Report No. 68399-SS Agricultural Potential, Rural Roads, and Farm Competitiveness in South Sudan May 23, 2012 Agriculture and Rural Development Unit Sustainable Development Department Country Department AFCE4 Africa Region Document of the World Bank ACRONYMS AND ABBREVIATIONS AEZ Agro-Ecological Zone ASARECA Association for Strengthening Agricultural Research in Eastern and Central Africa CLC Cropland Connectivity Index ESW Economic Sector Work FAO Food and Agriculture Organization GFRP Global Food Crisis Response Program GIS Geographic Information System GoSS Government of South Sudan HH High production potential and high population density HL High production potential and low population density IFPRI International Food Policy Research Institute LGP Length of Growing Period LH Low production potential and high population density LL Low production potential and low population density MAF Ministry of Agriculture and Forestry MARF Ministry of Animal Resources and Fisheries MDTF-SS Multi-Donor Trust Fund for Southern Sudan MH Medium production potential and high population density ML Medium production potential and low population density NBHS National Baseline Household Survey RAI Rural Accessibility Index SDG Sudanese Pound SSCCSE South Sudan Centre for Census, Statistics, and Evaluation US$ United States Dollar WFP World Food Programme Vice President: Makhtar Diop Country Manager/Director: Laura Kullenberg/Bella Deborah Bird Sector Manager: Karen McConnell Brooks Task Team Leader: Sergiy Zorya Co-Task Team Leader: Abel Lufafa ii All rights reserved: This volume is a product of the staff of the International Bank for Reconstruction and Development/The World Bank. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permission The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly.  For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, www.copyright.com  All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org iii CONTENTS ACKNOWLEDGMENTS .............................................................................................................................VIII EXECUTIVE SUMMARY .............................................................................................................................. IX 1. INTRODUCTION ...................................................................................................................................... 1 2. LAND USE, AGRICULTURAL POTENTIAL, AND POPULATION IN SOUTH SUDAN ............. 4 2.1. LAND USE AND LAND COVER ............................................................................................................... 4 2.2. POTENTIAL FOR AGRICULTURAL PRODUCTION AND POPULATION DENSITY ........................................ 8 3. AGRICULTURAL PRODUCTION ....................................................................................................... 12 3.1. HOUSEHOLD FOOD CONSUMPTION..................................................................................................... 12 3.2. CURRENT AGRICULTURAL PRODUCTION ESTIMATES ......................................................................... 14 4. AGRICULTURAL POTENTIAL .......................................................................................................... 17 4.1. METHODOLOGY ................................................................................................................................. 17 4.2. CROPLAND EXPANSION ...................................................................................................................... 19 4.3. POTENTIAL AGRICULTURAL PRODUCTION VALUES ........................................................................... 22 5. INVESTING IN ROADS ......................................................................................................................... 25 5.1. ROADS IN SOUTH SUDAN ................................................................................................................... 25 5.2. RURAL CONNECTIVITY: METHODOLOGY ........................................................................................... 26 5.3. ROADS FOR AGRICULTURAL DEVELOPMENT IN SOUTH SUDAN ........................................................ 27 5.4. BUDGET REQUIREMENTS ................................................................................................................... 32 5.5. REDUCING TRANSPORT PRICES AND ITS POTENTIAL EFFECT ON FOOD PRICES .................................. 35 6. AGRICULTURAL COMPETITIVENESS ........................................................................................... 37 6.1. PRICE COMPETITIVENESS ................................................................................................................... 37 6.2. FARM PRODUCTION COSTS ................................................................................................................. 42 6.3. COST-REDUCTION STRATEGIES .......................................................................................................... 45 7. CONCLUSIONS....................................................................................................................................... 49 8. REFERENCES ......................................................................................................................................... 51 9. ANNEXES ................................................................................................................................................. 53 iv ANNEXES Annex 1: Type of land use by 18 categories ...................................................................................................... 53 Annex 2: Type of land use by state .................................................................................................................... 54 Annex 3: Type of land use by livelihood zone .................................................................................................. 56 Annex 4: Population density and share of cropland by agricultural potential-population density typologies by state ................................................................................................................................................................... 58 Annex 5: Population density and share of cropland by agricultural potential-population density typologies by livelihood zone ................................................................................................................................................... 59 Annex 6: Share of food consumption by aggregated items for all households .................................................. 60 Annex 7: Type of rural households, with and without cereal consumption....................................................... 61 Annex 8: Livestock population by state: SSCCSE computed estimates, 2008 .................................................. 62 Annex 9: Quantity of crop production by state (tons)........................................................................................ 63 Annex 10: Cropland expansion by livelihood zones and typologies of agricultural potential areas (Scenario 1) . ................................................................................................................................................................... 64 Annex 11: Agricultural potential zones, areas of potential cropland expansion, and roads .............................. 65 Annex 12: Different types of roads across states by agricultural potential (km) ............................................... 67 Annex 13: Different types of roads across livelihood zones by agricultural potential (km).............................. 68 Annex 14: Matrix of distances between states in South Sudan (km) ................................................................. 69 TABLES Table 1: Area and share of aggregated land uses in total national land area ........................................................ 5 Table 2: Share of aggregated land uses by state (%) ............................................................................................ 6 Table 3: Share of cropland and other land uses by livelihood zone (%) ............................................................... 7 Table 4: Cropland, population, and population density by state ......................................................................... 10 Table 5: Population, population density, and cropland according to agricultural potential ................................ 10 Table 6: Share of various food items in household consumption (%) ................................................................ 12 Table 7: Estimated livestock population in South Sudan.................................................................................... 13 Table 8: Estimates of cereal production from the NBHS and WFP/FAO assessments ...................................... 15 Table 9: Value of agricultural production in South Sudan ................................................................................. 16 Table 10: Regional comparison of agricultural performance in 2009 ................................................................ 16 Table 11: Current and projected cropland area under Scenario 1 ....................................................................... 19 Table 12: Cropland and other land uses under moderate and high expansion scenarios .................................... 20 Table 13: Current and potential agricultural value due to cropland expansion .................................................. 22 Table 14: Current and potential agricultural value under increased cropland and yield/ha ................................ 23 v Table 15: Relationship between rural connectivity and realization of crop production potential in Sub-Saharan Africa .................................................................................................................................................................. 24 Table 16: Benchmarking South Sudan’s roads against other African countries ................................................. 25 Table 17: Benchmarking international freight for South Sudan’s road network against regional corridors ...... 25 Table 18: Different types of roads and their lengths (km) by state, South Sudan .............................................. 28 Table 19: Total length (km) of different types of roads by agricultural potential zone ...................................... 29 Table 20: Access to different roads by agricultural potential zone using a 2 km boundary ............................... 29 Table 21: Access to different roads by agricultural potential zone using a 5 km boundary ............................... 30 Table 22: Types and lengths of roads needed to meet rural connectivity targets ............................................... 31 Table 23: Roads distribution by state in high agricultural potential zone (%) .................................................... 31 Table 24: Roads distribution by livelihood zone in high agricultural potential zone (%) .................................. 31 Table 25: Cost of rehabilitation and reconstruction of two-lane inter-urban roads ............................................ 32 Table 26: Cost scenarios for road rehabilitation, construction, and maintenance in South Sudan ..................... 33 Table 27: Budget requirements for road investments under the base scenario (US$ million) ............................ 34 Table 28: Approved budget in 2010 and 2011 in South Sudan (SDG million) .................................................. 34 Table 29: Budget requirements for road investments under the pragmatic scenario .......................................... 35 Table 30: Measures and outcomes for reducing transport prices along the main transport corridors in Central and West Africa .................................................................................................................................................. 36 Table 31: Measures and outcomes for reducing transport prices along the main transport corridors in East Africa .................................................................................................................................................................. 36 Table 32: Actual and landed prices by import source, March 2011 (US$/ton) .................................................. 40 Table 33: Simulated impact of lower transport prices on maize prices in South Sudan (US$/ton) .................... 41 Table 34: Simulated impact of lower transport prices on sorghum prices in South Sudan (US$/ton) ............... 41 Table 35: Key elements of maize production costs and revenues in South Sudan, Uganda, and Tanzania ....... 42 Table 36: Labor costs for typical farm production activities in South Sudan ..................................................... 43 Table 37: Gross margins of sorghum production in South Sudan ...................................................................... 45 Table 38: Production costs per ha and ton of output ........................................................................................... 45 Table 39: Retail input prices in the selected East and Southern African countries, May 2011 (US$/ton) ......... 47 FIGURES Figure 1: Cereals balance in South Sudan ............................................................................................................ 1 Figure 2: Aggregated land use/cover map ............................................................................................................ 5 Figure 3: Livelihood zones in South Sudan .......................................................................................................... 7 Figure 4: Population density in South Sudan ........................................................................................................ 9 vi Figure 5: Spatial patterns of agricultural potential and population density ........................................................ 11 Figure 6: Illustration of cropland expansion at the pixel level............................................................................ 18 Figure 7: Cropland expansion under Scenario 1 ................................................................................................. 21 Figure 8: Cropland expansion under Scenario 2 ................................................................................................. 21 Figure 9: Different road types in South Sudan ................................................................................................... 28 Figure 10: Combination of roads, agricultural potential zones, and cropland areas ........................................... 32 Figure 11: Typical maize flows in South Sudan ................................................................................................. 38 Figure 12: Maize prices in Juba, Nairobi, and Kampala ..................................................................................... 38 Figure 13: Typical sorghum flows in South Sudan............................................................................................. 39 Figure 14: Sorghum prices in South Sudan and Kadugli (Sudan) ...................................................................... 39 Figure 15: Thick vegetation in Yambio .............................................................................................................. 44 Figure 16: Open fields in Malakal ...................................................................................................................... 44 vii ACKNOWLEDGMENTS This Economic Sector Work was prepared by a task team led by Sergiy Zorya (Senior Economist, ARD), Abel Lufafa (Agricultural Officer, AFTAR) and Berhane Manna (Senior Agriculturist, AFTAR) from the World Bank. The background studies on agricultural potential and road investments were carried out by Xinshen Diao, Renato Folledo, Liangzhi You, and Vida Alpuerto from the International Food Policy Research Institute (IFPRI). The analysis of farm production costs in South Sudan was undertaken by Severio Sebit (Consultant, AFTAR). Marie Claude Haxaire (Operations Analyst, ARD) prepared the transport and food prices matrix to analyze trade arbitrage and competition with imports from Sudan and Uganda. The task team is grateful to the Ministry of Agriculture, Forestry, Cooperatives and Rural Development, the Ministry of Animal Resources and Fisheries, and the South Sudan Center for Census, Statistics and Evaluation for the data and information, as well as for comments provided on this Economic Sector Work. Hyoung Wang (Economist, FEUUR) and Jeeva Perumalpillai-Essex (Sector Leader, EASTS) served as peer reviewers. Cecilia Briceño-Garmendia (Senior Infrastructure Economist, AFTSN) provided guidance on the methodology for developing rural infrastructure. William Battalie (Senior Economist, AFTP2) advised on fiscal sustainability of road investments and helped link this analytical work with other sector activities, including the South Sudan Development Plan. Tesfamichael Nahusenay Mitiku (Senior Transport Engineer, AFTTR) provided guidance on road investments, including unit costs and the priority framework. John Jaramogi Oloya (Senior Rural Development Specialist, AFTAR), Christine Cornelius (Consultant, AFTAR), and Mylinda Night Justin (Consultant, AFTAR) provided advice and comments throughout the report. Bella Deborah Bird (Country Director, AFCE4), Laura Kullenberg (Country Manager, South Sudan), Laurence Clarke (Country Director, AFCS2 and former Country Manager, Juba), Karen McConnell Brooks (Sector Manager, AFTAR), and Louise Scura (Sector Leader, AFTAR) supported the study and ensured that resources were available for its implementation. Amy Gautam, Hawanty Page, and Gbangi Kimboko edited the report. viii EXECUTIVE SUMMARY 1. South Sudan has a huge but largely unrealized agricultural potential. Favorable soil, water, and climatic conditions render more than 70 percent of its total land area suitable for crop production. However, less than 4 percent of the total land area is currently cultivated and the country continues to experience recurrent episodes of acute food insecurity. Limited use of productivity-enhancing technologies, capacity constraints, non-tariff barriers, high labor costs and poor infrastructure hinder progress and also constrain production, productivity and the competitiveness of South Sudan’s agriculture relative to its neighbors. This report presents information to guide planners and decision makers not only in addressing both short- and medium-term food security needs but also in positioning South Sudan’s agriculture sector to effectively compete with its neighbors. 2. Most analytical work conducted by the World Bank in the agriculture sector in South Sudan has so far focused on how to provide immediate responses to food security emergencies and price spikes. This includes the Bank’s input to the Government’s Development Plan and several agricultural value chain studies funded under the Multi- Donor Trust Fund for Southern Sudan. This analytical work is different in that it has a longer-term and forward looking perspective. Such an outlook is equally important at this time as it helps ensure that ongoing immediate responses are coherent and in sync with the overriding objective of agricultural policy which is to lower food costs, reduce poverty and increase the sector’s competitiveness at lowest costs. 3. The report assesses agricultural potential in South Sudan and the possibility of increasing agricultural production through increases in cropped area and per capita yield improvements. It highlights the importance and contribution of rural roads to improving agriculture production in South Sudan, identifies road networks that are necessary to accelerate expansion of cultivated land in areas that are considered to have high agricultural potential and provides estimates of the budgetary requirements for road investments in those areas. The report also assesses the implications of infrastructure investments on agricultural competitiveness and the scope for reducing production costs in South Sudan to enable producers to compete with food imports, especially from Uganda. 4. The value (realized agricultural potential) of total agricultural production in South Sudan was estimated at US$808 million in 2009. Seventy-five percent (US$608 million) of this value accrues from the crop sector, while the rest is attributed to the livestock and fisheries sectors. The average value of household production is US$628, of which US$473 is realized from crops. Average value of production per ha is US$299 compared to US$665 in Uganda, US$917 in Ethiopia, and $1,405 in Kenya in 2009. 