Report No. 49564-ET Ethiopia Diversifying the Rural Economy An Assessment of the Investment Climate for Small and Informal Enterprises October 6, 2009 Agriculture and Rural Development Unit Sustainable Development Network Africa Region Document of the World Bank CURRENCY EQUIVALENTS (Exchange Rate Effective September 29, 2009) Currency Unit = Ethiopian Birr (ETB) US$ 1.00 = 12.6 ETB FISCAL YEAR July 8 - July 7 WEIGHTS AND MEASURES Metric System ABBREVIATIONS AND ACRONYMS ADLI Agricultural Development-led Industrialization Strategy AFTAR Agriculture and Rural Development Unit, Africa Region AgSS Annual Agricultural Sample Survey CSA Central Statistical Agency CPI Consumer Price Index EA Enumeration area EAMHS Ethiopian Agricultural Marketing Household Survey EES Ethiopian Enterprise Survey ETB Ethiopian Birr GDP Gross Domestic Product HIECS Household Income, Expenditure and Consumption Survey ICA Investment Climate Assessment MFI Microfinance Institution MOFED Ministry of Finance and Economic Development MOLSA Ministry of Labor and Social Affairs MSE Micro and Small Enterprise NGO Non-Governmental Organization PASDEP Plan for Accelerated and Sustainable Development to End Poverty PSNP Productive Safety Net Program RICA Rural Investment Climate Assessment RICS Rural Investment Climate Survey SNNP Southern Nations, Nationalities and Peoples Region TFP Total Factor Productivity WRSI Water Resource Satisfaction Index WMS Welfare Monitoring Survey Vice President: Obiageli Katrya Ezekwesili Country Director: Kenichi Ohashi Sector Manager: Karen McConnell-Brooks Task Team Leaders: Josef Loening & Laketch Mikael Imru -iii- Acknowledgements We would like to thank the Ministry of Finance and Economic Development, the Ministry of Agricultural and Rural Development, the Ministry of Trade and Industry, the Ministry of Labor and Social Affairs, the National Bank of Ethiopia, and the Federal Micro and Small Enterprise Development Agency for comments on the draft report and background papers. In particular we would like to thank State Ministers H.E. Mekonnen Manyazewal, H.E. Aberre Deressa, and H.E. Tadesse Haile for detailed discussions. We are grateful to staff from these ministries and agencies, and to Getachew Adem and Samia Zekaria, for providing comments and suggestions for improvements during the course of the analysis. A team led by Josef Loening and Laketch Mikael Imru prepared this Rural Investment Climate Assessment. Mans Söderbom and Bob Rijkers (enterprise performance and rural- urban comparison), Kathleen Beegle and Gbemisola Oseni (food security and shocks), Elena Bardasi and Abay Gethahun (gender), Patricia Seex (policy options), and Isabel Günther and Markus Olapade (literature review and distributional effects) conducted analysis for the core background papers. Also Sjoerd Bakker and Anne Muir (qualitative analysis), Haddis Mulugeta (nonfarm institutions), Ram Ramaswamy (rural finance), as well as Jeffry Lecksell (maps), Tim Love, Angel Bennet, and Joanna Syroka (compilation of rainfall data), greatly supported the team. As part of a piloting effort to diversify the coverage of rural statistics in Ethiopia, the Central Statistical Agency conducted the Rural Investment Climate Survey. Yasin Mussa (managerial oversight), Habekiristos Beyene (survey team leader), and Biratu Yegezu (sampling and financial management) led the team. Esayas Muleta, Mengistu Kefale, Shemeles Mulugeta, Samuel Hailu, Zenebe Fikirie, Ayenew Legesse, and Ermyas Arega helped with design, analysis, data management and the statistical abstract. The core team was supported by Juan Muñoz (sampling), James Keough (questionnaires and pretest), Diane Steele (training of trainers and enumerators), Isabel Günther (sampling, data cleaning and analysis), Azeb Fissha (field work), Rose Mungai (tabulation plan), and in particular Kathleen Beegle (technical advice) and Gbemisola Oseni (technical documentation and statistical abstract). Financial support from Bank Netherlands Partnership Program and the Norwegian Trust Fund (rural survey), the Research Committee of the World Bank, the Swedish International Development Agency, and the Royal Netherlands Embassy in Ethiopia (background papers) is gratefully acknowledged. The team benefited from the excellent overall guidance of Kenichi Ohashi (Country Director for Ethiopia and Sudan), Karen McConnell-Brooks (Sector Manager, Agriculture and Rural Development Unit), Christine Cornelius (Program Coordinator, Agriculture and Rural Development Unit), and Jeeva Perumalpillai-Essex (Sector Leader, Sustainable Development Department). We also benefited from discussions with Derek Byerlee, Jonathan Baker, Ian Campbell, Luc Christiaensen, Catherine Dom, Achim Fock, Deepak Mishra, Stephen Mink, Catherina Ruggeri, and Paul Moreno-Lopez. Finally, we are grateful for the valuable advice of our peer-reviewers Magdi Amin, Mulat Demeke, Steve Haggblade, and Mona Sur. -iv- Table of Contents Executive Summary ............................................................................................................ 1 1. Introduction ..................................................................................................................... 9 2. Size and Basic Characteristics of the Rural Enterprise Sector ..................................... 19 3. Gender Differences in Enterprises Charactersitics ....................................................... 25 4. Enterprise Dynamics: Performance, Constraints, and Opportunities ........................... 35 5. Rural Enterprises, Food Security, and Distributional Effects ....................................... 45 6. The Untapped Potential of Rural Towns ...................................................................... 59 7. Policy Options For Promoting Rural Diversification ................................................... 65 References Annex 1: Selected Summary Tables ................................................................................. 79 Annex 2: Selected Results From Regression Analysis ..................................................... 92 Annex 3: Survey Methodology ......................................................................................... 99 Bibliography ................................................................................................................... 105 Background Documents .................................................................................................. 107 -v- Boxes, Figures, Maps, and Tables List of Boxes Box E1: Some Myths About Ethiopia's Rural Economy ..............................................................2 Box 1: Empirical Basis of this Report.......................................................................................10 Box 2: Complementarities of Rural and Urban ICAs ...............................................................12 Box 3: What is the Rural Nonfarm Sector? Definitions ...........................................................14 Box4: How Big is the Rural Nonfarm Enterprise Sector? Contribution to Income.................16 Box 5: What is new? A Guide to the Evidence on Informal Rural Enterprises in Ethiopia .....20 Box 6: Getachew, a Rural Manufacturer of Household Items Living by the Roadside............22 Box 7: Mintiwab, a Student Selling Spices in the Market ........................................................28 Box 8: Why is Female Participation in Nonfarm Employment so High in Rural Towns? .......29 Box 9: How do Enterprises in Ethiopia Compare with Tanzania? ...........................................36 Box 10: How to identify Successful Small Enterprises? Some Stylized Facts for Ethiopia.......37 Box 11: Theoretical and Empirical Framework for the Rural-Urban Comparison ....................61 Box 12: Allene, a Grain Trader from a Market Town ................................................................67 Box 13: Small Towns, Great Significance: Institutions Shaping Rural Enterprise Development in China ...................................................................................................70 Box 14: Rural Finance in Ethiopia: Limited Access and Variety of Products............................73 Box 15: Rural Enterprise Support in Ethiopia: A Crowded Landscape......................................74 List of Figures Figure E1: Relationship between Rural and Urban ICAs ..............................................................12 Figure E2: Informal Sector Urban vs. Rural Business Constraints, 2006-2007 ............................12 Figure 1: Agricultural and Overall GDP Growth, 2001/2-2008/9 ................................................15 Figure 2: Decomposition of CPI Growth, 1999-2008 ..................................................................18 Figure 3: CPI Growth, 1999-2008 ................................................................................................18 Figure 4: Localized Nature of Business, 2007 ..............................................................................22 Figure 5: Nonfarm Seasonality, 2007 ...........................................................................................23 Figure 6: Probabilities of Being Engaged in Nonfarm Employment, 2007 ..................................29 Figure 7: Distribution of Start-up Capital, 2007 ...........................................................................31 Figure 8: Distribution of Enterprise Revenue, 2007 .....................................................................32 Figure 9: Rural Business Constraints in Ethiopia and Tanzania, 2005-2007 ...............................36 Figure 10: Estimated Nonfarm Enterprise Participation Rates, 1998-2007 ..................................38 Figure 11: Enterprise Sales Growth, 2006-2007............................................................................43 Figure 12: Changes in Number of Employees since Start-up, 2007 ..............................................44 Figure 13: Changes in Labor Days and Hired Workers, Since Start-up, 2007 ..............................44 Figure 14: Enterprises Opened in the Last 3 Years Now Closed by Food Security Status of Wereda, 2007.................................................................................................................46 Figure 15: Households with Enterprise by Food Security Status of Wereda, 2007 ......................46 Figure 16: Amahara-Share of Enterprise Income, Among Household with Enterprises, by Food Security Status of Wereda, 2007 ...................................................................................53 Figure 17: Amahara-Income Distribution by Household Category, 2007 .....................................53 Figure 18: Size Distributions, 2007 ...............................................................................................59 Figure 19: Distributions of Capital Intensity, 2007 .......................................................................60 Figure 20: Distributions of Value Added, 2007.............................................................................62 -vi- List of Maps Map 1: Coverage of the Rural Investment Climate Survey, 2007 ...............................................10 Map 2: Nonfarm Enterprise Participation Rates by Geographical Zone, 2007 ...........................24 Map 3: Market Demand as 1st Major Business Constraint by Geographical Zone, 2007 ...........40 Map 4: Rural Finance as 2nd Major Business Constraint by Geographical Zone, 2007 ..............41 Map 5: Transport as 3rd Major Business Constraint by Geographical Zone, 2007 .....................42 List of Tables Table E1: Ethiopia's Policy Options to Promote Rural Entrepreneurship ......................................7 Table 1: Average Share of Rural Household Income by Source, 2000 and 2008 .......................16 Table 2: Rural Enterprise Participation Rates and Contribution to Income, 2007 .....................19 Table 3: Composition of Rural Enterprise Sector, 2007 .............................................................21 Table 4: Selected Enterprise Characteristics, 2007 .....................................................................21 Table 5: Distribution of Nonfarm Enterprises by Sector, Region, and Sex, 2007 .......................26 Table 6: Median Enterprises Sales by Sector and Sex of Owner, 2007 ......................................27 Table 7: Reason for Enterprise Start-up by Sex of Head, 2007 ...................................................27 Table 8: Importance of Nonfarm vs. Agriculture Employment by Sex, 2007 ............................28 Table 9: Sector Distribution of Nonfarm Businesses, 2007 ........................................................30 Table 10: Employment Characteristics of Nonfarm Enterprises, 2007 .........................................32 Table 11: Estimated Entry and Exit Rates, 2006-2007 ..................................................................39 Table 12: Major Business Constraints in Rural Areas, 2007 .........................................................39 Table 13: Business Constraints by Food Security Status of Wereda, 2007 ..................................47 Table 14: Source of Start-up Capital by Food Security Status of Wereda, 2007 .........................48 Table 15: Reason for Enterprise Start-up by Food Security Status of Wereda, 2007 ...................48 Table 16: Sector of Enterprise Start-up by Food Security Status of Wereda, 2007 ......................49 Table 17: Estimated Agricultural Income Change in the Last 3 Years, 2007 ..............................50 Table 18: Amhara-Characteristics of Households with and Without Enterprises, 2007 ...............51 Table 19: Amhara-Income Sources by Food Secure Status of Wereda, 2007 ..............................52 Table 20: Participation of Households in Income-generating Activities by Expenditure Quintile, 1998 ...............................................................................................................55 Table 21: Households Income by Source and Expenditure Quintile, 1998 ...................................56 Table 22: Gini-Decomposition by Income Source, 1998 ..............................................................57 Table 23: Transition Matrix: Urban Manufacturing Firms, 2006 ..................................................63 Table 24: Transition Matrix: Rural Manufacturing Firms, 2007 ...................................................63 -vii- Annex Tables Annex 1 Table 25: Participation Rates, Industry Type, and Mean Age of Enterprises, 2007......................79 Table 26: Percentage Distribution of Enterprises by Constraints that Prevent Operations and Growth, 2007 ..........................................................................................................80 Table 27: Percentage Distribution of Households by Constraints that Prevent Opening a Nonfarm Business, 2007.............................................................................................81 Table 28: Percentage Distribution of Enterprises by Main Reason for Starting an Enterprise, 2007 ...............................................................................................................................82 Table 29: Percentage Distribution of Enterprises by Main Source of Start-up Capital, 2007 .......83 Table 30: Percentage Distribution of Enterprises Closure, 2007 ...................................................84 Table 31: Enterprises by Number of Employees, Sales Growth, and Share of Profits in Household Income, 2007 ...............................................................................................85 Table 32: Average Distance to Agriculture Input and Output Markets and All-weather Roads, 2007 ...............................................................................................................................86 Table 33: Amhara-Number and percentage distribution of Socio-economic Characteristics of Enterprise Owners .........................................................................................................87 Table 34: Amhara-Percentage Distribution of Enterprises by Start-up Capital Category .............88 Table 35: Amhara-Number and Percentage Distribution of Households by Source of 100 Birr in Case of Emergency, All Households.........................................................................89 Table 36: Amhara-Number and Percentage Distribution of Households by Type of Shock during the Last 12 months .............................................................................................90 Table 37: Amhara-Number and Percentage of Household that Suffered from Food Shortages during the Last 12 months .............................................................................................91 Annex 2 Table 38: Probability of Rural Nonfarm Enterprise Ownership, 2007 ..........................................92 Table 39: Probability of Rural Nonfarm Enterprise Closures, 2007 .............................................93 Table 40: Determinants of Enterprise Profits, 2007 ......................................................................94 Table 41: Amhara-Enterprise Cobb-Douglas Production Function, 2007.....................................95 Table 42: Rural Urban Comparison Production Functions, OLS Regressions on separate samples ..........................................................................................................................97 Table 43: Rural Urban Comparison Production Functions, OLS Regressions on pooled small ...... manufacturing firms' sample.........................................................................................98 Annex 3 Table 44: Amhara-Household Characteristics for the Four Specific Zones in Rural Amhara ....104 -viii- EXECUTIVE SUMMARY A. OVERVIEW 1. Ethiopia's rural nonfarm sector is significant and participation is increasing. The sector is particularly important for women and poorer households. Nonfarm enterprises provide income-earning opportunities to those lacking alternative options and supplementary income for farming households. The returns to running a nonfarm firm are low, but there is tremendous heterogeneity in enterprise performance. Agriculture and the nonfarm sector are mutually reinforcing through market synergies. Markets are small, fragmented, and localized. Strengthening and developing small towns appears to be a promising area in support of rural development. As the Ethiopian economy develops the nonfarm sector will grow and become increasingly important as an alternative employer of labor and source of livelihood in rural areas. This suggests the policy priority should not be "either agriculture or the nonfarm sector" but a balanced approach. B. KEY FINDINGS 2. Enterprise activity is more prevalent in rural towns and is especially important for women. Nonfarm enterprise activity is highest in rural towns and lowest in remote rural areas. Proximity to markets and roads is also a strong predictor of participation. 3. The nonfarm enterprise sector makes an important contribution to rural income in Ethiopia. Approximately 25 percent of all households in rural Ethiopia own one or more nonfarm enterprises. Participation rates are rising. Despite high participation rates, very few households rely exclusively on nonfarm enterprise activity. Though it is difficult to measure enterprise profits and household income precisely, the 2007 Rural Investment Climate Survey (RICS) suggest that nonfarm enterprise profits account for approximately 40 percent of total household income for those households that run a nonfarm firm. Comparison of the three most recent Welfare Monitoring Surveys (WMS) suggests that participation in the sector is growing, though most of the existing firms do not expand their workforce. 