmind that some PMEACE have small landholdings and may be vulnerable to policy change. PACFS 246. There are around 186,000 PACFS. They cultivate one millionhectares with an average farm size o f 5.7 ha. It is difficult to generalize about their farming systems, but: - In irrigated areas, their income comes primarily from tree-crops (olives, dates and orchards: 50%) and livestock (25%), as well as market gardening. - Inrainfed farms inthe Northern andCentral areas cropping systems andincomes are often dominatedby `grandes cultures' [field crops] andtree-crops. - Southern rainfed farms have fewer field crops and more tree crops and small ruminants. PACFS treecrops are mostly olives, almonds and pistachios. Mixed farming modes o f cultivation include irrigated olives, market gardening and milk production. 247. PACFS depend on family labour with the occasional use o f seasonal hired labor. Women tend to work on livestock and poultry rearing, milking, artisanal 41PACFS are definedbythe following features: i)disposer d'unrevenue net d'exploitation (agricole) - RNE-inferieur a 3500 DT, seuil de viabilite de son exploitation agricole; ii) l'agriculture c o m e avoir activite principale (ou a defaut c o m e revenuprincipal) ;iii)Ne pas disposer d'une main d'aeuvre salariee permanente (non familialte). 42Interms of farm size, three types of farmers are generally recognized: small farmers (the 53% who farm less than 5 hectares); mediumsized farmers (farming between 5 and 10 hectares) representing 20% o f the farming population as well as those between 10 and 50 hectares (24% o f farmers); and large-scale farmers, who farm between 50 and 100 hectares and represent 3% o f the farming population, working 2% o f the land. - 70 - activities, food processing and specific field tasks such as hoeing. The growth o f the dairy andbeef sectors, noted earlier, has mainly been achieved through female labor. 248. PACFS, especially the poorest, use little formal credit. Physical access (distance) andbureaucratic procedures limit their access. But 87% o f PACFS say that they would use credit if it were more accessible and adapted to their needs. Informal loans between family members, neighbours and so on, are fairly common, especially among the better-off PACFS. 249. Specific studies o f PACFS list the constraints noted in earlier sections as facing the farm sector as a whole. These are: low capacity to invest, insufficient access to credit, weak fanner organization, poor access to extension, training, research, andmarket information, an ageing population andlandfragmentation. Managing the social aspects of institutional change 250. The previous sections o f this report have described how the relationship between Government and farmers could be realigned. The suggestion is that Government could re-engineer institutions and processes to make them more responsive to farmers' needs. This approach could apply to the extension service, research, farmers' organizations, budgetary management, and credit and land tenure institutions. 251. But we have seen that farmers are a very diverse group. As Government tries to become more responsive to their needs, there is a risk that the dialogue will be captured by larger farmers and men, to the exclusion o f smaller farmers, the landless andwomen. 252. A key role for Government, therefore, will be to ensure that fanner- Government dialogue (e.g. consultations andresearch) are structured so as to embody the diversity o f actors. The key principle is disaggregation by socio-economic status and gender. It should apply, for example, to the choice o forganizations consulted, the constitution o f focus groups, the design o f surveys, the definition o f targets and the selection o f monitoring indicators. The MARH can build upon its experience with farmer participation (e.g. in ODESYPANO) and the collection o f disaggregated survey data (e.g. the recent PACFS study). Managing the social impacts of market reform 253. The potential social consequences o f lowering tariffs and guaranteed cereals producer prices require thorough analysis. The issues are complex and only partly understood, so the following section identifies critical issues for further study. 254. Liberalization will affect different stakeholders via employment, prices (production, consumption and wages), assets and transfers (Chemingui et a12005). It affects the demand for labor, which will affect especially the welfare o f low income and more constrained households. The precise social impacts are complex to predict, - 71 - and will depend on the context, nature and pace o f the reforms themselves. At this stage we can only make general predictions. 255. Section 1II.A identified very roughly the winners and losers from market liberalisation. Seen from a social perspective, some important points emerge: - The farms that are not harmed by liberalisation are mostly in two o f the poorest regions o f the country, the Centre-West and South. The sub-sectors classified by this model as `winning' are livestock, tree crops and horticulture which are geographically dispersed andaccount for around 60% o f agricultural labor use. - The `losers' are the cereal producers, o f all sizes, in the north and north-west, specifically L e K e f and Beja. These Governorates are among the poorest in the country and include significant numbers o f PACFS growing cereals with family labour. 80% o f these cereal farmers are over 40 years o f age. Relatively few farmers there have non-farming sources o f income: 15% in L e Kef. A high proportion o f government farming operations in state lands [terre dominiales] are cereals-based and in the fertile North and North-West, and would therefore fall into this category. These farms employ a poor seasonal labor force, many o f whom are landless andvulnerable. - Liberalization may create better jobs for higher-skilled workers in h i t s , vegetables and agro-processing. - The poor will receive the greatest proportional benefit from lower food price. So, as a starting point, the initial focus o f analysis should be on the vulnerability o f small family owned farms and cereal laborers, inthe North andNorth-West. 256. Tunisia's unemployment rate is highat 15% (INS, 2000), and the employment impacts o f liberalization are a major policy issue (World Bank, 2004a). With market liberalisation labor will be `released' from agriculture to other sectors, and within agriculture from cereals to other products43. 257. Faced with partial liberalization 13,300 jobs would likely be affected (Ideaconsult, 2005a); the CGE model estimated that some 87,000 jobs would be `shifted' out o f agriculture inthe case o f total liberalization. Iffield crops (essentially cereals) alone were liberalized fewer farm jobs would be lost, approximately 9,000 because these crops have a relatively low labor requirement and it is also assumed that farmers would diversify into other crops with higher labor use per hectare. 258. Until now, however, the non-agricultural sector has not been able to draw labor out o f farming into more productive activities. The aggregate employment structure has remained fairly stagnant. Recent evidence from the World Bank's Employment Strate y for Tunisia shows that the target o f 80,000 jobs per year (outlined in the lot Development Plan) hinges largely on jobs in agriculture and 8 administration, neither o f which are necessarily long term viable options for 43Evidence from Tunisia suggeststhat workers shift from agriculture to construction when agriculture is doing badly, because construction also employs low-skilled labor. - 72 - employment creation (see section on rural land). There may be significant costs entailed inretraining displaced labor, some o fwhom might be `untrainable'. 259. So there i s a real risk that the potential gains from reforms will not be reaped because labour cannot shift to more competitive sectors o f the economy. Less skilled workers, and those unable to adapt, will be most vulnerable. Social tensions may increase and coping strategies (diversification o f income sources) would be strained (World Bank, 2001). 