5. Increasing cropland from the current 4 percent of total land area (2.7 million ha) to 10 percent of total land area (6.3 million ha) under a modest cropland expansion scenario would lead to a 2.4-fold increase in the value of total agricultural output relative to the current level (i.e., to approximately US$2 billion versus the current US$808 million). If coupled with a 50 percent increase in per capita yields, this cropland expansion would lead to a 3.5-fold increase in the value of total agriculture output (i.e., to US$2.8 billion) and would also increase the value of crop production per ha from US$227 to US$340. If per capita yields double, the value of total agriculture production under a modest cropland expansion scenario would increase to US$3.7 billion, and would outstrip the current value of agricultural production in neighboring Uganda. Increasing productivity threefold would increase the value of agricultural production to US$5.5 billion. ix 6. Investments to improve rural connectivity would not only have to first target areas identified as having high agricultural potential, but would also have to adopt a pragmatic approach towards the quality (type) of the roads given severe budget constraints and competing development needs, as well as the low capacity of the local construction industry. A pragmatic approach implies construction of lower quality roads (with lower unit costs) and larger boundaries for assessing roads coverage. This would reduce the capital requirement for rural roads from US$5 billion to US$2 billion and accelerate the achievement of rural connectivity. Full paving investments would be deferred to the future. These investments in roads have to be accompanied by other measures geared towards reducing transport prices, including the promotion of competition among transport service producers and abolishment of various non-tariff barriers to trade, both internal and at cross-border points if they are to translate into reduced food prices, improved food security and competitiveness. If investments in roads reduced current transport prices by half (from US$0.65 per ton-km to US$0.32 per ton-km), maize prices in Juba would fall from the current US$689 to US$628 per ton, or by 9 percent if other factors remain constant. If transport prices decline from US$0.65 to US$0.33 per ton-km, or by 49 percent, the derived sorghum prices in many markets would fall by 30 percent. 7. Improved rural connectivity, especially if combined with good transport policy and regulations, will be transformative, but in and of itself will not be sufficient to sustain the competitiveness of South Sudanese farmers. Neighboring countries still have lower production costs and will benefit from better roads by providing more affordable prices to South Sudanese consumers, especially in urban areas. Complementary productivity-enhancing investments and market-supportive regulations are therefore required to improve the competitiveness of South Sudan’s agriculture. In the short term, removing bottlenecks to using the available seed varieties in the East Africa region would increase access to improved germplasm, and would help narrow the current yield gap. Investments in mechanization to reduce drudgery and high costs associated with cropping would also allow South Sudanese farmers to increase production at relatively lower costs. Support for adaptive agricultural research would allow release of new and superior seed varieties and would also help overcome other constraints (e.g., pests and diseases) to yield increases. Advisory services will be essential to maximize farm returns from the use of improved inputs, including mechanization and the development of irrigation. For all of these public investments, it is important to ensure that they “crowd in� private investment rather than discouraging it. x INTRODUCTION 1. South Sudan has a huge but largely unrealized agricultural potential. The country is richly endowed with a good climate and fertile soils rendering more than 70 percent of its total land area suitable for crop production. In fact, a few decades ago – in the 1980s-, South Sudan was a net exporter of food commodities. However, the prolonged conflict in the intervening years mediated a breakdown in agricultural support services, institutions, infrastructure and work ethic leading to the near collapse of the country’s agricultural production systems. The country thus gained its independence amidst ongoing challenges in agriculture production and with a significant track record of negative food balances (Figure 1) which are typically addressed through food aid. Figure 1: Cereals balance in South Sudan 100 0 Cereal surplus MT(000) 100 200 300 400 500 Consumption year 600 05/06 06/07 07/08 08/09 09/10 10/11 11/12 Source: Data from WFP and FAO. 2. The agriculture sector will be key to the post-conflict recovery and development of South Sudan. A broad review of research (Brinkman and Hendrix, 2010) points to a nexus between food insecurity and conflict and concludes that food insecurity heightens the risk of civil and communal conflict. Therefore, South Sudan must immediately address its food security challenges if the country is to secure sustained peace and recovery and ensure legitimacy of the state. This would prevent the country from relapsing into conflict, as has happened in some post- conflict countries where the state was unable to provide food security for its citizens (Collier, 2007). Beyond food security, however, agriculture will be critical to the long term growth and development of South Sudan. Over 80 percent of the population in South Sudan depends on the agriculture sector as a source of livelihood, and there is a strong consensus in the Government of South Sudan (GoSS) that agriculture should be a vehicle for broad-based non-oil growth and economic diversification. The sector consequently, features prominently in South Sudan’s 2011- 2013 Development Plan. 3. Despite its high potential and the important role that agriculture will have to play in the stability and eventual development of South Sudan, it‘s performance is largely suboptimal. Production is primarily rain-fed, subsistence in nature, characterized by primordial 1 technology, high input costs, and low productivity. Where opportunities for surplus production exist, local producers have little or no incentive to produce for the markets because the poor status of roads limits connection to the centers of consumption. Retail markets in urban areas are hence mainly served by imports at very high prices, and with little secondary economic benefit to the rural areas that should otherwise be their natural supply. 4. In the short-term, lowering food prices and ensuring food security will hinge on progress in: (i) increasing agricultural productivity; (ii) creating and improving systems of agricultural services provision; and (iii) strengthening relevant institutions, policies and regulations. Through funding from the Multi-Donor Trust Fund for Southern Sudan (MDTF-SS) and the Global Food Crisis Response Program (GFRP), Trust Fund the World Bank is supporting GoSS in increasing the productivity and output of agricultural producers, strengthening agricultural institutions at both the central and state levels, and building human resource capacity in the agriculture sector. The Bank has also articulated policy options that the GoSS could adopt to lower the cost of food and promote farming with an eye towards future exports. 5. In the long-term however, beyond productivity gains, key to recapturing and realizing the full contribution of the agriculture sector to overall economic growth and diversification in South Sudan will be progress in resolving infrastructure (roads) bottlenecks to enable access to markets and distribution systems and implementing market-based measures to promote the country’s competitiveness relative to its neighbors. 6. The work described in this report is a first step to addressing the longer-term issues related to the competitiveness of South Sudan’s farmers in a regional context. It focuses on the options for increasing the amount and value of agricultural production in the crop sector, the potential contribution of rural roads to increasing crop production and how to sequence and prioritize rural road investments in a way that maximizes their contribution to realization of the country’s full agricultural potential, especially in light of the competing needs for resources, the very high construction and maintenance costs of rural roads, and the low capacity of the local construction industry.1 The report also explores possible ways of increasing the cost competitiveness of agriculture in South Sudan vis-à-vis its neighbors (Uganda and Sudan). 7. The core sections of the report include:  A presentation of basic information on land use and production potential in South Sudan.  An estimate and analysis of agricultural production in South Sudan.  An assessment of the potential for expanding cropland to increase agricultural production.  Assessment of the contribution and role of improved rural roads and enhanced access to markets in creating incentives for future expansion of cultivated land in areas with high agricultural potential. 1 The focus is on crop production because data required for similar analyses for livestock, fisheries, and forestry resources, including gum acacia, are not yet available in South Sudan The recent value chain studies on livestock and gum acacia financed by the Multi Donor Trust Fund provide useful information for policy makers on these subsectors; to avoid duplication, this information is not repeated here. 2  An estimation of budget requirements for road investments in areas with high agricultural potential.  An analysis of the implications of better road infrastructure for agricultural competitiveness, including an assessment of farm price and cost competitiveness vis- à-vis Uganda and Sudan, to highlight areas where costs can be reduced to enable South Sudan to compete with food imports, even if local marketing and logistics costs decline in the future. 3 LAND USE, AGRICULTURAL POTENTIAL, AND POPULATION IN SOUTH SUDAN 8. This section describes the current land use and land cover in South Sudan. It focuses on agricultural uses and outlines the extent and coverage of various land use/cover types in the different states and livelihood zones.2 Using the Length of Growing Period (LGP)3 as a proxy, the section also describes the potential for agricultural production in South Sudan as well as the relationship between agricultural production potential and population. 1.1. Land use and land cover 9. South Sudan is endowed with abundant virgin land under climatic conditions that are considered suitable for agriculture. According to (Diao et al., 2009), more than 70 percent of South Sudan has a LGP longer than 180 days and is therefore suitable for crop production. However, land use and land cover data (FAO, 2009) show that most of the land that is suitable for agriculture is still under natural vegetation. Only 3.8 percent (2.5 million ha) of the total land area (64.7 million ha) is currently cultivated, while the largest part of the country (62.6 percent) is under trees and shrubs (Table 1).4 This ratio (cropland to total land) is very low in South Sudan compared to Kenya and Uganda, where despite less favorable LGPs, cropland accounts for 28.3 percent and 7.8 percent, respectively, of total land area. 10. Most of the cropland in South Sudan is rainfed. A two-step sequential process was used to derive land use/cover data from a 295 land use types depicted in the FAO (2009) land cover map for South Sudan. First, the 295 land use types were resampled and aggregated into eighteen land use types (see Annex 1), thirteen of them agriculture-related (including trees and tree crops). In the second step, the thirteen agriculture-related land use types were further aggregated into six categories (Table 1): cropland, grass with crops, trees with crops, grassland, tree land, and flood land (Diao et al., 2011). Irrigated area is limited to only 32,100 ha, mainly in Upper Nile. Flood land used for rice production is also limited, at about 6,000 ha, and is located primarily in Northern Bahr el Ghazal (Figure 2). 2 The country is divided into seven livelihood zones that are defined based on climate conditions and farming systems (SSCCSE, 2006): Eastern Flood Plains, Greenbelt, Hills and Mountains, Ironstone Plateau, Nile-Sobat Rivers, Pastoral, and Western Flood Plains. Ironstone Plateau is the largest zone, accounting for 23.5 percent of total land area. The second largest zone is Eastern Flood Plains, which accounts for 20.4 percent of national land. The Western Flood Plains and Greenbelt account for 14.2 and 12.7 percent of total national land, respectively. 3 Length of Growing Period is the concept used in the Global Agro-Ecological Zone (AEZ) project led by International Institute for Applied Systems Analysis and the UN Food and Agriculture Organization. See Fisher et al. (2002) for details. 4 In this analysis, the total land area of South Sudan is estimated at 64.7 million ha, using the data from the most recent Land Cover Database (2009). 4 Table 1: Area and share of aggregated land uses in total national land area Land use Area (ha) Share of total land (%) Cropland 2,477,700 3.8 Grass with crops 325,100 0.5 Trees with crops 1,707,300 2.6 Grassland 9,633,800 14.9 Tree land 40,526,900 62.6 Flood land 9,497,600 14.7 Water and rock 482,700 0.7 Urban 37,000 0.1 Total 64,688,300 100 Source: Aggregated from Land Cover Database, FAO (2009). Figure 2: Aggregated land use/cover map Source: Modified from Land Cover Database, FAO (2009). 11. Most cropland is concentrated in five states: Upper Nile (19.0 percent of total crop land), Warrap (15.3 percent), Jonglei (14.3 percent), Western Equatoria (11.4 percent), and Central Equatoria (11.2 percent). As shown in Table 2, these five states account for 70 percent of 5 national cropland and 56 percent of national territory. Almost all irrigated crops (mainly rice) are in Upper Nile; rice on flood land is all in Northern Bahr el Ghazal (Annex 2). Fruit trees and tree plantations are exclusively in Western, Central, and Eastern Equatoria, most probably due to the suitable climatic conditions in these states. Table 2: Share of aggregated land uses by state (%) State Cropland Grass Trees Grassland Tree Flood Water Urban Total with with land land and crops crops rock Upper Nile 19.0 26.0 7.1 27.1 7.8 9.0 9.5 25.8 11.4 Jonglei 14.3 25.2 7.3 14.8 19.7 26.7 17.3 8.8 19.5 Unity 4.5 16.1 2.5 7.7 3.7 14.9 6.4 17.1 6.0 Warrap 15.3 8.1 14.9 5.2 3.5 11.4 1.8 0.9 5.6 Northern Bahr el Ghazal 9.8 1.1 4.2 1.0 4.7 7.3 15.3 3.2 4.7 Western Bahr el Ghazal 2.0 4.0 12.9 4.2 18.6 13.5 18.5 10.4 14.9 Lakes 9.9 0.6 2.7 5.6 7.1 9.0 4.3 5.1 7.0 Western Equatoria 11.4 7.5 19.9 9.0 15.7 1.4 17.5 3.7 12.5 Central Equatoria 11.2 8.6 21.4 4.5 7.7 2.4 3.7 22.1 6.9 Eastern Equatoria 2.6 2.7 7.1 21.0 11.6 4.4 5.6 2.8 11.4 National average 3.8 0.5 2.6 14.9 62.6 14.7 0.7 0.1 100.0 Source: Authors’ estimates based on FAO (2009). 12. The Western Flood Plains livelihood zone has the most cropland (34.2 percent of national cropland) (Figure 3). This zone has the highest ratio of cropland to total land, as cropland and grass with crops/trees with crops account for 8.5 and 5.4 percent of zonal territorial area, respectively (Table 3 and Annex 3). 6 Figure 3: Livelihood zones in South Sudan Source: SSCCSE (2006). Table 3: Share of cropland and other land uses by livelihood zone (%) Livelihood zone Crop Grass with Trees Grass Tree land Flood Water Urban Total land crops with land land and crops rocks Eastern Flood Plains 26.2 49.2 8.1 35.2 18.3 14.4 8.9 32.4 20.4 Greenbelt 17.6 13.9 28.0 8.3 15.4 1.2 18.4 4.0 12.7 Hills and Mountains 4.2 4.1 10.3 8.6 11.0 3.5 3.4 22.5 9.2 Ironstone Plateau 7.0 5.6 18.0 10.5 29.5 16.8 19.4 13.7 23.5 Nile-Sobat Rivers 10.0 10.9 4.8 5.3 5.4 30.7 26.3 8.8 9.4 Pastoral 0.8 4.5 4.2 20.2 10.3 6.5 5.1 0.9 10.6 Western Flood Plains 34.2 11.8 26.5 12.0 10.1 26.8 18.5 17.6 14.2 Source: Authors’ estimates based on the Land Cover Database FAO (2009). 7 1.2. Potential for agricultural production and population density 13. To a large extent, the suitability of an area for agriculture is a key determinant of the performance of production systems. A frequently used proxy for an area’s suitability for farming is the LGP, defined as the number of days when both moisture and temperature conditions permit crop growth. Depending on its LGP, an area may allow for no crops or for only one crop per year (e.g., in arid or dry semi-arid tropics where LGP is less than 120 days a year), or it may allow for multiple crops to be grown sequentially within one year. Classifying the aggregated land use by LGP shows that 27.3 percent of cropland in South Sudan is located in areas where agricultural potential is high (LGP more than 220 days) and another 41.5 percent in areas with medium agricultural potential (LGP between 180 and 220 days). 14. An association exists between population density and the potential for agricultural production in a given area. According to the 2008 population census, there are 8.2 million people in South Sudan. The actual distribution of this population is difficult to map since a large number of returnees continue to come back each year, and their settlement location is hard to continuously update. Figure 4 shows population density based on the 2008 population census data and the latest LandScan population distribution data for South Sudan. The majority of South Sudanese live in rural space, which is classified as “low density� (population less than 10 per km2) and “medium to high density� (population more than 10 per km2) areas. 15. The population density in South Sudan is very low compared to elsewhere in the region. Average population density is estimated at 13 people per km2 compared to 166 in Uganda, 70 in Kenya, 83 in Ethiopia, and 36 people per km2 for Sub-Saharan Africa in 2009.5 Two states have a population density of less than 10 people per km2: Western Bahr el Ghazal (3 per km2) and Western Equatoria (8 per km2), while five states have a density that lies between 10 per km2 and 20 per km2 (Table 4). Of these, Upper Nile has the largest cropland area nationally but a population density of 13 per km2. Three states, Warrap, Northern Bahr el Ghazal, and Central Equatoria, have a population density over 20 per km2. These three states also have relatively high cropland shares in total land; i.e., 8.8, 8.3, and 6.4 percent, respectively. 5 World Development Indicators, the World Bank. 8 Figure 4: Population density in South Sudan Source: Compiled from a combination of GRUMP and LandScan (2009). 16. There is a high spatial correlation between the potential for agricultural production and population density in an area. Areas with “high� and “medium� production potential based on LGP have the highest population density. According to Boserup (1965; 1981), 50 people per km2 is a threshold population that indicates the possibility of promoting agricultural intensification.6 In South Sudan, population density in the high agricultural potential areas is about 66 per km2, and 54 per km2 in the medium agricultural potential areas (Table 5). Overall, there is high to medium population density in areas of high and medium agricultural potential. These areas, however, have low per capita cropland values. 6 Rural population density varies positively with land productivity but only up to the point where overcrowding leads to land degradation. 9 Table 4: Cropland, population, and population density by state Cropland Grass/trees Total land Share of cropland in Population Population with crops total land (%) density State (ha ) Grass/trees (person/km2 Cropland (person) with crops total land) Upper Nile 470,100 206,100 7,658,500 6.1 2.7 964,353 13 Jonglei 354,800 205,800 12,106,300 2.9 1.7 1,358,602 11 Unity 110,900 95,500 3,729,600 3.0 2.6 585,801 16 Warrap 379800 280,100 4,329,100 8.8 6.5 972,928 22 Northern Bahr el Ghazal 243,600 74,700 2,946,500 8.3 2.5 720,898 24 Western Bahr el Ghazal 50,000 234,200 10,208,800 0.5 2.3 333,431 3 Lakes 245,600 47,200 4,375,400 5.6 1.1 695,730 16 Western Equatoria 281,400 364,300 7,780,100 3.6 4.7 619,029 8 Central Equatoria 276,300 393,900 4,315,200 6.4 9.1 1,103,592 26 Eastern Equatoria 65,100 130,700 7,238,800 0.9 1.8 906,126 13 National total 2,477,700 2,032,500 64,688,300 3.7 3.1 8,260,490 13 Source: Authors’ estimates based on LandScan (2009) and SSCCSE (2010). Table 5: Population, population density, and cropland according to agricultural potential Agricultural potential defined by LGP High Medium Low LGP>220 180-220 <180 Total days days days Population 25.4 33.8 15.8 75.1 Population density 66 54 51 57 High-medium Land 4.8 7.8 3.9 16.6 Population density Cropland area 15.3 26.7 17.9 59.9 Cropland ha per capita 0.18 0.23 0.33 0.24 Population 8.7 11.9 4.4 24.9 Population density 3 4 3 4 Low Land 31.5 35.2 16.7 83.4 Cropland area 12.0 14.9 13.2 40.1 Cropland ha per capita 0.41 0.37 0.89 0.48 Population 34.1 45.7 20.2 100.0 Population density 12 13 12 13 Total Land 36.4 43.0 20.6 100.0 Cropland area 27.3 41.5 31.1 100.0 Cropland ha per capita 0.24 0.27 0.46 0.30 Source: Authors’ estimates based on LandScan (2009) and SSCCSE (2010). 