4. Nonfarm enterprises provide self-employment opportunities, yet virtually no wage labor opportunities. Almost all nonfarm firms are small and own very little capital; the median capital stock is roughly 194 Birr (approximately US$ 16). The overwhelming majority of enterprises are one-person enterprises and less than 1 percent of all enterprises employ more than three workers. The most prominent nonfarm enterprise activities are trading and wholesale, closely followed by manufacturing and services. While the miniscule scale at which enterprises operate is striking, enterprises do not seem to operate at a sub-optimal scale. -1- Box E1: Some Myths about Ethiopia's Rural Economy There are a number of widely held but mistaken views, or myths, about Ethiopia's rural economy, which can be attributed partly to a shortage of information. Findings from the Rural ICA help to shed light on certain aspects of Ethiopia's rural nonfarm economy: Myth 1 ­ In rural Ethiopia all households engage in agriculture: there are no enterprises. Some 25 percent of rural household participate in some form of nonfarm enterprise activity. For about 8 percent of rural households, their enterprise is the dominant source of income. Myth 2 ­ Nonfarm enterprises are economically unimportant in rural Ethiopia. While enterprise activity if often concentrated in the low return sector, it is nevertheless an important source of income, particularly for women and food insecure households. Enterprise households in Ethiopia generate on average 42 percent of their income from nonfarm activities. Myth 3 ­ Manufacturing and grain-milling activities dominate the nonfarm sector. The dominant sector is trade, engaging more than 50 percent of rural enterprise households. Myth 4 ­ It is more important to support agriculture than nonfarm enterprises. Agriculture and the nonfarm sector are mutually reinforcing, because rural nonfarm enterprises are an essential part of agricultural input and output markets, and agricultural service delivery in general. Myth 5 ­ Governance and land policy are the main constraints for rural enterprises. Constraints to rural enterprise in Ethiopia are spatially quite heterogeneous. On average, however, enterprises appear to be much more constrained from the demand side than the supply side and the most important supply-side constraints are access to financial services suitable for rural business and high transport costs due to remoteness. Myth 6 ­ Support to the nonfarm sector is futile from a policy perspective. An important finding of this study is that the investment climate in rural towns can support comparable productivity performance to those of urban informal microenterprises. This suggests that supporting nonfarm enterprises in small rural towns can yield high returns ­ and mutually benefit the agricultural and other sectors through production, consumption, and labor market linkages. 5. Average firm performance is rather stagnant. Even though the returns to capital are high at the margin, very few firms invest or expand their workforce. No more than 8 percent of all firms have increased the number of employees and only 30 percent have increased the total number of labor days used per annum since start-up. A mere 20 percent of firms have re-invested since they started. The lack of investment is due to the high-risk environment that entrepreneurs face, the high cost of and limited access to capital in rural areas, and diminishing returns to capital. The likelihood of investing falls as uncertainty (proxied by the variability in agricultural performance induced by rainfall volatility) increases. Investment is also negatively correlated with the household's ability to access emergency finance, suggesting that households with better insurance or access to credit are more likely to invest. 6. Markets are small and localized. For example, more than 90 percent of entrepreneurs walk to the market and very few firms sell to customers outside their own community. Because of high transport and transaction costs, most firms are local monopolists and even if they are not, they have substantial market power, further limiting their incentives to invest. Market fragmentation seems to be the most important constraint hampering the performance of nonfarm enterprises. This is borne out by the impressions of firm managers, who consider a lack of demand, transport, and inadequate access to credit their most important problems. Market fragmentation limits demand and helps -2- explain the heterogeneity in the returns to capital and labor, as well as why firms do not invest and expand. 7. Enterprise activity is worthwhile when other opportunities are lacking. The returns to running a nonfarm firm are very low. On average about Birr 5.6 per day (less than US$ 0.5) and even lower for enterprises managed by women. These marginal returns are much lower than the agricultural wage rate for casual workers. Enterprise activity is highly countercyclical with agriculture, which suggests that nonfarm enterprise activities are most appealing when the opportunity cost of labor is low. 8. While many enterprises are not very profitable, there is tremendous heterogeneity in enterprise performance both across and within locations, which is indicative of rural market fragmentation. For example, enterprises located in rural towns are almost twice as profitable as enterprises located in very remote rural areas. Enterprises engaging in trading or wholesale activities are the most profitable, perhaps reflecting the existence of arbitrage opportunities. Those engaging in manufacturing activities are the least productive. Enterprise productivity is about 50 percent higher for firms with a male manager. Even after controlling for activity choice, capital intensity, and other differences between enterprises managed by men and women remain. 9. Enterprise sales are also strongly correlated with the agricultural performance of local and adjacent communities. The reason appears to be that demand for nonfarm products is much higher when agricultural performance is strong. In addition, uncertainty regarding agricultural performance limits incentives to invest, at least in the short run. Moreover, income from agricultural activities is the most important source of start-up capital for the overwhelming majority of entrepreneurs. Very few enterprises invest and start-up capital, determined by access to finance, is a strong determinant of future profitability. 10. Women play a very important role in Ethiopia's nonfarm enterprise sector. Women are more likely to be engaged in nonfarm activities than men, especially in small towns. Women tend to take-up nonfarm activities because they face constraints in other domains, especially agriculture, and not necessarily because they are well positioned to exploit profitable market opportunities. By contrast, men are able to exploit complementarities between nonfarm activities and agriculture. Activities in which women engage in are often limited, and typically concentrated in low-profitability sectors requiring little training and skills. High female participation despite substantially lower returns attests to the underprivileged position of women in the Ethiopian labor market. 11. The nonfarm economy can be an important source of additional income for food insecure households. In a setting with limited agricultural potential or highly variable weather, income from nonfarm activities can augment and smooth income flows for rural households. At first sight it appears that a substantial number of nonfarm activities in Ethiopia only provide limited opportunities. But they could be very important from a food security point of view. This is especially relevant to Ethiopia where an estimated 4.6 million people periodically require emergency food assistance and as many -3- as 7.3 million chronically food insecure people receive a cash or food transfer through a Productive Safety Net Program (PSNP). 12. Because of its complementarity with agriculture, nonfarm enterprise activity does not significantly reduce the supply of labor to agricultural activities. Nonfarm enterprise activity is much lower during the peak agricultural season, reflecting household labor allocation decisions to prioritize agriculture. Conversely, the nonfarm enterprise sector does not seem to suffer a labor shortage. The low marginal productivity of labor in combination with the fact that very few enterprises hire workers suggests that supplying more labor to (existing) nonfarm enterprises might simply not be worthwhile at the margin. If anything, the nonfarm sector absorbs labor that cannot be gainfully employed elsewhere, rather than "pulling" people away from agricultural activities. 13. Rural market integration would enhance enterprise productivity and stimulate firm growth. The study compares the performance of rural nonfarm manufacturing enterprises with manufacturing enterprises in rural towns and in major urban centers, where markets are better integrated. Though urban firms are much larger than rural firms on average, the focus is on comparing microenterprises since these constitute the most appropriate comparison group. Also, an exclusive focus on manufacturing enterprises minimizes the differences due to sectoral affiliation, though urban enterprises produce a much broader range of products than rural ones. 14. The urban investment climate for small informal enterprises differs from the rural one. Urban enterprises are more capital intensive, have a better-educated workforce and are less reliant on household labor. In addition, they do not exhibit seasonality. Urban enterprises typically have much better access to credit, urban infrastructure is far superior to that in rural areas, and competitive pressure is higher in urban centers. While these differences affect factor intensity and business size, the technologies used by urban and rural informal enterprises are often similar. Interestingly, although total factor productivity (TFP) of nonfarm enterprises located in rural areas is much lower on average than enterprises in urban areas, TFP of nonfarm enterprises in rural towns is on a par with micro enterprises located in urban centers. C. POLICY IMPLICATIONS 15. Where possible, policymakers should capitalize on the complementarities between agriculture and the nonfarm enterprise sector. It is likely that policy reforms that benefit nonfarm enterprises also benefit the agricultural sector and vice versa. Better access to credit, upgraded transport facilities, and improved insurance, for example, would benefit farmers and entrepreneurs alike. Moreover, enhanced agricultural performance is likely to stimulate the performance of nonfarm enterprises, while improved off-farm performance might stimulate agricultural growth, by acting as a "pull" factor. 16. Promoting market integration through the formation of small market towns is a particularly promising policy option. Market integration can be enhanced through improvement of transport and information systems, increasing competition, and the -4- removal of market failures in credit markets. Since the returns to market integration seem to be highest at the lowest levels of market integration, promoting rural market towns might be a good way to enhance the productivity of the nonfarm sector. But the overall slow dynamic performance of rural nonfarm enterprises suggests that rural towns themselves might need to be better integrated into the economy to foster sustained growth. More generally, the results from this assessment show that improving agricultural performance is essential to stimulate rural growth. Increased agricultural productivity would not only benefit the vast majority of rural households by boosting their incomes, but also benefit the nonfarm enterprise sector by raising the demand for nonfarm goods and encouraging factor accumulation.1 17. Limited access to finance is a crosscutting constraint in the effort to build rural livelihoods and to support smallholder farming. It is, however, currently addressed in an uneven manner. Despite recent growth in services and market penetration, banks, micro-finance institutions (MFI) and multipurpose cooperatives cover less than the total demand. One approach to help address this gap is to build grassroots institutions to expand outreach of financial services to rural areas. In addition to micro- finance institutions, rural savings and credit cooperatives could be promoted in areas where they do not currently operate. 18. Promotion and selective support to groups of businesses with market potential seems to be a promising area of intervention. Support would include supply chain reviews and problem solving on an activity by activity basis. Support would probably need to focus on activities with market potential outside the immediate area and promotional efforts focus on matching local resources to external, even international, consumers. Such selective support may be politically or technically difficult for the government agencies to provide, and is therefore seen more as an opportunity for Non- Governmental Organizations (NGOs). 19. Skills and education are positively associated with enterprise start-up and participation, but formal education remains at a very low level in rural Ethiopia. Service delivery for both skills development and introduction of new technology is likely to remain in the public domain for the near future. Internationally, there are significant successes in public provision of services related to rural nonfarm enterprise, especially in the area of technology development and dissemination. However, less successful examples also abound. On balance, experience suggests that such efforts must (a) focus on key widely produced products/services; (b) link with local input suppliers to ensure sustained and affordable access to the necessary inputs; and (c) provide short-term assistance in facilitating the transition of small firms to new technologies and possibly also to new marketing channels. 20. Investment climate and enterprise development policies should be mindful of the different needs and constraints experienced by women entrepreneurs. However, if targeted appropriately, some of the highlighted program areas--access to finance, supply chain reviews, and skills development--appear to be particularly relevant. 1 Dercon and Hodinott (2005) argue that small towns are key to improve welfare of rural Ethiopians. -5- Targeting female entrepreneurs would be in particular of interest at the project level, considering government or donor supported investments that aim to enhance rural entrepreneurship. 21. Policies seeking to address food insecurity in rural Ethiopia should consider the potential contribution of the rural nonfarm enterprise sector. Even low-return nonfarm activities may prove to be important from a welfare point of view, although not necessarily a substitute for higher-return activities such as wage labor. In food insecure rural areas, the nonfarm sector could potentially play a very important role in ensuring rural livelihoods. D. SUMMARY 22. The rural nonfarm sector provides income-earnings opportunities to those lacking alternative options and in the low seasons for farming. It is sizable and significant. Nonfarm enterprise activities are particularly important for women and food insecure households. While nonfarm enterprises make an important contribution to rural income and employment, running a nonfarm enterprise in Ethiopia is predominantly a means to complement agricultural income, rather than an alternative pathway out of poverty. On average, the returns to enterprise are low and few firms increase their workforce or capital stock. 23. The main constraints appear to operate on the demand side. Supply-side constraints also exist, notably in finance and infrastructure, but are geographically diverse. Markets are small and highly localized. Strong local agricultural performance affects nonfarm enterprise performance through increased demand. Within this context, the main supply-side investment climate constraints--access to finance and transportation--appear to "bite" less than in other countries. This suggests that a two- pronged approach is appropriate. This should include agriculture development and market town development in addition to selective, geographically targeted, investment climate interventions that address the major supply-side constraints.2 24. The question of how to achieve rural income diversification is likely to become increasingly important in Ethiopia over the coming years. As the Ethiopian economy develops, with higher productivity and better performance in agriculture, the nonfarm sector will also grow and become increasingly important as an alternative employer of labor and source of livelihood in rural areas. This suggests the policy priority should not be "either agriculture or the nonfarm sector" but a balanced approach focusing on the spillovers between the sectors, particularly production, consumption and labor market linkages. This will include ensuring that rural nonfarm enterprises are not constrained by the rural investment climate in responding to new opportunities. 2 Using a simulation model Diao et al. (2007) finds Ethiopia's exclusive focus on agriculture--or insufficient attention to non-agriculture--may be counterproductive. While consumption linkages are much stronger than production linkages, a combination of agricultural growth combined with nonagricultural growth would be most beneficial to reduce rural poverty. -6- Table E1: Ethiopia's Policy Options to Promote Rural Entrepreneurship: Short to Medium-Term ISSUE SHORT-TERM MEDIUM-TERM Overall strategic approach on demand side Capitalize on the complementarities with Continued emphasis on agricultural development Policies to promote rural entrepreneurship need to agriculture as a major pre-requisite for interventions in support take into account the inter-relationships with of the rural nonfarm sector agriculture and heterogeneity of the rural nonfarm sector Interventions should aim to maximize spillover from related support (for example extension) Rural Market Town Development Small enterprises in town exhibit significant Stakeholder consultation and consensus on a Prioritization exercise for investment in transport productivity potential with beneficial linkages regional pilot program to stimulate small market infrastructure and other public goods in small to the agricultural output and input markets town development, private enterprise growth, and market towns based on spatial economic analysis rural-urban linkages and any local economic and business development strategies Some basic spatial master planning to prioritize and manage investment in infrastructure within rural towns Improving Access to Rural Finance Access to finance suitable for nonfarm Review current efforts to improve access to credit Invest in grassroots financial institutions and enterprises unavailable in rural areas focusing on the need to increase supply chains relevant to the rural nonfarm coverage and to promote more flexible product enterprises lines Development and piloting of financial instruments Feasibility analysis for market potential of urban feasible for small entrepreneurs other than group and semi-urban/rural mobile-banking taking into lending consideration infrastructure and regulatory constraints Pilot for mobile-banking schemes in urban and semi-rural areas -7- ISSUE SHORT-TERM MEDIUM-TERM Providing support to entrepreneurs Institutional support is uneven Review of strengths and weaknesses and Establish a monitoring team to supervise agreed measures implemented by line ministries and implementation arrangements by line ministries regional governments and regional governments Consider extending the scope of extension Consider developing local economic and business services to include nonfarm enterprise development strategies Support is having limited impact Review of experiences by NGOs and public service Take successful experiences in delivery of delivery systems including cost-benefit analysis of services (skills development and advisory interventions services, technology dissemination) to scale as appropriate Identification of groups of businesses with market potential and collective constraints General market development efforts through the identification and delivery of a limited number of key missing ingredients along supply chains most relevant to the rural nonfarm economy Considering gender is important Ensure women are targeted appropriately at the project level Monitoring of trends Development of nationally representative database Rural Socio-Economic Survey on rural and semi-rural income diversification patterns with ability to monitor trends and programs Rural/urban classification allowing disaggregation by settlement size and identification of rural Refinement of rural/urban classification in multiple market towns surveys conducted by the Central Statistical Agency (CSA) Incorporation of rural income diversification, entrepreneurship, and private sector development Consultation with stakeholders on knowledge gaps issues into national research programs Addressing food insecurity Even low-return nonfarm activities may be Consider the potential contribution of the rural Address why participation is currently lower in important from a welfare point of view nonfarm enterprise sector to food security insecure areas, particularly among women Study the interaction and contribution of labor- Ensure access to external markets not vulnerable based safety nets and engagement in nonfarm on local agricultural performance enterprises -8- 1. INTRODUCTION A. OVERVIEW 25. Understanding the opportunities and constraints in Ethiopia's rural nonfarm enterprise sector is of crucial importance. The economy remains highly dependent and vulnerable on the performance of the agricultural sector. Ongoing population growth and land degradation increases the need for income diversification strategies. The Plan for Accelerated and Sustainable Development to End Poverty (PASDEP) considers the promotion of nonfarm enterprise activity as an additional catalyst for rural development, though in practice promoting nonfarm activities has had a limited role, partly because of the little knowledge of the sector in Ethiopia, where it is often believed that rural equals agriculture. 