260. Moreover, as noted above, women tend to move out of agriculture at a slower pace than men. The dependency ratio o f rural populations could be further exacerbated with young girls and women being left behind to care for elderly relatives. Inother countries, as males migrate out o f agriculture women increasingly become subsistence farmers, because they do not have enough labour for more intensive agriculture. This is often accompanied by increasing poverty. Socialpolicy options 261. Social measures will need to be as diverse as the rural population itself, reflecting the age, gender and geographic location o f the affected farmers. O f particular concern are those who lose out from reforms and find it difficult to adapt. This group is likely to include the poorest and most vulnerable, including the landless, the old and sick, seasonallaborers etc. Becausethe dynamics o f change will only become clear as it unfolds, implementation should be gradual and accompanied by a thorough Poverty and Social Impact Analysis (PSIA) with the full, participatory involvement o f stakeholders. 262. The targeting o f active labor market programs (ALMP) will need to be improved. While the Government plays a central role in social protection, through largepublic sector employment, heavy regulation o f the labor market, tight control of vocational training, active labor market programs (ALMP's) (World Bank, 2004a) and social safety nets, targeting of these programs is inadequate (World Bank, 2 0 0 4 ~ )Job.creation schemes, mainly micro-credit and youth interventions, are the ~ ~ most important category o f ALMP spending, representing90% o f total expenditures. The majority of these programs benefit post-secondary graduates who account for only 6% o f the unemployed and are mostly urban. However, the unemployed are dominated by those with primary schooling or less (86% o f the unemployed). In addition, despite high spending, only a small share o f the workforce participates in ALMP's - 5.3% o fthe labor force in2002 (World Bank 2004~). 263. The misalignment o f ALMP programs i s illustrated by the micro-finance programs which in principle are targeted towards unemployed and at-risk workers. Less than one-third o f these programs are inrural areas and one study o f three micro- The GOThas shifted its focus of late from untargeted social assistance programs (i.e. universal consumer food subsidies) to targeted assistance programs (regional development programs, cash transfer programs etc). See World Bank, Republic o f Tunisia, Social Conditions Update - 73 - finance programs45 noted that the beneficiaries are predominantly male (75-80%), between 20-29 years old and have some education and training qualifications (half have secondary schooling). The programs also mainly serve urban areas (6943%) and are heavily skewed towards self employment inmanufacturing. 264. A comprehensive contributive social security system exists, which covers urban wage-earners, agriculture workers and the self-employed. However, evasion among agriculture workers is high and, as noted above, the majority o f agricultural workers are family-based labor and are not covered (World Bank, 2000a). 265. The GOT could therefore consider introducing targeted social protection schemes, re-employment programs and passive income support programs (World Bank,2004a). ALMP's could emphasize support forjob searches andretraining. 266. Significant agricultural employment data gaps exist and the quality o f information systems to monitor and evaluate employment data is poor. Specifically, employment surveys and impact evaluations are o f uneven quality and analysis and the disseminationo f agriculture employment information (fulltime, seasonal andpart time) is inadequate and insufficiently disaggregated by gender (World Bank, 2004a). Existing surveys (household and enterprise-level) ought to supply better quality information on labor demand and supply on a regular basis and improvements are recommended inmethodologies and definitions used (e.g. labor force projections). It is also recommended that collaboration among concerned ministries be strengthened to better exploit existing information and evaluate the labor-market situation as well as monitor programs, and adapt policies according to market needs. Dissemination o f information (micro-data from surveys) to various stakeholders and analysts could be improved to more effectively analyze the data and to buildconsensus on employment policy issues. 267. Farmers will still need high quality services. As mentioned above, any strategy to make research, extension and land tenure services more responsive should focus on the needs o f the most affected and vulnerable groups. Government could strengthen and expand well-established community development organizations in rural areas, such as ODESYPANO, a regional development agency assisting rural populations in diversifying their agriculture base and accessing micro-credit. As it happens, ODESYPANO is located inthe Northwest, where the small cereals farmers most vulnerable to price reform are concentrated. 268. Other countries such as Mexico (see Box 4) and Turkey are now de-linking farm support from production. In this way, they achieve the political objective o f a controlled and predictable income transfer to rural society - but without the market distortions and fluctuations intransfer values that go with production subsidies. 45BTS(TunisianSolidarityBank's), 26-26 andFNE21-21(WorldBank World Bank Republic ofTunisia EmploymentStrategy2004, p. 82) - 74 - Box 4. De-linking income support fromproductioninMexico PROCAMP0 - Programof direct supportto rural areas PROCAMPOwas introduced in 1993 to compensate farmers for the deprotection associated with the North American Free Trade Agreement (NAFTA) and the GATT Accords. Objectives were: (a) political: to help make the free trade agreement acceptable to farmers; (b) economic: to provide farmers with liquidity to adjust production to the new set o f relative prices; and (c) social: to compensate farmers for lost income, alleviate poverty impacts and reduce out- migration. The program provides farmers (defined as those with the legal usufruct rights over the land) with a fixed payment per hectare that is de-linked from current production. The number o f eligible hectares was determined by area each farmer had under the crops covered (initially nine) duringa three year reference period (1991-93). Since there are two agncultural cycles, payments are made twice a year for the area that had been planted in the corresponding cycle. There is a ceiling o f 100 ha eligible per farmer. The rate per hectare is the same nationwide. In 1997, payments averaged US$329 per recipient and US$68 per hectare (compared to US$83 per hectare for 2002). Only producers registered at the start o f the program in 1994 are eligible, and they have to re-register at the start o f each crop cycle. Although since 1995 the aid is decoupled (i.e. not linked to the cultivation o f specific crops), the land must be used for either crops, livestock or forestry activities or as part o f an approved environmental program. The intention i s to make payments only to active farmers and to keep up the level of rural activity (and so avoid out- migration). Payment for each crop cycle is made by check, collected from one o f the 700 local agncultural extension posts (CADER or Centros de Apoyo a1 Desarrollo Rural). PROCAMPO "qualification certificates" can be used as collateral for banks or input retailers, giving beneficiaries more timely access to their PROCAMPOresources duringthe planting season. The program i s intended to be phased out once the fifteen year transition period provided for Mexico's agriculture sector under NAFTA i s concluded, i.e. in2008. Alianza para el Campo-Alliancefor the countryside o f agncultural sector restructuring. To support this, a second program - Alianza - was introduced Inorder to makethe mostout o fthe NAFTA-drivenreforms, Mexico decidedonabroadprogram in 1996to improveproductivity andto promote the transitionto higher value crops. The program comprises: matching mants to finance productive investments; and sup~ortservices for a wide range o f agricultural sub-programs (24 in2002) designed to transfer modern technology, promote