10 17. Figure 5 shows the spatial patterns of agricultural potential and population density according to the six possible permutations of population density (High-medium and Low) and agricultural potential (High, Medium, and Low). This spatial presentation expands information presented in Table 5. High agricultural potential/high-medium population density areas (HH), high agricultural potential/low population density areas (HL), and medium agricultural potential/high-medium population density areas (MH) are the ones best positioned to generate quick wins and development benefits from public and private investments, and thus should be prioritized for agricultural development programs in the country. Annex 4 and Annex 5 provide details of population density and cropland by agricultural potential by state and livelihood zones, respectively. Figure 5: Spatial patterns of agricultural potential and population density Source: Authors’ presentation. Note: HH: LGP >220 days per year and population density >=10 per km2; HL: LGP >220 days per year and population density <10 per km2; MH: LGP between 180 and 220 days per year and population density >=10 per km2; ML: LGP between 180 and 220 days per year and population density < 10 per km2; LH: LGP < 180 days per year and population density >=10 per km2; LL: LGP < 180 days per year and population density <10 per km2. 11 AGRICULTURAL PRODUCTION 18. There are no official agricultural production statistics in South Sudan. But there are data on household consumption that can be used to derive production estimates, given the predominance of subsistence agriculture in the country. In this study, household consumption data from the 2009 National Baseline Household Survey (NBHS) were used to derive food production estimates. This section begins with a presentation of household food consumption and then estimates current agricultural production based on household consumption. 1.3. Household food consumption 19. Cereals, primarily sorghum and maize, are the dominant staple crops in South Sudan. According to the NBHS, more than 75 percent of rural households in the country consume cereals (Annex 6). At the state level, the percentage of rural households that consume cereals varies from 62 percent in Western Bahr el Ghazal to as high as 95 percent in Northern Bahr el Ghazal. There are four states in which more than 80 percent of rural households consume cereals, and five states in which 60 to 65 percent of rural households consume cereals. 20. For the country as a whole, cereal consumption accounts for 48 percent of total primary food consumption in value terms (Table 6). The share of cereals in total primary food consumption increases to 52 percent when only rural households are considered. 7 When non- cereal consuming rural households are excluded, this share further increases to 57 percent, indicating that cereals are the most important staples in rural households’ food consumption bundle (Annex 7). At the state level, the share of cereals in total rural households’ primary food consumption ranges from 63 to 81 percent in four states (Unity, Warrap, Northern Bahr el Ghazal, and Lakes), and is more than 55 percent in Jonglei. Table 6: Share of various food items in household consumption (%) State Cereals Roots Pulses & oil seeds Other crops Livestock Fish Upper Nile 26.7 2.0 6.1 31.4 30.8 3.0 Jonglei 55.1 0.2 1.5 3.5 38.8 0.9 Unity 76.7 0.8 1.4 11.7 8.3 1.1 Warrap 74.7 0.0 6.4 3.8 11.6 3.5 Northern Bahr el Ghazal 60.3 0.2 2.6 5.5 23.2 8.2 Western Bahr el Ghazal 24.0 1.2 5.3 17.5 40.3 11.7 Lakes 68.5 1.2 2.6 4.9 12.9 9.9 Western Equatoria 34.6 5.5 6.8 16.9 27.8 8.4 Central Equatoria 35.8 4.6 3.8 21.5 31.8 2.5 Eastern Equatoria 43.2 0.9 2.1 7.9 44.0 1.9 National total 48.0 1.8 3.8 12.7 29.7 4.0 Source: Estimated from NBHS (2009). 7 Food is defined as all crops including processed crop products (such as cereal flour and root flour), livestock (i.e., meat, milk, eggs), and fish products. 12 21. While cereals are the most important food crops for the country as a whole, almost a quarter of rural households do not consume cereals at all, depending instead on other staples (Annex 7, column 6). Thirty-five to thirty-seven percent of households in five states (Central Equatoria, Western Equatoria, Lakes, Western Bahr el Ghazal, and Upper Nile) and only 5 percent of households in Northern Bahr el Ghazal and 8.5 percent in Eastern Equatoria fall under this category. 22. Livestock is another important food source in South Sudan. Although estimates differ by source, South Sudan is known to have one of largest livestock herds in Africa. According to FAO’s 2009 estimates, South Sudan has a cattle population of 11.7 million, 12.4 million goats, and 12.1 million sheep (Table 7). Using these estimates, South Sudan ranks 6th in Africa in terms of livestock population size, but these numbers are considered conservative in the country. Livestock population estimates generated from the 2008 Sudan Census show a cattle population of 35.5 million, 20.8 million goats, and 27.3 million sheep (Annex 8). Table 7: Estimated livestock population in South Sudan Population (head) Share in national total (%) State Cattle Goats Sheep Total Cattle Goats Sheep Total Upper Nile 983,027 439,741 640,209 2,062,977 8.4 3.5 5.3 5.7 Jonglei 1,464,671 1,207,214 1,400,758 4,072,643 12.5 9.7 11.6 11.2 Unity 1,180,422 1,754,816 1,487,402 4,422,640 10.1 14.1 12.3 12.2 Warrap 1,527,837 1,369,005 1,290,045 4,186,887 13.0 11.0 10.7 11.6 Northern Bahr el Ghazal 1,579,160 1,630,361 1,285,231 4,494,752 13.5 13.1 10.7 12.4 Western Bahr el Ghazal 1,247,536 1,120,095 1,265,977 3,633,608 10.6 9.0 10.5 10.0 Lakes 1,310,703 1,464,421 1,232,282 4,007,406 11.2 11.8 10.2 11.1 Western Equatoria 675,091 1,153,283 1,169,705 2,998,079 5.8 9.3 9.7 8.3 Central Equatoria 878,434 1,153,283 1,265,977 3,297,694 7.5 9.3 10.5 9.1 Eastern Equatoria 888,278 1,132,541 1,025,297 3,046,116 7.6 9.1 8.5 8.4 National total 11,735,159 12,424,760 12,062,883 36,222,802 Source: FAO Livestock Population Estimates Oct 2009. 23. Nationally, livestock account for 30 percent of total primary food consumption in value terms, a share which is similar across rural and urban households (Table 6). In three states, livestock products account for close to or more than 40 percent of rural households’ primary food consumption (39 percent in Jonglei, 40.3 percent in Western Bahr el Ghazal, and 44 percent in Eastern Equatoria) as shown in Table 7. When measured by quantity of red meat consumption, only Jonglei and Eastern Equatoria have an average meat consumption (i.e., 32 kg and 47 kg per capita, respectively) that is significantly higher than the national average (17 kg per capita).8 24. Fish accounts for 4 percent of food consumption at the national level. It is, however, relatively more important in four states: Northern Bahr el Ghazal, Western Bahr el Ghazal, Lakes, and Western Equatoria, where the share of fish in total household consumption is 8.2 8 Total consumption of red meat is estimated at 145,000 tons, assuming 17 kg per capita consumption as reported in the NBHS (2009). This is four times more than that reported in Musinga et al. (2010), who estimated total annual meat production to be 41,124 tons. 13 percent, 11.7 percent, 9.9 percent, and 8.4 percent, respectively (Table 6). When households that consume cereals and/or roots and tubers are excluded, the share of fish products in total food consumption increases to 12 percent for rural households (NBHS, 2009). 1.4. Current agricultural production estimates 25. There are geographical differences in food consumption among rural households in South Sudan. This heterogeneity manifests itself in spatial patterns, considered here to be indicators of heterogeneity in production. Therefore, NBHS food consumption data are used to estimate the current spatially disaggregated agricultural production9. It is assumed that, with the exception of cereals, all agricultural products consumed in South Sudan are produced domestically. For these products, total consumption as outlined in the previous section is assumed to equal domestic production. Since South Sudan imports significant amounts of maize from Uganda and sorghum from Sudan, cereal production is estimated separately. 26. A multi-step process was used to estimate cereal production. First, cereal flour consumption was converted into grain, assuming that it takes 1.25 kg of grain to produce 1 kg of flour. Second, post-harvest losses (the difference between gross and net production in Table 8) are estimated at 20 percent, following the assumption used by FAO/WFP. Third, it is assumed that only 55 percent of grain purchased by rural households is produced locally, while the rest is attributed to imports. For urban households, all market purchases are assumed to come from imports. Local grain production is then defined as the consumption met by households’ own production, household stocks, and 55 percent of total rural households’ purchases. The computations at the state and national levels are reported in Table 8. 27. These estimates of cereal production are higher than those reported in the FAO/WFP annual assessments, with the exception of 2008/09. The divergence mainly arises from differences in per capita consumption assumptions. In Eastern Equatoria, for example, per capita grain consumption is estimated at 247 kg in the NBHS (2009) and 124 kg by FAO/WFP (2011). At the national level, per capita grain consumption is estimated at 108 kg by FAO/WFP, versus 157 kg in the NBHS. As shown in Table 8, the ratio of net cereal production to consumption is 0.64 at the national level, while it is 1.05 and 0.70 in the FAO/WFP assessments for 2008/09 and 2010/11, respectively. State level cereal production is also different in these two data sets. For example, Western Equatoria is ranked the largest cereal producing state in the FAO/WFP assessment, while according to the NBHS, Eastern Equatoria, Jonglei, and Warrap all produce more cereal than is estimated for Western Equatoria in the FAO/WFP assessment. 28. From these production estimates (for both cereals and other agricultural products that are considered to be domestically produced), the value of current agricultural production is calculated at the state level. The calculation considers both quantity of consumption and production for individual crops (Annex 9) and their corresponding prices. The prices used in the calculation are averaged from individual households’ reports in the NBHS. When the price for a specific product in a state is extremely low or high compared to the other states, the national average price is used. If the price for a particular product is not available in the survey or is extremely low compared to that in neighboring countries, the lowest relevant price from neighboring countries is used. 9 The accuracy of our estimates in turn depends on the accuracy of the NBHS data. 14 Table 8: Estimates of cereal production from the NBHS and WFP/FAO assessments NBHS (2009) Gross production Net production Consumption Ratio of net production to consumption National 1,019,341 849,451 1,320,468 0.64 Upper Nile 64,419 53,682 93,745 0.57 Jonglei 190,810 159,008 258,476 0.62 Unity 41,715 34,763 51,151 0.68 Warrap 140,688 117,240 180,927 0.65 Northern Bahr el Ghazal 101,361 84,467 136,776 0.62 Western Bahr el Ghazal 16,331 13,609 24,987 0.54 Lakes 112,972 94,144 152,881 0.62 Western Equatoria 68,462 57,052 79,087 0.72 Central Equatoria 72,441 60,367 130,303 0.46 Eastern Equatoria 210,142 175,118 212,313 0.83 FAO/WFP (2008/09) Gross production Net production Consumption Ratio of net production to consumption National 1,252,230 1,001,785 953,204 1.05 Upper Nile 49,278 39,421 64,788 0.61 Jonglei 101,594 81277 103,623 0.78 Unity 46,253 37,001 59,815 0.62 Warrap 247,415 219,534 189,505 1.16 Northern Bahr el Ghazal 83,605 66,884 118,436 0.56 Western Bahr el Ghazal 68,409 54,728 54,337 1.01 Lakes 136,215 108,972 91,823 1.19 Western Equatoria 273,218 218,574 96,822 2.26 Central Equatoria 132,363 105,890 82,399 1.29 Eastern Equatoria 86,880 69,504 91,656 0.76 FAO/WFP (2010/11) Gross production Net production Consumption Ratio of net production to consumption National 873,823 695,226 986,230 0.70 Upper Nile 61,234 48,985 86,429 0.57 Jonglei 104,844 83,874 158,133 0.53 Unity 29,647 23,714 57,710 0.41 Warrap 117,496 93,998 104,216 0.90 Northern Bahr el Ghazal 80,256 60,378 87,378 0.69 Western Bahr el Ghazal 42,205 33,765 41,465 0.81 Lakes 82,843 66,274 84,181 0.79 Western Equatoria 140,103 112,080 87,903 1.28 Central Equatoria 115,968 92,775 156,655 0.59 Eastern Equatoria 99,227 79,380 122,160 0.65 Sources: Authors’ estimates based on NBHS and compared with FAO/WFP (various years). Note: NBHS production is calculated by consumption met by own products, stocks, and 55 percent of food purchases in rural areas. 29. The value of total agricultural production in South Sudan is estimated to have been US$807.7 million in 2009. Crop production only is estimated at US$607.6 million (Table 9). This agricultural value represents the presently realized agricultural potential in South Sudan. For the country as a whole, the average household’s agricultural production value is US$628, of which US$473 is from crops. Western Equatoria has both the highest total and crop agricultural values, accounting for 18.4 and 22.2 percent, respectively, of national values. Measured by household agricultural value, Western Equatoria is also the richest state. Central Equatoria has the second largest total agricultural and crop value, accounting for 17.5 and 18.9 percent of national totals, respectively, and also ranks second in terms of agricultural value per household. 15 30. Western Bahr el Ghazal (2.5 percent of national total) and Unity (3.3 percent) have the lowest values of agricultural production and are among the states with the lowest agricultural values per household. Table 9: Value of agricultural production in South Sudan Total value (‘000 US$) Percentage Per household (US$) State Inc. livestock Crop only Inc. livestock Crop only Inc. livestock Crop only & fish & fish & fish National 807,694 607,617 100 100 628 473 Upper Nile 87,373 49,860 10.8 8.2 627 358 Jonglei 112,535 72,446 13.9 11.9 598 385 Unity 26,512 18,092 3.3 3.0 385 263 Warrap 67,188 56,660 8.3 9.3 401 338 Northern Bahr el Ghazal 48,450 36,475 6.0 6.0 370 279 Western Bahr el Ghazal 20,376 12,657 2.5 2.1 354 220 Lakes 63,448 51,800 7.9 8.5 703 574 Western Equatoria 148,473 135,024 18.4 22.2 1,284 1,168 Central Equatoria 140,999 114,857 17.5 18.9 801 653 Eastern Equatoria 92,340 59,744 11.4 9.8 611 395 Source: Estimated based on NBHS (2009). 31. The agricultural output value in South Sudan in 2009 is low compared to that in neighboring countries. The value of agricultural output per ha in South Sudan was less than half of the agricultural value added in Tanzania and Uganda, a third of that in Ethiopia, and less than one quarter of that in Kenya (Table 10). The gap in agricultural value added per capita is smaller because of the smaller population in South Sudan. It is worth noting, however, that the comparison is between South Sudan’s agricultural output and agricultural value added in other countries,10 meaning that the actual difference is even larger than that presented in Table 10. Table 10: Regional comparison of agricultural performance in 2009 Country Agricultural value added Agricultural value added Agricultural value added (current US$ million) per ha per capita (current US$) (current US$) Ethiopia 13,632 971 165 Kenya 7,304 1,405 184 Tanzania 5,563 618 127 Uganda 3,658 665 112 South Sudan 808 299 99 Source: World Development Indicators for East African countries and NBHS 2009 for South Sudan. 10 There are no data on variable costs in South Sudan to calculate agricultural value added. On the other hand, there are no recent data on agricultural output in 2009 denominated in US$ for countries in question to compare the values of agricultural output. 16 AGRICULTURAL POTENTIAL 32. As outlined in the previous section, the current agricultural production and its attendant value in South Sudan are low. Given the abundant land and favorable climatic and soil conditions, there is considerable scope to increase production. At a fundamental level, agriculture production in South Sudan can be increased through two approaches that can be mutually reinforcing: increasing the area of cropped land and increasing the amount of production per unit area. This section estimates the potential agricultural value that would accrue from expanding cropland area and increasing crop productivity. The value of other subsectors, e.g., livestock and fisheries, is assumed to remain constant. 1.5. Methodology 33. Although current cropland is limited, there is abundant unutilized land that is suitable for crop production in South Sudan. Presently, this land is mainly under natural vegetation, such as grass and trees, but could be converted into cropland if it became profitable for its users. Based on LGP, population density, and current land use/cover, potential cropland expansion is estimated with five and ten year horizons. The precision and accuracy of the cropland expansion projections are hindered by lack of additional location specific information and the inability to ground truth the estimates. In addition, realizing the agricultural potential of new cropland depends on many other factors, such as public policies and investment, which are not considered in the projections here. 34. The cropland projections are based on the land use/cover data presented in Section 2. First, it is assumed that the ratio of crop area to the total area under “grass with crops� and “trees with crops� land uses is 10 percent. Current cropland is then derived from land use coverage in Table 1 and is computed as the sum of land use area under “cropland� and 10 percent of land use area under “grass with crops� and “trees with crops� (Table 11). From this computation, it is estimated that cropland area is 2.7 million ha or 4.1 percent of total land area in South Sudan. Anecdotal information indicates that currently, cropland in South Sudan is mainly expanding into areas with trees (see Section 6). Hierarchically in this cropland expansion model, therefore, all land currently under “trees with crops� (2.6 percent of total land) is the first to be converted into cropland. Once this potential for expansion is exhausted, further cropland expansion occurs at the expense of “tree land� (currently accounting for 62.6 percent of total territory). There is considerable uncertainty as to the condition of forests in South Sudan, and the quality of forests unfortunately cannot be captured by the GIS data available for this analysis. Ideally, cropland expansion would need to occur in low value forests, to avoid the loss of communities’ access to forest resources, upon which their livelihood depends, and for environmental conservation purposes. To prevent farmland expansion into high value productive forests and gazetted areas, it is critical for the GoSS to develop a coherent policy, regulatory, and strategic framework for the sector that reconciles the twin goals of conservation and livelihood support, for example by promoting participatory forest and woodland management, and enhancing forest-related environmental and other services.11 11 See World Bank (2010) for further details. 17 35. Land under “grass with crops� and “grassland� is unlikely to become cropland due to unfavorable climatic and soil conditions and is therefore assumed not to be converted. Other land uses in Table 1 also remains in non-crop use in the modeled time frame. The crop expansion model uses raster-based GIS neighborhood analyses in which a pixel (with a resolution of 1 km2) is the basic unit of land and is assumed to be under a single land use. Two scenarios are modeled based on the rate of expansion: (1) a moderate expansion rate scenario, and (2) a high expansion rate scenario. 