26. The rural and agricultural strategy incorporated within the PASDEP acknowledges that both agricultural and nonfarm income generating possibilities should be emphasized especially in drought prone areas. Although the strategy introduces important new approaches to enhance rural economic growth, very little is known about the basic characteristics, the constraints, and the performance of the rural nonfarm enterprise sector in Ethiopia. This report is an attempt to fill some of these gaps. Given the previous lack of information on the nonfarm economy, the assessment contributes to a better understanding of the rural nonfarm economy in Ethiopia. It may therefore be an input for the next phase of PASDEP, which needs to be assessed and renewed in 2010. 27. What does the rural investment climate assessment measure? The following chapter argues that assessing the rural investment climate measures the "economic environment" of the poor. By assessing supply-side and demand-side constraints of the local economy, one can identify critical areas of reform and prioritize public investments. Change in rural incomes and diversification is largely determined by the performance of the rural economy. Private entrepreneurs in these areas are of particular importance because they create beneficial links between the nonfarm economy and agriculture. In this context, rural nonfarm enterprises contribute to alleviating rural poverty, and may be of growing significance. 28. This report is organized into seven chapters. The first chapter lays the analytical groundwork for assessing the rural investment climate in Ethiopia and establishes a broader context for the empirical findings. The second chapter analyzes size and basic enterprise characteristics. The third chapter sheds light on the role of women in rural entrepreneurship. The fourth chapter analyzes enterprise dynamics: start-up, closure, and growth. The fifth chapter is dedicated to the welfare effects of rural enterprises, in particular their impact on food security and distributional effects. The sixth chapter compares rural and urban informal enterprise performance and considers the role of small market towns. The final chapter summarizes the findings and offers reflections for policy. -9- Box 1: Empirical Basis of this Report The empirical basis for this report is a Rural Investment Climate Survey (RICS). Ethiopia's Central Statistical Agency (CSA) fielded the survey during December 2006 and January 2007. The household-based survey consisted of two complementary efforts: 1. The RICS-AgSS was carried out in conjunction with Ethiopia's Annual Agricultural Sample Survey (AgSS). Covering about 14,000 households and 3,500 enterprises, it attaches a three-page nonfarm enterprise module to the existing agricultural survey. It is minimalistic in terms of the collected level of detail for enterprise and households, but still sufficient to draw analytical conclusions. It fully covers all four major regions of Ethiopia: Tigray; Oromia; Southern Nations, Nationalities and Peoples (SNNP) and Amhara. The RICS-AgSS is thus representative for these 4 regions or about 90 percent of Ethiopia's population of 77 million. A limitation of the survey is that it does not cover information for the remaining regions, in particular the pastoral areas. 2. The RICS-Amhara was carried out as a complementary exercise, following models implemented by Tanzania, Sri Lanka, Nicaragua, Indonesia or Benin. Covering about 2,900 households, 760 enterprises and 180 communities, it captures very detailed information for about one-half of Amhara's population of 18 million. It covers both food secure and food insecure areas. Data can be matched with RICS-AgSS. It also covers the nonfarm wage sector in small rural market towns. The RICS-Amhara is considered as pilot exercise and is not representative at the national level. How reliable is the data? A technical manual prepared by CSA (2008b) in cooperation with the World Bank team documents methodologies and procedures. The manual also assesses the quality of RICS-Amhara. Household assets and basic demographic characteristics are compared with the Welfare Monitoring Surveys for 2000 and 2004. Such a comparison reveals a very close fit for selected indicators. After completing the interviews, based on their comparative experience, the enumerators where also asked to assess the quality of subjective constraints and sales levels reported by entrepreneurs. For some 95 percent of the sample, the enumerators believed that the answers are realistic. Qualitative field evidence complements the survey data throughout the report. Looking beyond pure survey data, background studies from Bakker (2007) as well as Muir and others (2007) provide detailed insights into sectoral constraints and rural livelihoods decisions in rural Ethiopia, thus verifying often perception-based constraints. In addition, drawing from a wide range of sources, Günther and Olapade (2007) comprehensively review the earlier evidence in the nonfarm sector for Ethiopia. Map 1: Coverage of the Ethiopia Rural Investment Climate Survey, 2007 -10- B. WHAT IS THE RURAL INVESTMENT CLIMATE? Assessing the economic environment of the poor 29. A country's "investment climate" is its environment for private sector activity. The quality of the investment climate is determined by the risks and transaction costs of investing in and operating a business, which in turn are primarily determined by the legal and regulatory framework, barriers to entry and exit, and conditions in markets for labor, finance, information, infrastructure services, and other productive inputs. Governments influence the quality of their country's investment climates through policies, institutions, and their relationship with the private sector. The quality of the investment climate is linked to poverty reduction by the impact of better investment climates on private sector activity, and thus on economic growth and employment. 30. Investment climate refers to the opportunities and incentives for firms to invest productively, create jobs, and expand (World Bank, 2004a). Among others, the investment climate includes factors that are incentives or disincentives for starting and running a business, including financial services, infrastructure, governance, regulations, taxes, labor, and conflict resolution. The investment climate is recognized as important to improve output, employment, and enterprise productivity (Dollar and others, 2005), all of which hold the potential to stimulate employment growth and reduce poverty. Micro- entrepreneurs in rural areas create jobs needed to increase income. They provide goods and services and often pay taxes needed to (partly) fund local investments, but the size of their contribution largely depends on the environment in which private business can operate. Both risks and barriers can undermine rural entrepreneurship, hence, it is important to understand the conditions necessary to develop rural nonfarm enterprises. 31. The Ethiopia Rural Investment Climate Assessment (RICA) is among the first to take a comprehensive look at the--overwhelmingly informal--business environment in rural areas.3 The majority of Investment Climate Assessments (ICA) has not considered the heterogeneity of the investment climate across different areas and sectors. The standard approach is heavily biased toward registered (bigger) enterprises in the manufacturing sector, which are typically located in urban areas. Rural areas have lower population densities, making infrastructure and many services costly to maintain. Transaction costs are high, there are relatively more market failures, and the rural economy has distinct seasonality and employment patterns. Most important is that the rural population typically works on farms or in micro-enterprises. In Ethiopia, where some 85 percent of employment is in the rural areas, it is thus essential to conduct comparable analyses in rural areas. 3 As part of a larger World Bank initiative, these piloting RICAs cover Sri Lanka, Nicaragua, Tanzania, Indonesia, Benin, and Ethiopia. Two related studies were also carried out in Bangladesh and Pakistan. An urban-focused ICA for Ethiopia was conducted by the World Bank in 2008. -11- Box 2: Complementarity of Rural and Urban ICAs In companion to the Rural ICA the World Bank also conducted a standard urban-focused ICA. Both assessments have different purposes. The Urban ICA looks at the business environment in 14 major urban centers, focusing on formal and bigger enterprises in the manufacturing and service sectors. The Ethiopian Development Research Institute conducted a survey of 610 enterprises in 2006. It also covers the urban informal sector in these cities, though with a small sample of about 120 firms. Compensating for a bias towards registered or bigger enterprises, the Rural ICA considers the trading sector and takes into consideration the heterogeneity of the investment climate across geographic areas. Data is from a large rural survey covering more than 14,000 households executed by CSA in four major regions. It considers welfare and food security effects. The Urban and Rural ICAs are largely complementary, as evidenced in the figure below. The Rural ICA finds that the rural Figure E1: Relationship between Rural and Urban ICAs enterprise sector is sizeable and economically significant. This is contrary to the common belief that there is no Formal diversification beyond agriculture in rural Urban ICA Ethiopia. Though agriculture is the dominant source of income and nonfarm activities are mostly low return, about 25 percent of rural households are engaged Rural ICA in some sort of entrepreneurial activity, For about 8 percent of rural households, Informal nonfarm enterprises income is relatively more important than agriculture. On average, nonfarm enterprise profits account for 42 percent of total income Rural Urban among households owners that run an enterprise. Moreover, households with nonfarm enterprises are more likely to be food secure. The sector is particularly important for women. Most enterprises engage in trading agricultural Figure E2: Ethiopia ­ Informal Sector Urban vs. Rural commodities. Business Constraints, 2006-2007 The Urban ICA finds that the investment 40 Percent of firms reporting major constraints climate in Ethiopia has improved over the 35 past 5 years (World Bank, 2009). Nonetheless, productivity levels remain 30 to operations and growth low when compared with peer groups, and Ethiopian products remain uncompetitive 25 in international markets. The Urban ICA 20 suggests that productivity is held back by a number of structural and economic factors 15 that combine to make the economy less 10 flexible and responsive. Beyond efficiency at the firm level, enterprises in urban 5 Ethiopia appear to be inefficient in its allocation of resources across firms. The 0 ec nt W t n c h io n gy m nd r es n financial sector, land policies, industrial Fi ts ty Tr n ce r or ty bo ov a te io t io E l me fe ke ci lo sp Re Tax Co La Te cat tit La Sa ra na tri no ar pe an n policy, patterns of inter-firm contracting, st i er M un gi m and the state of market institutions G m r) co ai le contribute to a lack of flexibility. These nf Te (U factors appear to limit competition in such Urban Informal Rural Remote Rural Town a way that the most productive urban enterprises cannot systematically increase Source: 2006/7 RICS-Amhara; 2006/07 PICS for 14 major cities. their market shares. Numbers are illustrative and are not nationally representative. The informal sectors are the fastest growing segment of the private sector, due to the flows of labor from rural to urban Ethiopia and the absence of alternative ways to generate incomes. A naïve look at perceived business constraints in the informal urban sector in 14 major cities and several rural market towns with up to 10,000 habitants reveals illustrative findings. In major urban areas of Ethiopia, informal entrepreneurs feel constrained about access to finance, land and taxes. In rural areas and small towns informal firms lack market demand, and feel constrained by access to finance and a variety of infrastructure services. -12- Understanding constraints of rural enterprises 32. Both supply and demand constraints affect rural nonfarm enterprises. In Ethiopia, demand constraints for rural enterprises are mainly related to agriculture. Profits from agricultural production, income earned from nonfarm enterprises, and demand generated outside the rural economy can all contribute to effective demand for the goods and services produced by rural entrepreneurs. Which of these sources of demand is the most important depends on the local environment and the degree of development in which the enterprise operates. A virtuous cycle of development can arise through the interaction of farm and nonfarm activities (Evans, 1992). Agricultural and nonfarm activities are linked in several ways--through consumption (demand for final products), production (backward and forward supply of inputs among businesses), finances (remittance and savings channeled through urban institutions), and labor market links. 33. On the supply side, a wide variety of factors determine the ability of rural enterprises to produce goods and services. Supply constraints also affect the cost of goods and services that may include the state of local infrastructure, ability to access finance and the cost of doing so, cost and quality of labor, quality of the local regulatory environment, and extent of competition, knowledge of market opportunities, and stability and security in the area. If enterprises use old and highly labor-intensive technologies to deliver goods and services, unit costs can be high and productivity low. Under such circumstances, it is only profitable for enterprises to serve a local clientele because of high transaction costs. 34. What is the role of the investment climate in this context? First, private entrepreneurs are needed in the creation of the beneficial links between the nonfarm economy and local farmers, for example, through agricultural input and output markets. However, unjustified risks, transaction costs, or other barriers to business operations can undermine rural entrepreneurship. Second, the investment climate not only affects rural nonfarm entrepreneurs but also farm activities. For example, poor access to rural finance and infrastructure hits both farm and nonfarm activities. This RICA may therefore be useful in a broader context. Assessing the economic opportunities and constraints of rural firms sheds light on the general factors pertinent to poverty and rural development. This assessment can help to prioritize rural investments. -13- Box 3: What is The Rural Nonfarm Sector? Definitions Nonfarm activities include all economic activities in rural areas except agriculture, livestock, fishing and B. hunting. Processing of farm products and then selling them is defined as nonfarm activity. The nonfarm sector thus includes all secondary and tertiary activities independent of their scale and technological sophistication. Nonfarm activities can be full or part-time, formal or informal, and of seasonal or periodic character. Nonfarm activities can take place at home, a specific business location, or be performed by itinerant traders. Unpaid production of goods and services for home consumption and unproductive economic activities such as begging and gambling are excluded. A technical manual explains these and other definitions in more detail (CSA 2007b). Due to extensive training of the enumerators and through the provisions of an indicative example list for different type of nonfarm activities, no major difficulties were encountered in correctly identifying and distinguishing nonfarm activities during the fieldwork of the RICS. Nonfarm sector Agricultural sector Self-employment Self-employment farms Self-employment (small enterprises) (small farms) RICS-AgSS Self-employment enterprises RICS-Amhara Wage employment Wage employment (hired nonfarm labor) (hired farm labor) RICS-Amhara Wage-employment enterprises "Off-farm" employment (outside own farm) Wage-employment An important distinction is between self- and wage-employment. Nonfarm activities include both self- employment (firms) and wage-employment (hired labor). The main focus of the RICS-AgSS is on nonfarm enterprises, which excludes wage employment in the nonfarm sector. The RICS-Amhara compensates for this information gap. Information on nonfarm wage employment is collected but tends to be very small. An important consideration is that households typically earn income from multiple sources. Therefore, households relying predominantly on agricultural income can engage in the nonfarm sector, even though the scope of the activity is small. This report does not use the term "off-farm" employment. It is sometimes used in agriculturally focused studies. Sometimes the term is also confused with or used as a synonym for the nonfarm sector. Off-farm employment typically means employment "outside the owner's farm" and includes nonfarm self- employment, nonfarm wage employment, and agricultural wage employment. Nonfarm employment is thus smaller than off-farm employment. The official definition of "rural" is very narrow in Ethiopia. It typically includes population settlements below a threshold of 2,000 habitants. This official definition is also used in RICS. This is relatively low compared with many other countries, where official definitions often refer to concentrations of some 5,000-25,000 people. Moreover the rural definition in Ethiopia is not always strictly used. For example, in special cases, population settlements that include an important market, school or serves as administrative capital may well be classified as urban, even though the population density is much below 2,000 habitants. The rural nonfarm sector tends to be underestimated. This is because many activities are typically concentrated among small rural towns that according to the official definition are classified as urban. The urban definition in Ethiopia is thus rather broad and includes population groups that in other countries could well be classified as rural. To correct for this bias (and because of important functional linkages between small towns and surrounding rural areas) the RICS-Amhara includes on a pilot basis small market towns with a population size of up to 10,000 habitants. This allows a unique distinction between rural areas and small towns in the Amhara region. -14- C. SNAPSHOT OF ETHIOPIA'S RURAL ECONOMY Recent increase in agricultural growth 35. Ethiopia is one of Africa's largest countries with about 77 million people. Ethiopia has among the highest dependence on agriculture of any country in the world. Ethiopia's agricultural sector is a major contributor to the economy and is central to food security and poverty reduction. Agriculture accounts for an estimated 44 percent of national gross domestic product (GDP), almost 86 percent of exports, and 80 percent of employment. Nearly 90 percent of the poor depend on agriculture for their livelihood. Additionally, agriculture determines the country's overall food security at an aggregate level and is crucial for reducing the food deficit for an estimated 12 million people who are chronically or transiently food insecure. 