investment ininfrastructure, improve livestock health, and support integratedrural development. Administration and decision making for Alianza para el Campo i s decentralized to the states. Matching funds are required both from state governments and from farmers. According to the different programs, producers contribute an average o f 50 percent, the federal government 32 percent, and the state governments 19 percent. Farmers submit requests to the Rural Development Districts (DDR), the Support Centers for Rural Development (CADER), and the state coordinators. The requests are approved by committees which manage each o f the Alianza sub-programs. Farmers present the approval document to their supplier or contractor. The farmer pays his share and the supplier or contractor claims the balance from Alianza. Certain aspects o f the program could be more pro-poor. Some sub-programs require group participation, which may - 75 - be difficult for the A voucher-based program would allow poorer farmers to buy their inputslocally instead o f traveling to the government-certified distributors. Also, the program is rather cumbersome and expensive to administer. More decisions - for example on the nature o f sub-programs - could be decentralized. Oportunidades(formerlyPROGRESA) Opportunities - After several years implementation o f PROCAMPO, it was evident that rural poverty was persisting. Therefore, PROGRESA was introduced in 1997to alleviate poverty through monetary and in-kind benefits, and to invest in education, health and nutrition. PROGRESA was re- launched as Oportunidades in 2002. PROGRESA has as its basic objective to improve the education, health and nutrition o f poor families, particularly children and their mothers, by providing education and health services, monetary assistance and nutrition supplements. The program provides cash transfers to families in exchange for regular school attendance and visits to healthposts. The payments are provided directly to mothers or the female head o f households. The program is expensive to run, and as it relies on government services and the road infrastructure may exclude very poor households in remote areas. A second problem is the quality o f education provided inschools. There may be considerable improvements to be attained by linkingbenefits to performance, such as grantingbonuses to encourage successful completion o f a grade. 269. A key point, however, is that well-designed rural social programmes should not j~&be seen as mitigatory measures to accompany agricultural reforms. They are desirableper se. We have seen inearlier sections that subsidizing outputs, credit, loan arrears and water is inefficient and socially regressive, because most o f the benefits go to better-off farmers. So, as well than saying that social measures are needed as mitigation, we can also say that social policy reform will achieve Government's social objectives more effectively than market distortions. Next Steps 270. A more precise understanding o f the social impacts of reform is needed, so a participatory Poverty and Social Impact Analysis (PSIA) is the most appropriate startingpoint. Critical areasto be studied include: The political costs of liberalization. If the Office des CCr6ales' operations are reformed, are there any unforeseen impacts on food prices. If so what are the implications, particularly for the urbanpoor? What are the precise spatiaYgeographica1 consequences ? Which regions will be most impacted andhow? While the above analysis has pointed out the potentially vulnerable groups (women left behind by male out-migration, - 76 - landless farmers, unemployed youth, the elderly etc) how are these groups distributed geographically andhow will they be affected? 0 How will non-farming agriculture-related activities be affected? In other words what will the knock-on effects for rural households be? 0 How will liberalization affect social organization and cohesion? Will the social fabric o f agriculture households and communities be affected? If so how? 0 What is the relationship between agriculture liberalization and land tenure? Who has access to what and how will access to assets affect decision makingaround agriculture? - 77 - IV. OPTIONS FORTHE FUTURE 271. The preparation o f the llth 12th Plans i s an historic opportunity for and Government to establish new policy directions. When faced with such a major turning-point, it i s important to keep the big picture inview. It i s easy to lose sight o f the overall vision, and to become distracted by the minutiae o f detailed policy adjustments. 272. The vision suggested by this report i s one o f competitiveness throunh responsiveness. Growth would come from a private sector47 that is empowered to respond to consumers' demandfor quality food. The role o f Government would be to respond to the private sector's demand for first-rate public goods4*. Governance changes in producer groupings and management changes in the MARHwould make them in turn more responsive to the needs o f farmers and traders. Targeted social programmes would address the needs o f the poor and vulnerable more effectively thanthe farm subsidies that they wouldreplace. Table 19 illustrates how the themes o f competitiveness and responsiveness appear in the report's recommendations on the different areas o f agricultural policy-making. The remainder o fthis section will look at these areas inmore detail49. 273. Although the policy options are presented below in an a la carte" list, those " that involve removing price distortions are more important than the rest. Price liberalisation i s an "apex" reform. This is because farmers can only take full advantage o f other reforms if the price system rewards them for doing so. So each individual recommendation below should be understood as part o f an integrated hierarchy o f actions. Inorder to illustrate the linkages between them, Box 5 presents a more detailed roadmap for the phasing andprioritisation o f reforms. 47Producers, processors, traders, exporters etc 48for example quality regulation 49The background papers also provide a large quantity o f detailed recommendations on specific crop supply chains. - 78 - Natural resources Quantitative physical investment targets Understanding and managing the incentives that dnve natural resource use MARHresources Quantitative physical investment targets Responsiveness to farmers' perceived needs Finance subsidies I Incentivising lenders Rural land II Directedand Titling consolidation II Understanding and managing the incentives that drive resource use Social aspects I Agricultural subsidies II programmes Targeted active and passive socialI A. POLICY AND INSTITUTIONAL FRAMEWORK (1) Trade Policy The suggested strategy is that trade policy should be used to encourage farmers to reallocate land, labour and investment in response to domestic and international market demand.This would initiallyinvolve: (a) A substantial phased reduction -- beginning immediately and completed in a few years -- o f tariff protection for hard and soft wheat, accompanied by a matching reduction in the guaranteed producer price (for as long as that instrument remains in place). More analysis would be needed o f the best design and sequencing o f tariff- reduction, price-reduction andprivatization o f the Office des CCrCales' functions. (b) The total liberalisation o f barley imports, and also o f livestock concentrate and its components. This would not only benefit livestock producers, but would also encourage farmers to move out o f a crop inwhich they are not competitive. (2) Domestic Policies The proposed strategy here is to allow supply chains to respond to domestic consumer demand. This would involve: (a) Measuring the incidence o f benefits from consumer wheat price subsidies and, ifthey are large enough, identifying how to provide such resource transfers alongside a liberalised domestic cereals market; (b) Transferring Office des CCrCales commercial operations to the private sector. Trade in cereals would be undertaken by private enterprises. A privately-governed interprofessional organisation would be responsible for supporting the physical exchange o f products, creating a