36. The cropland expansion pattern will vary based on climatic conditions, soil characteristics, and population density but is likely to follow the logic schematically presented in Figure 6 and detailed below for the moderate expansion scenario (Scenario 1):  In a high production potential/high population density (HH) area, if a pixel C (current cropland) is surrounded by pixels under tree land, then the eight immediate adjoining pixels (identified with 1s in Figure 6), the sixteen pixels (identified with 2s) immediately surrounding the pixels identified with 1s, and the twenty-four pixels (identified with 3s) immediately adjacent to those identified with 2s are assumed to become cropland in the next five to ten years (all the 1s, 2s, and 3s in Figure 6 are candidates).  For HL and MH areas, cropland expansion will be more modest; only the eight pixels (identified with 1s in Figure 6) immediately adjoining pixel C and the sixteen pixels (identified with 2s) are assumed to become cropland in the future if they are currently covered by tree land.  In ML and LH areas, the expansion is even lower; only the eight pixels immediately adjoining pixel C are assumed to become cropland in the future if currently covered by tree land. 37. It is assumed that any land that is currently not under crops in LL areas will not become cropland in the future. Thus in the moderate expansion scenario, for each square kilometer of current cropland, the maximum possibility is to convert another 48 km2 into cropland in HH areas, 24 km2 into cropland in HL and MH areas, and 8 km2 in ML and LH areas. Figure 6: Illustration of cropland expansion at the pixel level 5 5 5 5 5 5 5 5 5 5 5 5 4 4 4 4 4 4 4 4 4 5 5 4 3 3 3 3 3 3 3 4 5 5 4 3 2 2 2 2 2 3 4 5 5 4 3 2 1 1 1 2 3 4 5 5 4 3 2 1 C 1 2 3 4 5 5 4 3 2 1 1 1 2 3 4 5 5 4 3 2 2 2 2 2 3 4 5 5 4 3 3 3 3 3 3 3 4 5 5 4 4 4 4 4 4 4 4 4 5 5 5 5 5 5 5 5 5 5 5 5 Source: Authors’ illustration. 18 38. The high expansion scenario (Scenario 2) doubles the cropland in the moderate expansion scenario. The results of this scenario occurring in the next five to ten years are based on the following assumptions:  Pixel sets 1, 2, 3, 4, and 5 surrounding pixel C (current cropland) in a HH area are assumed to become cropland if they are currently covered by tree land.  In HL and MH areas, pixel sets 1, 2, 3, and 4 surrounding pixel C and currently covered by tree land are assumed to become cropland in the future.  In ML and LH areas, only pixel sets 1, 2, and 3 are assumed to become cropland if currently covered by tree land. 1.6. Cropland expansion 39. The rate of expansion of cropland will be area - and context-specific. The actual extent of expansion will be determined by access to markets, land and forest policy and regulations, and access to tools and labor required for land clearing and tree cutting. In Scenario 1, other factors being constant, cropland will increase by 2.3 times, from the current 2.7 million ha to 6.3 million ha (Table 11 and Figure 7). 40. The expansion is likely to take place through a conversion of tree land into cropland, yet with low relative decline in forested areas. The share of tree land in total land area would decline from 62.6 percent to 59.5 percent (Table 12). The largest expansion of cropland area is expected in Western Bahr el Ghazal (from a very low base) and the three Equatorial states. It is projected that Western and Central Equatoria would account for 20 percent and 19 percent, respectively, of the new cropland, with the shares in Warrap, Upper Nile, and Jonglei at 10 to 13 percent. About 20 percent (and above) of the total land area in Warrap, Central Equatoria, and Western Equatoria would be cultivated as a result of the cropland expansion under Scenario 1. Table 11: Current and projected cropland area under Scenario 1 State Current Expanded Increase Share of cropland in total state cropland* cropland from base area (%) (ha) (ha) (x times) Current Expanded Upper Nile 504,900 683,700 1.4 6.6 8.9 Jonglei 373,600 636,100 1.7 3.1 5.3 Unity 119,500 167,900 1.4 3.2 4.5 Warrap 405,400 723,600 1.8 9.4 16.7 Northern Bahr el Ghazal 247,600 394,100 1.6 8.4 13.4 Western Bahr el Ghazal 73,100 447,000 6.1 0.7 4.4 Lakes 248,200 431,200 1.7 5.7 9.9 Western Equatoria 317,000 1,294,700 4.1 4.1 16.6 Central Equatoria 313,900 1,192,300 3.8 7.3 27.6 Eastern Equatoria 77,600 296,700 3.8 1.1 4.1 TOTAL 2,680,900 6,267,400 2.3 4.1 9.7 Source: Authors’ estimates. Note: *Current cropland area includes 10 percent of “grass with crops� and “trees with crops.� 19 41. As expected, most cropland expansion is projected in areas with high agricultural potential. The Greenbelt would increase its share of cropland from 18.2 percent to 25.7 percent of total cropland in South Sudan (Annex 10). Significant cropland expansion is also projected in the Ironstone Plateau (from 7.6 percent to 17.4 percent) and Hills and Mountains (from 4.6 percent to 8.5 percent) livelihood zones. The areas with high to medium production potential and population density, i.e., HH, HL, and MH, would expand from the current 52.7 percent to 64.9 percent of total cropland area. 42. An increase in cropland would result in larger farm sizes under the moderate expansion scenario. If the expansion occurs in the next five years, per capita cropland size would increase from 0.32 ha to 0.67 ha, assuming a 2.5 percent annual population growth. If expansion takes ten years, per capita cropland size would increase to 0.59 ha. 43. While the rate of cropland expansion is already rapid in Scenario 1, the per capita cropland endowment would still be lower than in neighboring countries. A scenario that doubles the rate of expansion under Scenario 1 results in a 3.5-fold increase in cropland compared to the current cropland area (Table 12). Cropland area would increase to 9.2 million ha, or 14.3 percent of national land. As a result, the share of tree land in total land would decline from the current 62.6 percent to 54.9 percent. The per capita cropland area under this scenario increases from 0.32 to 0.99 ha if expansion takes place within the next five years and to 0.87 ha if expansion occurs over a ten year period. 44. Figure 7 and Figure 8 show the spatial patterns of land expansion under the two scenarios. Table 12: Cropland and other land uses under moderate and high expansion scenarios Area (ha) Share of total land (%) Land use category Current Scenario 1 Scenario 2 Current Scenario 1 Scenario 2 Cropland 2, 477,700 6,267,400 9,237,400 3.8 9.7 14.3 Grass with crops 325,100 292,600 292,600 0.5 0.5 0.5 Trees with crops 1,707,300 0 0 2.6 0.0 0.0 Grass land 9,633,800 9,633,800 9,633,800 14.9 14.9 14.9 Tree land 40,526,900 38,477,100 35,507,100 62.6 59.5 54.9 Other land use* 10,017,300 10,017,300 10,017,300 15.5 15.5 15.5 Total 64,688,300 64,688,300 64,688,300 100.0 100.0 100.0 Source: Authors’ estimates. Note: Other land use includes Flood land, Water and rock, and Urban as categorized in Table 1. 20 Figure 7: Cropland expansion under Scenario 1 Source: Authors’ estimates. Figure 8: Cropland expansion under Scenario 2 Source: Authors’ estimates. 21 1.7. Potential agricultural production values 45. The increase in cultivated area through cropland expansion would lead to higher agricultural output and correspondingly, to a higher value of agricultural production. Even the modest cropland expansion (Scenario 1) would lead to a 2.4-fold increase in the value of total agricultural output (crops, livestock and fisheries) compared to the current estimated output value (Table 13). Potential agricultural production may reach US$2 billion, up from the current US$808 million, which is still far below the level of output produced in neighboring countries (Table 10). The largest increase is expected in the three Equatorial states, Western Bahr el Ghazal, and Warrap. 46. Improvements in agricultural productivity are necessary if South Sudan is to increase production to levels comparable to those observed in the region. Average cereal yields in South Sudan are estimated at 0.8-0.9 tons per ha (FAO/WFP, 2011). Real obtained yields could actually be lower than these averages since the cropland area used in the FAO/WFP (2011) assessments is much lower than that observed in the FAO land cover map (FAO, 2009). 47. These average cereal yields are lower than those in Uganda (1.6 tons per ha), where there is minimal use of tradable inputs, and much lower than in Kenya (2 tons per ha) and Ethiopia (3 tons per ha), where more tradable inputs are used. The wide gap between actual and biophysically attainable yields per unit area (Fisher et al., 2002) in South Sudan points to an immense scope for increasing the average cereal yields. Table 13: Current and potential agricultural value due to cropland expansion Potential agricultural value due to land State Current Current agricultural value expansion (‘000 US$) cropland (ha) (‘000 US$) Scenario 1 Scenario 2 Upper Nile 504,900 87,373 105,027 174,381 Jonglei 373,600 112,535 163,443 282,402 Unity 119,500 26,512 33,839 54,902 Warrap 405,400 67,188 111,662 176,754 Northern Bahr el Ghazal 247,600 48,450 70,018 113,642 Western Bahr el Ghazal 73,100 20,376 85,112 183,642 Lakes 248,200 63,448 101,630 162,464 Western Equatoria 317,000 148,473 564,908 893,758 Central Equatoria 313,900 140,999 462,360 789,355 Eastern Equatoria 77,600 92,340 261,019 530,365 TOTAL 2,668,000 807,694 1,959,028 2,796,474 Source: Authors’ estimates. Note: The estimate of potential agricultural value assumes changes in the value of crop production due to the expansion of cropland, keeping the values of livestock and fisheries output constant. 48. Several levels or magnitudes of possible yield improvement are considered. Average cereal yields per ha are assumed to increase by 50 percent to reach the average level in Uganda, by 100 percent to attain the level in Kenya, and by 200 percent to achieve the average yields in Ethiopia. A 50 percent yield increase would translate into a 3.5-fold increase in the current value of agricultural production in South Sudan. This increase in agricultural value would also be 45 percent higher than an increase accruing from Scenario 1 of land expansion at current yield 22 levels (Table 14). The value of crop production per ha would grow from US$227 to US$340. If yields can increase to the average levels obtained in Kenya, the value of total agricultural production in South Sudan would outpace the current value in Uganda (compare with Table 10) and crop value per ha would reach US$453. A 200 percent increase (to match levels in Ethiopia) in yield per unit area would increase crop value to US$1,020 per ha. Table 14: Current and potential agricultural value under increased cropland and yield/ha Potential agricultural value (‘000 US$) Current Land expansion Land expansion Land expansion Land expansion agricultural only (Scenario 1) (Scenario 1) (Scenario 1) States value (Scenario 1) With 50% yield With 100% yield With 200% yield (‘000 US$) increase increase increase Upper Nile 87,373 105,027 138,783 172,540 240,054 Jonglei 112,535 163,443 225,120 286,797 410,151 Unity 26,512 33,839 46,549 59,259 84,678 Warrap 67,188 111,662 162,229 212,796 313,930 Northern Bahr el Ghazal 48,450 70,018 99,040 128,061 186,104 Western Bahr el Ghazal 20,376 85,112 123,824 162,525 239,929 Lakes 63,448 101,630 146,622 191,613 281,596 Western Equatoria 148,473 564,908 840,637 1,116,366 1,667,825 Central Equatoria 140,999 462,360 680,469 898,578 1,334,796 Eastern Equatoria 92,340 261,019 375,232 484,444 717,868 TOTAL 807,694 1,959,028 2,838,504 3,717,979 5,476,930 Source: Authors’ estimates. 49. Realization of the projected agricultural potential will hinge on many factors and the appropriate resolution of a number of constraints. Some of the factors are institutional (such as land ownership) while others are policy related (e.g., decisions on investment in public goods that support agriculture growth). The GoSS has made considerable progress towards formulating policies that positively contribute to increases in agriculture production and has also attempted to lessen the impacts of a number of constraints to increased production. However, rural connectivity is still a binding and overriding constraint to increased production. Without improved connectivity and reduced transport costs, the agricultural potential of South Sudan will not be realized and food insecurity will not be effectively ameliorated. Table 15 presents the findings of a recent study on Sub-Saharan Africa showing that the realization of agricultural potential (column 4) depends on access to markets (columns 1 and 2). An area that is nine hours away from the market, for example, realizes only 8 percent of its agricultural potential, compared to 46 percent for an area only four hours away from the market. Thus, to realize agricultural potential in South Sudan as discussed above, public investments are required to “reduce the distance� between production and consumption areas. The next section presents a strategy for investing in roads to maximize their contribution to the realization of agricultural potential in South Sudan and reducing food prices. 23 Table 15: Relationship between rural connectivity and realization of crop production potential in Sub-Saharan Africa Travel time Distance to ports (km) Total crop production Crop production relative to (hours) (US$ million) potential production (%) 1.7 470.0 12,469 41.1 3.0 527.7 10,168 45.6 4.1 569.2 7,823 46.6 5.1 607.5 6,959 33.2 6.3 656.0 4,594 20.2 7.6 696.0 3,479 16.3 9.3 741.4 2,580 8.2 11.7 762.6 2,031 5.9 15.4 770.9 1,316 4.7 24.8 716.1 1,405 2.9 Source: Dorosh et al. (2008). Note: Agricultural potential reported in Table 15 is estimated by IFPRI using the same methodology as in this report. 24 INVESTING IN ROADS 1.8. Roads in South Sudan 50. The transport system in South Sudan is characterized by low levels of accessibility, dilapidated infrastructure, and high transport cost. South Sudan’s road network is one of the worst in Africa, ranking far below other African countries in all aspects (Table 16). Less than 5 percent of the existing 7,171 km of primary roads are in good condition, and with the exception of the newly constructed urban paved roads and the Juba-Nimule road, the entire network is gravel, dilapidated, and mainly inaccessible during the rainy season. Table 16: Benchmarking South Sudan’s roads against other African countries Indicator South East Resource-rich Low income Middle income Sudan Africa countries countries countries Classified road density (km per 1,000 sq- 15 101 57 88 278 km of arable land area) Primary network paving ratio (% roads) 2 n/a 82 72 32 Unpaved road traffic (vehicles per day) 53 47 54 39 75 Condition of national and regional roads (% 5 59 80 86 n/a in good or fair condition) Source: World Bank (2011a). 51. Freight tariffs in South Sudan are very high and at least twice those found in the main African corridors and even in Sudan (Table 17). The price differential is explained by very poor quality of the road network and the asymmetry of trading patterns. Poor infrastructure forces trucks to carry small loads and face much longer travel times. Small loads over long distances automatically increase the average unit cost of transportation. For instance, limitations along the Juba Bridge preclude trucks from carrying more than 45 tons (World Bank, 2011c). Furthermore, South Sudan’s trading is concentrated in the south with its East African neighbors and follows a very asymmetric pattern that essentially doubles transport costs faced by trucking companies. Trucks enter South Sudan with import goods but return empty to Uganda and Kenya (World Bank, 2011b). Table 17: Benchmarking international freight for South Sudan’s road network against regional corridors South Sudan West African Central East Southern Sudan Corridor African African African Corridor Corridor Corridor Freight tariff (US cents/ton-km) 20 8-10 8 13 7 5 Roads in good condition (%) 5 26 72 49 82 100 Source: World Bank (2011a). Note: South Sudan and Sudan figures include only regional and national roads. 52. Transport prices for domestic routes are even higher than those for regional routes. From Yei to Juba, for example, transport prices reach US$0.65 per ton-km. In other locations, they are even higher. In Uganda and Kenya, average transport prices are about US$0.15-0.20 on primary roads (World Bank, 2009). 25 53. The fragmented and sparse transport infrastructure networks, enormous travel time, and high transport prices impede access to rural and agricultural production areas. Road density is only 15 km per 1,000 km2 of arable land area, below the average in the rest of Africa (Table 16). Large parts of the economically productive areas in the country are isolated from markets and are vastly underutilized. Except for those living along the interstate roads, most of the rural population has no access to markets during the rainy season, which spans over five to seven months. 54. Underdeveloped road infrastructure amidst competing demands for limited resources present significant trade-offs in the spatial allocation of road investments. While the current stage of roads development in South Sudan is such that upgrading the core interstate roads network to an acceptable standard is essential before embarking on feeder roads development, such investments, if not accompanied by corresponding investments in feeder roads that enhance access to agriculturally important areas, will not effectively contribute to agriculture growth and will not necessarily yield the best possible return on investment. Resource constraints dictate that any feeder roads be developed with enormous selectivity, and coordinated and sequenced with interventions in trunk roads. Ideally, geographic areas or clusters of agriculture areas with the highest potential as identified in Section 4 (and with fewer infrastructure hurdles) should be prioritized first for feeder roads to more rapidly link productive areas and markets. This section details the prioritization of road investment to achieve the highest connectivity in agriculturally important areas at least cost. The next subsection describes the methodology used. 1.9. Rural connectivity: methodology 55. Rural connectivity can be computed and measured in various ways. One frequently used measure is the Rural Accessibility Index (RAI), which measures the share of the rural population living within 2 km of an all-weather road. RAI is principally a social measure of rural connectivity (Carruthers et al., 2009). It does not factor in “economic� differences of rural areas, and is often criticized for its use of a two km boundary as a threshold of accessibility. Another approach to measure rural connectivity focuses on the market accessibility of agricultural production zones and is described as a market measure of rural connectivity. The African Infrastructure Country Diagnostic (AICD) studies (2009) used this approach to estimate the road network required to ensure that areas accounting for certain predefined percentages of total value of current and potential national agricultural output were connected to specified regional and national road networks. 56. In this ESW, both social and market connectivity measures are used. The latter, however, is modified into an adjusted market connectivity measure which adds more flexibility and pragmatism to the approach used in AICD (2009), due to the rich data available for South Sudan compared to the continent-wide less detailed dataset used in the AICD study. The adjusted market connectivity measure:  Combines agricultural potential and population density, emphasizing the need to invest in more populated areas. Priority is accorded to areas with “high production potential and high population density� (HH), “high production potential and low population density� (HL), and “medium production potential and high population density� (MH). Together, these areas are regarded as having high agricultural potential. 26  Aims to connect cropland area ranked by production potential and population density rather than the value of agricultural production, to achieve the highest Cropland Connectivity (CLC) index.  Presents the calculations for 2 km and 5 km boundaries or catchment areas. While a 2 km catchment area can provide easier access to markets than a 5 km boundary, in many countries this difference is insignificant (Starkey, 2007). In Uganda, for example, only after a 4.5 km threshold is less household consumption found to be correlated with distance to markets and distance to a tarmac road (Merotto and Verbeek, 2010). In the current study, therefore, cropland connectivity is computed for both 2 km and 5 km boundaries, the latter representing a pragmatic scenario designed to more affordably connect rural agricultural areas. 