36. According to official data, Figure 1: Ethiopia ­ Agricultural and Overall Ethiopia has made some impressive GDP Growth, 2001/2-2008/9 development gains in the past few 20 years. This data suggest that GDP 15 growth averaged 11.6 percent while agriculture has grown on average by 10 13 percent between 2003/04 and 5 2007/08 (Figure 1). Economic growth 0 has raised the living standard of 2008/09 forecast 2002/03 2003/04 2004/05 2006/07 2007/08 estimate 2001/02 2005/06 millions of people from a very low -5 base. The official poverty headcount -10 Agriculture and allied activities ratio has declined from an estimated Total GDP growth 44 percent in 1999/2000 to 38 percent -15 in 2004/2005, and may have continued Source: National Bank of Ethiopia. to fall given the high levels of growth. 37. The Government's commitment to agricultural development is reflected in its emphasis on agriculture in the budget. Including rural infrastructure, expenditure for agriculture and rural development has been around 25 percent of total spending, one of the highest shares in the world. As a result, the country's physical and social infrastructure has expanded rapidly: over the past seven years, the federal paved road network has increased by 43 percent, power generation capacity has nearly doubled, primary school enrollment has increased from 5.2 million to 13 million, and most health indicators have shown steady improvements. According to official data, recent growth of the agriculture sector, supported by several consecutive years with good weather, has been the driving force behind Ethiopia's growth performance. 38. The recent growth in agriculture, although impressive, is from a low base. Ethiopia's agricultural sector thus continues to be of subsistence nature. Land is fragmented, often highly degraded and predominantly rain-fed. Smallholder and subsistence-oriented farmers continue to produce over 98 percent of the agricultural output. Small and fragmented farm sizes, coupled with low-level technology and soil degradation have reduced the capacity of small farmers to undertake long-term -15- investments and innovation. As employment opportunities within agriculture are unable to keep up with growth in the labor force, there is a widely recognized need for diversifying rural incomes. 39. The government's primary focus in its approach to rural development has been on the intensification of agricultural production within the context of the Agricultural Development-led Industrialization Strategy (ADLI). More recently, however, as elaborated in PASDEP, the rural development strategy is broadened beyond the initial focus on agricultural intensification with recognition, though little sustained action, of the need to stimulate nonfarm growth. Box 4: How Big is the Rural Nonfarm Enterprise Sector? Contribution to Income Accurate estimates of household income are difficult to attain in any survey and no less so in the RICS. Indeed, the CSA has recently ceased providing income data from its Household Income, Expenditure and Consumption Survey (HIECS) and Welfare Monitoring Survey (WMS) due to concerns over reliability. However, several approaches were taken during the preparation of the RICA which all provide a similar estimate. 1. The RICS-AgSS survey in four rural regions asked all households who had an active enterprise about their monthly sales revenue, their monthly operating costs, and what percent of total household income during the most recent year was from the enterprise's total sales. This showed that enterprise net income contributed 42 percent of total income among households that run an enterprise. Given a participation rate of 25 percent, this implies contribution to total rural household income of just over 10 percent. 2. The RICS-Amhara survey asked for details of annual income from a full range of sources including sales of agricultural produce, wages and salaries, social benefits, gifts and remittances, and enterprise sales income net of operating costs. This showed that net enterprise income contributed 44 percent of total income among households that run an enterprise. In the Amhara region, which has below average participation, nonfarm enterprise income represents just less than 9 percent of total household income. 3. The most recent data on rural household income sources comes from the 2008 Ethiopian Agricultural Household Marketing Survey (EAHMS). Early results suggest nonfarm self- employment contributed almost 13 percent of rural household income in 2008. Comparison of the latest data with the latest HIECS / WMS for which data is available shows that the nonfarm enterprise's contribution to income has grown in recent years. Table 1: Ethiopia ­ Average Share of Rural Household Income by Source, 2000 and 2008 Agriculture Non-farm self Non-farm wages Remittances Other (including employment and salaries public sector) 2000 2008 2000 2008 2000 2008 2000 2008 2000 2008 74.9 72.7 9.1 12.8 4.9 3.6 3.9 1.5 7.1 9.4 Note: Other include rent income, pension and insurance income and any other source not included in the labeled categories, including public sector wage labor income. Data for 2008 is representative for Tigray, Amhara, SNNP and Oromia only. Moreover, it shows that nonfarm self-employment income makes the second largest contribution to household income after agriculture (72.7 percent), with nonfarm wage employment, remittances, and other sources including public sector wages accounting for the remainder. Taken together, these data point to a plausible estimate for the contribution of nonfarm enterprise income to total rural household income of at least 10 percent. Source: RICS-AgSS, RICS-Amhara, HIECS / WMS 2000, EAHMS 2008. -16- Shift from food aid to cash transfers; and rural cooperatives 40. Addressing structural sources of chronic food insecurity requires long-term interventions. Given the high levels of chronic food insecurity in Ethiopia, and following a severe drought in 2002/2003, the Government developed the New Coalition for Food Security, a long-term policy framework for reducing hunger and food insecurity in Ethiopia. As part of this initiative, it launched the Food Security Program, which combined a series of complementary interventions designed to enable food insecure households to acquire sufficient assets and income to "graduate" out of food insecurity and improve their resilience to shocks. The shift from food aid to cash transfers is an essential part of the Government's strategy. The Government decided that an alternative to food aid was needed to support the consumption needs of chronic, predictably food- insecure households, as well as to address some of the major underlying causes of food insecurity. 41. The Government started implementation of an employment-based Productive Safety Net Program (PSNP) in 2004/2005. The PSNP replaced the emergency humanitarian appeal system as the chief instrument for assisting chronically food- insecure people in rural Ethiopia. It was scaled up to reach 7.3 million people. The PSNP provides cash and in-kind resources to chronically food insecure households, largely via wage labor-intensive public works. The focus of the public works program is on soil and water conservation activities, developed within an integrated watershed management- planning framework. 42. Another important recent development supported by the Government is that of farmers' cooperatives and their amalgamation into cooperative unions. The Government's aim is that cooperative services are extended throughout the country to supply inputs to smallholders and market the output. The support for cooperatives has been in place since 1994, but received renewed attention in the last few years. As of 2005-06 cooperative coverage was already estimated at 35 percent of all Kebeles4. Given the strong government support, the number and membership of cooperatives have grown rapidly since then. The cooperatives are serving to amalgamate smallholders for purposes of finance and marketing of inputs and outputs. 4 The fourth administrative layer in Ethiopia. -17- D. RECENT ECONOMIC SHOCKS 43. In early 2008 Ethiopia was hit hard by the global food crisis, and Figure 3: Ethiopia ­ CPI Growth, July 1998 to August 2009 (y/y changes) possibly had one of the highest food price inflation rates in Africa. The high food price inflation is mainly attributed to the sharp rise in the cereal prices (Figure 3). The overall consumer price index (CPI) inflation rate reached a historical peak of over 60 percent in July 2008, before falling sharply to -3.9 percent by August 2009 (Figure 2). Due to accompanying macroeconomic imbalances, such as the lack of Source: calculations based on CSA data. foreign exchange and pressure on the balance of payments, Ethiopia has faced a deeper crisis than many other countries in the Africa region. 44. The food crisis Figure 2: Ethiopia ­ Decomposition of CPI Growth, January 1999 to August 2009 (y/y fundamentally revealed that changes) Ethiopia's impressive official growth rate has not removed the long- standing problem of pervasive food insecurity, the absence of alternative sources of income other than agriculture to diversify risks, and may point to structural weaknesses of the economy, in particular its severe vulnerability to price shocks. As food accounts for 57 percent of total household consumption Source: calculations based on CSA data. expenditure, high food prices have during 2007-2008 caused severe hardship for the people, especially the most vulnerable segments of the population (Loening and Oseni, 2008). 45. The question of how to achieve rural income diversification is likely to become increasingly important. As the Ethiopian economy develops, with higher productivity and better performance in agriculture, the nonfarm sector will also grow and become increasingly important as an alternative employer of labor and source of livelihood in rural areas. This suggests the policy priority should not be "either agriculture or the nonfarm sector" but a balanced approach focusing on the spillovers between the sectors, particularly consumption linkages. This will include ensuring rural nonfarm enterprises are not constrained in responding to new opportunities by the rural investment climate. -18- 2. SIZE AND BASIC CHARACTERISTICS OF THE RURAL ENTERPRISE SECTOR A. OVERVIEW 46. Ethiopia's nonfarm enterprise sector is sizable and significant. Women are important actors in the sector. Nonfarm enterprise activity is highly seasonal and complementary to agriculture. For a small minority, often women, it is a crucial alternative to agriculture. B. THE NONFARM ENTERPRISE SECTOR 47. Ethiopia's nonfarm enterprise sector is sizable and significant. About 25 percent of rural households participate in the nonfarm enterprise sector. There are differences in participation rates across regions. The percentage of households participating ranges from only 20 percent in Amhara to 37 percent in the SNNP. Nonfarm enterprise profits account on average for 42 percent of total income among households that run an enterprise. The majority of nonfarm enterprises are run part-time, either in parallel with agriculture, or periodically as a substitute for agriculture. Less than 3 percent of rural households rely exclusively on income from nonfarm enterprises. Table 2: Ethiopia ­ Rural Enterprise Participation Rates and Contribution to Income, 2007 (percent) Category Tigray Amhara Oromia SNNP Total Households owning nonfarm enterprise 22 20 23 37 25 Male households owning enterprise 18 12 13 25 15 Female households owning enterprise 29 35 42 52 41 Enterprise as sole source of income 2 2 3 2 3 (no income from agriculture) Enterprise as major source of income 19 7 10 6 8 (agriculture being less important) Estimated enterprise profits to 40 44 46 38 42 household income (owners) Source: 2006/7 RICS-AgSS. 48. The relatively high level of participation in nonfarm activities is somewhat surprising. The conventional wisdom, based on limited data, was of little diversification beyond agriculture in rural Ethiopia (Günther and Olapade, 2007). Although some small studies have found participation in nonfarm enterprise activities to be in the region of 30 percent, the most comprehensive survey prior to the RICS found a participation rate of 9 percent (MOLSA, 1997a). Not only is participation higher than previously thought, it is comparable with the average across Africa, estimated to be 20-25 percent in terms of simple participation (Haggblade et al, 2007). Moreover, the total participation rate is probably an underestimate because the RICS sample excludes the pastoral regions, and with the exception of the Amhara region, does not cover small rural towns, which are conventionally classified as urban in Ethiopia. -19- 49. Women are important actors in the sector. Female-headed households own nearly one-half of all enterprises. Yet, women head only one-fourth of the households. This implies that almost every second household headed by a woman operates a nonfarm enterprise, while only about 1 in 6 households headed by men own a nonfarm enterprise. Furthermore, nonfarm enterprise income tends to be more significant as a share of total income for female-headed households. They are more likely to engage in nonfarm enterprise as a primary activity. Women tend to work in nonfarm activities because they face constraints in other domains, especially agriculture, and not necessarily because they are well positioned to exploit profitable market opportunities. More details on gender differences are given in the subsequent chapter. Box 5: What is new? A Guide to the Evidence on Informal Rural Enterprises in Ethiopia The Ethiopia Rural ICA is probably one of first studies to systematically look at Ethiopia's nonfarm sector and small informal enterprises in rural areas. Günther and Olapade (2007) extensively review more than 50 publications on Ethiopia's rural labor market over the past decade, covering formal publications, reports from development and government agencies, and several doctoral and master theses. Their main finding are inconclusive. Neither the size nor basic sectoral patterns are known. Much of the evidence comes from experimental surveys in selected Weredas. In addition, income data yields inconclusive results. A few pieces, though outdated, stand out: · The Ministry of Labor and Social Affairs (MOLSA) conducted a pilot wage and nonfarm-labor survey in 1996. The subsequent report documents empirical findings (MOLSA 1997a) and focuses on appropriate nonfarm technologies in Ethiopia (MOLSA, 1997b). · Woldehanna's (2000) doctoral thesis provides an in-depth analysis of a few Weredas in Tigray. Though based on a small and non-representative survey, it is a groundbreaking empirical analysis for Ethiopia. · Pernille Sørenson (2003) on food security and Jonathan Baker (1986) on anthropological aspects of rural-urban linkages provide excellent qualitative analyses for the Amhara region. Mulat Demeke (2001) has focused on policy aspects of income diversification. -20- C. ENTERPRISE ACTIVITIES AND CHARACTERISTICS 50. Trading, in particular in agricultural commodities, is the dominant activity. In all regions surveyed, except Amhara, more than half of the enterprises are in the trade sector, followed by manufacturing and lastly services. In Amhara, most enterprises are in manufacturing, closely followed by trade. Trade is heavily dominated by retail sale via stalls and markets and the retail sale of food and beverages, followed by wholesale trade in agricultural products. Wholesale trade, however, contributes not more than 6 percent of all nonfarm trading activities. Table 3: Ethiopia ­ Composition of Rural Enterprise Sector, 2007 (percent) Sector Tigray Amhara Oromia SNNP Total Manufacturing 30 43 35 32 36 Food, beverages, brewing, distilling 13 20 23 17 19 Grain milling 3 1 1 1 1 Other manufacturing 13 22 12 14 15 Trade 56 41 52 58 51 Wholesale trade 10 4 4 8 6 Retail trade via stalls and markets 19 22 25 31 26 Other retail trade 28 15 23 19 20 Services 14 16 13 11 13 Hotels and restaurants 7 6 5 6 6 Transport services >1 1 1 >1 1 All other services 7 9 8 4 7 Source: 2006/7 RICS-AgSS. 51. The main manufacturing activity is home brewing and distilling of alcohol, followed by the textile businesses, mainly dominated by weaving. In the service sector the sale of food and beverages dominates, as one-half of services are composed of hotel, restaurant and bar services; followed by community services, such as sewage, disposal, and sanitation activities; transportation; and hairdressing. Table 4: Ethiopia ­ Selected Enterprise Characteristics, 2007 (percent) Characteristics Tigray Amhara Oromia SNNP Total Firm age 6.3 7.4 5.8 5.6 6.1 Average number of workers 1.7 1.3 1.4 1.5 1.4 1 employee 71 77 73 69 73 2 or 3 employees 26 22 26 29 26 4 to 9 employees 3 >1 1 2 1 10+ employees >0.1 >0.1 >0.1 >0.1 >0.1 Source: 2006/7 RICS-AgSS. -21- 52. Most enterprises are young and Figure 4: Ethiopia ­ Localized Nature of small. About ¾ of all enterprises are Business, 2007 one-person firms while the remaining ¼ employ only two or three workers. Location of nonfarm enterprise Only one percent of all enterprises In community employ more than three workers. The Same Woreda, different FA EA likelihood of owning an enterprise rises with education up to 5 years of Same FA, different community EA, schooling, which is considerably Same Zone, different Woreda higher than the average. Same Region, different Zone 53. Few enterprises operate on a Different Region full-time basis. Nonfarm enterprises are set up primarily as a complement 0 10 20 30 40 50 60 70 to agriculture, providing an alternative source of income in periods when the Location of nonfarm customers level of activity in agriculture is low. More than 95 percent of enterprises are Local consumers or passers-by owned by a sole proprietor. Only 3 Market percent are registered with any Traders government office (CSA, 2008a). Others Bigger enterprises, however, are more NGOs likely to be registered compared to smaller ones. Cooperatives Government 54. Economic activities are highly 0 10 20 30 40 50 localized. Enterprises tend to be located in, or close to, the community Source: 2006/7 RICS-AgSS. where the individual owner lives (Figure 4). Local consumers or passers-by are the most important customers for more than 40 percent of the firms. Box 6: Getachew, a Rural Manufacturer of Household Items Living by the Roadside Getachew is married, has one son and lives in a rented place. He works alone but his wife is also active in business: she sells local beer at home. Six months ago, he moved to the area and started manufacturing household items: small containers and charcoal stoves, produced from metallic iron scraps and old containers. He has a prime location for selling his items, as he is located along a main transport road. He has no working premises and does his work "under the sun" in an open space. Given the fact that he sells his products at the same place where he produces them, there are no additional transport costs. He uses only small hand tools, as he does not have any machine. Raw materials are brought from about 270 km away. He is the only supplier of those kinds of products in the area. His customers are villagers and people who are passing by for marketing and other purposes. Sometimes people buy on a credit basis. As he has a very small capital, he was not requested by local authorities to get a license and he does not need to pay any taxes. His business is profitable and he has the idea of expanding. He thinks 3,000 Birr would be enough to scale up his business, allowing him to buy tools and produce more. However, none of the existing credit sources are available to him. Private moneylenders charge very high interest rates (10 percent per month). A group loan was not an option as he is new to the area and therefore is not easily accepted. Source: Bakker (2007). -22- Nonfarm enterprises are often complementary to agriculture 55. Enterprise activity is highly Figure 5: Ethiopia ­ Nonfarm Seasonality, 2007 seasonal. As indicated, the majority of nonfarm enterprises are run as a 25 Percentage of firms indicating month complementary activity to agriculture 20 either in parallel with agriculture, or periodically as a substitute for 15 as most active agriculture. Therefore seasonality is a 10 sign of the close countercyclical interaction between agricultural and 5 nonfarm activities. Figure 5 shows that 0 activities peak during the month of ay ne em r Au uly ch br y st ril Ja b e r ct r r M y ec be O be ov e Fe uar r gu Ap N ob ua ar October until December, dropping to M Ju J D em m n p te Se their low point during the planting and harvesting seasons. On average, some 44 Source: RICS-AgSS. percent of households with nonfarm enterprises operate them on a highly seasonal basis, and a further 19 percent only operate their enterprise during the three peak months per year. These numbers indicate that enterprises are dormant over long periods. 