system o f forward pricing, risk management and the standardisation o f contracting practices. (c) Establishing a small strategic cereals reserve, with releases to be targeted at vulnerable consumers in times o f exceptional crisis. This reserve would not be used for market intervention. - 79 - (c) Inpartnership with private sector organisations, developing a food safety and animal health assurance system, applied to products for export and domestic consumption in an integrated manner. (d) Removing controls on retail margins and other price controls. Competition in food markets will ensure reasonable margins and the Ministry should focus its interventions on identifyingand attacking collusive behaviour. (e) Developing appropriate norms and standards inconjunction with (privately-governed) inter-professionalorganisations. (f) Ensuring that the objectives defined for the Ministry o f Agriculture under Performance-Based Budgeting emphasise quality. - 80 - Box 5. Phasing andprioritisation of reforms The experience o f other reforming countries helps us to distinguish between reforms that are urgent, high-priority and relatively easy, and those that need to be implemented more gradually. Of course, a much more detailed analysis i s needed to map out the best path for Tunisia. However, one can still sketch inbroad terms how reformmightbeprioritised. Priority actions that arepolitically less diflcult (1) The allocation by transparent public auction o f import quotas (e.g. for cereals, sugar) and export quotas (e.g. olive oil). Ths would greatly improve transparency and efficiency, and encourage private investment and competition. The fiscal impact could be positive ifprivate bidders can use quotas more efficiently than public boards. (2) The elimination o f export controls, except for subsidisedproducts. (3) The privatisation o fthe Cereals Board's transport, port andretail operations. (4) The elimination of retail margin controls (assuming, as CNEA has reported, that they are widely circumvented. Ifnot, this recommendationjoins those below.) Actions to be implemented more gradually (1) The reduction of tariffs and the increase o f quotas for cereals, accompanied by a temporary programme of area-based cash transfers. The experience o f Mexico, Turkey and Romania is that such programmes canbe set up quickly. (2) The abolition o fthe Cereals Board's domestic purchasing. a. The fust phase would involve sub-contracting its purchasing to private traders. During thls phase, transport subsidies for isolated areas would maintain the single domestic price. At this stage, the producer would feel no negative effects. b. The second phase would involve the gradual removal of the transport subsidies, which would bring about price liberalisation. l%s i s when the temporary cash transfer programme would be introduced. c. The h r d phase would open the way for direct sales by private companies to millers. It will be essential that the Cereals Board's operations are not allowed to discourage privatetrading. d. In the final phase, the Cereals Board would stop its commercial operations altogether. Whilst it is emphasised that social and economic analysis does NOT justify cereals consumption subsidies, Government may wish to maintain them by via subsidies to mills and/or private traders. B. PROFESSIONALORGANIZATIONSAND SUPPORT SERVICES The suggested approach is to encourage producer groupings to become more responsive to the perceived needs o f their members. Elements of this strategy would include: - 81 - (a) Redefiningthe governance arrangements for UTAP and other associations (GIPs, GDA, GIC etc..) to encourage the perceptionthat they are owned by farmers and not arms o f government. (b) Create a diversity o f legal frameworks for farmer organisations and avoid imposing the GDAP model as a standard solution. Whilst maintaining the new SMSA concept, develop a legal basis for farmer-governed, independent, profit- making cooperatives, updating cooperative legislation and regulations to be in line with best practice defined by the international cooperative movement. (c) Redefining Government's role as providing a facilitating framework rather than becoming directly involved in management. A semi-public agency with majority farmer ownership would promote farmers' groupings and agricultural extension. This would primarily involve capacity-building, linking different types o f organisation andhelping cooperatives form unions and apex organizations. (d) Using actual Government services, rather than the associative sector, to perform core government functions like regulating phytosanitary standards and food safety. c. MANAGEMENT OFWATER The broad approach in the irrigation sector would be to make the use-value o f water, rather than physical investment targets, the driver o f policies and plans. This would involve: (a) A shift inemphasis from water-mobilisation to integratedwater management. (b) The use o f cost-benefit analysis to determine least-cost ways o f achieving the country's water mobilisation objectives. (c) Strengthening the capacity o f the central services o f the MARH,with emphasis on strengthening the use o f economic analysis. (d) The extension o f the two-part tariff. It would also be useful to consider the introduction o f replacement costs inthe calculation o f the fees. (e) The phasing out o f subsidies for the replacement o fwater-saving equipment. (f) Endingthe policyo fprovidingwater at a reducedprice for cereals. (g) Continuing the transfer o f downstream water management to user groups. (h) Monitoring the conversion of GICs into GDAPs to see whether these multi- purpose groupings can really serve as water-management bodies cooperatives at the same time. D. RESOURCINGTHE AGRICULTURAL SECTOR The broad approach to the financing o f the agricultural sector would be to adjust the incentives framework underwhich the financial sector and the MARHitselfoperates, so as to make lenders and the MARHbudget more responsive to supply chains' needs. This would involve: - 82 - (a) The MARH should prepare for and commit itself fully to Perfonnance-Based Budgeting.This will involve: creatinga PBBmanagement cell withinthe MARH, bringingtogether the functions o f strategic planning, administration, finance and investment; establishing an implementation plan and mobilizing the required technical assistance; launching a communication and training campaign for MARHstaffat all levels; anddeveloping along-term staffing andhumanresource development plan, taking account o f the MARH's new role and the high rate o f retirements expected incoming years. (b) A study ofthe demand for credit anddecision-makinginthe credit market. (c) Adopting international best practices in micro-finance, such as: allowing micro- finance institutions to set rates that cover their costs, the use o f subsidies for start- up costs only, the dismantling o f competing "social" credit schemes and legislation facilitating the creation and growth of micro-finance institutions. (d) Recastingthe relationship between Government and the BNA so as to clarify the incentives framework for BNA management. (e) Shifting the drought response away from ad hoc input subsidies towards transparent transfers. The fiscal impact o f such programmes could be mitigated by the purchase o f drought re-insurance. E. RURALLAND The overall suggestion is that Government should study fanners' land-use strategies and support the creation o f a flexible market inwhich these strategies can be pursued more easily. (a) It is essential to understand phenomena such as absenteeism, fragmentation, frozen titles and an ageing farming population as part o f Tunisia's current social and cultural transition. (b) Government should study the relationship between landtenure and landuse. (c) Government will then be able to adopt specific policies in the light o f the study's findings. They could include: - The creation o f a simplified single landcode; -- The creation o f local one-stop shops for landtransactions; Legislation and communication campaigns to promote recognition of the -- Certificat de Possession; Fiscal incentives to landconsolidation; Utilisation o f the Carte Agricole to inform land-use