1.10. Roads for agricultural development in South Sudan12 57. Roads considered in this study are those needed to move consolidated agricultural output to the nearest market center. This covers the existing: (i) core interstate primary roads; (ii) other primary roads; (iii) secondary roads; and (iv) tertiary roads. It does not include roads needed to connect fields to the nearest village, which need not be all-weather, as often a track suitable for people, motorcycles, or carts is sufficient. It also does not include any new roads in addition to the existing network, realizing the great need and priority to focus first on upgrading and rehabilitating existing roads. There are about 15,764 km of roads in South Sudan, most of which are in poor condition. The road network consists of 2,696 km of “interstate primary roads� (connecting all state capitals plus major cross-border corridors); 4,475 km of “other primary roads�; 6,292 km of secondary roads; and 2,301 km of tertiary roads (Table 18 and Figure 9). Secondary and tertiary roads, as well as some primary roads, are considered “rural.� About 10,200 km, or 65 percent of the total road network, are located in areas with high agricultural potential (HH, HL, and MH) (Table 19). 12 The estimates for roads investments used in this section are taken from a report commissioned for this ESW (see Diao et al., 2011). 27 Figure 9: Different road types in South Sudan Source: WFP maps. Table 18: Different types of roads and their lengths (km) by state, South Sudan State Interstate Other primary Secondary Tertiary Total Upper Nile 506 311 984 0 1,801 Jonglei 49 1,056 833 589 2,527 Unity 326 323 55 0 704 Warrap 215 323 559 0 1,096 Northern Bahr el Ghazal 130 239 567 0 936 Western Bahr el Ghazal 316 364 790 0 1,470 Lakes 369 123 357 3 853 Western Equatoria 335 533 688 538 2,095 Central Equatoria 312 891 187 561 1,950 Eastern Equatoria 139 312 1,271 610 2,332 Total 2,696 4,475 6,292 2,301 15,764 Source: Authors’ estimates based on the WFP maps. 28 Table 19: Total length (km) of different types of roads by agricultural potential zone Agricultural Interstate Other primary Secondary Tertiary Total potential zone HH 389 1,249 1,004 887 3,529 HL 485 641 1,570 1,416 4,112 MH 582 874 1,121 0 2,577 ML 276 939 1,193 0 2,408 LH 443 373 535 0 1,350 LL 522 400 862 0 1,783 Total 2,696 4,475 6,292 2,301 15,764 Source: Authors’ estimates based on the WFP maps. 58. Focusing on areas with the highest agricultural potential and population density would have the highest development impact. It would yield the highest payoff to investments in rural roads, allowing farmers to compete with food imports in the short run and to also conquer cross-border markets in the medium to long run. Cropland (current and potential from the expansion scenarios in Section 4.2) and roads data were used to compute requirements to meet cropland connectivity targets, conservatively estimated at 60 percent of current cropland and 50 percent of potential expanded cropland areas in high agricultural potential areas. 59. At this stage in the reconstruction of South Sudan, the GoSS’s investments are likely focused primarily on completing the interstate primary roads and interconnecting the state capitals. But investments in interstate roads will only marginally improve rural connectivity. The completion of all interstate primary roads across the country will provide access to roads to 18 percent of the population (using the RAI) and 7 percent of the current cropland in high agricultural potential areas (based on the CLC index) within a 2 km boundary (Table 20). Table 20: Access to different roads by agricultural potential zone using a 2 km boundary Total high potential HH HL MH Total RAI zones Interstate primary roads Current cropland 0.06 0.05 0.09 0.07 0.06 0.18 Cropland under expansion Scenario 1 0.04 0.03 0.08 0.05 0.04 0.12 Cropland under expansion Scenario 2 0.04 0.03 0.08 0.04 0.04 0.11 Interstate and other primary roads Current cropland 0.28 0.16 0.17 0.20 0.15 0.39 Cropland under expansion Scenario 1 0.24 0.12 0.17 0.16 0.14 0.34 Cropland under expansion Scenario 2 0.22 0.08 0.17 0.13 0.11 0.32 Primary and secondary roads Current cropland 0.34 0.21 0.27 0.27 0.22 0.47 Cropland under expansion Scenario 1 0.32 0.15 0.26 0.23 0.20 0.43 Cropland under expansion Scenario 2 0.31 0.11 0.26 0.19 0.17 0.41 Primary, secondary, and tertiary roads Current cropland 0.43 0.30 0.42 0.39 0.32 0.58 Cropland under expansion Scenario 1 0.41 0.23 0.40 0.33 0.29 0.54 Cropland under expansion Scenario 2 0.39 0.17 0.38 0.27 0.25 0.51 Source: Authors’ estimates. 60. Investing in other roads is necessary to achieve a higher rural connectivity. Completing all primary roads will increase the CLC index to 20 percent in high agricultural potential areas and 15 percent in the whole country. The RAI will be 39 percent. The maximum share of current cropland that can be accessed through the existing roads (primary, secondary, 29 and tertiary), once fully rehabilitated, is 39 percent for high agricultural potential areas and 32 percent for the country as a whole, under a 2 km boundary assumption. The highest RAI would be 58 percent (Table 20). 61. If cropland expansion occurs according to the two expansion scenarios, rural connectivity will actually decline. With the existing roads, the rural connectivity index in high agricultural potential areas will decline from 39 percent to 33 percent under expansion Scenario 1 and to 27 percent under expansion Scenario 2 (Table 20). Correspondingly, the RAI will decline to 51 percent compared to the current 58 percent. 62. A more pragmatic approach to road investments is to increase the catchment area from 2 to 5 km as discussed above. When this wider boundary is considered, the CLC index for current cropland rises to 64 percent in high agricultural potential areas compared to 39 percent within a 2 km boundary (Table 21). Even under the high crop expansion scenario, about 51 percent of total cropland (with 71 percent of the population) in high agricultural potential areas will be connected to roads. This coverage is deemed sufficient to provide the necessary impetus for long term agricultural growth in the country. Table 21: Access to different roads by agricultural potential zone using a 5 km boundary Total high potential HH HL MH Total RAI zones Interstate primary roads Current cropland 0.10 0.09 0.17 0.13 0.11 0.27 Cropland under expansion Scenario 1 0.06 0.06 0.15 0.09 0.09 0.20 Cropland under expansion Scenario 2 0.06 0.06 0.15 0.08 0.08 0.19 Interstate and other primary roads Current cropland 0.46 0.32 0.32 0.36 0.28 0.54 Cropland under expansion Scenario 1 0.39 0.25 0.30 0.30 0.26 0.49 Cropland under expansion Scenario 2 0.36 0.20 0.29 0.26 0.24 0.46 Primary and secondary roads Current cropland 0.55 0.39 0.48 0.47 0.39 0.64 Cropland under expansion Scenario 1 0.51 0.33 0.45 0.42 0.37 0.60 Cropland under expansion Scenario 2 0.50 0.28 0.44 0.37 0.34 0.58 Primary, secondary, and tertiary roads Current cropland 0.69 0.56 0.66 0.64 0.54 0.77 Cropland under expansion Scenario 1 0.64 0.49 0.64 0.58 0.53 0.74 Cropland under expansion Scenario 2 0.62 0.42 0.61 0.51 0.49 0.71 Source: Authors’ estimates. 63. More than 11,000 km of existing roads need to be rehabilitated to meet the rural connectivity targets in high agricultural potential areas; i.e., a CLC index of 60 percent of the current cropland and 50 percent of the expanded cropland areas. This would account for 72 percent of the existing total road network in South Sudan, without building any new roads (Table 22). The share of “rural roads� (secondary and tertiary) in this requirement is estimated at 52 percent. 30 Table 22: Types and lengths of roads needed to meet rural connectivity targets Km required to connect to market: 60% of current cropland Total road network (km) and 50% of expanded cropland Roads needed to satisfy market-access criterion 11,458 15,759 Of which: Core interstate primary roads 2,696 2,696 Other primary roads 2,764 4,475 Secondary roads 3,695 6,285 Tertiary roads 2,303 2,303 Source: Authors’ estimates. 64. Most roads will have to be completed in the three Equatorial states and in Jonglei. These four states account for 79 percent of the roads network required to meet the rural connectivity targets (Table 23), or about 11,000 km (Annex 12). From a livelihood zone perspective, most roads are located in the Greenbelt (34 percent) (Table 24 and Annex 13). This zone also has the longest network of rural roads, together with the Hills and Mountains zone. These are the areas with the highest agricultural potential in terms of favorable climate and population density and thus they should be prioritized for earlier investments to provide the fastest stimulus to agricultural growth in the country (Figure 10). Table 23: Roads distribution by state in high agricultural potential zone (%) State Interstate primary Other primary Secondary Tertiary Total Upper Nile 0.0 3.1 4.7 0.0 2.5 Jonglei 3.1 20.0 17.9 25.6 18.1 Unity 0.0 2.6 0.4 0.0 0.8 Warrap 4.8 4.7 7.7 0.0 4.7 Northern Bahr el Ghazal 7.4 2.6 3.4 0.0 3.0 Western Bahr el Ghazal 10.6 2.7 3.8 0.0 3.6 Lakes 21.3 2.5 6.3 0.1 6.0 Western Equatoria 21.8 18.5 18.6 23.4 20.1 Central Equatoria 21.4 32.2 5.1 24.4 19.1 Eastern Equatoria 9.5 11.0 32.3 26.5 22.0 Source: Authors’ estimates. Table 24: Roads distribution by livelihood zone in high agricultural potential zone (%) Livelihood zone Interstate primary Other primary Secondary Tertiary Total Eastern Flood Plains 2.0 7.4 9.7 11.3 7.0 Greenbelt 39.4 27.6 46.3 21.1 33.5 Hills and Mountains 15.4 22.4 20.0 36.3 21.6 Ironstone Plateau 16.2 2.0 7.2 18.9 7.6 Nile-Sobat Rivers 0.7 2.3 3.1 1.3 2.0 Pastoral 21.3 28.8 7.9 11.1 21.2 Western Flood Plains 5.0 9.6 5.8 0.1 7.0 Source: Authors’ estimates. 31 Figure 10: Combination of roads, agricultural potential zones, and cropland areas Source: Authors’ presentation. 1.11. Budget requirements 65. The unit cost of road construction in South Sudan is among the highest in Africa and extremely onerous by any standard (Table 25). It is well recognized that in post-conflict economies, prices tend to escalate, due to political instability and insecurity, and also to construction booms, where high demand for reconstruction meets an inelastic supply response. In the case of South Sudan, the high cost situation is worsened by the shortage of skilled operators and technicians and the extraordinarily high cost of living and hardship for the mobilized labor force (World Bank, 2011c). Table 25: Cost of rehabilitation and reconstruction of two-lane inter-urban roads South Sudan DRC Ghana Mozambique Nigeria Ethiopia Malawi Average unit cost 1,000-1,300 229 261 279 330 388 421 ('000 US$/km) Source: World Bank (2011c). 32 66. The domestic construction industry is very underdeveloped. Developers have limited or no information on the potential for infrastructure developments and upcoming investments, and procurement practices are poor. Costs are further escalated because construction materials are not available locally, costs associated with shipping materials to the site of construction are enormous, there is almost non-existent competition in the construction market, and there is widespread incidence of land mines that need to be cleared prior to construction. 67. High quality roads are critical for economic development in South Sudan. However, given the many urgent competing demands on government resources, the significant length of uncompleted roads, and the low capacity of the domestic construction industry, pragmatic decisions are required to develop roads in stages. In this analysis, two investment options are presented. The first is a base scenario, with desirable investments to achieve the highest standards of road rehabilitation and construction. Under this scenario, all interstate and other primary roads are upgraded to two-lane paved roads with double surface asphalt treatment, at an average unit cost of US$1,150,000 per km (Table 26)13. Secondary roads are upgraded to two- lane gravel standard with seal or wearing course. All tertiary roads are upgraded to two-lane gravel roads designed for fifty vehicles a day. Annual road maintenance is estimated to be US$30,000 per km, to cover spot improvements and repair works in addition to regular maintenance. Table 26: Cost scenarios for road rehabilitation, construction, and maintenance in South Sudan Road type Base scenario Pragmatic scenario Interstate primary US$1,150,000 per km US$1,150,000 per km roads Paved asphalt two-lane road Paved asphalt two-lane road Other primary roads US$1,150,000 per km US$370,000 per km Paved asphalt two-lane road Gravel two-lane road with seal or stabilized gravel wearing course Secondary roads US$370,000 per km US$200,000 per km Gravel two-lane road with seal or stabilized Gravel two-lane road designed for 50 vehicles a gravel wearing course day, with adequate drainage structures and pavement Tertiary roads US$200,000 per km US$100,000 per km Gravel two-lane road designed for 50 vehicles Gravel two-lane road designed for 30 vehicles a a day, with adequate drainage structures and day, with critical drainage structures and basic pavement surfacing and variable road width Road maintenance US$30,000 per km US$15,000 per km Including spot improvement and repair works Routine maintenance only Source: Authors’ estimates based on input from the World Bank Transport Sector staff. 68. In addition, once capital investments are made, regular maintenance would require US$344 million annually in high potential zones and US$473 million for the total roads network, adding another 15 to 21 percent of the 2010 public expenditure (Table 28).14 13 The road construction costs that are presented are conservative estimates and could be higher especially in areas far away from State Capitals. 14 At the exchange rate of 2.5 SDG per US$1, the total budget in 2010 was about US$2,252 million. The total expenditure on transport and roads equaled US$192 million. 33 Table 27: Budget requirements for road investments under the base scenario (US$ million) Road type Roads in high potential zone Total roads network Interstate primary roads 3,100.6 3,100.6 Other primary roads 3,178.1 5,146.3 Secondary roads 1,367.2 2,325.4 Tertiary roads 460.6 460.6 Total capital spending 8,106.5 11,032.8 Road maintenance 343.7 472.8 Source: Authors’ estimates. Table 28: Approved budget in 2010 and 2011 in South Sudan (SDG million) 2010 2011 Budget items for all sectors Salaries 2,234 2,433 Operating expenses 2,258 2,076 Capital expenditure 1,138 1,258 Total budget 5,630 5,767 Budget items for transport and roads Salaries 15 13 Operating expenses 9 6 Capital expenditure 456 496 Total budget for transport and roads 480 515 Source: GoSS budget estimates. 69. Although the government’s fiscal position has improved after independence, still these costs are very high in light of other needs in the country, and therefore a more pragmatic approach/scenario is recommended. In this scenario, while all interstate primary roads are upgraded to the same standard as in the base scenario, other primary roads are constructed at the lower gravel standards (Table 26). Secondary roads are upgraded to class A rural roads designed for fifty vehicles per day, with adequate drainage structures and pavement; tertiary roads are designed for thirty vehicles per day, with critical drainage structures and basic surfacing. Lower standard feeder roads designed for ten vehicles per day or fewer are unlikely to be common in the high potential agricultural areas, though they may be a pragmatic solution in other rural areas. 70. In the case of South Sudan, high end networked infrastructure services are not a feasible option in the short and medium term. Adopting low-cost modern technologies could substantially reduce the cost of expanding access to roads, and help make the transitional period and the potential funding gap manageable. Initially adopting a gravel road standard – perhaps with some light asphalt stabilization or locally available sealing as discussed above – could help accelerate the achievement of rural connectivity, with full paving investments deferred to a later date. It is estimated that careful choice of technology and targeting feeder road interventions to the highest quality agricultural land could reduce the transport sector spending needs by 40 percent thus freeing up resources for other equally important investments. Under the pragmatic scenario, the budget needs for roads to meet rural connectivity targets are estimated at US$5.1 billion (including US$2 billion for rural roads), compared to US$8.1 billion (including US$5 billion for rural roads) under the base scenario (Table 29). 34 Table 29: Budget requirements for road investments under the pragmatic scenario (US$ million) Road type Roads in high potential zone Total roads network Interstate primary roads 3,100.6 3,100.6 Other primary roads 1,022.5 1,655.8 Secondary roads 739.1 1,257.0 Tertiary roads 230.3 230.3 Total capital spending 5,092.4 6,243.6 Road maintenance 171.9 236.4 Source: Authors’ estimates. 71. The largest share of the roads budget would be spent on interstate primary roads. They are expensive and expenditure would need to be twice as large as the entire 2010 capital investment budget. It is therefore critical to reduce the unit costs of interstate primary roads, not only to reduce the overall budget envelope, but also to be able to turn quickly to construction of rural roads (i.e., secondary and tertiary roads) that are critical for rural connectivity. Rural roads are estimated at 6,000 km, and would require a budget of US$1.0 billion and US$1.8 billion, respectively, under the pragmatic and base scenarios. 1.12. Reducing transport prices and its potential effect on food prices 72. Investments in roads will reduce transport costs and transport prices in South Sudan and should also reduce food prices and improve food security. However, the extent of reductions will depend on policies on competition in the trucking sector, regulations, non-tariff barriers, and the functioning of the food collection and distribution systems among other measures. It is important to ensure that these policies complement the value of roads investment, rather than reducing it. The objective is to ensure: (i) that better roads result in lower transport costs for the trucking industry (through lower use of fuel and tires, and lower maintenance and other costs), and (ii) that the transport cost savings resulting from road improvements are passed on to producers and consumers. 73. The critical precondition for this is competition among transporters. Concerns about the competitive nature of transport operators have long been recognized, most recently in a study on international corridors in Africa (Teravaninthorn and Raballand, 2009). In a monopoly environment, investments in roads reduce transport costs but those cost savings are not usually transferred to end users through lower transport prices and reduced food prices. In other words, the lower transport costs increase profits of the trucking industry but do not reduce costs for producers and consumers. This has happened in many Western and Central African countries, for example, where strong cartels of transport firms oppose opening of the sector, resulting in an insignificant pass-through of any cost savings to end users of transport services (Table 30). The situation is different in competitive environments such as in East Africa, where a reduction in transport costs eventually led to a reduction in transport prices. 35 Table 30: Measures and outcomes for reducing transport prices along the main transport corridors in Central and West Africa Decrease in transport Increase in sales (%) Decrease in transport Measure costs (%) price (%) Rehabilitation of corridor from fair to good -5 Not substantial (NS) +/0 20% reduction in border-crossing time -1 +2/+3 +/0 20% reduction in fuel price -9 NS +/0 20% reduction of informal payment -1 NS +/0 Source: Teravaninthorn and Raballand (2009). Table 31: Measures and outcomes for reducing transport prices along the main transport corridors in East Africa Decrease in transport Increase in Decrease in transport Measure costs (%) sales (%) price (%) Rehabilitation of corridor from fair to good -15 NS -7/-10 20% reduction in border-crossing time -1/-2 +2/+3 -2/-3 20% reduction in fuel price -12 NS -6/-8 20% reduction of informal payment -0.3 NS +/0 Source: Teravaninthorn and Raballand (2009). 74. It is imperative for South Sudan, therefore, to promote competition among transporters to achieve results similar to those in East African countries. Transport prices and costs in Kenya and Uganda are lower than in Central and Western Africa. The competitive nature of their transport industry results in the significant pass-through of cost savings, from improved roads to lower transport prices for end users (Table 31). Thus, if South Sudan promotes competition in the transport sector, better roads will translate into reduced food prices for most of the population and would produce nation-wide benefits in terms of food security. 