56. Policies to promote rural income diversification in Ethiopia needs to take into account theses seasonal patterns. They have a significant impact on the incentives and capabilities of households to engage in nonfarm activities. This is because seasonality may act as a constraint to rural enterprise growth: firms may experience an ebb and flow of workers that hampers continuity and ability to upgrading skills. Moreover, as it is often not worthwhile or risky to establish the business on a permanent basis, seasonal demand fluctuations can drive entrepreneurs into informality. Finally, and here in particular in the manufacturing and construction sectors, seasonality often implies an additional need for short-term capital, which cannot be easily met. 57. Enterprise participation and characteristics varies across and within regions. The proportion of households participating in the sector ranges from 20 percent in Amhara to 37 percent in the SNNP region, but exceeds 50 percent in the Burji and Gedeo zones in Oromia. Enterprise is the most important source of household income for 19 percent of households in Tigray, which has a relatively low overall participation rate of 22 percent. -23- Map 2: Ethiopia ­ Nonfarm Enterprise Participation Rates by Geographical Zone, 2007 2007 -24- 3. GENDER DIFFERENCES IN ENTERPRISES CHARACTERSITICS A. OVERVIEW 58. Women play a very important role in Ethiopia's nonfarm enterprise sector. Nonfarm income diversification is especially important when women do not have sufficient access to agricultural land, or are widowed or divorced. Women are more likely to be engaged in nonfarm activities than men, especially in small towns. Women tend to take-up nonfarm activities because they face constraints in other domains, especially agriculture, and not necessarily because they are well positioned to exploit market opportunities. By contrast, men are better able to exploit complementarities between nonfarm activities and agriculture. Activities which women engage in are often limited, and typically concentrated in low-profitability sectors requiring little training and skills. B. BASIC CHARACTERISTICS Women are more likely to participate in nonfarm activities than men 59. There are important gender differences in the propensity of engaging in nonfarm enterprises. Female-headed households are much more likely to rely on nonfarm enterprises as the only or an additional source of household income. Overall, more than 40 percent of female-headed households report a nonfarm activity, while only 17 percent of male-headed households do (Table 2). Furthermore, more than 5 percent of female- headed households report a nonfarm activity as the only activity compared with less than 1 percent of male-headed households. 60. A reason for the high participation of women in nonfarm employment is that work roles are often segregated according to sex. Men are traditionally responsible for agricultural tasks, such as plowing and cutting seeds. Women perform a wide variety of agricultural tasks, such as weeding, preparing and carrying manure, helping with harvesting, grinding seeds, vegetable growing and the management of small livestock ­ but they do not undertake plowing, which is reserved for men. As a result, it can be very difficult for unmarried, divorced or widowed women to be independent farmers. Single women therefore need to generate additional income through nonfarm activities in their own community, or migrate to rural towns. -25- Women tend to concentrate in activities with low earnings 61. There are also important gender differences in the type of nonfarm activities (Table 5). For example, the production and sale of alcohol is a typical female activity. Women also predominate as owners of bars, hotels, and restaurants, but working in such establishments requires regular contacts with a male clientele. Generally, only independent (unmarried, divorced, or widowed) women can undertake these activities. Men, on the other hand, are more active in retail trade, an activity that implies higher mobility. Table 5: Ethiopia ­ Distribution of Nonfarm Enterprises by Sector, Region, and Sex, 2007 (percent) Tigray Amhara Oromia SNNP Sector of enterprises Men Women Men Women Men Women Men Women Manufacturing of food 4 >1 4 1 2 3 1 5 Manufacturing of alcoholic beverages >1 40 3 30 7 31 2 41 Manufacturing of textiles 9 12 20 9 7 3 8 5 Other manufacturing 3 3 10 9 8 9 7 7 Wholesale 10 >1 9 1 4 2 11 5 Retail 65 25 46 32 59 43 61 28 Restaurants 3 16 1 11 1 8 4 9 Other Services 5 6 7 7 12 2 7 >1 Source: 2006/7 RICS-AgSS. 62. The composition of women's nonfarm enterprise activities is similar across regions. There are not large differences across regions in either the composition of nonfarm activities or the relative importance of these activities for women and men. Women are disproportionately represented in the manufacturing of alcoholic beverages such as beer and araqé.5 Men are disproportionately represented in wholesale and retail trade. However, the percentage of women who engage in retail is not insignificant, particularly in Oromia. 63. Women tend to concentrate in activities with relatively lower revenues than those of men. The manufacturing of alcoholic beverages, accounting for one third of women's nonfarm activities, has the lowest median sales. Women also earn less revenue than men do within the same sector (Table 6). In retail trade, a sector that accounts for more than one-half of men's employment, men's median sales are three times larger than those of women (Bardasi and Getahun, 2008). In the restaurant sector, with a prevalence of women, the median sales of women's enterprises is one fourth of men's. Overall women are disproportionately found in lower revenue activities and earn less within the same sector. 5 Araqé is a homemade distilled drink originating in the highland areas, based on germinated grains. The preparation of araqé is labor-intensive. -26- Table 6: Ethiopia ­ Median Enterprises Sales by Sector and Sex of Owner, 2007 (Birr/month) Ratio Sector All Men Women women/men Manufacturing of food 629 812 214 0.26 Manufacturing of alcoholic beverages 110 142 92 0.65 Manufacturing of textiles 136 179 39 0.22 Manufacturing of other 185 311 59 0.19 Wholesale 847 844 471 0.56 Retail 546 703 235 0.33 Restaurants 310 611 160 0.26 Other Services 176 216 57 0.26 Source: 2006/7 RICS-AgSS. 64. Male and female-headed households start rural nonfarm enterprises for similar reasons. Moreover, the patterns of start-up motives are similar across all four major regions in Ethiopia. But men and women differ with respect to the options available. Lack of access to agricultural land is more important for women. Less women consider agricultural income attractive as a means to invest in nonfarm enterprises. Some 43 percent of women were "pushed" into nonfarm activities by factors such as low or volatile agricultural income, rather than being "pulled" by profitable opportunities (see Table 7 for categorization of push and pull factors). Although this percentage is slightly higher than for men, what is striking is the high percentage of both men and women engaged in nonfarm activities because of constraints they experienced elsewhere, most notably in agriculture. Overall, agriculture is the sector of choice of both men and women, but nonfarm activities complement agriculture when the returns from this sector are lower than expected. Table 7: Ethiopia ­Reason for Enterprise Start-up by Sex of Head, 2007 (percent) Men-headed Female-headed Reasons for enterprise start-up households households Push Household lost wage earnings 1.1 2.7 No access to agricultural land 9.3 13.4 Low or volatile agricultural income 29.0 27.3 Pull Means to invest agricultural earnings 50.3 43.6 Markets opportunity 3.4 3.1 Other Support from NGO or cooperative >0.1 0.3 Advice from relatives/friends 3.5 2.0 Social and economic independence 1.9 1.3 Other 1.4 6.4 Source: 2006/7 RICS-AgSS. -27- 65. The type of engagement of men and women in nonfarm activities and the characteristics of women and men's activities are similar across Amhara, Oromia, SNNP and Tigray (Bardasi and Gethahun, 2008). The following section will use information from the RICS-Amhara for a more detailed analysis, which also allows directly identifying the sex of the owner (rather than relying on the sex of the household head). These two characteristics more than compensate for the disadvantage of a narrower geographical focus. The evidence presented so far indicates that the reality that we are going to describe for Amhara with respect to men's and women's engagement in nonfarm activities is similar in the other Ethiopian regions. Box 7: Mintiwab, a Student Selling Spices in the Market Mintiwab is a full-time student in 9th grade. Mintwab's father was a teacher and died six months ago. There are three children in the family. She helps her mother in the market when she does not have classes. They trade in three market places in the Wereda and also at home, despite the fact that they have no separate shop there. In the market, Mintiwab has a plastic shade. Since there are not many spice traders, she typically has good sales. A problem is that the spice is brought from up to 180 km away and transport costs are high. Mintiwab's mother has traded spices since 1991 and uses credits from different sources. She started the business when she saw that many people, particularly husbands, were dieing and leaving families facing problems. She was very concerned about what would happen to her family when the same fate would happen to her. Source: Bakker (2007). C. GENDER DIFFERENCES FOR THE AMHARA REGION 66. Women in Amhara are more likely to be engaged in nonfarm activities than men6. This finding is similar to national patterns. But for women nonfarm enterprises are more likely to be the only activity (Table 8). Almost 50 percent of women who participate in the nonfarm enterprise sector have the activity as their only activity. In comparison, 63 percent of men with a nonfarm activity have it as their secondary activity besides agriculture. Thus, it appears that men more than women are able to exploit complementarities between nonfarm activities and other activities. Men are more likely to engage in a nonfarm activity when they see an opportunity to diversify. Table 8: Amhara ­ Importance of Nonfarm vs. Agriculture Employment by Sex, 2007 (percent) Of those engaged in nonfarm activities Men Women All Nonfarm employment is the only activity 29.0 46.5 38.9 Nonfarm is more important than agriculture 5.1 19.1 13.0 Nonfarm is less important than agriculture 63.2 30.2 44.5 Nonfarm and agriculture are equally important 2.8 4.3 3.7 Source: 2006/7 RICS-AgSS. Sample includes all individuals aged 10+ who are employed. 6 Participation rates differ from previous chapters where household nonfarm participation is considered. In this chapter, in order to understand gender differences, individual participation rates are calculated. -28- Box 8: Why is Female Participation in Nonfarm Employment so High in Rural Towns? Being a woman head of household increases the probability of being engaged in nonfarm activities. Using regression analysis, Bardasi and Gethahun (2008) find that being a woman has a positive and significant impact on nonfarm participation in the four zones surveyed through RICS-Amhara. The gender effect is very robust, as it does not disappear even when controlling for a large set of personal, household, and area characteristics. In particular, the gender effect varies in relation to the marital status. Women with limited access to land and who are separated or divorced also have a significantly higher probability of engaging in nonfarm activities than other women. The biggest effect occurs when women are located in rural towns. In his seminal work on rural-urban interactions, Baker (1986, 1990) notes that one of the most salient features in Ethiopia is that women outnumber men in almost all small towns. This is because of significant migration from the countryside. While the causes for rural migration to small towns are multiple, a fundamental source of rural-urban female migration in Ethiopia is related to access to land and marriage. In the case of divorce or widowhood, females are often forced to migrate to small towns because of limited opportunities in agriculture for single adult households, often related to cultural practices that segregate agricultural activities between men and women. Moreover, among many ethnic groups, husbands have typically kept the land upon dissolution of a marriage in the case of divorce (Fafchamps and Quisumbing, 2002 and 2005).a Figure 6: Amhara ­ Probabilities of Being Engaged in Nonfarm Employment (percent) 80 70 60 50 40 30 20 10 0 H n d gh n ld d lth e er n tio H re ce ea as w ho ou ea ng i ld ca To in -h or B se pl ou rw ch iv en du on 1. ou 6. o /d Y he om re N .N e rh ed or 7. ig 2. he 10 w at M ge H ig ar e 4. ig 5. or H ep .B M 3. S 11 8. 9. Men Women 1. Base case: Rural household head, 40 years old, married, with some elementary school, no children less than 6, not a migrant, household of 5 members, household has a plough, food secure area, no access to finance, average daily wage rate in agriculture is Birr 8. 2. Headship: like 1, but individual is not head of the household. 3. Education: like 1, but high school instead of elementary. 4. More children: like 1, but 3 children less than 6 instead of 0. 5. Wealth status: like 1, but highest asset tercile instead of lowest. 6. Location: like 1, but resides in town instead of rural area. 7. Age: like 1 but age is 20 instead of 40. 8. More women in the household: like 1 but 2/3 of adults are women instead of 1/3. 9. Marital status: like 1 but divorced instead of married. 10. Plough: like 1 but the household has no plough instead of having one. 11. Household size: like 1, but the household has 8 members rather than 5. Source: Computations based on 2006/7 RICS-Amhara. a The now ongoing land certification program has as objective to secures or gives land ownership title for both males and females, and they have equal rights on land management. -29- Location and access to land affect women's participation even more than men's 67. Women have significantly higher participation rates in small towns. Measured on an individual rather than household basis, the gender difference in participation is much larger in rural towns of up to 10,000 habitants than it is in remote rural areas. In remote rural areas, about 9 percent of women are engaged in nonfarm activities, as compared to about 6 percent of men. But in rural towns nonfarm activities absorb up to 76 percent of women, compared to a much lower 44 percent of men. 68. In Amhara, manufacturing is more common for women, while trade is less common. Similar to national patterns, women are less likely to be involved in trade activities, both in rural and in small town areas. They are substantially more likely to be involved in manufacturing. A detailed classification of industrial sectors shows that men and women's nonfarm businesses are different. They also vary in relation to small towns or rural area. The largest concentration of women is in the manufacturing of alcoholic beverages, both in rural areas and, especially, in the small towns. In rural areas men are mostly found in the manufacturing of textile and leather articles. Men concentrate also in trade, both wholesale and retail, especially in small towns. Women, on the other hand, are more likely to operate hotels and restaurants in small towns. Table 9: Amhara ­ Sector Distribution of Nonfarm Businesses, 2007 (percent) Rural towns Rural areas Total Sectors Men Women Men Women Men Women Food processing 2.3 3.0 4.2 0.0 3.8 0.9 Alcoholic beverages 1.9 42.9 2.9 35.5 2.7 37.6 Textile and leather 21.6 17.3 32.5 27.2 30.2 24.4 Other manufacture 3.8 1.3 10.8 9.6 9.3 7.3 Wholesale 14.4 1.3 12.0 2.8 12.5 2.4 Retail sale 44.1 15.8 29.7 12.8 32.7 13.7 Restaurants 2.3 16.5 2.2 6.4 2.2 9.2 Others 9.8 1.9 5.7 5.6 6.6 4.6 Source: 2006/7 RICS-Amhara. Sample includes all individuals who are owners or managers of a nonfarm business. 69. Male and female entrepreneurs experience the same constraints but with different intensity. There are only small differences between men and women's perceptions on the most important constraint to enterprise operations and growth. Both report access to markets as the most important constraint to their business, followed by rural finance and transportation. However, there are gender differences in the intensity of each constraint (Bardasi and Gethahun, 2008). Some constraints are perceived more intensely by women --in particular access to water, low demand, access to informal sources of credit, and fear of not being able to repay the loan. Men by contrast are more likely to complain about problems related to access to markets, market information, and material inputs. -30- 70. Education appears not to be important for female nonfarm participation. Regression estimates show that education has a significant and positive effect on the probability of male participation in nonfarm activities (Bardasi and Gethahun, 2008). However, education does not have a significant effect for women's participation in nonfarm enterprises. Access to education differs for men and women in the Amhara region, 81 percent of all women received neither formal nor informal education as compared to 69 percent of men. This finding may be an indication that the type of nonfarm activities accessible to women (irrespective of their level of education) are not as remunerative as those chosen by men--and not as remunerative as alternative activities in the farm sector. Therefore, for women with higher education nonfarm employment is not necessarily more remunerative than for women with lower education. 71. More women in the household leads to higher participation in nonfarm enterprises. The effect of household size in general on nonfarm participation is not significant for the Amhara region. However, for women, the probability of engaging in nonfarm activities is positively related to the proportion of women among adult household members, suggesting that women are better able to participate in nonfarm employment when other women in the household can provide labor to the farm or remain in the home to take care of children and engage in domestic tasks. The presence of children itself has a negative impact on the probability of nonfarm employment, confirming that there could be a conflict between the type of nonfarm activities taken up by Ethiopian women and the need to provide care for children in the household. 72. Women tend to engage in Figure 7: Amhara ­ Distribution of Start-up Capital, nonfarm activities that require low start- 2007 up capital. The median start-up capital of .3 male-owned nonfarm enterprises is five Log start up capital Kernel estimates times higher than of female-owned ones. The Kernel density function shows the .2 different distribution of start-up capital by location (rural areas and small towns) .1 and sex of the business owner. After controlling for location, women's activities have a lower start-up capital 0 than men's activities. A large proportion 0 2 4 x 6 8 10 of female-owned nonfarm enterprises in Rural women Urban women rural areas had very low start-up capital. Rural men Urban men By contrast, nonfarm enterprises with the largest start-up capital were mostly Source: 2006/7 RICS-Amhara. NB: Urban here refers to rural towns. male-owned small town activities. 