planning; - 83 - - Creation o f an effective institutional mechanism to enforce land-use planning - decisions; Provisions permittingemphyteotic and other long-term leases for foreigners. F. SOCIAL PERSPECTIVES AGRICULTURAL LIBERALIZATION ON The approach suggested is to deploy targeted social programmes, which will achieve Government's social objectives more effectively and at less cost than farm subsidies. Special elements o f this approach would include : Structuring farmer-Government consultations (e.g. on research and extension) andthe governance o f farmers' organisations so as to ensure diversity o frepresentation. Increased emphasis on the provision o fregular extension services to women. A Poverty andSocial ImpactAssessment (PSIA), focusing onthe distributional impacts o fproposed policy reforms on the well-being o f different stakeholder groups, with particular attention to the poor and vulnerable. Improved dissemination and inter-ministerial collaboration on the analysis o f labour market data. The improvedtargeting o fActive Labour Market Programs on the rural poor, with emphasis onjob searches andretraining. Strengthened and expanded community development organisations e.g. ODESYPANO. - 84 - References Aksoy, M.A. and J. Beghin, 2005. Global Agricultural Trade and the Developing Countries, Washington, DC: World Bank. Baffes, J. H.de Gorter, 2005. "Experience with DecouplingAgricultural Support," Chapter 5 in M.A.Aksoy and J.C. Beghin, eds., GlobalAgricultural TradeandDeveloping Countries,World Bank. Bazza M., and Ahmad M. (2002) A comparative assessment of links between irrigation water pricing and irrigation performance in the Near East. Conference on irrigation water policies, Agadir, Morocco 15-17 June, 2002. Ben Ammar B. (December 2005) Le secteur agricole en Tunisie Power Point Presentation by Directeur GCnCral Etudes et DCveloppement Agricole, Ministbe de 1'Agriculture et des RessourcesHydrauliques, Tunis, Caser0.P.and Varoudalus A. (2004) Growth,private business and the cost of doing business in Tunisia.A comparative perspective. World Bank discussion paper. CEPI (2000) Cahier du CEPIn.4 -Etude de positionnement de la branche huilerie Chemingui, M, A., C. Thabet (Preliminary Draft June 2005) Agricultural Trade Liberalization and Poverty in Rural Areas in Tunisia: Micro-simulation in a general equilibriumframework. CNEA (2005a) Etude de la Filidre CLrLales. Unpublishedmimeo. CNEA (2005b) Etude de la Filidre Fruits et Lkgumes. Unpublishedmimeo. CNEA (2006) Etude dufinancement du secteur agricole. Unpublishedmimeo, inpreparation. Codron J-M., Bouhsina Z., Fort F., Coude1E. and Puech A. (2004) Supermarkets in Low-income Mediterranean Countries: Impacts on Horticulture Systems. Development Policy Review, Vol. 22, No. 5, pp. 587-602, September 2004 Diaz-Bonilla, E., X. Diao, and S. Robinson, forthcoming (April 2006), "Thinlung Insidethe Boxes: Protectioninthe Development and Food Security Boxes versus Investmentsinthe Green Box", inE.Diaz-Bonilla, S.E. Frandsen, and S. Robinson, eds., WTONegotiations and Agricultural TradeLiberalization: the Efect of Developed Countries' Policies on Developing Countries, Oxford: CAB1Publishing. Echevema, R. G., and H. Elliot. 2002. "Financing Agricultural Research by Competitive Funding." InD. Byerlee and R. G. Echevema (eds.), Agricultural Research Policy in an Era of Privatization. Wallingford: CAB1Publishing. Feder G., Zijp W. and Willett A. (2000) Apcultural extension - generic challenges and some ingredientsfor solutions. Gharbi M.(1998) Private, collective and state tenure inTunisia. Land Reform 1998:l - 85 - Gibson C., Williams A., and Ostrom E. (2004) Local enforcement and better forests. World Development 33:2,273 -284. Gill G.J. and Camey D. (1999) Competitive agncultural technology funds in developing countries. Natural ResourcePerspectives no. 41. London: OverseasDevelopment Institute. Guellouz M. (2004) SCcuritC alimentaire : contraintes et dCfis Auto-suffisance : mythe ou rCalitC. PrCsentt B l'atelier((AssuringFoodandNutrition Security inAfrica by 2020. Kampala. Ideaconsult (2005a) Actualisation de 1'dtude de la compdtitivitk du secteur agricole. Unpublished mimeo. Ideaconsult, (2005b) Etudedes Concessions Tarifaires Accordees par 1' EUa m Exportations Agricoles Tunisiennes, Phase 2 Report. Ideaconsult-IRAM-FERT-AMCIDA(2006) Etudesur les structures professionnelles et interprofessionnelles ainsi que sur les services de soutien ci 1'agriculture. IMF(2004) Tunisia: 2004 Article IVconsultation. StaffReport. IMFCountry Reportno. 041359 I N S (2000) EnquZte nationale sur le budget, la consommation et le niveau de vie des mdnages MARH I GTZ (2005) Changements climatiques: efets sur 1'dconomie tunisienne et stratkgie d'adaptation pour le secteur agricole et les ressources naturelles. Ministhe de L'Agriculture. Direction GCnCrale de la Planification, du DCveloppement et des Investissements Agricoles (1996) EnquZte sur les structures des exploitations agricoles, 1994- 1995 Mtimet A. (2004) Gateway to land and water information: TunisiaNational Report. Publishedon http:Nwww.fa0.org Rochegude A. (2005) Etude sur les questionsfonci&-s en Tunisie.Mimeo. Rouchiche S., and Abid H.(2002) R6le desplantationsforesti2res et des arbres horsfor& duns l'arnbnagernentforestier durable -btude de cas de la Tunisie.;FAO, Tunis mars 2002, p. 14 Seale J., Regmi A. and Bernstein J. (2003) International Evidence on Food Consumption Patterns. U S Department o f Agnculture Technical Bulletinno. 1904. Sellami H.(2005) Investing in Agriculture. China InternationalBusiness 01 December 2005. ShepherdA. (2005) Theimplications of supermarket developmentfor horticulturalfarmers and traditional marketingsystems in Asia. Mimeo. FAO, Rome. Traill, W.B. (2006) The Rapid Rise of Supermarkets ? Development Policy Review 242, 163- 174. Valdes A. and Foster W. (2003) The positive externalities o f Chilean agriculture: the significance of its growth and export orientation. A synthesis o f the roles o f agriculture Chile case study. - 86 - World Bank(1996) From universalfood subsidies to a self-targetedprogramme: a casestudy in Tunisian reform. World Bank(1999) MENA rural finance: performance, constraints andoptions. Reporton Egypt, MoroccoandTunisia. World Bank (2000a) Republic of Tunisia: Social Conditions Update, Volume I. World Bank (2000b) Republic of Tunisia :agricultural competitiveness study. Reportno. 20883- TUN. World Bank(2000~)Tunisia: Private Sector Assessment Update. World Bank (2001) Republic of TunisiaAgricultural Competitiveness: apolicy note on transition issues in theprocess of agricultural liberalization. World Bank(2002) Reaching the Rural Poor: A Rural Development Strategyfor the Middle East and North Africa Region. World Bank(2003) Rkpublique Tunisienne: Mise a Jour de la Pauvretd World Bank(2004a) Republic of TunisiaEmployment Strategy, Volumes Iand II. World Bank (2004b) Gender and Development in the Middle East and North Africa: Women in the Public Sphere World Bank (2004~)Tunisia CountryAssistance Strategy ReportNo. 