75. Non-tariff barriers should be eliminated to ensure that investments in roads provide benefits to farmers and consumers. There are many reports pointing to a number of non-tariff barriers in South Sudan, ranging from road blocks and security checks to ambiguous collection of local taxes and various fees (Selassie, 2009; Asebe, 2010; World Bank, 2011b). On the route to Juba from the two border posts of Kaya and Nimule, trucks transporting goods are typically stopped to pay various fees every 7 to 15 km, or five to ten times (World Bank, 2011b). For large trucks, the total amount paid is often not large compared to the transport costs, but the main concern is the high opportunity cost of wasted time.15 For smaller traders, however, the monetary costs of various fees are significant. Non-tariff barriers on certain routes can be a high proportion of transport costs, as is likely to be the case for trade between Lira, Uganda, and Juba, where the difference in maize price (US$550 per ton in April 2011) is only partially (32 percent) explained by transport fees (US$177).16 Besides not bringing revenues to the budgets, these additional costs also reduce the value for money of roads investments and hurt agricultural competitiveness. 15 In a recent World Bank report (2011b), the amount of such payments per ton-km was found to range from US$0.012 per ton- km for a 40 ton truck to US$0.046 per ton-km for a 10 ton truck. Total payment is estimated at SDG 200, using an exchange rate of 2.3 SDG per US$1. The distance between Kaya and Juba is 233 km, and between Nimule and Juba, 193 km. 16 Transport prices between Lira and Juba are estimated using the following assumptions: the distance between Lira and Nimule (border post) is 212 km, with a transport price of US$0.25 per ton-km. The distance between Nimule and Juba is 193 km, with a transport price of US$0.65 per ton-km. 36 AGRICULTURAL COMPETITIVENESS 76. Are investments in roads sufficient first to increase and then maintain competitiveness of South Sudan’s agriculture? Especially if complemented with good transport policy and regulations, roads will surely be transformative but not sufficient. The analysis below suggests that other productivity-inducing public investments are still needed if South Sudan is to compete with neighboring countries (e.g., Uganda and Sudan) that are currently very competitive in South Sudanese markets. These countries have lower production costs, and improved roads as analyzed in Section 5 would make them even more competitive by reducing the “distance� between their own farmers and South Sudanese consumers. This section, deals with price and cost competitiveness of farms in South Sudan vis-à-vis Uganda and Sudan, assessing the current situation and identifying farm cost-reduction strategies. 1.13. Price competitiveness 77. Staple food prices in South Sudan are very high, at least double those in the major markets in Sudan and Uganda. White maize is imported mainly from Uganda, as shown in Figure 11, and consumed in the southern part of the country. Ugandan maize prices are the lowest in East Africa, and thus very competitive; the price gap between Kampala and Juba can reach as high as US$800 per ton in some months. Sorghum, another key staple, is mainly imported from Sudan (Figure 13) and the import parity prices imputed from the prices in Kadugli, a border town in Sudan, are also much lower than in the major markets in South Sudan (Figure 14).The price wedge between Kadugli and Juba can reach US$500-600 per ton. 37 Figure 11: Typical maize flows in South Sudan Source: www.fews.net. Figure 12: Maize prices in Juba, Nairobi, and Kampala Source: www.fews.net and www.ratin.net. 38 Figure 13: Typical sorghum flows in South Sudan Source: www.fews.net. Figure 14: Sorghum prices in South Sudan and Kadugli (Sudan) Source: www.fews.net. 39 78. Import prices have been setting local prices on many markets. Ugandan maize affects local prices in the markets closest to the Ugandan border. The landed maize prices from Lira are actually lower than local prices in some markets there (Table 32).17 The same applies to sorghum that comes from the North. Although the “law of one price� cannot be strictly applied in South Sudan, due to very poor roads and many non-tariff barriers, the food imports exert and will continue exerting significant pressure on local prices.18 Table 32: Actual and landed prices by import source, March 2011 (US$/ton) Maize Sorghum Market Current price Simulated price Current price Simulated price Juba, Central Equatoria 759 689 843 1,433 Aweil, Northern Bahr el Ghazal 843 909 843 1,140 Bentiu, Unity 943 995 843 729 Bor, Jonglei n/a 517 943 1,186 Kuajok, Western Bahr el Ghazal 843 878 843 846 Malakal, Upper Nile 1,686 1,846 716 1,211 Rumbek, Lakes 1,138 664 927 1,058 Torit, Eastern Equatoria 843 449 421 1,409 Wau, Warrap 674 812 674 1,043 Yambio, Western Equatoria 421 671 421 1,370 Source: Authors’ estimates based on the distances between markets presented in Annex 14. 79. Competitive pressure is likely to increase once market connectivity is improved in South Sudan. In many markets, landed import prices will be lower than in the past once transport prices are reduced. Currently, the average transport price in South Sudan is about US$0.65 per ton-km. If investments in roads reduced these prices by half (from US$0.65 per ton- km to US$0.32 per ton-km), imported maize prices in Juba would fall from the current US$689 to US$628 per ton, or by 9 percent (Table 33). In Rumbek, the price reduction would be more dramatic, due to the longer distance to the Ugandan border. The largest output price effect of lower transport prices is expected in Yambio, assuming Ugandan maize flows through Juba. If transport prices in South Sudan decline to the level of average transport prices in Uganda, maize price reduction is expected to range from 12 percent in Juba to 57 percent in Yambio. If transport prices in South Sudan declined to the current level along major transport corridors in Africa (Table 17), the local price reduction would be even sharper. 17 The landed price is estimated as output price at the source of imports (Lira for maize and Kadugli for sorghum) plus transport (distance times unit costs, $0.25 per ton-km in Uganda and $0.65 per ton-km in South Sudan) plus fixed non-tariff fees on the route from Lira to Juba. 18 Under the law of one price, the price difference between a pair of markets equals the transport cost between the markets. 40 Table 33: Simulated impact of lower transport prices on maize prices in South Sudan (US$/ton) Juba Rumbek Torit Yambio Derived prices (at transport price of US$0.65/ton-km) 689 964 749 471 Derived prices (at transport price of US$0.33/ton-km) 628 768 658 271 Derived prices (at transport price of US$0.22/ton-km) 607 700 627 203 Derived prices (at transport price of US$0.10/ton-km) 584 626 593 128 Price reduction, simulation 1 -9% -20% -12% -42% Price reduction, simulation 2 -12% -27% -16% -57% Price reduction, simulation 3 -15% -35% -21% -73% Source: Authors’ estimates. 80. The decline in sorghum prices is expected to be even larger than that of maize prices if market connectivity in South Sudan is improved. If transport prices decline from US$0.65 to US$0.33 per ton-km, or by 49 percent, the derived sorghum prices in many markets are expected to fall by 30 percent, compared to 9 to 20 percent for maize (Table 34). This large food price effect comes from the longer distances between the source of imports, Sudan, and markets in South Sudan, and thus a bigger share of transport expenses in wholesale sorghum prices (see Annex 14 with the distance matrix). Table 34: Simulated impact of lower transport prices on sorghum prices in South Sudan (US$/ton) Juba Aweil Rumbek Wau Derived prices (at transport price of US$0.65/ton-km) 1285 992 910 895 Derived prices (at transport price of US$0.33/ton-km) 829 680 638 631 Derived prices (at transport price of US$0.22/ton-km) 672 573 545 540 Derived prices (at transport price of US$0.10/ton-km) 358 358 412 406 Price reduction, simulation 1 -36% -31% -30% -30% Price reduction, simulation 2 -48% -42% -40% -40% Price reduction, simulation 3 -72% -64% -55% -55% Source: Authors’ estimates. 81. South Sudan therefore cannot just invest in roads, but should also invest in other productivity-enhancing public goods to improve its competitiveness. This is particularly important due to the high production costs in South Sudan, which prevent most farms from increasing food production even with very high current food prices in consumption areas. The high production costs are the result of low investments in land, high labor costs, high tradable input prices, and high upfront land clearing/tree uprooting costs. The next section looks at farm production costs and farm margins in detail. 41 1.14. Farm production costs19 82. Farm production costs in South Sudan are much higher than those in most of its neighboring countries. They are especially high compared to Uganda, where production costs and food prices are the lowest in East Africa. South Sudan lags behind in all key cost elements (Table 35) facing:  Higher labor requirements, mainly due to the need for land clearing after many years of no land cultivation;  Higher labor costs, ranging from US$5.2 per man-day in Kajo-Keji, Morobo, and Yambio, to about US$10.3 per man-day in Malakal, compared to US$1.0 in Uganda and US$2.3 in Tanzania;  Lower yields; and  Higher prices of tradable inputs and lower efficiency of their use. Table 35: Key elements of maize production costs and revenues in South Sudan, Uganda, and Tanzania South Sudan Uganda Tanzania Average yield (kg/ha) 800 1,200 1,120 Farm-gate price (US$/kg) 0.50 0.15 0.20 Labor requirements (man-days/ha) 72 47 52 Labor cost (US$/man-day) 7.50 1.00 2.31 Use of seeds (kg/ha) 10.5 5.4 7.0 Seed price (US$/kg) 1.57 0.92 1.35 Source: Authors’ estimates based on various data sources and field surveys done by the World Bank. 83. The largest contributor to farm production costs in South Sudan is labor, even where tractors are used for some operations. Labor costs for sorghum production range from US$304 per ha in Kajo-Keji in Central Equatoria to US$565 per ha in Yambio, Western Equatoria (Table 36).20 High labor cost is a result of: (i) high labor requirements for preparing land for cultivation; and (ii) high daily wage rates. High wages are, however, partially offset by low land rents (since land is typically available at no cost). 84. Decades of conflict prompted farmers to flee their land, allowing regeneration and progression of vegetation towards climax formations (mainly forests and shrubs). In most areas, therefore, significant upfront work (mainly cutting, uprooting, and removing trees) is required to clear the climax vegetation formations before the land is cultivable. In Morobo and Kajo-Keji areas, for example, sixteen to twenty man-days per ha are required just to uproot trees (Table 36). Such work is among the main cost disadvantages of South Sudan vis-à-vis its neighbors, where initial land clearing was completed many years ago. Labor is typically hired for this work, but is reported to be expensive and in short supply, especially during the planting and harvesting campaigns, making cropland expansion an expensive undertaking. Mechanized 19 The data on farm production costs were collected in February-March 2011 in various states by visiting farms, NGOs, and farmer groups. Primary data were collected for subsistence versus mechanized farms. Details are in Sebit (2011). 20 All primary data for farm costs were collected in SDG and feddans. To convert all the data into US$/ha, the exchange rate used is 2.9 SDG per US$1, and 1 feddan equals 0.42 ha. 42 activities are only seen in Kajo-Keji and Morobo, in addition to the large mechanized operations in Malakal, Upper Nile, but are usually limited to tillage, harrowing, and planting, while most other operations are carried out manually using family labor. Farming in Yambio, Western Equatoria appears to be the most labor intensive due to the dense forestation formations, the need for frequent weeding (due to high rainfall and incipient soil fertility which promotes weed growth), and harvesting challenges in areas with many trees and shrubs (e.g., thick vegetation in Yambio; see Figure 15) versus harvesting in open fields in Malakal (see Figure 16). Table 36: Labor costs for typical farm production activities in South Sudan Kajo-Keji, Morobo, Yambio, Yei, Tonj North, Central Central Western Central Warrap Equatoria Equatoria Equatoria Equatoria Hired labor Tree cutting (man-days/ha) 19.85 15.88 17.00 5.66 13.9 First tillage using hand hoes (man- days/ha) n/a n/a 23.82 11.92 n/a First tillage/plowing (man-days/ha) 7.9 7.1 n/a n/a n/a Harrowing (man-days/ha) 7.9 7.1 n/a n/a n/a Manure application (man-days/ha) n/a n/a n/a n/a 23.82 Total man-days/ha 35.65 30.08 40.82 17.58 37.72 Daily rate (US$) 5.17 5.17 5.17 6.89 10.34 Hired labor (US$/ha) 184 156 211 121 390 Family labor Land clearing/slashing (man-days/ha) 5.32 3.19 12 14.89 2.66 Harrowing and ranking (man-days/ha) n/a n/a 5.32 2.99 n/a Planting (man-days/ha) 2.93 12.77 5.32 4.79 2.13 First weeding (man-days/ha) 4.52 2.66 10.64 11.98 1.33 Second weeding (man-days/ha) 4.52 2.66 7.98 n/a 1.33 Harvesting (man-days/ha) 4.26 12.77 21.28 3.35 2.66 Post-harvest activities, drying and threshing (man-days/ha) 1.6 0.71 15.96 9.98 1.2 Total man-days/ha 23.15 34.76 78.5 47.98 11.31 50% hired labor cost (US$/ha)* 60 90 203 165 58 100% hired labor costs (US$/ha) 120 180 406 331 117 Total costs (US$/ha)** 304 335 617 452 507 Source: Authors’ estimates based on various data sources and field surveys done by the World Bank. Notes:*Family labor is assumed to be half as expensive as hired labor. ** Family labor is priced at the same rate as hired labor in computing total costs. 85. Daily wage rates in South Sudan are extremely high compared to that elsewhere in the region. In the Equatorial states, wage rates are about US$6 per man-day, while in Warrap and Upper Nile they can reach US$10 per man-day. It is important to note, however, that in many villages, a working day is only four hours compared to the norm of eight hours. The effective wage rate, therefore, could be twice as high as indicated above if computed based on an eight hour work day. The true opportunity cost of family labor is not known in South Sudan, and though it does not cost as much as hired labor, its cost is not zero even in remote areas. In the analysis below, family labor is calculated at full and half of hired labor costs to estimate net farm margins. 43 86. Labor cost is the largest, but not the only, element of farm production costs. Other costs include seeds, hand tools, and tractor services. The use of tradable inputs is typically limited to seeds, often self-produced recycled seeds. Instances of fertilizer and agricultural chemical use are very rare, with the exception of the mechanized irrigation scheme in Malakal, Upper Nile. When these costs are added, they are often higher than the revenues generated from farm production output. Table 37 presents the gross and net margins of typical farms in various areas of South Sudan. Gross margins, estimated as revenue less variable costs, are positive in most areas, mainly due to high output prices. Many farms compensate for low yields with high output prices, but that advantage may disappear once the connectivity of urban consumption centers with imported food is improved. Further deducting the costs of family labor makes farm profits (i.e., net margins) very small, and in most instances, gross margins are not sufficient to cover labor costs valued at market wage rates. Figure 15: Thick vegetation in Yambio Figure 16: Open fields in Malakal Source: Sebit (2011). 44 Table 37: Gross margins of sorghum production in South Sudan Kajo-Keji, Malakal, Morobo, Tonj North, Yambio, Yei, Central Upper Central Warrap Western Central Equatoria Nile**** Equatoria Equatoria Equatoria Output price (US$/kg) 0.69 0.57 0.34 1.30 0.69 0.41 Yield (kg/ha) 952 429 1,000 952 1,000 1,000 Gross revenue (US$/ha) 657 244 340 1,238 690 410 Variable costs Hired labor (US$/ha)* 184 114 311 390 211 242 Seeds (US$/ha) 16 12 6 15 27 23 Hand tools (US$/ha) n/a 164 41 21 82 25 Draft power (tractor) (US$/ha) n/a 57 148 n/a n/a n/a Gross margin (US$/ha) 456 -103 -10 812 370 241 Family labor (US$/ha) 120 n/a 180 117 406 331 Man-days (8 hour day) 23 n/a 35 11 79 48 Daily rate (US$/man-day) 5.17 n/a 5.17 10.34 5.17 6.89 Net margin 1 (US$/ha)** 337 -103 -190 695 -36 -90 Net margin 2 (US$/ha)*** 396 -103 -100 754 167 76 Source: Authors’ estimates based on the field survey, February-March 2011. Notes: *Major hired labor activities are: tree uprooting (in all locations, particularly in the Equatorial states), first tillage with hand hoes (Yei and Yambio), tractor services for planting and harrowing (Kajo-Keji, Malakal and Morobo), manure application (Tonj North). **Net margin 1 assumes the cost of family labor is the same as that of hired labor. ***Net margin 2 assumes that family labor costs half as much as hired labor. ****All operations in Malakal are typically carried out by hired labor. 87. Even when production costs are lower than revenues, they are still too high to compete with farm gate prices prevailing in Uganda and Sudan and with landed import prices in South Sudan. If sorghum prices in Table 37 are reduced to US$0.2 per kg (the prevailing farm-gate price in neighboring countries), even gross margins (value added) would become negative. Over time, food prices in South Sudan are expected to decline due to the increased investments in roads and improved security dividends in terms of greater cross-border trade and higher domestic production. In anticipation of lower output prices in South Sudan, farmers need to raise yields to generate profits, because at the current low yields, farm profits that cover both variable and fixed costs can be generated only at farm prices ranging from US$334 per ton in Yei to US$523 in Yambio (Table 38). Table 38: Production costs per ha and ton of output Kajo-Keji, Malakal, Morobo, Tonj Yambio, Yei, Production costs Central Upper Central North, Western Central Equatoria Nile Equatoria Warrap Equatoria Equatoria Variable costs (US$/ha) 201 347 350 426 320 169 Total costs (US$/ha) 261 347 440 484 523 334 Variable costs (U$/ton) 211 810 350 447 320 169 Total costs (US$/ton) 274 810 440 508 523 334 Source: Authors’ estimates based on the field survey, February-March 2011. 1.15. Cost-reduction strategies 88. Given the high cost of living in South Sudan and the experience of other natural resource-dependent countries, it is unlikely that labor wages – the most significant component of overall farm production costs – will decline appreciably in the short to 45 medium term. Reductions in farm production costs in South Sudan would therefore have to accrue from a combination of increased land and labor productivity. Examples abound in the country where mechanization of some part of the production process has led to significant cost savings. Sebit (2011) shows that when some operations were conducted using tractors, 23 percent less labor was used in the production of sorghum than when all production-related activities were carried out using manual labor. Similarly, when tractors were used, 16 percent less labor was used in producing maize compared to situations in subsistence farmer holdings where only manual labor was used. The use of ox-ploughs in Yambio was shown to reduce the labor requirements for primary tillage by at least six days. Access to and greater use of mechanization will therefore help reduce overall farm production costs. 89. South Sudan is in the incipient stages of formulating an agricultural mechanization policy that will help improve the use and efficiency of agricultural tools, implements, and machinery in agricultural production and value addition operations. It is critical that the approach adopted to stimulate mechanization in the country takes into consideration lessons and experiences in other developing countries. For example, ambitious and politically motivated tractor schemes became fiscal burdens to both the governments and farmers without necessarily raising productivity. It is equally important to be aware that the same predicament befell schemes in countries where mechanization was heavily subsidized through the provision of government- planned and -operated machinery services. These experiences point to a general failure of government-run services to provide timely and profitable mechanization inputs to farmers. The government has to recognize that the private sector is better placed to provide mechanization services and should strive to create conditions for largely self-sustaining development of mechanization with minimal direct intervention. In South Sudan, successful private sector-driven models already exist in Upper Nile, Unity, and Central Equatoria. Other measures that can be used to reduce labor costs include the use of conservation tillage where feasible, reliance on herbicides where the skills for use are available, reliance on draught power, and other labor saving equipment, e.g., ox-ploughs. 