73. Most enterprises employ unpaid family labor. Women's businesses do not differ that much from men's businesses in terms of the quantity of labor that they employ. Only a small proportion of enterprises employ paid workers. -31- 74. Women's activities are much smaller than men's activities as measured by enterprise revenue. The exception is the production of alcoholic beverages, hotels, and restaurants in which females dominate. In the textile sector--a sector that absorbs a quarter of all women business owners--both the median value added per worker7 and the median revenue of men's enterprises is more than ten times larger than the corresponding figures for women's. In retail sales, a sector where almost 14 percent of women entrepreneurs are engaged, both the median value added per worker and the median revenue of men's businesses are four times larger than women's (Table 10.) Table 10: Amhara ­ Employment Characteristics of Nonfarm Enterprises, 2007 Nonfarm enterprises with Men Women All Paid workers 6.5 1.7 3.8 Non-family hired labor 5.2 0.3 2.5 Unpaid family labor 99.1 99.2 99.2 One unpaid family member 74.2 79.6 77.3 Two or more unpaid family members 24.9 19.7 21.9 Source: 2006/7 RICS-Amhara. 75. Revenues of male-owned enterprises are higher than those of women. The median revenue and the median value added of rural, female-owned enterprises are 3.5 and 2.2 times smaller than those of female- owned enterprises in small town Figure 8: Amhara ­ Distribution of Enterprise Revenue, 2007 areas. They are also 5.6 and 3.7 .4 times smaller than male-owned enterprises in rural areas, Log sales Kernel estimates .3 respectively. Figure 8 shows the distribution (Kernel densities) of .2 the revenue of nonfarm enterprises, by location and sex of the business .1 owner. Rural female-owned nonfarm businesses have lower 0 average revenues compared to the 4 6 8 10 12 x remaining three groups. Men Rural women Urban women business owners in small towns, Rural men Urban men and also in rural areas, have higher revenues than women do, while Source: 2006/7 RICS-Amhara. women in small towns have larger NB: Urban here refers to rural towns. revenues than women in rural areas. 7 Value added refers to the difference between sales revenues and material input costs. -32- D. SUMMARY 76. Women's enterprises are smaller and less profitable than men's, but offer an important opportunity for employment and income generation, especially for those in vulnerable situations such as single women and others without access to land. 77. Although the relatively high participation of women in non-farm activities indicates that they do not face disproportionately high entry-barriers, policy support to non-farm activities should take into consideration the gender-specific nature of those activities. In particular, women face certain constraints more intensely than men, which are only partially related to the different type of activities they engage in. Some constraints that women disproportionately judge as major ones--such as access to credit, or fear of not repaying a loan--indicate that women, irrespective of the sector they are involved in and despite the fact that they generally engage in small scale activities, have greater difficulties than men in solving the basic operational problems of their enterprises. 78. There are other, broader constraints for which policy measures are not easy to find. These are related to the custom, tradition, culture, and other social norms that dictate women's role in the economic sphere, such as the extent of their engagement in agriculture and other domains, their ability to act as economic agents, and even their freedom of movement outside the house. The fact that these constraints cannot be easily solved does not mean that they do not exist or that the potential of women's role in the non-farm economy cannot be higher than it is currently. -33- 4. ENTERPRISE DYNAMICS: PERFORMANCE, CONSTRAINTS, AND OPPORTUNITIES A. OVERVIEW 79. Overall, the profits from nonfarm enterprise are low and very few firms invest and grow. However, some perform much better than the average, others much worse, and there are some promising sectors. The performance of local agriculture affects productivity, probably because of an increase in local demand. Enterprises in rural towns perform better than those in remote rural areas, suggesting that demand-side problems, because of small fragmented markets and a poor investment climate, in remote areas are the main constraints. 80. Our data indicate that few enterprises add to their capital stock after start-up, and very few increase their labor input. Nevertheless, there is some evidence the sector has grown over the last decade, due to net entry into the sector. Policies facilitating the integration of markets would make nonfarm enterprises less dependent on the local rural economy, which may help these enterprises develop beyond supplying a small and volatile local market with low value-added products. B. ENTERPRISE PERFORMANCE Overall, nonfarm enterprise profits and productivity are low 81. Most enterprises are young, very small, and static. The average age of enterprise is six years. Only 1 percent of all enterprises employ more than three workers. In terms of employment within enterprises, there is very little growth: only 8 percent of firms have expanded their labor force since start-up, while about 3 percent have shed workers. 82. Few enterprises operate on a full-time basis. Furthermore, enterprise activity is highly seasonal and countercyclical with agriculture. Thus few households appear to specialize in nonfarm enterprise activities. Instead, it seems nonfarm enterprises are set up primarily as a complement to agriculture, providing an alternative source of income in periods when the level of activity in agriculture is low. Firms operate on average 8 months per year and 14 days per active month. 83. The estimated average daily profit is 5.6 Birr per workday, or less than a dollar per workday8. Profits are highest in Tigray, where the estimated daily return to working in a nonfarm enterprise is 8.7 Birr, and lowest in Amhara, where the corresponding return is 5.0 Birr. The average monthly profit in an active month is 55 Birr or US$ 4.5. The average annual profit, averaging across inactive and active periods, is 340 Birr, or approximately US$ 27. In fact, profits per workday are lower, on average, than the daily wage rate for casual agricultural workers ­ around 9 Birr in 2007. 8 Profit refers to sales revenue less material inputs and labor costs. -35- 84. Of course, there is a lot of heterogeneity across firms: some perform much better than the average, others much worse. Enterprises run by male-headed households are twice as profitable per workday as enterprises run by female-headed households, whose higher participation rates suggest they lack alternative earnings options. The relationship between the age of the household head and profitability is inverse U-shaped. Young entrepreneurs become more profitable over time. Beyond 40 years, profitability falls with age. 85. In addition, enterprises engaged in trading yield, on the average, higher returns than enterprises engaged in services. The high returns to trading activities could reflect arbitrage opportunities due to limited economic integration. Manufacturing enterprises yield the lowest returns. Mobile enterprises or those that operate close to a market are more profitable than others. Box 9: How do Enterprises in Ethiopia Compare with Tanzania? Rural enterprises characteristics in Figure 9: Rural Business Constraints in Ethiopia and Tanzania are similar to Ethiopia. Tanzania, 2005-2007 Some 28 percent of rural Market demand households in Tanzania reported that at least one member was working in a nonfarm business in Registration Finance 2005. This compares to about 25 percent in Ethiopia in 2007. Similarly, while the overall landscape of nonfarm enterprises in Tanzania is diverse, the Crime Transport predominant entrepreneurial activity of rural nonfarm enterprises is trading. Nonfarm enterprises in rural Tanzania are very small, heavily affected by seasonality, and Governance Utilities the majority is operated by one Rural Ethiopia Rural Tanzania person. While the rural nonfarm sector in Tanzania is equally a low- Source: 2007 RICS-AgSS; 2005 Tanzania RICS (Kidiane and return sector that is struggling to Loening, 2008). Perceived major business constraints on a scale compete in a difficult business from zero to 50 percent. environment, there are a number of marked differences. Tanzanian enterprises generate about US$1.5 on sales revenue per working day, compared to only US$ 0.6 in Ethiopia. The sector is also more dynamic, with about one-third of rural enterprises growing relatively fast. Due to a rapidly growing agricultural sector in recent years, limiting demand-side constraints, rural enterprise constraints in Tanzania operate mainly from the supply-side. Access to finance, road infrastructure and rural cell phone communication are correlated with enterprise growth. This contrasts with Ethiopia, where the biggest constraint is related to market demand. These findings are confirmed both with multivariate regression analysis but also by simply plotting perceived enterprise constraints. -36- Box 10: How to Identify Successful Small Enterprises? Some Stylized Facts for Ethiopia Identifying rural enterprises which have a dynamic potential and are relatively profitable is important for supporting the nonfarm sector. But given the large heterogeneity of small enterprises across sectors and space, a simple categorization of dynamic enterprises in Ethiopia is less straightforward than expected. The following typologies have some merit in describing aggregate patterns, but may not work well at the level of the individual enterprise because of exceptions. However, taken together they can yield some insights: · Firm being pulled into the nonfarm sector have higher potential for growth. Enterprises that are "pulled" into the nonfarm sector because of market opportunities and means to invest agricultural earnings are likely to have more dynamic potential. The RICS indeed suggests that many pull enterprises are typically more capital intensive, more productive and more profitable than push enterprises. Push enterprises tend to cluster in low-productivity manufacturing activities and are more likely to be operated by women, often lacking alternative earnings opportunities. But this categorization is not perfect: the data also finds that some push enterprises are highly profitable, while some pull enterprises are not very productive. · Firms selling tradable goods can generate local growth. Growth in output of non-tradables is ultimately constrained by local demand, while growth in the output of tradables is predominantly constrained by supply. Consequently, growth in tradables output can be an engine of economic growth, with positive multiplier effects on the non-tradables sector through consumption linkages, while growth in the non-tradables is unlikely to lead to sustained economic growth. While appealing, in Ethiopia only a small minority of goods are tradables sold outside the locality or Wereda of production. · Initial capital, location, and manager characteristics matter. Other stylized facts suggest that a number of enterprise characteristics matter for dynamic potential and enterprise profitability. Background papers for the Ethiopian Rural ICA find that the amount of capital is a particularly important determinant of enterprise profitability. Enterprises which are registered are far more profitable than enterprises which are not. Enterprises in rural towns tend to outperform enterprises in rural areas. Enterprises operated by women are less profitable. The education of the manager is convexly correlated with the productivity of the enterprises. Yet all of these enterprise characteristics only partially explain rural enterprise performance. · Enterprise productivity varies with activity. Returns, factor requirements and household characteristics vary strongly with activity of enterprises. In Ethiopia, enterprises engaged in trade are typically much more profitable than manufacturing or service firms. Managers of trade enterprises are typically better educated. Small manufacturing enterprises seem to provide income opportunities for those lacking other options. The sectoral composition of nonfarm enterprise activity also varies geographically, as well as with the level of economic activity in the community. Despite these patterns, the background papers for the Ethiopian Rural ICA, using econometric techniques, find that activity choice alone only explains a proportion of the total variation in enterprise performance, and other unobservable factors are important as well. Source: Summarizing evidence based on Beegle and Oseni (2008); Bardasi and Gethahun (2008); and Loening, Rijkers and Söderbom (2008). -37- The sector has grown through increased participation 86. Although existing enterprises tend not to grow their labor force the nonfarm enterprise sector as a whole has been growing in recent years due to increased household participation. Households mainly engaged in rural nonfarm activities rose from 4.5 percent in 1998 to 7.7 percent in 2006. Simple participation rates are more volatile, but also tend to show an increasing trend, rising from 23.0 percent in 1998 to 24.6 percent in 2006 (Figure 10). Figure 10: Ethiopia ­ Estimated Nonfarm Enterprise Participation Rates, 1998-2007 Percentage of households owning nonfarm enterprise 30 27.0 24.6 25 23.0 22.0 20 15 10 7.7 6.5 4.5 4.6 5 0 1998 2000 2004 2007 Nonfarm enterprise as main and secondary activity Nonfarm enterprise as main activity Source: 1998 -2004 WMS; 2007 RICS - AgSS. Numbers for 2007 are representative for Amhara, Oromia, SNNP and Tigray regions only. All figures exclude the Gambela region. 87. In 2006-2007 the gross entry rate, defined as the percentage of new firms in the population of firms, was 17 percent. This is high, indicating that every one in six firms in the sector has been operating for less than a year. Some of these firms may have been re- opened after a temporary seasonal closure, but most survey respondents would probably not consider a re-started enterprise as a new enterprise. This entry rate is therefore probably best contrasted with the permanent exit rate of 8 percent, and not the total closure rate, including seasonal closure and permanent exit, of 25 percent, which is very high. In either case, there is a lot of churning in the sector, which is consistent with the low average of firm age (Table 11). -38- Table 11: Ethiopia ­ Estimated Entry and Exit Rates, 2006-2007 (percent) Classification New Entry Permanent Exit Seasonal Closure All enterprises 17 8 17 Sector Manufacturing 16 7 13 Trade 20 9 21 Services 23 8 9 Region Tigray 15 15 16 Amhara 16 9 13 Oromia 21 8 18 SNNP 17 6 17 Source: 2006/7 RICS-AgSS. Numbers are approximations due to the use of recall data and seasonality. C. CONSTRAINTS TO OPERATIONS 88. Table 12 summarizes self-reported data on the most severe constraint to running and starting-up an enterprise. Credit, markets and to a lesser extent transportation are the most commonly cited constraints for all groups. However, as the following maps show there is some significant variation across and within regions. In Tigray, Amhara and Oromia lack of market demand is the most commonly cited constraint. However, in the SNNP region problems accessing finance are by far the most commonly cited. Table 12: Ethiopia ­ Major Business Constraints in Rural Areas, 2007 Market Access to Infrastructure Lack of Government Labor Classification demand finance and Transport technology regulation availability Perceived main constraints to enterprise operations and growth All enterprise owners 39 38 16 2 4 2 By sector Manufacturing 47 30 15 4 2 3 Trade 31 45 16 1 5 1 Services 39 38 16 2 4 2 By region Tigray 42 29 21 3 6 1 Amhara 44 28 17 3 6 2 Oromia 41 36 16 3 3 2 SNNP 33 49 13 >1 2 2 Perceived main constraints at enterprise start-up Enterprise Owners 23 47 10 8 2 11 Non-Enterprise Owners 24 47 10 7 2 10 Source: 2006/7 RICS-AgSS. -39- Map 3: Ethiopia ­ Market Demand as 1st Major Business Constraint by Geographical Zone, 2007 32°E 34°E 36°E ( ( 38°E . ,0 ' ,:'. .,42°E ,,,,, ,- ETHIOPIA ERITREA PERCEIVED CONSTRAINTS I TO NONFARM BUSINESS '(# ,,4 \. , _ / OPERATIONS AND GROWTH BY GEOGRAPHICAL lONE, 2007 "\ q MARKETS t7 Two '- 0 <10 '" ........ ,,'>. \.. 010-19 0 2 0 - 29 I ,.. \ ~ 30-39 r '/' SUDAN Four ,' AiFAR --' 0" ( ,/ / UJII(YU . ,60 0 .50-60 - - - - - lONE BOUNDARIES 40 - 49 (Ii , ,-,-,-,/ REGION BOUNDARIES ( INTERNATIONAL 80UNDARIES ~ 44°E Jlpga , .......... ,..' , \_ ... '- -.- - SOMALIA " lkgelrabur -- ,, -,..., - -- ......' _ . _ . _/ I J"'" " ",,,,'" ..... ----,..--"'''' \ \ Warder / / -_J/ S:OMA LI '., / , SOUTHERN NATIONS.~ " ,, Karalre \ ., .- / / NATlONALITES ANOPEOPLES , ;, , . Gade \........ .... ............ . / / , "' . . . . . _ . . - . . --- ____ . . .J '''\ / Afrkr ...... _-, / 0 50 100 150 200 ..,.,.,.."'7 _._ . _._~ V I tt I I 7 , , I I Lwen ,/ / ( o 50 100 150 Miles , /.~ ,.. '\ , - ,' n,;J mop WOJ produced by ~ Mop DeJ;gn Un;loF The world Bonk. =:;'7:~c;:;;z,rtO!oo:e~:1t::v~~hkoo ..... / ../: ,/ / 05 '~ KH H A .... I / Groop, ony jvdg.->I 00 ~ /ego! JIOM of OIly terr;lory, orony eMorJMlenl or ""~eptonc" of w<;n boundori", '" o - -' ---.r,- .r / ~ V\..7MI '4UM 34¢E 38¢E 40¢E / 42' E 44°E 46¢E 48°E : : : : : i 05 '-.. J'V -40- Map 4: Ethiopia ­ Rural Finance as 2nd Major Business Constraint by Geographical Zone, 2007 32°E 34°E 36°E ( ( 38°E . ,0 ' ,:'. .,42°E ,,,,, ,- ETHIOPIA ERITREA PERCEIVED CONSTRAINTS I TO NONFARM BUSINESS OPERATIONS AND GROWTH BY GEOGRAPHICAL lONE, 2007 "\ q CREDIT t7 Two " '" D. \.. ~ D 10-19 20-29 .,50 I ,.. \ ~ 30-39 Four ,' AiFAR r '/' . 40-49 / --' 0" ,/ SUDAN ( UJII(YU - - - - - lONE BOUNDARIES (Ii , ,-,-,-,/ REGION BOUNDARIES ( INTERNATIONAL BOUNDARIES ~ 44°E Jlpgo , .......... ,..' , \_ ... '- -.- - SOMALIA " lkgehabur - -- ......' -,..., , - , _ . _ . _/ I J"'" " ",,,,'" ..... ----,..--"'''' \ \ Warder / / -_J/ S:OMA LI '., / , SOUTHERN NATIONS. ~ " ,, Korohe \ ., .- / / NATlONALITES ANOPEOPLES , ;, , . Gode \........ .... ............ . / / , "' . . . . . _ . . - . . --- ____ . . .J '''\ / Afrkr ...... _-, / 0 50 100 150 200 ..,.,.,.."'7 _._ . _._~ V I tt I I 7 , , I I Lwen ,/ / ( o 50 100 150 Miles , /.~ ,.. '\ , - ,' n,;J mop WOJ produced by ~ Mop DeJ;gn Un;loF The world Bonk. =:;'7:~c;:;;z,rtO!oo:e~:1t::v~~hkoo ..... / ../: ,/ / 05 '~ KH H A .... I / Groop, ony jvdg.->I 00 ~ /ego! JIOM of OIly terr;lory, orony eMorJMlenl or ""~eptonc" of w<;n boundori", '" o - -' ---.r,- .r / ~ V\..7MI '4UJ-\ 34¢E 38¢E 40¢E / 42' E 44°E 46¢E 48°E (; 05 '-.. tJl -41- Map 5: Ethiopia ­ Transport as 3nd Major Business Constraint by Geographical Zone, 2007 32°E 34°E 36°E ( ( 38°E . ., ' ,:'. .,42°E , ,,,, --(I ETHIOPIA ERITREA REP, PERCEIVED CONSTRAINTS I TO NONFARM BUSINESS OF ,< YEMEN OPERATIONS AND GROWTH BY , GEOGRAPHICAL lONE, 2007 14"N \ q t7 TRANSPORTATION Two " '" ,,> Dt on ~ legel 110M of OIly terrilory, orany tp ~ ::: UGANDA -;: KENVA "","- "'''',., ., ' '- ' /' endor""""nt or ""~eptonc" of w<;n boundari", 8 __ OOC 32 E -L____________ ~~~~~ 3.N . ______ ~LL~________~3'r'E~__________~'f~~E______~~__~'~__________~~____~~~~~~~~~__~~____ .. I , ' ...... . f / 42 E ,UoE 46¢E 48°E ~ V. -42- D. OTHER FINDINGS ON ENTERPRISE PERFORMANCE 89. Very few firms in the sample invest in equipment or machinery. Only 19 percent of all firms have made any investment since start-up. Moreover, the firms that do invest typically invest only very small amounts. For the overwhelming majority of enterprises, the most important source of investment finance is non-agricultural sales. Agricultural sales are also an important source. Funds from financial institutions are not. Firms that started with higher amounts of initial capital, and older firms are more likely to invest in capital stock. This could be because, as time goes on, upgrading the capital stock becomes more important. Alternatively, it could be that young firms face higher uncertainty regarding the prospects of the enterprise, which may lead to caution in investment. 