28791-TUN World Bank(2004d) RepublicofTunisia CountryAssistance Evaluation.OED. World Bank (2004e). Tunisia: ExportDevelopmentProjectII., PAD(June). World Bank (20040.Promoting agro-enterprise and agro-food systems development in developing and transition countries :towards an operational strategyfor the WorldBank Group World Bank(2005a) Managingagriculturalproductionrisk: innovationsindevelopingcountries. World Bank(2005b) Whereis the wealth of nations ? World Bank (2005~)TheStatus and Progress of Women in theMiddleEast and NorthAfrica Region. World Bank(2005d) TheStatus and Progress of Women in theMiddleEast and NorthAfrica Region. World Bank (2005e) Beyond the city: the rural contribution to development. World Trade Organization(2005) TradePolicy Review: Report by the Secretariat, Tunisia - 87 - ANNEX 1.THE ECONOMICS OFLIBERALISATION Highproducerpricesmight appearbeneficialina countrythat is concerned about rural welfare. Economic theory, on the other hand, indicates that there are important costs attached to farm protection. Except invery special circumstances, these costs always outweigh the benefits: a) Farm protection encourages farmers to undertake activities for which the costs (labour, water, purchased inputs andcapital) are actually greater thanthe value to the nation o fthe output, interms o f saved imports. Another way o f lookingat the same process is to say that support for agriculture discourages resources such as labour, water and investment capital from beingallocatedto other sectors, even if they can be usedmoreproductively inindustry and services. b) Differences inthe levelo fprotection encourage the farmer to shiftresources out o f less protected activities into more protected activities, even though the value to the nation interms o f saved imports is less. c) The apparent gain to the economy from higher farm prices is not actually a gain at all. This is because the higherprices are actually being paid for by other Tunisians when they pay for agricultural produce. Tunisian consumers pay higher prices for fruit, vegetables anddairy produce. The Tunisian Government andtaxpayer pay more for the wheat that is then passed on to the consumer at a subsidised price. The Tunisian livestock farmer pays a higher price for fodder. Several other important economic issues are not directly captured by the models: Farmprice support andinput subsidies are ineffective tools for protecting rural social welfare, because the benefits o f support are proportional to a farm's productive capacity. The worst-off receive least benefit. Highprices encourage producers to accelerate the rate at whichtheyuse sustainable natural resources, such as water, rangelands and fish stocks, as well as to overuse inputs such as fertilizer andother chemicals, which often has negative environmental impacts. Distortedprice signals cause investors -- bothpublic andprivate -- to devote too much capital to over-protected importableproducts andtoo little to exportables. Inadynamic context, the negativeeffects ofthis misallocation ofinvestment on productivity growth accumulate over time, resulting inever-increasing costs to national income. This study estimated the impact o f artificially high farm prices upon the Tunisian economy. Four types o f analysis were used to estimate the impacts o f agricultural price - 88 - protection in Tunisia (Ideaconsult, 2005): consumer budget simulations, computable generable equilibrium models, linear programming farm models and domestic resource cost analysis. The methodology was based upon an earlier Bank study (World Bank, 2000), andthe scenarios were tailored to the most recent data andpolicy questions. Consumer budget simulations. Using national survey data (INS, 2000) it was possible to estimate the effect o f agricultural price liberalisation upon consumers' welfare. The model answered the question: what change in consumers' incomes would compensate them for agricultural price liberalisation, firstly if their purchases remained the same and secondly ifthey could adjust their purchases to take account o f changed relative prices? A Computable General Equilibrium (CGE) model simulated the national economy by a series o f equations calibrated to fit observed Tunisian data. The model assumes that the economy has a fixed supply o f labour and investment capital, which will be allocated between agriculture and other sectors on the basis o f their rates o f return. A high rate o f protection for agriculture will therefore encourage the allocation o f labour and capital to farming, even though they could contribute more to the national economy in industry and services. The model's predictions take the form o f a one-off jump in GDP resulting from liberalisation, which sets the economy on a new growth path. The total benefits are presented as the difference between Tunisia's with- and without-liberalisation annual GDPs, summed over 25 years. Such CGE models assume that industry and services will reabsorb any investment capital and labour released by agricultural sector liberalisation. However, the findings o f separate studies (Casero and Varoudakis 2004, World Bank 2004) on financial and labour markets suggest that this assumption should be accepted with caution. Inthe financial market, the lack o f credit ratinginformation and a culture o f lending against collateral mean that the availability o f security, rather than the supply o f capital, may be the binding constraint upon non-farm investment. Unemployment rates (17% inTunis in2001) also suggest that crude labour supply i s not a binding constraint upon non-farm employment. To take account o fmarket imperfections, a "cost o f adjustment" element was built into the simulations. It i s intended to reflect the cost o f retraining labour and upgrading investment, and i s therefore a function o f the size o fthe adjustment. In the case o f full liberalisation, for example, it is estimated to bejust under 1billion DT over 25 years. A variant simulation modelled the liberalisation o f field crop [grandes cultures] (essentially cereals and oilseeds) prices alone. It lets labour and capital be reallocated within agriculture, and thus relies much less on the assumption that industry and services will absorb them. Domestic resource cost analysis. This compares the economic cost o f producing a commodity with its price on the world market. The ratio o f the former to the - 89 - latter i s a measure o f competitiveness. A domestic resource cost ratio greater than 1means that the product is definitely uncompetitive. A ratio lower than 1means that it mi& be competitive ifquality and other non-price issues canbe addressed. Linear programming simulations. The impacts o f farm price liberalisation were simulated for sixty specimen farm types. Each farm model was based upon the technical coefficients and input prices that were collected during the preparation o f the "Carte Agricole". For realism, the technical coefficients are based upon actual observations, not upon ideal farming techniques, and shadow [Cconomique] prices were used for labour and water. Farmbehaviour was then modelled on the assumption that profit-maximising farmers will reallocate land andwater between crops inresponse to relative price signals. Although the analysis did not allow us to simulate how factor markets as a whole will respond to price changes, it did highlightwhich farmtypes, crops andzones will bewinners andlosers. - 90 - ANNEX 2. THE "CARTE AGRICOLE" The Carte Agricole, recently completed at a cost o f 5 million DT, is a Geographical Information Systems (GIS) application that identifies the ago-economic potential o f Tunisia's land resource to a high degree o f detail. It was the basis for the analysis o f Tunisia's agricultural competitiveness insection 1II.A above. The challenge now facing the MARHis o f ensuring that this considerable investment is put to good use, that its data are kept up-to-date and that the greatest possible number o f users has access to it. Indeed, there is a risk that the model's integrity might be compromised without an effective system for updating its database. The MARH has chosen a decentralized approach to database management, where each CRDA is responsible for maintaining its own data. Although this will help keep the CRDAs committed and involved, it increases the risk that data quality will deteriorate. At the same time, the practical value o f the Carte Agricole will depend largely upon its dissemination outside the MARH, for example among inter-professional groupings, banks, universities inTunisia and abroad, investors, the MDCI, the Ministere duPlanand institutions involved inland-use planning. An information management planfor the Carte Agricole would be the first step towards ensuring that it remains a useful resource management tool inthe longrun. -91 - ANNEX 3. TERRES DOMANIALES Around 0.5 million ha o f prime arable land belong to the state estate, and are managed through one o f four systems: operation by parastatal "agro-combinats", operation by the Ministry o f Agriculture itself (around 100,000 ha), leases to private individuals on the basis o f recommendations made by regional and national "commissions d'attribution des terres domaniales" composed of government and union representatives, or usufruct agreements with unitks coopkratives de production agricole or young agricultural