90. In tandem with mechanization, South Sudan has to pursue other productivity enhancing measures if it is to reduce farm production costs. Key to this will be the use of tradable inputs and the provision of advisory services on technology and other production related activities. Production in South Sudan is predominantly based on local cultivars or land races of the main staple crops. The genetic potential of these land races is very low and they are generally unresponsive to improved crop management practices. Therefore, regardless of the other agronomic measures used, yields of these crops will still be low. Attempts should therefore be made to remove bottlenecks to the use of improved varieties. As the policy landscape on seeds evolves, and the necessary infrastructure is put in place, efforts should be made to ensure that South Sudan accesses seeds from neighboring countries. Seeds traded in the Association for Strengthening Agricultural Research in Eastern and Central Africa (ASARECA) countries, for example, should be approved for sale in South Sudan without further regulatory approval other than truth-in-labeling. In the medium to long term, support will be needed to improve seed supply to farmers through investments, training, and technical assistance at several levels of the seed chain, from breeder seed through farmer-based seed production. Programs to upgrade the capacity of selected public research stations to produce and store breeder seed for targeted species through investments in irrigation, cold storage, other equipment, and operational support will also be needed. Further, support is needed to strengthen the enabling environment for seed trade and improving the capacity of seed traders. Availability of seed can also be increased by 46 assisting producers active in the informal seed market. Some of the informal seed producers could be helped to expand their markets and encouraged to graduate into seed enterprises in the long term. 91. Realization of yield potentials for improved varieties requires a significant increase in the level of fertilizer use in South Sudan. As elsewhere in Sub-Saharan Africa, fertilizer use is currently very low in the country. Comprehensive data are not available but on the basis of cultivated area, South Sudan uses less than 3 kg of plant nutrients per ha, most of which is used in the irrigated areas in Upper Nile and Unity. A synthesis of studies on factors that have undermined demand for fertilizer in Africa (Morris et al., 2007) indicates favorable incentives (strong fertilizer response and favorable price relations, i.e., input/output price ratios) for maize and sorghum, the key crops in South Sudan. Fertilizer use should therefore be profitable if accessed at reasonable prices. Possible options to increase use of fertilizers include: (i) adopting favorable taxes and tariffs on fertilizer imports; (ii) improving access to finance for private sector investors in the fertilizer business; (iii) considering linking up with other fertilizer importers in the region for regional procurement purposes; and (iv) supporting the development and scaling up of networks of input dealers. In the long term, given the oil endowments in the country, the GoSS should assess the economic viability of local fertilizer production. 92. Due to high transport costs and poor distribution infrastructure, prices of tradable inputs in South Sudan in general are expected to be above those prevailing in Uganda, Kenya, and Tanzania, and close to the level in other land-locked countries such as Rwanda, Burundi, and Zambia (Table 39). These high input prices call for serious attention to the efficient use of inputs (e.g., through improvements in soil and moisture management) to enhance yield response to fertilizer. There will also be a need to promote small-scale irrigation to reduce the risk associated with rainfall variability and to increase the profitability of investments in fertilizer adoption. Table 39: Retail input prices in the selected East and Southern African countries, May 2011 (US$/ton) NPK 17-17-17 Urea 46-0-0 DAP 18-46-0 Maize hybrid seeds Maize OPV seeds Burundi 940 780 1,080 n/a 540 Kenya 700 660 880 1,820 n/a Malawi n/a 860 n/a 2,710 2,560 Rwanda 680 580 820 2,500 530 Tanzania 820 600 960 2,310 1,450 Uganda 860 720 1,000 2,040 920 Zambia n/a 900 n/a 4,000 n/a Source: AMISTA, www.amista.org. 93. Intensifying production is a knowledge intensive activity as it requires greater management of a wider range of factors. Therefore, the GoSS will have to support the improvement of farmers’ skills and knowledge through the provision of advisory services to help increase productivity and lower production costs. The public extension system is still dysfunctional after many years of conflict, and an overarching model of service provision has not yet evolved. Private parties, especially NGOs, dominate the agricultural extension system. The GoSS can take advantage of this situation and, in line with current best practice, develop a pluralistic advisory service system under which private extension providers are either funded to 47 provide extension field services or are incorporated in some way into the public sector extension system. The GoSS can consider promoting grassroot command of the extension system by devolving fiscal responsibility to the lowest possible level of authority, consistent with organizational competencies and the efficient use of funds. Technologies for dissemination can be drawn from those used in Ethiopia, Uganda, and Kenya, which have similar ecosystems and consumer tastes, and well-developed technologies. 94. Reducing post-harvest losses through post-harvest management can lower the costs of production in South Sudan. Besides training farmers in post-harvest management and government-led rehabilitation and upgrading of storage facilities, the government should promote private ownership and operation of storage facilities alongside those of the government. 95. Promotion of rainwater harvesting and irrigation is also important. Given the high cost of irrigation, irrigation development must promote benefits among as many beneficiaries as possible, including support to the emergence of forward and backward linkages between irrigated agriculture and markets through the private sector. South Sudan’s vast water resources are not sufficiently developed to smooth overall variability or the impact of droughts. Currently, areas are mostly cultivated by subsistence-oriented smallholder farmers practicing rainfed agriculture. A strategy for irrigation development aimed at defining a set of medium to long term measures or action plans is important. An institutional framework for South Sudan’s water sector has been developed, and a policy document was published in 2007. A more detailed strategic framework for the country’s water policy is now needed, enabling the country to enact more specific laws on the provision of water for industry, agriculture, and the population. 96. Production costs will not go down in the short term, but they can be reduced in the medium term. Lower farm costs are preconditions for competitiveness, economic growth, and poverty reduction in South Sudan. It is important to establish a division of labor from the very beginning, such that the public sector creates conditions, via regulations and investments, for the private sector to invest and generate profits. Public goods, as discussed in this ESW, are numerous and are critical to spur agricultural growth. If these public goods are provided, South Sudanese farmers should be able to feed the nation and provide food to neighboring countries that are less endowed in terms of agricultural potential. If these public goods are not provided, South Sudan will continue to experience high levels of poverty and dependence on food aid. 48 CONCLUSIONS 97. South Sudan has a huge but largely unrealized agricultural potential. Favorable soil, water, and climatic conditions render more than 70 percent of its total land area suitable for crop production. For several reasons, however, less than 4 percent of the total land area is currently cultivated. Infrastructure bottlenecks, non-tariff barriers, high labor costs, and limited use of productivity-enhancing technologies hinder progress and also constrain the competitiveness of South Sudan’s agriculture relative to its neighbors. This report proffers and describes possible strategies through which South Sudan’s agricultural potential can be realized and its regional competitiveness improved to foster more inclusive growth. 98. Using household consumption data from the NBHS and a GIS-based model, the report estimates current agricultural production in South Sudan. It also assesses the potential for increasing agricultural production (and the respective attendant value) by increasing cropped areas and per capita yields. The report identifies rural roads that are necessary to accelerate expansion of cultivated land in areas that are considered to have high agricultural potential and provides estimates of the budgetary requirements for road investments in those areas. The report also assesses the implications of infrastructure investments on agricultural competitiveness and the scope for reducing production costs in South Sudan to enable producers to compete with food imports, especially from Uganda. 99. The value (realized agricultural potential) of total agricultural production in South Sudan was estimated at US$808 million in 2009. Seventy-five percent (US$608 million) of this value accrues from the crop sector, while the rest is attributed to the livestock and fisheries sectors. The average value of household production is US$628, of which US$473 is realized from crops. Average value of production per ha is US$299 compared to US$665 in Uganda, US$917 in Ethiopia, and $1,405 in Kenya in 2009. 100. Increasing cropland from the current 4 percent of total land area (2.7 million ha) to 10 percent of total land area (6.3 million ha) under a modest cropland expansion scenario would lead to a 2.4-fold increase in the value of total agricultural output relative to the current level (i.e., to approximately US$2 billion versus the current US$808 million). If coupled with a 50 percent increase in per capita yields, this cropland expansion would lead to a 3.5-fold increase in the value of total agriculture output (i.e., to US$2.8 billion) and would also increase the value of crop production per ha from US$227 to US$340. If per capita yields double, the value of total agriculture production under a modest cropland expansion scenario would increase to US$3.7 billion, and would outstrip the current value of agricultural production in neighboring Uganda. Increasing productivity threefold would increase the value of agricultural production to US$5.5 billion. 101. Improved rural connectivity is necessary for land expansion, yield improvements, and the resultant increases in the value of agriculture output. Required investments in rural roads would not only have to first target areas identified as having high agricultural potential, but would also have to adopt a pragmatic approach towards the quality (type) of the roads given severe budget constraints and competing development needs, as well as the low capacity of the local construction industry. A pragmatic approach implies construction of lower quality roads (with lower unit costs) and larger boundaries for assessing roads coverage. This would reduce the capital requirement for rural roads from US$5 billion to US$2 billion and accelerate the achievement of rural connectivity. Full paving investments would be deferred to the future. 49 These investments in roads have to be accompanied by other measures geared towards reducing transport prices, including the promotion of competition among transport service producers and abolishment of various non-tariff barriers to trade, both internal and at cross-border points. 102. Improved rural connectivity, especially if combined with good transport policy and regulations, will be transformative, but in and of itself will not be sufficient to sustain the competitiveness of South Sudanese farmers. Neighboring countries still have lower production costs and will benefit from better roads by providing more affordable prices to South Sudanese consumers, especially in urban areas. Complementary productivity-enhancing investments and market-supportive regulations are therefore required to improve the competitiveness of South Sudan’s agriculture. In the short term, removing bottlenecks to using the available seed varieties in the East Africa region would increase access to improved germplasm, and would help narrow the current yield gap. Investments in mechanization to reduce drudgery and high costs associated with cropping would also allow South Sudanese farmers to increase production at relatively lower costs. Support for adaptive agricultural research would allow release of new and superior seed varieties and would also help overcome other constraints (e.g., pests and diseases) to yield increases. Advisory services will be essential to maximize farm returns from the use of improved inputs, including mechanization and the development of irrigation. For all of these public investments, it is important to ensure that they “crowd in� private investment rather than discouraging it. 50 REFERENCES AICD. 2009. "Africa Infrastructure: A Time for Transformation. Part 2 - Sectoral Snapshot." Africa Infrastructure Country Diagnostic. World Bank, Washington, D.C. Asebe, E. 2010. "Regional Trade and Transportation Facilitation Assessment in Southern Sudan and Northern Great Lakes Region." Draft report prepared for the World Bank, September 15, 2010. Boserup, E. 1965. The Condition of Agricultural Growth. New York: Aldine Publishing. Boserup, E. 1981. Population and Technological Change: A Study of Long Term Trends. Chicago: University of Chicago Press. Brinkman, H.-J. and C. Hendrix. 2010. "Food Insecurity and Conflict: Applying the WDR Framework." Background Paper for the World Development Report 2011. World Bank, Washington, D.C. Carruthers, R., R. Krishnamani and S. Murray. 2009. "Improving Connectivity: Investing in Transport Infrastructure in Sub-Saharan Africa." Background Paper 7 for Africa Infrastructure Country Diagnostic Project. World Bank, Washington, D.C. Collier, P. 2007. "Post-Conflict Recovery: How Should Policies be Distinctive?" Centre for the Study of African Economies, Oxford University. Diao, X., V. Alpuerto, R. Folledo, C. Guvele and L. You. 2009. "Assessing Food Security and Development Opportunities in Southern Sudan." Paper prepared by Development Strategy and Governance Division of IFPRI for US Agency for International Development. IFPRI, Washington, D.C. Diao, X. and R. Folledo. 2011. "Indices for Population and Cropland Access to the Road and Cost Estimation of Road Construction in Southern Sudan." Background Paper prepared for the World Bank by the Development Strategy and Governance Division, IFPRI, Washington, D.C. Diao, X., L. You, V. Alpuerto and R. Folledo. 2011. "Current Condition and Agricultural Potential in Southern Sudan." Background Paper prepared for the World Bank by the Development Strategy and Governance Division, IFPRI, Washington, D.C. Dorosh, P., H. Wang and L. You. 2008. "Crop Production and Road Connectivity in Sub- Saharan Africa: A Spatial Analysis." Washington, D.C.: IFPRI and the World Bank Spatial and Local Development Team, FEU, SDN. FAO/WFP (various years). "FAO/WFP Crop and Food Security Assessment Mission to Southern Sudan." Special Report. www.fao.org/giews. Fisher, G., H. van Velthuizen and F. Nachtergaele. 2002: "Global Agroecological Assessment for Agriculture in the 21st century: Methodology and Results." RR-02-02. Laxenburg, Austria: International Institute for Applied System Analysis. Merotto, D. and J. Verbeek. 2010. "Uganda: Public Expenditure Review. Strengthening the Impact of the Roads Budget." World Bank, Washington D.C. Morris, M., V. Kelly, R. Kopicki and D. Byerlee. 2007. "Fertilizer Use in African Agriculture. Lessons Learned and Good Practice Guidelines." World Bank, Washington, D.C. 51 Musinga, M., J. Gathuma, O. Engorok and T. Dargie. 2010. "The Livestock Sector in Southern Sudan: Results of a Value Chain Study of the Livestock Sector in Five States of Southern Sudan Covered by MDTF with a Focus on Red Meat." The Netherlands Development Organization for MDTF. NBHS. 2009. "Southern Sudan: National Baseline Household Survey." Southern Sudan Centre for Census, Statistics and Impact Evaluation, Juba. Sebit, S. 2011. "Analysis of Farm Production Costs in Southern Sudan." Background Paper prepared for the World Bank, Washington, D.C. Selassie, Z. 2009. "Southern Sudan: Non-Oil Revenue Study." Report for the African Development Bank. African Development Bank, Côte d’Ivoire. SSCCSE. 2006. "Southern Sudan Livelihood Profiles. A Guide for Humanitarian and Development Planning." Southern Sudan Centre for Census, Statistics and Evaluation, Juba. Starkey, P. 2007. "A Methodology for Rapid Assessment of Rural Transport Services." Sub- Saharan Africa Transport Policy Program Working Paper 87 (A). World Bank, Washington, D.C. Teravaninthorn, S. and G. Raballand. 2009. "Transport Prices and Costs in Africa: A Review of the Main Trade Corridors." World Bank, Washington, D.C. World Bank. 2009. "East Africa: A Study of the Regional Maize Market and Marketing Costs." AFTAR Report 49831, World Bank, Washington, D.C. World Bank. 2010. "A Legal and Institutional Policy Framework for Sustainable Management of Forest Resources in South Sudan." Policy Note, Environment and Natural Management Department, Africa Region. Washington, D.C.: World Bank. World Bank. 2011a. "AICD: South Sudan Infrastructure: A Continental Perspective." Draft April 2011. World Bank, Washington, D.C. World Bank. 2011b. "Behind the Recent Growth of South Sudan's Cross-Border Regional Trade with Uganda." Draft Report, PREM Africa Region, June 2011, World Bank, Washington, D.C. World Bank. 2011c. "Policy Note on Infrastructure for GoSS Development Plan." Africa Sustainable Development Network, World Bank, Washington, D.C. 52 ANNEXES Annex 1: Type of land use by 18 categories Land use Area Share of Land use Area Share of total land total land (sq km) (%) (sq km) (%) Rainfed crop 23,793 3.7 Shrubs or tree with crop 17,030 2.6 Irrigated crop 321 0.0 Grass 96,338 14.9 Rice on flood land 60 0.0 Shrubs 205,066 31.7 Fruit crop 1 0.0 Tree with shrubs 176,949 27.4 Tree crop, plantation 62 0.0 Woodland with shrubs 23,254 3.6 Shrubs, tree and woodland Rainfed crop on post flood land 254 0.0 94,976 14.7 with flooded land Rainfed crop on temporary flood land 285 0.0 Water 3,501 0.5 Grass with crop 3,251 0.5 Rock 1,326 0.2 Shrubs with crop 43 0.0 Urban 370 0.1 Source: Aggregated from Land Cover Database, FAO (2009). Land use/cover aggregated into 18 categories Source: Authors’ presentation based on Land Cover Database, FAO (2009). 