90. Factors which determine enterprise performance include the characteristics of the manager, the sector of enterprise activity, the performance of the agricultural sector, and the location of enterprise activity (Loening, Rijkers and Soderbom, 2008. See also Table 40 and Table 41 in the Annex 2). 91. Firms with a male head are more Figure 11: Ethiopia ­ Enterprise Sales Growth, productive than those with a female 2006-2007 head. Productivity initially declines with 60 additional years of manager's education, Percentage of firms reporting change but starts to increase after 5 years of 50 education. Manufacturing activities are in sales 2006 to 2007 among the least productive activities 40 while trading activities, such as 30 wholesale and retail, are very productive. 20 92. The prospect of a good crop 10 raises productivity among nonfarm enterprises, probably because of higher 0 Ethiopia Tigray Amhara Oromia SNNP local demand. Enterprises in rural towns are more productive than enterprises in Increase Non change Decrease more remote rural areas9. Source: 2006/7 RICS-AgSS. 9 For further information see Chapter 6 and Rijkers, Soderbom, and Loening, 2009. -43- E. SUMMARY 93. The survey data strongly suggest that market fragmentation due to remoteness is a Figure 12: Ethiopia ­ Changes in Number of Employees since Start-up, 2007 key impediment to the performance of the nonfarm enterprise sector. Market 2 fragmentation manifests itself most Average number of employees obviously in low and localized demand for nonfarm enterprise produce. 1 94. The lack of enterprise growth across the board could suggest that nonfarm enterprises are already close to their optimal size , despite operating at a very small scale. This is consistent with the idea that demand 0 for nonfarm enterprise products is limited Ethiopia Tigray Amhara Oromia SNNP and indicates that incentives for expansion At start-up Current may be lacking. Consistent with this, our production function estimates indicate that Source: 2006/7 RICS-AgSS. enterprises located in rural towns are significantly more productive than Figure 13:Ethiopia ­ Changes in Labor Days and Hired Workers, Since Start-up, 2007 enterprises located in other rural areas. These findings point to demand-side problems in other rural areas. Increase 95. By contrast, supply-side investment No change climate variables, such as telecom, water, electricity, land and buildings, security or Decline bribes, are not as strongly correlated with either productivity or investment. (Tables 40 0 20 40 60 80 100 and 41, Loening, Rijkers and Soderbom, Percentage of firms reporting change in 2008). Taken together, these findings Number of labor days Number or hired workers suggest that it is possible supply-side constraints do not "bite" --in the sense that removing the constraints would improve Source: 2006/7 RICS-AgSS. performance--if demand is low. 96. It seems that the combination of poor infrastructure and remoteness result in high transaction costs, as a result of which markets are small and highly localized. Consequently, demand for nonfarm enterprise products is low, which limits incentives to invest and expand and helps explain why most enterprises remain small. Policies facilitating the integration of markets would make nonfarm enterprises less dependent on the local rural economy, which may help these enterprises develop beyond supplying a small and volatile local market with low value-added products. -44- 5. RURAL ENTERPRISES, FOOD SECURITY, AND DISTRIBUTIONAL EFFECTS A. OVERVIEW 97. This chapter focuses on food security in rural Ethiopia and the role of nonfarm enterprises in reducing the effects of food shortage. Ethiopia is considered one of the most food insecure countries in the world. Within sub-Saharan Africa, it is one of the seven countries that constitute half of the region's food-insecure population (Feleke, 2005). 98. The nonfarm economy can be an important source of additional income for food insecure households. In a setting with limited agricultural potential or highly variable weather, income from nonfarm activities can augment and smooth income flows for rural households. At first sight, as evidenced in the previous chapters, it appears that a substantial number of nonfarm activities in Ethiopia only provide limited opportunities. But they could be very important from a food security point of view. This is especially relevant in Ethiopia where an estimated 4.6 million people periodically require emergency food assistance and as many as 7.3 million chronically food insecure people receive a cash or food transfer through the PSNP. 99. The focus of this chapter is on nonfarm enterprises, which constitute the largest share of nonfarm income opportunities. The analysis is divided into two parts. The first section covers the four main regions (Tigray, Amhara, Oromia and SNNP) accounting for about 90 percent of Ethiopia's rural population. The second section gives a more detailed analysis of the Amhara region, which is not representative for Ethiopia as such, but provides additional insights. The third section includes wage employment and assesses distributional impacts. The results show that nonfarm enterprises are associated with food security. In the Amhara region, the finding is concentrated amongst female-headed households. Nonfarm enterprise activity also tends to reduce rural inequality. B. FOOD SECURITY AND NONFARM ENTERPRISE IN ETHIOPIA Rural nonfarm enterprise activity is associated with food security 100. Ethiopia is considered one of the most food insecure countries in the world. There are many reasons why countries experience food insecurity. In Ethiopia, food insecurity can be attributed to low agricultural productivity and agricultural input market constraints. Further contributing to food insecurity is the exposure of households to shocks such as drought and variable rainfall. -45- 101. Responses to food Figure 14: Ethiopia ­ Enterprises Opened in the Last 3 insecurity in Ethiopia have Years Now Closed by Food Security Status of Wereda, conventionally been emergency 2007 food-based interventions. 45 However, since many households 40 Percent of enterprises are not "transiently" but rather 35 30 "chronically" food insecure, the 25 Food secure woreda aid has not been deemed effective. 20 Food insecure woreda This led to the initiation of the 15 PSNP by the Ethiopian 10 5 government in 2004/2005 with the 0 main objective of reducing Tigray Amhara Oromia SNNP household vulnerability, improving household and community Source: 2006/7 RICS-AgSS resilience to shocks, and breaking the cycle of dependence on food aid. The PSNP targets the poorest of the poor through providing predictable and timely employment in public works and direct support. In this sense, it is targeting chronic food insecurity (and poverty) rather than transient food insecurity, which continues to affect many food secure areas. 102. Even for households that Figure 15: Ethiopia ­ Households with Enterprise by are primarily engaged in farming, Food Security Status of Wereda, 2007 the nonfarm economy can be an 45 important source of additional 40 Percent of households household income. Especially in 35 settings with limited potential to 30 25 Food secure woreda expand agricultural productivity or 20 Food insecure woreda in the face of highly variable 15 weather, income from nonfarm 10 activities can augment and smooth 5 0 income flows for rural households. Tigray Amhara Oromia SNNP At first sight, it appears that a substantial number of nonfarm Source: 2006/7 RICS-AgSS activities in Ethiopia provide low income and appear stagnant. But they are very important from a food security point of view. 103. The presence and income from nonfarm activities can help households cope better with shocks and be more food secure. This suggests that even low-return nonfarm activities may prove to be important from a welfare point of view, although not necessarily a substitute for higher-return activities, such as wage labor. In food insecure rural areas, the nonfarm sector could potentially play a very important role in ensuring rural livelihoods. -46- 104. Rural nonfarm enterprise activity is positively associated with food security. Overall, nonfarm enterprises are more common in food secure Weredas10 than in food insecure ones. Households in food insecure Weredas are less likely to currently have a nonfarm enterprise or to have operated one in the last 3 years. These differences are statistically significant overall, and within three regions. In the SNNP region, however, enterprise ownership is higher among households in food insecure Weredas. Multivariate regression analysis found that remoteness and other controlled for factors do not explain these differences (Beegle and Oseni, 2008). Controlling for the distance of the community to the nearest market and all-weather road, and for socio-demographic characteristics, food security continues to be positively and significantly associated with nonfarm enterprise activity. However, a main factor determining entry into the nonfarm sector is favorable rainfall, which proxies for strong agricultural performance. Business constraints are slightly more severe in food insecure Weredas 105. Access to markets and credit are the main constraints for both food-secure and insecure Weredas and for all types of enterprise. Table 13:Ethiopia ­ Business Constraints by Food Security Status of Wereda, 2007 (in percent) Households in Households in Main perceived constraint to enterprise All food secure food insecure operations and growth households Wereda Wereda Electricity, Telecommunication, Water 2.7 2.2 3.2 Transport 12.9 14.1 11.3 Interest rates, ability to pay back loans 17.0 17.8 15.8 Access to markets, low demand 38.7 36.7 41.4 Government (corruption, restrictive laws) 2.0 2.8 1.1 Safety (criminality, theft) 0.8 0.6 1.1 Lack of technology, access to information 2.2 3.0 1.1 Registration and permits 0.7 0.8 0.5 Lack of financing or ability to borrow 21.1 20.4 22.0 Lack of knowledge 1.0 1.3 0.6 Other 1.0 0.3 1.8 Source: 2006/7 RICS-AgSS. 106. Perhaps because of a more challenging business environment, exit from nonfarm enterprise is more likely to occur in food insecure areas. Figures 14 and 15 show the gaps between food insecure and food secure areas in the four major regions. With the exception of Amhara, enterprises in food insecure areas are less likely to still operate. Overall, using multivariate regression analysis controlling for region, distance to markets, and distance to roads, food insecurity continues to be significantly associated with a higher probability of closure (Beegle and Oseni, 2008). In Tigray and Oromia, enterprises that started in food insecure areas were more likely to close. The likelihood that an 10 The third administrative layer in Ethiopia. -47- enterprise will close is exceptionally high in food insecure areas of Oromia, where almost 2 out of 5 firms that started in the last 3 years were no longer in operation. There are differences in the probability of firms closing by sector. Firms in retail, food and beverage production, and manufacturing have higher probability of closure in food insecure areas (Beegle and Oseni, 2008). Table 14:Ethiopia ­ Source of Start-up Capital by Food Security Status of Wereda, 2007 (in percent) Source of enterprise All Food secure Food insecure start-up capital during the last 3 households Wereda Wereda years Agriculture 59.2 61.2 56.8 Nonfarm self-employment 8.5 7.9 9.3 Wage or salary 1.2 1.3 1.1 Remittances 0.3 0.3 0.2 Sale of assets 0.6 0.7 0.6 Bank of cooperative loans 1.8 1.2 2.7 Family or friends 11.5 10.1 13.4 Private moneylenders 9.9 9.8 10.0 Other 6.9 7.7 6.0 Source: 2006/7 RICS-AgSS. Table 15: Ethiopia ­ Reason for Enterprise Start-up by Food Security Status of Wereda, 2007 (in percent) All Food secure Food insecure Reasons for enterprise start-up households Wereda Wereda Push (insurance) Household lost wage earnings 1.8 2.2 1.4 No access to agricultural land 11.2 13.0 8.9 Low or volatile agricultural income 29.3 28.8 30.0 Pull Means to invest agricultural earnings 47.6 47.4 47.6 Markets opportunity 2.7 2.1 7.3 Other Support from NGO or cooperative 0.1 >0.1 0.1 Advice from relatives/friends 2.7 3.0 2.2 Social and economic independence 1.8 1.1 2.8 Other 2.8 2.5 3.2 Source: 2006/7 RICS-AgSS. 107. The sources of start-up capital do not differ by food security status. Likewise, the main reasons motivating the firm start-up do not appear to differ across food insecure and food secure areas. Table 14 shows the source of start-up capital for enterprises which were operated in the last three years for all households and by food security status. -48- Income from agriculture is the main source of capital for over half of all enterprises. Other common sources of income are family or friends in the community, nonfarm self- employment income, and private moneylenders. Very few households receive start-up capital from wage or salary income or loans from banks. This is indicative of low participation in wage employment for rural households and inadequate access to credit. Table 16: Ethiopia ­ Sector of Enterprise Start-up by Food Security Status of Wereda, 2007 Sector of enterprise start-up All Food secure Food insecure during last three years households Wereda Wereda Distilling of spirits, wines and other food manufacturing 19.1 25.6 * 10.8 Hotel and restaurant 5.6 4.5 * 7.0 Retail trade via stalls and markets 26.1 25.7 26.6 Retail (not stalls and markets) 20.0 14.3 * 27.3 Wholesale trade 5.7 4.2 * 7.7 Manufacturing (excluding food and beverage) 14.8 16.1 ** 13.0 Services 6.8 7.1 6.4 All other: grain milling, transport, communications, real estate, business, public services, other personal services 2.0 2.5 * 1.3 Source: 2006/7 RICS-AgSS. Statistical significance in difference between food insecure and secure Weredas: *= 1%, **=5%. 108. There are strong differences in sector of firms operated across food secure and insecure Weredas. Households in food secure Weredas are significantly more likely to participate in manufacturing (food and non-food) and service sectors than those in food- insecure Weredas. The most common nonfarm activity for households in food insecure Weredas is retail trade, which in most cases is low income. The retail sector, especially via stalls, usually requires less capital. Thus poorer households, which are more likely to be food insecure, are expected to participate more in such activities. The statistical significance of the differences between the nonfarm sector for food secure and insecure Weredas are not a function of remoteness of communities (Beegle and Oseni, 2008). After controlling for distance to markets and roads, the difference between food secure and insecure Weredas remains. 109. Household nonfarm enterprise activity in food insecure areas is associated with increases in agricultural income in the last 3 years. About 45 percent of these households reported that their agricultural income has increased in the last 3 years. This is consistent with a positive correlation between wealth and income shares from enterprises, as indicated in the next section for Amhara. But it is not possible to disentangle the causality to identify if nonfarm enterprises result in high farm income (say, through facilitating the purchase of improved inputs) or if higher agricultural earnings result in households venturing into nonfarm self-employment. As noted by Woldehanna and Ellis (2005), farm and nonfarm earnings "reinforce each other for improving livelihoods." -49- Table 17:Ethiopia ­ Estimated Agricultural Income Change in Last 3 Years, 2007 (in percent, as reported by households) Food secure Weredas Food insecure Weredas Households with Households with Households with no Households Agricultural no enterprise enterprise enterprise with enterprise income change in last 3 years in last 3 years in last 3 years in last 3 years Increase 37.3 34.4** 33.7* 44.7* No change 26.7 27.9 25.3 20.8* Decrease 36.0 37.7 41.0* 34.6 Source: 2006/7 RICS-AgSS. Statistical significance in difference between categories, relative to base category of households in food secure Wereda with no enterprise: *= 1%, **=5%. C. ADDITIONAL INSIGHTS ON FOOD SECURITY AND RURAL ENTERPRISES FROM AMHARA REGION 110. A number of interesting findings emerge from the Amhara region. A more in- depth analysis was conducted for the Amhara region where more information from the RICS-Amhara survey was available on households with and without nonfarm enterprises (Beegle and Oseni, 2008). There is also unique information on households in remote rural areas versus those in small rural market towns. Women are less likely to have an enterprise in food insecure areas than secure ones 111. Regardless of location and food security status, female-headed households are more likely to operate an enterprise. In the Amhara region, rural households in food secure areas are generally more likely to operate a nonfarm enterprise, confirming the earlier findings. Controlling for multiple other factors (such as education, female headship, wealth quartile, household demographics, distance to services, and past shocks), the difference in likelihood is very close to the rate found for the four major regions. Among households in Amhara, the probability of having a nonfarm enterprise is about 4 percentage points lower if the household is in a food insecure Wereda. But this effect is concentrated solely among female-headed households. For male-headed households, there is no difference in likelihood. -50- Table 18: Amhara ­ Characteristics of Households With and Without Enterprises, 2007 (in percent) Rural remote Rural town Food secure Weredas Food insecure Weredas Food secure Weredas Food insecure Weredas No With No With No With No With Variables enterprise enterprise enterprise enterprise enterprise enterprise enterprise enterprise Head's education: no education 79.5 68.9*** 78.4 77.2 40.6 53.1** 27.4 66.0*** Head's education: some primary 18.1 29.0*** 20.6 21.0 17.5 35.2*** 7.7 23.1*** Head's education: more than primary 2.4 2.1 1.0 1.8 41.9 11.7*** 64.9 10.8*** Female-headed household 16.9 36.2*** 21.5 41.9*** 31.5 43.3** 42.8 58.1** Household head is migrant (has not always lived in this community) 24.6 36.1*** 22.1 32.8*** 83.5 74.9** 71.6 51.1*** Kilometer (km) to nearest food market 7.9 6.8** 11.6 7.9*** 0.5 0.4 0.8 0.9 Crop shock in 2006 21.2 16.8 41.6 34.5* 7.2 3.6 17.3 18.5 Crop shock in 2005 16.0 13.1 37.1 32.9 3.3 2.3 11.5 9.2 Livestock shock in 2006 13.2 9.3 21.8 16.6 0.5 4.7** 5.9 4.6 Livestock shock in 2005 10.5 5.1** 18.7 6.7*** 1.9 1.7 2.6 2.7 Other economic shock in 2006 4.3 7.0 5.5 10.8** 4.9 10.9* 26.3 23.1 Other economic shock in 2005 3.5 5.7 5.7 5.6 3.0 4.2 10.0 7.4 Illness or death in 2006 18.1 13.8 24.0 27.1 13.4 24.7** 15.8 28.0** Illness or death in 2005 14.7 9.5* 20.5 25.5 9.0 15.0 5.2 17.9*** Community has a bank 11.9 9.9 12.3 13.6 19.8 13.3 15.5 7.7* Community has microfinance institution 23.8 18.2 32.6 32.9 84.0 86.7 91.1 93.2 Number of households 1,007 177 1,012 139 129 175 136 134 Source: 2006/7 RICS-Amhara. Statistical significance in difference between households with and without enterprises: ***= 1%, **=5%, *=10%. With exception of km to food market, all variables are binary (1 if true, else 0). -51- 112. The prevalence of shocks is higher for female headed households in food insecure Weredas. Generally, households in food insecure areas are more likely to be female headed and have higher rates of economic shocks than their counterparts in food secure areas. However, the prevalence of past economic shocks (crop loss, livestock shock, illness or death of household members) is not different between households with and without enterprises. One exception to this is among rural market towns: households with enterprises have higher rates of mortality and morbidity. Households with illness or death of family members may subsequently have to start nonfarm enterprises as a means of survival. 