graduates. Government policy and current legislation rule out the privatisation o f such land except in exceptional circumstances, such as the re-issuing o f land that has been expropriated duringlandconsolidationexercises. Government's management o f the terres domaniales reflects a number o f conflicting objectives: 0 keeping a body o fprime agricultural landunder the control o fthe state; 0 creating employment opportunities for young farmers and agricultural graduates; 0 greater national self-sufficiency in cereals, for example via the attribution o f 62,000 ha to cereal-production (CNEA 2005a); 0 while not necessarily maximising government revenue from the estate, at least minimisinggovernment subventions to those who farm it. It was beyond the scope o f this study to examine the efficiency with which this substantial land resource is being managed. There is limited information in the public domain on this subject.50 Based upon average cropland value estimates for Tunisia (World Bank 2005), it is estimated that this lands are capable o f producing a surplus for reinvestment or remittance to government o f at least USD 60 millionper year, or 0.3% of GDP. Ifthe Government wishes to mobilise the state estate for economic development, a first step might be to benchmark the profitability o f its farms against their potential. This would give the Government an indication o f the cost o f pursuing the multiple objectives listed above. The standard budgets o f the Carte Agricole (see below) can provide a point o freference. 50 A study in2002 of 55 Socittts de Mise enValeur et de DeveloppementAgricole, groups ofagricultural graduates withusufruct agreements, found that only 29% had "bonnes perspectives". 78% had yields which were equal to or less than the regional averages, despite highquality landholdings and preferential access to credit. - 92 - ANNEX 4. FORESTS AND RANGELANDS 1. Forests and rangelands cover an estimated area o f 5.5 million hectares, or a third o f the country. The forest area is 1.2 million hectares, o f which 1millionhectares are forest in the strict sense o f the term (including 515,000 ha o f plantations) and about 170,000 ha consists o f scrubland. Three quarters of the forests are in the north- and centre-west o f the country. The natural forests are made up o f oak (cork, zeen and green), pine (Aleppo and maritime) and thuya; the artificial forests mostly consist o f eucalyptus, acacia and pine ("pignon", Aleppo and radiata). 2. Forests and rangelands are an important foundation o f the rural economy, especially in the relatively poor areas in the west. Overall, they meet 15-20% o f the fodder needs o f the national herd" and about 15% o f energy needs. The population livinginforestry and grazing areas is estimated to be around 1 million, or 10% o f the total population and 25% o f the rural population. Their main activities are livestock, subsistence agriculture, and forestry work. Rangelands cover nearly 4.3 million hectares, o f which 2.5 million hectares are collective land, 1.1million hectares private land, and 743,000 hectares under esparto grass. As well as their role as a factor o f production, forests and rangelands provide essential environmental functions: they protect against soil erosion and dam siltation. 3. The loth Environment Plan provides an integrated structure and serves as a Tunisia's commitment to preserving its environmental resources is well known. framework for the detailed and ambitious Five Year Plan objectives (see table below). For example, between 1990 and 2005 Tunisia's area under forests increased by 4.3%, at a rate o f 11square kilometres per year, while it fell by 0.1% in North Africa and the Middle East overall (World Bank, Little Green Data Book, 2006). Challenges 4. Unfortunately, population growth and a shift towards sedentary settlement are leading to the degradation o f forests and rangelands. Three million hectares have been degraded by cultivation, overgrazing (including esparto land), water and wind erosion, illegal land-clearance, charcoal production, fires and pests. For forests, cultivation, widespread overgrazing and fires are leading to the loss o f 2,600 ha per year, a loss valued at 0.1% o f GNP (World Bank, Environment at a 51Underamillion cattle, mostly inthe north, and about eight million sheep and goats concentrated inthe centre and south. - 93 - Glace: Tunisia 2004). Rangeland degradation is equally apparent, though no detailed studies are available. It appears as a reduction o f soil coverage and the nutritional value o f plant growth. Around 8 million ha in the south-west are affected by desertification. The area under esparto is falling by around 1.5% per year, and yields have dropped from 4.5 Q/ha to 3.0 Q/ha. 5. However, it i s clear that the economic value o f forest products is well below potential. For example, many studies report that income from non-wood forest products could increase significantly if there were a favourable regulatory framework for the private sector to invest. One study estimates that they have a potential value o f DT 125 millionper year. 6. So the challenge facing Tunisia can be summed up inone question: how can one encourage the rural population to conserve and even invest in fragile forests and rangelands, both for income-generation and environmental reasons ? To pursue this analysis, one hasto distinguishbetweentwo cases: a. Where the state, its agent or a private operator manages use-rights on its own account (e.g. commercial operation o f a state forest, plantations and private rangelands); b. Where there are common property rights (e.g. public and collective rangelands, wood collection rights), and use is governed by the informal decisions o f the user. In this case, the state can obviously only influence users' decisions indirectly, by puttingan appropriate incentives framework inplace. The efforts and successes o f the MARH have until now been concentrated inthe first o f the two cases above. However, one can observe the first steps towards the encouragement o f local communities to manage common property rights efficiently and sustainably. 7. Although the data below are provisional, they illustrate the successes and failures o f the loth Plan. Ingeneral, where a public or private operation is managing the resource on its own account, the Plan objectives have been met. But where the public environmental interest is the main concern, they have not. - 94 - _. Iable. Forest amiraiiwlancIacliic\.rmcnts of`the 10'' Plaii (hectares) 2002f`i 1 Strategy IO'" Ptan (2002/06) Objectives Objectives 2002 - 2005 Forest~ ~ a ~ ~ a ~1~ 0~ ~ ~~~9 , ~0~ ~40~3 ~, ~ ~ ~ In state forest lands 7 0 , 0 ~ ~ .12,000 16,000 Roadsidesand urban areas 20,~00 6,000 1,000 Besidewater bodies 20,000 10,000 4,400 ~ i n d ~ ~ ein kirrigation schemes a s 30,000 10,000 1,700 Private plan~atio~s 50,~00 16,000 26,000 ~ a n ~ ~ ~ a n d ~ 210,000 56,000 38,000 In state forest lands 100,000 25,000 9,000 e o ~ l ~rangel~nds,governed byforest law ~ ~ i ~ e 4~,000 10,000 4,000 On private land 70,000 21,000 24,800 8. The logic of classic deve~op~ne2itp ~ ~ n n jwith~ the G ~ i , o ~ ~ fixing~ ~ e ~ i ~ o b j ~ c t i ~ ~~en~a ,n c ~ no gb j e c t ~ ~ands co~itrolljn~the res~iltingasset, has proved ~ itself well suitcd to p~iblic~ l a ~ t a t ~ onnstate lands. The Plan's a c h ~ ~ ~in~ ~ ~ ~ n t s o s this regard have becn 125% of the target. Moreover, b u d g c t ~allocat~o~s to forests and ran~~~ands an ~ n c r e a s ~apprec~ationof their i n i p o ~ a ~The. show ~ g c ~ ratc of growth was 27% ~etweenthe IOth and 1lth Plans, c o n ~ p a r ~with 5% for d ~ u b l i cin~estnientoverall and 16% for the water sector. In2003, total ~ ~ ~ ~ s t n i ~ in forests and ran~e~ands was 13.9% of all investmen~s~compar~dwith 4.7% in 1987. 