53 Annex 2: Type of land use by state A: By 18 types of land use categories Upper Jonglei Unity Warrap N. Bahr W. Bahr Lakes Western Central Eastern National (sq km) Nile el Ghazal el Ghazal Equatoria Equatoria Equatoria total Rainfed crop 4227 3219 982 3458 1999 449 2171 2577 2505 574 22161 Irrigated crop 127 3 0 0 0 0 0 5 0 0 135 Rice on floodland 0 0 0 1 61 0 0 0 0 0 62 Fruit crop 0 0 0 0 0 0 0 0 0 1 1 Tree crop, plantation 0 0 0 0 0 2 0 24 37 0 63 Rainfed crop with temporarily flooded land 4 50 46 29 37 12 53 3 13 11 258 Rainfed crop on post flooding land 0 17 0 33 161 1 52 0 6 17 287 Grass with crop 856 830 531 266 36 133 19 247 282 90 3290 Shrub with crop 38 0 0 0 0 0 0 1 5 0 44 Shrub or tree with crop 1193 1254 436 2571 721 2239 459 3442 3703 1234 17252 Grass 26189 14334 7408 4991 1006 4039 5409 8688 4399 20270 96733 Shrub 21030 63817 13096 8033 3788 5873 13450 22395 16602 37099 205183 Tree with shrub 10359 15484 1842 5158 15257 60861 14040 31578 13906 9745 178230 Woodland with shrub 392 892 24 1087 1 8964 1254 9831 957 183 23585 Tree, shrub and other vegetation on flood land 8505 25213 14125 10753 6902 12795 8487 1354 2253 4186 94573 Water 315 803 299 83 701 537 194 122 111 237 3402 Rock 129 9 2 0 17 329 7 697 61 27 1278 Urban 97 33 64 3 12 39 19 14 83 11 375 B: By 8 aggregated categories (sq km) Cropland 4358 3289 1028 3520 2258 464 2276 2609 2561 603 22966 Grass with crop 856 830 531 266 36 133 19 247 282 90 3290 Shrub with crop 1231 1254 436 2571 721 2239 459 3443 3709 1234 17297 Grass 26189 14334 7408 4991 1006 4039 5409 8688 4399 20270 96733 Shrub and tree 31781 80193 14962 14279 19046 75699 28714 63804 31465 47027 406970 Tree, shrub and other vegetation on flood land 8505 25213 14125 10753 6902 12795 8487 1354 2253 4186 94573 Water and rock 444 811 301 83 717 865 201 818 173 264 4677 Urban 97 33 64 3 12 39 19 14 83 11 375 Total 73461 125957 38855 36466 30698 96273 45584 80977 44925 73685 646881 C:% of total national land by state (18 types of land use categories) Rainfed crop 19.1 14.5 4.4 15.6 9.0 2.0 9.8 11.6 11.3 2.6 100 Irrigated crop 93.5 2.6 0.0 0.0 0.0 0.0 0.0 3.9 0.0 0.0 100 Rice on floodland 0.0 0.0 0.0 1.4 98.6 0.0 0.0 0.0 0.0 0.0 100 Fruit crop 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 100.0 100 Tree crop, plantation 0.0 0.0 0.0 0.0 0.0 2.8 0.0 38.0 59.2 0.0 100 Rainfed crop with temporarily flooded land 1.7 19.2 17.9 11.1 14.4 4.7 20.6 1.0 5.1 4.4 100 Rainfed crop on post flooding land 0.0 6.1 0.0 11.5 56.1 0.3 18.2 0.0 2.1 5.8 100 Grass with crop 26.0 25.2 16.1 8.1 1.1 4.0 0.6 7.5 8.6 2.7 100 Shrub with crop 86.1 0.0 0.0 0.0 0.0 0.0 0.0 2.0 11.9 0.0 100 Shrub or tree with crop 6.9 7.3 2.5 14.9 4.2 13.0 2.7 20.0 21.5 7.2 100 54 Grass 27.1 14.8 7.7 5.2 1.0 4.2 5.6 9.0 4.5 21.0 100 Shrub 10.2 31.1 6.4 3.9 1.8 2.9 6.6 10.9 8.1 18.1 100 Tree with shrub 5.8 8.7 1.0 2.9 8.6 34.2 7.9 17.7 7.8 5.5 100 Woodland with shrub 1.7 3.8 0.1 4.6 0.0 38.0 5.3 41.7 4.1 0.8 100 Tree, shrub and other vegetation on flood land 9.0 26.7 14.9 11.4 7.3 13.5 9.0 1.4 2.4 4.4 100 Water 9.3 23.6 8.8 2.4 20.6 15.8 5.7 3.6 3.3 7.0 100 Rock 10.0 0.7 0.1 0.0 1.3 25.7 0.5 54.6 4.8 2.1 100 Urban 25.8 8.8 17.1 0.9 3.2 10.4 5.1 3.7 22.1 2.8 100 D: of total national land by state (8 aggregated land use categories) Cropland 19 14.3 4.5 15.3 9.8 2.0 9.9 11.4 11.2 2.6 100 Grass with crop 26 25.2 16.1 8.1 1.1 4.0 0.6 7.5 8.6 2.7 100 Shrub with crop 7.1 7.3 2.5 14.9 4.2 12.9 2.7 19.9 21.4 7.1 100 Grass 27.1 14.8 7.7 5.2 1.0 4.2 5.6 9.0 4.5 21.0 100 Shrub and tree 7.8 19.7 3.7 3.5 4.7 18.6 7.1 15.7 7.7 11.6 100 Tree, shrub and other vegetation on flood land 9 26.7 14.9 11.4 7.3 13.5 9.0 1.4 2.4 4.4 100 Water and rock 9.5 17.3 6.4 1.8 15.3 18.5 4.3 17.5 3.7 5.6 100 Urban 25.8 8.8 17.1 0.9 3.2 10.4 5.1 3.7 22.1 2.8 100 Total 11.4 19.5 6.0 5.6 4.7 14.9 7.0 12.5 6.9 11.4 100 55 Annex 3: Type of land use by livelihood zone A: By 18 types of land use categories (sq km) Eastern Greenbelt Hills & Ironstone Nile-Sobat Pastoral Western National Flood Plains Mountains Plateau Rivers Flood Plains total Rainfed crop 5861 3957 924 1527 2170 181 7426 22046 Irrigated crop 102 5 0 0 4 0 0 111 Rice on floodland 0 0 0 0 0 0 62 62 Fruit crop 0 0 1 0 0 0 0 1 Tree crop, plantation 0 59 1 4 0 0 0 64 Rainfed crop with temporarily flooded land 4 0 20 23 105 0 109 261 Rainfed crop on post flooding land 18 0 19 42 0 4 210 293 Grass with crop 1610 454 136 183 357 147 388 3275 Shrub with crop 39 6 0 0 0 0 0 45 Shrub or tree with crop 1368 4848 1793 3121 841 731 4598 17300 Grass 34136 8071 8327 10144 5092 19529 11583 96882 Shrub 57811 23240 34503 20270 13853 32804 22907 205388 Tree with shrub 16405 31565 9679 86222 7827 8883 17371 177952 Woodland with shrub 420 7901 775 13653 179 104 659 23691 Tree, shrub and other vegetation on flood land 13648 1106 3312 15917 29004 6188 25384 94559 Water 281 157 113 573 1202 215 846 3387 Rock 129 687 42 314 1 16 3 1192 Urban 123 15 85 52 33 4 67 379 B: By 8 aggregated categories (sq km) Cropland 5984 4022 966 1595 2278 185 7807 22837 Grass with crop 1610 454 136 183 357 147 388 3275 Shrub with crop 1406 4854 1793 3121 841 731 4598 17344 Grass 34136 8071 8327 10144 5092 19529 11583 96882 Shrub and tree 74636 62706 44956 120144 21858 41792 40937 407029 Tree, shrub and other vegetation on flood land 13648 1106 3312 15917 29004 6188 25384 94559 Water and rock 409 844 154 887 1203 231 849 4577 Urban 123 15 85 52 33 4 67 379 Total 131953 82073 59728 152042 60666 68807 91613 646882 C:% of total national land by state (18 types of land use categories) Rainfed crop 26.6 17.9 4.2 6.9 9.8 0.8 33.7 100 Irrigated crop 92.0 4.8 0.0 0.0 3.2 0.0 0.0 100 Rice on floodland 0.0 0.0 0.0 0.0 0.0 0.0 100.0 100 Fruit crop 0.0 0.0 100.0 0.0 0.0 0.0 0.0 100 Tree crop, plantation 0.0 93.1 1.4 5.5 0.0 0.0 0.0 100 Rainfed crop with temporarily flooded land 1.7 0.0 7.8 8.8 40.1 0.0 41.7 100 Rainfed crop on post flooding land 6.0 0.0 6.7 14.2 0.0 1.2 71.9 100 Grass with crop 49.2 13.9 4.1 5.6 10.9 4.5 11.8 100 Shrub with crop 86.1 13.9 0.0 0.0 0.0 0.0 0.0 100 Shrub or tree with crop 7.9 28.0 10.4 18.0 4.9 4.2 26.6 100 Grass 35.2 8.3 8.6 10.5 5.3 20.2 12.0 100 56 Shrub 28.1 11.3 16.8 9.9 6.7 16.0 11.2 100 Tree with shrub 9.2 17.7 5.4 48.5 4.4 5.0 9.8 100 Woodland with shrub 1.8 33.4 3.3 57.6 0.8 0.4 2.8 100 Tree, shrub and other vegetation on flood land 14.4 1.2 3.5 16.8 30.7 6.5 26.8 100 Water 8.3 4.6 3.3 16.9 35.5 6.4 25.0 100 Rock 10.8 57.7 3.5 26.4 0.1 1.3 0.2 100 Urban 32.4 4.0 22.5 13.7 8.8 0.9 17.6 100 D: of total national land by state (8 aggregated land use categories) Cropland 26.2 17.6 4.2 7.0 10.0 0.8 34.2 100 Grass with crop 49.2 13.9 4.1 5.6 10.9 4.5 11.8 100 Shrub with crop 8.1 28.0 10.3 18.0 4.8 4.2 26.5 100 Grass 35.2 8.3 8.6 10.5 5.3 20.2 12.0 100 Shrub and tree 18.3 15.4 11.0 29.5 5.4 10.3 10.0 100 Tree, shrub and other vegetation on flood land 14.4 1.2 3.5 16.8 30.7 6.5 26.8 100 Water and rock 8.9 18.4 3.4 19.4 26.3 5.1 18.5 100 Urban 32.4 4.0 22.5 13.7 8.8 0.9 17.6 100 Total 20.4 12.7 9.2 23.5 9.4 10.6 14.2 100 57 Annex 4: Population density and share of cropland by agricultural potential-population density typologies by state High agricultural potential Medium agricultural potential Low agricultural potential High/medium Low population High/medium Low High/medium Low Population density population density population population population population Total (person/km2) density (Type HL) density density density density (Type HH) (Type MH) (Type ML) (Type LH) (Type LL) Upper Nile 0 0 46 5 65 3 12 Jonglei 48 4 42 7 47 6 11 Unity 0 0 58 5 49 5 15 Warrap 0 0 50 7 51 4 22 N. Bahr el Ghazal 0 0 87 3 42 4 24 W.Bahr el Ghazal 46 1 101 2 25 0 3 Lakes 30 3 59 4 0 0 16 Western Equatoria 68 3 46 3 0 0 8 Central Equatoria 89 5 38 5 0 0 25 Eastern Equatoria 51 5 29 3 19 3 12 National 66 3 54 4 51 3 13 Cropland share (%) Upper Nile 0.0 0.0 1.3 0.7 7.6 15.1 24.7 Jonglei 0.4 0.4 3.6 3.1 2.9 2.3 12.8 Unity 0.0 0.0 0.8 0.4 1.4 1.4 4.1 Warrap 0.0 0.0 8.8 4.0 1.1 0.2 14.1 N. Bahr el Ghazal 0.0 0.0 3.6 1.1 3.5 0.4 8.6 W.Bahr el Ghazal 0.0 0.0 0.6 1.6 0.0 0.4 2.6 Lakes 0.1 0.1 5.7 2.7 0.0 0.0 8.6 W. Equatoria 5.4 5.4 0.1 0.2 0.0 0.0 11.0 C. Equatoria 7.1 3.8 0.0 0.0 0.0 0.0 10.9 Eastern Equatoria 1.2 1.5 0.0 0.0 0.0 0.0 2.7 National 14.1 11.1 24.7 13.7 16.6 19.7 100 Source: Authors’ estimates based on NBHS (2009) and LandScan (2009). 58 Annex 5: Population density and share of cropland by agricultural potential-population density typologies by livelihood zone High agricultural potential Medium agricultural potential Low agricultural potential Population density High/medium Low High/medium Low High/medium Low Total (person/km2) population population population population population population density density density density density density Eastern Flood Plains 28 5 38 6 45 3 11 Greenbelt 77 3 20 1 0 0 14 Hills and Mountains 63 4 0 4 0 0 17 Ironstone Plateau 41 3 75 2 25 0 5 Nile-Sobat Rivers 112 8 66 6 85 5 18 Pastoral 59 4 41 5 16 3 6 Western Flood Plains 34 5 59 6 41 5 26 National 66 3 54 4 51 3 13 Cropland share (%) Eastern Flood Plains 0.0 0.0 4.1 3.2 8.0 15.8 31.2 Greenbelt 11.2 5.9 0.0 0.0 0.0 0.0 17.1 Hills and Mountains 1.5 2.7 0.0 0.0 0.0 0.0 4.2 Ironstone Plateau 0.8 1.7 1.6 2.5 0.0 0.4 7.0 Nile-Sobat Rivers 0.3 0.2 1.7 1.1 2.8 2.7 8.8 Pastoral 0.2 0.6 0.0 0.1 0.0 0.0 1.0 Western Flood Plains 0.1 0.0 17.2 6.9 5.7 0.9 30.7 National 14.1 11.1 24.7 13.7 16.6 19.7 100.0 Source: Authors’ estimates based on NBHS (2009) and LandScan (2009). 59 Annex 6: Share of food consumption by aggregated items for all households Pulse Other Per HH Per capita Cereals Roots & oil Livestock Fish crops (US$/yr) (US$/yr) seeds National total 48.0 1.8 3.8 12.8 29.7 4.0 377 58 Upper Nile 26.7 2.0 6.1 31.3 30.8 3.0 466 61 Jonglei 55.1 0.2 1.5 3.5 38.8 0.9 415 65 Unity 76.7 0.8 1.4 11.6 8.3 1.1 242 31 Warrap 74.7 0.0 6.4 3.7 11.6 3.5 306 43 Northern Bahr el Ghazal 60.3 0.2 2.6 5.6 23.2 8.2 310 50 Western Bahr el Ghazal 24.0 1.2 5.3 17.4 40.3 11.7 255 47 Lakes 68.5 1.2 2.6 4.9 12.9 9.9 344 46 Western Equatoria 34.6 5.5 6.8 16.9 27.8 8.4 331 60 Central Equatoria 35.8 4.6 3.8 21.4 31.8 2.5 439 70 Eastern Equatoria 43.2 0.9 2.1 7.9 44.0 1.9 477 84 Rural total 51.9 1.4 3.5 9.8 29.4 4.0 341 53 Upper Nile 27.9 1.1 5.6 29.0 32.4 3.9 408 55 Jonglei 55.0 0.2 1.5 3.2 39.2 0.9 405 64 Unity 80.9 0.4 1.3 9.2 7.1 1.2 225 29 Warrap 77.3 0.0 6.3 2.7 10.3 3.5 293 42 Northern Bahr el Ghazal 63.0 0.0 2.5 4.3 21.4 8.8 281 46 Western Bahr el Ghazal 31.6 0.5 2.4 8.3 42.3 14.9 176 34 Lakes 68.0 1.2 2.5 4.5 13.5 10.2 320 43 Western Equatoria 37.0 6.6 7.8 17.0 22.3 9.3 286 53 Central Equatoria 40.7 5.6 3.4 17.5 30.7 2.1 322 53 Eastern Equatoria 43.0 0.6 1.8 6.3 46.9 1.4 469 83 Urban total 34.7 2.9 4.9 23.0 30.6 3.9 594 84 Upper Nile 23.8 4.1 7.2 36.6 27.2 1.0 694 83 Jonglei 56.1 0.3 1.2 7.2 34.3 1.0 622 82 Unity 50.7 3.7 2.5 26.5 15.9 0.6 446 45 Warrap 54.1 0.2 7.9 11.9 22.1 4.0 454 56 Northern Bahr el Ghazal 44.9 1.1 3.1 13.1 33.3 4.5 731 102 Western Bahr el Ghazal 19.3 1.7 7.1 23.1 39.1 9.7 351 62 Lakes 71.9 0.5 3.0 7.7 8.9 8.0 660 70 Western Equatoria 28.5 2.7 4.3 16.9 41.5 6.1 549 89 Central Equatoria 30.8 3.6 4.2 25.4 33.0 3.0 697 103 Eastern Equatoria 44.7 3.5 4.5 20.3 20.7 6.3 548 89 Source: Authors’ estimates based on NBHS (2009). 60 Annex 7: Type of rural households, with and without cereal consumption Cereal consuming households Without cereal consuming households States Total Subsistence Buyers In-between Total With Without root Without root or root/tuber but with livestock/fish consumption livestock/fish consumption consumption Number of rural households Total rural 850897 289542 411699 149656 249221 59881 127004 62337 Upper Nile 72561 16243 50248 6071 38595 2212 35876 507 Jonglei 150249 59141 71806 19303 28951 849 21673 6429 Unity 52553 20513 25873 6168 11074 1110 8892 1072 Warrap 119752 25144 69904 24704 35117 3182 14274 17661 N. Bahr el Ghazal 116450 27422 57996 31033 6137 300 4835 1002 W. Bahr el Ghazal 19838 6312 10785 2741 11931 1203 7855 2873 Lakes 53923 10953 25672 17298 29916 1174 15673 13069 Western Equatoria 62431 34955 17905 9572 33417 28233 2534 2650 Central Equatoria 78549 14207 51357 12985 42453 20656 7316 14481 Eastern Equatoria 124591 74655 30153 19783 11632 962 8076 2594 % of total rural households (total rural households =100) Total rural 77.3 26.3 37.4 13.6 22.7 5.4 11.5 5.7 Upper Nile 65.3 14.6 45.2 5.5 34.7 2 32.3 0.5 Jonglei 83.8 33.0 40.1 10.8 16.2 0.5 12.1 3.6 Unity 82.6 32.2 40.7 9.7 17.4 1.7 14 1.7 Warrap 77.3 16.2 45.1 16.0 22.7 2.1 9.2 11.7 N. Bahr el Ghazal 95.0 22.4 47.3 25.3 5.0 0.2 3.9 0.8 W. Bahr el Ghazal 62.4 19.9 33.9 8.6 37.6 3.8 24.7 9.0 Lakes 64.3 13.1 30.6 20.6 35.7 1.4 18.7 15.6 Western Equatoria 65.1 36.5 18.7 10.0 34.9 29.5 2.6 2.8 Central Equatoria 64.9 11.7 42.4 10.7 35.1 17.1 6 12.0 Eastern Equatoria 91.5 54.8 22.1 14.5 8.5 0.7 5.9 1.9 % of different types of households (type of rural households = 100) Total rural 100 34.0 48.4 17.6 100 24.0 51.0 25.0 Upper Nile 100 22.4 69.2 8.4 100 5.7 93.0 1.3 Jonglei 100 39.4 47.8 12.8 100 2.9 74.9 22.2 Unity 100 39.0 49.2 11.7 100 10.0 80.3 9.7 Warrap 100 21.0 58.4 20.6 100 9.1 40.6 50.3 N. Bahr el Ghazal 100 23.5 49.8 26.6 100 4.9 78.8 16.3 W. Bahr el Ghazal 100 31.8 54.4 13.8 100 10.1 65.8 24.1 Lakes 100 20.3 47.6 32.1 100 3.9 52.4 43.7 Western Equatoria 100 56.0 28.7 15.3 100 84.5 7.6 7.9 Central Equatoria 100 18.1 65.4 16.5 100 48.7 17.2 34.1 Eastern Equatoria 100 59.9 24.2 15.9 100 8.3 69.4 22.3 Source: Authors’ calculation based on NBHS (2009) Subsistence households are those were more than 90% of cereals consumed are from own produce Cereal buyers are households were less than 10% of cereals consumed are from own produce In-between households are those that are neither subsistence nor buyers 61 Annex 8: Livestock population by state: SSCCSE computed estimates, 2008 Population (head) Share in national total (%) Cattle Goats Sheep Total Cattle Goats Sheep Total Upper Nile 1,609,631 999,985 1,108,949 3,718,565 4.6 4.9 4.2 4.5 Jonglei 8,487,911 3,430,424 4,016,443 15,934,778 24.1 16.7 15.2 19.4 Unity 1,828,848 872,765 1,031,150 3,732,763 5.2 4.3 3.9 4.5 Warrap 3,065,690 1,377,243 1,977,304 6,420,237 8.7 6.7 7.5 7.8 N. Bahr el Ghazal 894,005 621,693 783,539 2,299,237 2.5 3.0 3.0 2.8 W. Bahr el Ghazal 241,920 82,066 206,902 530,888 0.7 0.4 0.8 0.6 Lakes 1,777,980 530,298 846,906 3,155,184 5.0 2.6 3.2 3.8 Western Equatoria 71,665 50,272 303,772 425,709 0.2 0.2 1.2 0.5 Central Equatoria 1,333,768 757,960 1,406,283 3,498,011 3.8 3.7 5.3 4.3 Eastern Equatoria 15,964,247 11,793,401 14,690,631 42,448,279 45.3 57.5 55.7 51.7 National total 35,275,665 20,516,107 26,371,879 82,163,651 % of FAO Upper Nile 164 227 173 180 Jonglei 580 284 287 391 Unity 155 50 69 84 Warrap 201 101 153 153 N. Bahr el Ghazal 57 38 61 51 W. Bahr el Ghazal 19 7 16 15 Lakes 136 36 69 79 Western Equatoria 11 4 26 14 Central Equatoria 152 66 111 106 Eastern Equatoria 1,797 1,041 1,433 1,394 National total 301 165 219 227 Source: Table 2.6 in Musinga et al. (2010). 62 Annex 9: Quantity of crop production by state (tons) Category Crop National Upper Jonglei Unity Warrap N. Bahr W. Bahr Lakes Western Central Eastern Nile el Ghazal el Ghazal Equatoria Equatoria Equatoria Cereals Maize 181292 84787 39194 13424 28622 4676 2216 27098 7252 22009 31314 Millet 40445 4618 692 10 649 1542 35 2330 9892 817 19861 Rice 8560 1110 2608 198 116 103 15 1064 1886 912 548 Sorghum 784391 48729 150892 27261 110558 94794 14001 82353 49353 48200 158249 Wheat 4653 1475 424 823 743 246 64 127 80 503 169 Roots and Tubers Cassava 1253367 243 2882 500 1027 0 20453 37765 692223 453829 44445 Plantain 4994 0 0 0 0 0 0 0 4072 550 373 Potatoes 4773 1344 143 192 11 121 280 31 0 1725 925 Sweet potato 10821 2377 321 37 33 116 145 1272 2231 1931 2357 Yams 334 57 0 0 43 0 0 0 0 0 233 Groundnuts and Pulses Beans/pulses 11574 1795 908 133 1570 110 482 259 944 4142 1231 Chick pea 2616 11 31 21 10 0 18 20 30 2475 0 Groundnuts 40853 393 1281 72 9325 1893 1393 11507 10663 3215 1112 Lentils 17343 4231 2861 835 2007 1054 770 268 467 1351 3500 Fruits Apples 264 77 0 6 14 8 8 60 4 118 25 Avocado 1092 0 14 0 0 0 0 0 0 948 130 Local banana 6868 2165 50 147 14 8 316 116 1745 1761 545 Dates 2377 1162 72 350 112 83 115 23 14 89 358 Mangoes 145276 2018 376 780 700 266 1645 2635 88430 39065 9362 Oranges 4131 1212 48 10 86 9 66 178 318 1451 752 Pineapples 5392 236 0 9 0 16 43 55 2396 2306 331 Papaya 9603 404 939 60 266 0 174 62 3806 3052 839 Vegetables Cabbages 7042 30 34 0 19 0 0 80 55 3404 3420 Carrots 122 3 0 0 0 7 0 0 0 102 10 Cucumber 748 167 0 23 5 1 10 7 8 82 445 Okra 25205 6637 1739 290 598 705 1061 1848 2031 5893 4403 Onions 28495 10582 222 1547 482 1290 1480 222 1761 7157 1750 Pumpkins 6299 474 1889 0 2593 653 0 516 0 113 62 Tomatoes 4883 1421 20 62 40 131 118 6 157 2161 767 Other high value crops Cocoa 5 5 0 0 0 0 0 0 0 0 0 Local coffee 6343 2048 404 244 148 379 161 730 709 1506 15 and tea Sesame 837 88 29 20 112 37 66 55 75 275 81 Sugar 7070 1685 549 102 181 0 35 24 308 2412 1775 Tobacco 1475 759 110 153 85 41 3 19 0 76 230 Source: Authors’ estimates based on NBHS (2009). Notes: * Cereal production = consumption from own products + stocks + 55% of rural purchased; other production = consumption from own products + stocks + purchased; ** Cereal flours and cassava flour are converted to corresponding grains and cassava tuber using ratios of 1:1.25 and 1:6, respectively***Grains and roots are further converted from net production to gross production using ratios of 1:1.2 and 1:2, respectively. 63 Annex 10: Cropland expansion by livelihood zones and typologies of agricultural potential areas (Scenario 1) Area Current cropland Cropland after expansion (sq km) HH HL MH ML LH LL Total HH HL MH ML LH LL Total Eastern Flood 0 0 926 712 1,776 3,525 6,940 0 0 1,684 1,494 2,047 4,105 9,330 Plains Greenbelt 3,202 1,665 0 0 0 0 4,867 7,489 8,617 0 0 0 0 16,106 Hills and 436 792 0 0 0 0 1,228 1,795 3,546 0 0 0 0 5,341 Mountains Ironstone 246 495 472 717 0 111 2,040 1,060 4,563 1,047 4,099 0 121 10,890 Plateau Nile-Sobat 85 51 501 309 836 790 2,572 216 275 812 985 1,061 883 4,232 Rivers Pastoral 67 183 4 32 0 0 286 455 680 30 158 0 0 1,323 Western Flood 18 12 4,963 1,988 1,651 245 8,877 84 36 8,277 4,482 2,222 354 15,453 Plains National 4,053 3,198 6,865 3,759 4,263 4,671 26,809 11,098 17,717 11,850 11,218 5,329 5,462 62,674 Share in Current cropland Cropland after expansion national total HH HL MH ML LH LL Total HH HL MH ML LH LL Total (%) Eastern Flood 0.0 0.0 3.5 2.7 6.6 13.1 25.9 0.0 0.0 2.7 2.4 3.3 6.5 14.9 Plains Greenbelt 11.9 6.2 0.0 0.0 0.0 0.0 18.2 11.9 13.7 0.0 0.0 0.0 0.0 25.7 Hills and 1.6 3.0 0.0 0.0 0.0 0.0 4.6 2.9 5.7 0.0 0.0 0.0 0.0 8.5 Mountains Ironstone 0.9 1.8 1.8 2.7 0.0 0.4 7.6 1.7 7.3 1.7 6.5 0.0 0.2 17.4 Plateau Nile-Sobat 0.3 0.2 1.9 1.2 3.1 2.9 9.6 0.3 0.4 1.3 1.6 1.7 1.4 6.8 Rivers Pastoral 0.2 0.7 0.0 0.1 0.0 0.0 1.1 0.7 1.1 0.0 0.3 0.0 0.0 2.1 Western Flood 0.1 0.0 18.5 7.4 6.2 0.9 33.1 0.1 0.1 13.2 7.2 3.5 0.6 24.7 Plains National 15.1 11.9 25.6 14.0 15.9 17.4 100 17.7 28.3 18.9 17.9 8.5 8.7 100 Source: Authors’ estimates based on Land Cover Database, FAO (2009). 64 Annex 11: Agricultural potential zones, areas of potential cropland expansion, and roads 65 66 Annex 12: Different types of roads across states by agricultural potential (km) State Interstate Other primary Secondary Tertiary Total Upper Nile High potential areas 0 86 172 0 258 Total 506 311 981 0 1,798 Jonglei High potential areas 46 553 663 589 1,850 Total 49 1,056 833 589 2,527 Unity High potential areas 0 73 14 0 86 Total 326 232 55 0 704 Warrap High potential areas 71 129 283 0 482 Total 215 323 559 0 1,096 Northern Bahr el Ghazal High potential areas 107 73 125 0 305 Total 130 238 567 0 935 Western Bahr el Ghazal High potential areas 155 76 139 0 370 Total 316 364 789 0 1,469 Lakes High potential areas 310 70 232 3 614 Total 369 123 357 3 853 Western Equatoria High potential areas 317 511 688 538 2,055 Total 335 533 688 538 2,095 Central Equatoria High potential areas 312 891 187 561 1,950 Total 312 891 187 561 1,950 Eastern Equatoria High potential areas 139 303 1,193 610 2,245 Total 139 312 1,268 610 2,329 Total roads network High potential areas 1,456 2,764 3,695 2,303 10,218 Total 2,696 4,475 6,285 2,303 15,759 Source: Authors’ estimates based on LandScan and WFP road maps. 67 Annex 13: Different types of roads across livelihood zones by agricultural potential (km) State Interstate Other primary Secondary Tertiary Total Eastern Flood Plains High potential areas 20 165 289 258 731 Total 411 647 1,392 258 2,708 Greenbelt High potential areas 392 616 1,384 483 2,875 Total 392 616 1,392 483 2,883 Hills and Mountains High potential areas 154 499 599 831 2,083 Total 154 509 599 831 2,093 Ironstone Plateau High potential areas 161 46 215 433 854 Total 1,027 685 1,341 433 3,485 Nile-Sobat Rivers High potential areas 7 50 94 30 182 Total 121 372 475 30 999 Pastoral High potential areas 212 642 235 254 1,343 Total 214 840 272 254 1,579 Western Flood Plains High potential areas 50 214 174 3 440 Total 289 726 645 3 1,663 Source: Authors’ estimates based on LandScan and WFP road maps. 68 Annex 14: Matrix of distances between states in South Sudan (km) State Central Northern Unity Jonglei Western Upper Lakes Eastern Warrap Western Equatoria Bahr el Bahr el Nile Equatoria Equatoria Ghazal Ghazal City Juba Aweil Bentiu Bor Kuajok Malakal Rumbek Torit Wau Yambio (Gogrial) Juba 0 637 923 203 792 2,234 415 133 643 426 Aweil 0 636 994 743 1,947 379 919 153 858 Bentiu 0 706 587 1,313 510 1,050 486 989 Bor 0 946 2,016 618 320 846 629 Kuajok (Gogrial) 0 1,494 330 869 104 809 Malakal 0 1,821 2,362 1,797 2,300 Rumbek 0 544 228 481 Torit 0 770 553 Wau 0 708 Yambio 0 Central N. Bahr Unity Jonglei W.Bahr Upper Lakes Eastern Warrap Western Equatoria el el Nile Equatoria Equatoria Ghazal Ghazal Country Border Juba Aweil Bentiu Bor Kuajok Malakal Rumbek Torit Wau Yambio City Sudan Kadugli 1,427 976 343 1,046 523 1,085 850 1,390 827 1,329 Uganda Nimule 191 991 1,123 388 943 2,433 614 283 842 625 Source: Authors’ estimates based on www.google.maps.com. 69