9. C o ~ ~ e r c i aoperations on state forest lands appear to meet G o ~ e ~ n ~ e ~ r l expectat~ons~The Rigie d'Exploitation ~ o ~ e s t(REF),e which comes dircctly ~ ~ r , m ~ n the sale ~o f sstanding timber, cork p r o d u c t i ~and ~ ~ ~ REF's income is a nper year, of\h/ DT 6 nijll~o~ifor cork is ~ ~ , 9tannes in 2004) and DT 4 ~ ~ O i for wood. The ~~c~~~~~ ~ ~ ~ ~ ~ de lao ~ ~ ~ o Cellulase et du Papier Alfa (SNCPA) provides incomes o f around DT 2.3 million far pickers, and the expart of esparto paper brings inaround DT `12 miffion ~~~~~ per year. But is the state*s direct ma cnt o f statc forests optimal ? Firstly, the e c ~ ~ o ~ ~ ~ sts has not been b e n c h ~ ~ kagaid ` ~ al, It seems, for e ~ ~ i p(see abm l e that n ~ ~ -forest~products~arcd ~ r ~ seriously ~n~er-perf5 . At first sight, t current levels of marketing are still far short o f their p o t e ~ ~forael ~~ a ~in~wood for the c o n s ~ ~ cindustry and ~ l e t ~ 5 ~ aromatic and medicinal plants. ~ e c o n d ~thc~ ~ o ~ ~ ~ financial y ~ e ~and ' s t personnel capacity i s already ilconstra ,ifone considers that only2 7 0 , ha,~ ~ ~ less than a third of the state forest, are ~ ~by up-to-datee~ ~ ~r i ~ ~gplans ~~ e n t ~ e ~ - 95 - and these plans only cover basic operations (cutting, infrastructure and clearings). The impression is that the State does not have the operational flexibility to make the most o f its forest resources. 11.Although private sector management can be more flexible and responsive to the market, its involvement in the management o f the state forest is limited. In the past, the Forest Code required a process o f use-rights allocation that ruled out any long-term planning and investment. The only way that the private sector could become involved was by obtaining government subsidies for reforestation on private landand through temporary occupation contracts for forestry and livestock operations. However, changes to the code in 2005 have made it possible for private operators and GFICs to obtain renewable concessions for 5 years. 12.Although it is only 9% o f total investment, private investment in forests and rangelands is rising. The Plan's area target was 118% achieved. Private grazing improvements are attractive, mostly thanks to the subsidies from the Office de 1'Elevage et des Paturages (OEP). The same goes for rangeland improvements, where 76,000 ha o fthe 100,000 ha target have beenrealized. Finally, 33,800 ha o f thornless cactus, out o f a target o f 40,000, have been created as a source o f h i t andforage. The management of common use rights 13. However, the MARHhas been less successful in creating an effective incentives framework for ensuring the participation o f rural communities inthe management o f common environmental resources. Seen in its historic context, the degradation o f such resources i s partly due to the weakening o f the collective social groupings that used to be their main users (see section 1II.E on rural land). The state has been unable to revive their role insustainable resource management. 14.For the 4.5 million hectares o f rangelands o f all kinds, the main problems have been relationships between users and ambiguities in the land tenure situation. A complicating factor is that many users are primarily farmers, and only take an interest in natural resource management when a profitable opportunity arises. Conseils de Gestion des Parcours Collectifs have been set up at Governorate level to address the issue o f user organisation. These councils have limited resources, and do not always appear genuinely representative o f all rangeland users. Finally, there are the controversial subsidies for animal feed, which are blamed for over- stocking. 15. To combat the over-exploitation o f non-private forests, Government relies almost entirely upon regulation. The forest code, for which the MARH's forests Directorate General is responsible, recognises open access rights for local populations regarding dead wood and brush, grazing, the gleaning of certain forest products and the cultivation o f specific landholdings. These rights are limited to the needs o f the user's household. The state assumes responsibility for enforcingthese regulations. - 96 - 16. However, it is evident that the state's ability to enforce its regulations is less than its ambition. There are institutional constraints. The CRDAs' forestry sections do not have the required human and financial resources for enforcement, a specific example o f the MARH's general resource shortage (see section 1II.D). This problem is likely to become more serious in future, given the highretirement rate andoperatingbudget constraints. 17. This observation reflects the new general consensus on natural resource management. Both in developed and developing countries, it is recognised that user communities are often better socio-politically adapted to natural resource management than the state. The classic work on this theme, Elinor Ostrom's "Governing the Commons"52, analyses management systems for forests, pastures and irrigation water in Switzerland, the Philippines, Spain and Japan. On this basis, it proposes certain principles as necessary for the effective control o f resource over-exploitation: a. Clearly demarcated physicalboundaries; b. That use should be monitored either by users or people designated by users; c. A system o fprogressive penalties for rule-breaking; d. That users should govern conflict resolution andrule amendment.53 18. The MARH i s moving towards community forest and rangeland management. The Code Forestier was revised in 1988 to enable community participation. A decree o f 1996 allows for the creation o f GICs for forestry activities (GFICs - there are currently 41). Two promising initiatives are underway with the DGF: The IntegratedForestry Management Project (with Japanese co-financing) andthe Project for Agro-Pastoral Development and Promotion o f Local Initiatives in the South (with IFAD co-financing). The DGF is accumulating in this way some highlyuseful experience inparticipatoryresource management. Optionsfor thefuture 19. A first objective would be to make more o f the commercial potential o f the state forest. The MARH could commission a "benchmarking" study to compare actual performance with potential and with the productivity o f other comparable forest areas (e.g. Australia, South Afkica, Andalusia, Greece). Then the MARH could consider how better to mobilise the management advantages o f the private sector. Three types o f action are recommended in this regard: (i) developing simple and operational concession models (ii) selecting zones for concessions for specified purposes e.g. non wood forest products (iii) actively and transparently marketing concessions to private operators and GFICs. 52 E. Ostrom (1990) Governing the Commons. CambridgeUniversityPress. 53 Ostromconcludes that "ifthis study does nothing more than shatter the convictions o fmany policy analysts that the only way to solve common pool resource problems is for external authorities to impose f i l l private property rights or centralized regulation, it will have accomplished one major purpose." - 97 - 20. The long term disappearance o f forests and rangelands would be a catastrophe. It is clear that the regulatory approach has reached its limits and will now need to give way to the gradual introduction o f community resource management. The experiences o f the two pilot projects should be consolidated and scaled up as soon as possible. The MARH should consider how better to incentivised resource conservation in the collective interest. In this regard, Biocarbon Financing could complement fiscal resources. In Nicaragua, Costa Rica and Colombia, for example, a World Bank forestry and livestock project provides financial rewards to farmers for reforestation, which is measured using satellite imaging. In the long-term improvements in the welfare o f forest communities is a precondition for conservation. 21. In short, the MARHcould concentrate on its core functions (policy formulation, defining norms, overseeing development projects) and share or delegate other activities. Such an approach, however, would require the DGF's human resource mix to change, with strengthening needed particularly in non-technical areas (natural resource management, rural sociology, microeconomics, financial management and so on). The forthcoming implementation o f performance-based budgeting inthe MARHshould be seized as an opportunity to redefine the DGF's objectives from first principles, especially where its own institutional capacity is concerned. - 98 -