113. The likelihood of migration differs by food security status. In rural areas, the likelihood of being a migrant household (defined as the household head not having "always lived in this community") is higher among enterprise households regardless of food security status. This might reflect lack of access to land among rural migrant households. But in rural market towns, migrant households are more common among households without enterprises regardless of food security status. Overall, there are more migrant households in food secure Weredas than in food insecure Weredas regardless of enterprise ownership. Table 19:Amhara - Income Sources by Food Secure Status of Wereda, 2007 (in percent of households with income by source) Rural remote Rural town Food Food secure insecure Food Food Food Food Income source secure insecure secure insecure Agriculture (crops and livestock)a/ 87.0 83.7** 92.2 87.2*** 21.8 24.5 Wages and salaries 10.4 11.7 8.5 10.7* 34.3 29.2 Nonfarm enterpriseb/ 16.8 14.9 13.7 12.9 55.8 47.8* Social benefits 6.1 63.9*** 6.3 65.7*** 3.9 34.0*** Gifts/remittances 15.6 13.3* 14.1 11.6* 34.0 41.0* Other (rent and pension) 3.3 2.8 2.3 1.9 15.9 17.0 Any income from nonfarm enterprise as well as income from agriculture or 11.9 8.9*** 11.3 8.4** 18.8 16.7 wages No income reported from any 2.4 0.9*** 2.5 0.8*** 1.3 0.5 category No income reported from farm, 4.2 7.1*** 3.5 6.1*** 12.4 22.8*** wages, and nonfarm enterprise Source: 2006/7 RICS-Amhara. a/ Agriculture income does not include household own-consumption and thus we do not have complete estimates on total agricultural production. b/ Households with any nonfarm enterprise are considered to have income from the enterprise, whereas other income sources are based on questions about any income from source in last 12 months. Statistical significance in difference between food insecure and secure Weredas by income source: *= 10%, **=5%, ***=1%. -52- Better access to finance in food insecure Weredas 114. Availability of financial services is higher in food insecure Weredas. Contrary to what may be expected, households in food insecure areas are more likely to have a microfinance institution in their community than households in food secure areas. Moreover, in Amhara, access to financial services is significantly associated with enterprise start-up for households in food insecure areas, but not for households in food secure Weredas. These findings might reflect the success of government policies and NGO initiatives aiming to promote credit and microfinance schemes to rural households in food insecure areas. Enterprise ownership is associated with higher total household income 115. Social benefits are a major Figure 16: Amhara ­ Share of Enterprise Income, source of income in food insecure Among Household with Enterprises, by Food Security Weredas (Table 19). For the Amhara Status of Wereda, 2007 region, households in food insecure 1 Weredas are much more likely to .8 receive some assistance from social Share of income from NFE benefits, which reflects the targeting .6 approach of the food and food-for- work programs, which is, at the first .4 stage, at the Wereda level. It was also found that households in food insecure .2 areas are significantly more likely to 0 rely exclusively on un-earned income 4 6 8 10 12 (Log of) total household income sources, specifically social benefits Insecure woredas Secure woredas and remittances. This is the case in both rural and rural market towns. Source: 2006/7 RICS-Amhara. 116. About ¾ of these social Figure 17: Amhara ­ Income Distribution by Household benefits are from food-for-work Category, 2007 activities. For households in rural 9 [Log of] Total HH Income market towns, where less than 15 percent of income comes from 8 farming, nonfarm enterprise earnings are more significant. Even in food 7 insecure areas, 25 percent of total income comes from these activities. 6 This is more than the portion of Rural, Insec, Ent Rural, Sec, No Ent Rural, Sec, EntEnt Town, Insec, No Ent Rural, Insec, No Ent Town, Insec, Ent Town, Sec, No Ent Town, Sec, EntEnt income from social benefits or remittances. The contribution of these enterprises is more pronounced in the Rural/Food Security/Enterprise Category food secure rural towns, where, on The figure shows mean, 25%tile, and 75%tile. average, 39 percent of income comes from nonfarm enterprise activities. Source: 2006/7 RICS-Amhara. -53- 117. Enterprise ownership is associated with higher total household income. The share of enterprise income decreases in wealth in rural food secure Weredas, but increases in rural market towns. In rural food secure areas in the Amhara region, the contribution of enterprise income in terms of income shares varies by wealth levels. In remote rural areas, the share of enterprise income in total income is decreasing in wealth. The opposite is true for rural market towns. In food insecure areas, there is no significant variation between towns and more remote rural areas in the share of enterprise income as a function of wealth. 118. Among poor households with an enterprise, income from the nonfarm enterprise constitutes a larger share of total household income in food secure Weredas. That is, for poor households in food insecure areas, enterprise income is not as significant, which could be partly a reflection of higher income share from social benefits. Overall, with the exception of food secure towns which have better opportunities for rural wage labor, enterprise ownership is associated with higher total household income, both at the mean as well as at the 25th percentile and 75th percentile of the distribution. 119. In rural food-secure areas, a higher level of education of the household head is strongly associated with enterprise ownership. However in rural market towns, this pattern is reversed. Households operating a nonfarm enterprise in rural towns are more likely to have heads with no schooling. This suggests that for rural market towns in food secure areas, the better educated household heads tend to be engaged in the wage sector. However, even controlling for farm income (as a proxy for land ownership since land holdings are not available in the data sets from the RICS-Amhara) in rural market towns is negatively associated with enterprise operation. This suggests that enterprises are more a fallback option for less educated households in rural market towns. Households with more education have alternative income opportunities. D. DISTRIBUTIONAL EFFECTS OF NONFARM ENTERPRISES Low-return nonfarm activities prevail 120. This section considers distributional effects of nonfarm activities, both in the rural wage and enterprise self-employment sector. Rural nonfarm activities are presented by their relative return. Nonfarm activities yielding an amount higher than the average monthly agricultural income, as calculated from the WMS 1998, receive the label "high- return" activity. If the revenues earned from such activity are below this threshold, they are considered as "low-return" nonfarm activity11. Such a breakdown reveals that low- return nonfarm activities prevail in Ethiopia. In 1998, according to Olapade (2007), the overall nonfarm participation calculated from rural income sources is in the order of 17 percent. Of these, 14 percent are classified as low-return activity and only about 3 percent 11 A few drawbacks need to be mentioned. The latest nationally representative income and expenditure survey in Ethiopia is from 2005. However, the 2005 HIECS income module has not been released to the public. Similarly, income data from the 2000 HIECS suitable for the present analysis is not available. The 1998 WMS does not directly furnish information on individual participation in nonfarm activities, but participation can be deduced from household's income sources. -54- as high-return (yielding higher incomes than agriculture). Public wage employment in rural areas accounts only for 2 percent. 121. The participation in high-return nonfarm activities is strictly increasing with household wealth. In Table 20 the participation rates by per-adult expenditure are tabulated. This is a first step in analyzing the importance of nonfarm activities for different wealth strata. It shows that, regardless of the income quintile, agriculture has a high participation rate of between 83 and 91 percent. The lowest rates of farm participation are observable in the poorest and in the richest quintile. The low agricultural participation of the richest quintile is attributable to the access of households to high- return nonfarm activities (5 percent) and public employment (5 percent). Table 20:Ethiopia ­ Participation of Households in Income-generating Activities by Expenditure Quintile, 1998 (percent) Per adult Nonfarm self and wage-employment activities Public wage equivalent Agriculture employment expenditures All Low-return High-return and other 1-low 82.9 20.7 19.3 1.5 1.1 2 87.9 18.2 16.1 2.1 1.5 3 89.6 15.7 13.5 2.3 1.5 4 90.6 14.4 11.5 3.0 2.2 5-high 84.9 17.2 12.2 5.1 5.0 Source: 1998 WMS. 122. The low agricultural participation in the bottom quintile is offset by a high participation in nonfarm activities (21 percent), predominantly low-return activities. Nonfarm participation declines with increasing expenditure: from 21 percent in the poorest quintile to 14 percent in the fourth quintile. For the top quintile, one observes an increase in nonfarm participation (17 percent). The participation in low-return activities shows a similar picture of a decrease from the bottom to the fourth quintile followed by an increase in the top quintile (12 percent). The participation in high-return activities is strictly increasing with expenditure. 123. A shift from low to high-revenue activities occurs as the household wealth level increases. Table 21 shows that the income structure by expenditure quintiles follows a U- pattern with regard to the share of income from non-agricultural activities. Households with the lowest expenditure have the highest share of nonfarm income in total income (16 percent). This share decreases with increasing per-adult expenditure to 10 percent for the fourth quintile. It is only the top quintile that has an elevated share of 13 percent. Breaking down income from nonfarm activities by high- and low-return activities shows that the top expenditure quintile, compared to the other quintiles, has a relatively low share in low-return activities and the highest share of income generated from high-return activities. This finding suggests access to high-return activities is more open to wealthier households. The share of income from high-return activities is relatively unimportant for the four lowest expenditure quintiles. -55- Table 21:Ethiopia ­ Households Income by Source and Expenditure Quintile, 1998 (percent) Per adult Nonfarm self and wage-employment activities Public wage equivalent Agriculture employment expenditures All Low-return High-return and other 1-low 70 17 15 2 13 2 76 13 11 2 11 3 79 11 9 2 10 4 79 10 7 3 12 5-high 73 13 8 5 14 Source: 1998 WMS. 124. The share of income generated from agriculture shows patterns of an inverted U. The poorest households have the lowest share of income from agriculture with 70 percent. This share increases until almost 80 percent for the quintiles three and four and declines as expenditure reaches the top quintile. This pattern suggests that households in the lowest quintile pursue nonfarm activities as a survival strategy to supplement agricultural income, while households in the top quintile are able to complement or abandon agriculture for nonfarm activities more lucrative than farming (Loening, Rijkers, and Söderbom, 2008). 125. Nonfarm activities are important for younger, female-headed, and landless households. Olapade (2008) shows that nonfarm activities are more important for households with young heads, mostly in the low-return activities. The reason for the low- income share generated by means of agriculture by younger household heads can be attributed to their difficulty in obtaining any land or sufficient land for livelihood generation. This might force them to fall back on nonfarm activities with mostly low- return character. Similarly, a lower share of agricultural income is generated by female- headed households compared to their male counterparts. The low share of agricultural income is off-set by income from low-return nonfarm activities. Landlessness is rare in rural Ethiopia. But for those without access to land income from non-agricultural activities nonfarm activities, both in the high and low-return sectors, appear to be a refuge. Rural nonfarm activity decreases inequality 126. Gini estimations are applied to two different populations. Table 22 shows the results of the Gini-decomposition, following the methodology proposed by Lerman and Yitzhaki (1985), for the total sample and for a sample restricted to households engaging in nonfarm activities only. As one would expect for a farm economy such as rural Ethiopia, with a coefficient of 0.69, agriculture is the most important and most equitably distributed income source. Nevertheless, an increase in agricultural income increases inequality in both samples (0.016 and 0.060), all else equal. This might seem surprising at -56- a first glance, but the fact that participation in agriculture is the lowest among the poorest quintiles supports this result12. 127. In contrast to agriculture, rural non-farm income only accounts for 10 percent of total income and 8 percent of inequality and, unsurprisingly, has a high Gini-coefficient of 0.93. Even though the elasticity is relatively low (due to the low overall incidence of nonfarm activities), the elasticity between non-farm income and inequality is negative for both samples (-0.014 and -0.064). An increase in non-farm income reduces inequality. This is consistent with the descriptive results. Participation in non-farm activities is relatively higher in the poorest quintiles, so an increase in income from this source is likely to benefit this group and decrease overall inequality. The results suggest that increasing access to non-farm activities, especially among disadvantaged groups, is not only a pro-poor development policy, reducing agricultural dependence, but also reduces inequality. Table 22: Ethiopia ­ Gini-Decomposition by Income Source, 1998 Complete sample (including Restricted sample (only households agriculture and all other income engaging in nonfarm self and wage sources) employment) Share of Gini-Index Source Share of Gini-Index Source total by income elasticity total by income elasticity income source of total income source of total Income source inequality inequality Agriculture 0.78 0.69 0.016 0.40 0.78 0.060 Nonfarm self and 0.10 0.93 -0.014 0.51 0.59 -0.064 wage employment Public wage 0.08 0.93 -0.015 0.07 0.92 -0.003 employment Other sources 0.05 0.98 0.013 0.02 0.99 0.006 Total income 0.62 0.54 Source: 1998 WMS. 128. These findings are in line with other evidence. Restringing their analysis for the Oromia region, the largest state in Ethiopia in terms of both area and population, van den Berg and Kumbi (2006) find that entry barriers are for nonfarm activities are low and the general growth of the sector will benefit the poor. Opportunity-led (high-return) activities are likely to have a low effect with regard to poverty reduction as they are mostly performed by wealthier parts of the population. Survival-led (low-return) activities are likely to decrease income inequality as they provide the poorest with additional income sources. Stimulating growth of the nonfarm sector could therefore be achieved without compromising equality. 12 This finding reflects the result of statistical analysis which looks at the likely impact of an increase in agricultural income assuming everything else remains constant. In reality it may not be the case that everything else does remain constant, for example the policy environment or other factors that affect income inequality may change. -57- E. SUMMARY 129. The analyses presented here show that there are some limited differences between enterprises operated by households that are in food secure areas compared to those in food insecure areas. Households with non-farm enterprises are more likely to be located in a food secure Weredas. Food security remains positively associated with non-farm enterprise activity when we control for geographical factors such as distance to markets and road, and for socio-demographic characteristics. In the Amhara region, this finding is concentrated amongst female-headed households; that is female headed households in food insecure areas are much less likely to have an enterprise than those in food secure areas. 130. Non-farm participation is more important for poorer households who derive a higher proportion of their income from it. The results show that an increase in non-farm earnings leads to a small decline in overall inequality. This is not surprising since non- farm activities are less important for richer households. The results suggest that increasing access to non-farm activities, especially among disadvantaged groups, is not only a pro-poor development policy, reducing agricultural dependence, but also reduces inequality. -58- 6. THE UNTAPPED POTENTIAL OF RURAL TOWNS: FINDINGS FROM A RURAL-URBAN COMPARISON OF ENTERPRISE PERFORMANCE A. OVERVIEW 131. Rural and urban firms operate in distinctly different investment climates. Rural firms operate in isolated and fragmented markets, selling almost exclusively to local markets, where competition is low, while urban firms serve relatively well-integrated markets, where competition is fierce. Urban firms also have much better access to utilities and better and cheaper access to credit. Rural firms consider markets, credit and transport as their major constraints, while access to credit and land, taxes, and competition are the most important problems for firms located in urban areas. Thus, a rural-urban comparison of enterprise performance provides a method to assess the impact of market integration and the investment climate on firm performance. B. COMPARING RURAL AND URBAN ENTERPRISE CHARACTERISTICS Urban firms are larger, more capital intensive, and more productive 132. Comparing informal urban and rural enterprises reveals large differences in size, factor usage, and total factor productivity (TFP). Urban firms are larger on average than firms in rural town and remote rural areas. Large urban manufacturing firms have roughly 29 employees on average, urban microenterprise have 3 employees, firms located in rural towns have 0.8, and firms located in remote rural areas have 0.6. Figure 18 demonstrates the extreme differences in the size distribution across rural and urban areas by plotting kernel densities on a log-scale for manufacturers and non-manufacturers. The density plots illustrate that there are virtually no large firms in rural areas, while large-scale activity is common in urban areas. Figure 18: Size Distributions, 2007 Manufacturers Non-Manufacturers .8 .6 .6 .4 .4 .2 .2 0 0 -2 0 2 4 6 8 -4 -2 0 2 4 6 Log of the full-time equivalent workforce Log of the full-time equivalent workforce Large Urban Urban Microenterprises Large Urban Rural Town Rural Town Rural (other) Rural (other) Source: 2006/07 RICS-Amhara and 2006 Ethiopian Enterprise Survey (EES) -59- 133. There are marked differences between urban and rural firms in the composition of the workforce. Rural nonfarm enterprises rely almost exclusively on unpaid household labor, while such labor only accounts for a small minority of the workforce in urban areas. In other words, rural enterprises provide self-employment opportunities, while urban enterprises provide wage labor opportunities. The vast majority of urban enterprises are exclusively managed by men, while most rural enterprises are headed by women. Managers of urban enterprises typically have at least secondary school education, while the overwhelming majority of rural enterprise managers have no education at all. 134. The sectoral composition of Figure 19: Distributions of Capital Intensity, 2007 enterprise activity differs across rural and urban areas, and is more diverse in urban Manufacturers areas. Processing of food and garments is a .25 more prominent manufacturing activity in .2 rural areas than in urban areas. Wholesale is .15 a more common urban activity. Moreover, .1 the activities urban firms engage in are often .05 technologically more sophisticated than the 0 activities of firms in rural areas. -10 -5 0 5 10 Log of the Capital-Labour Ratio Large Urban Urban Microenterprises 135. Informal urban firms use much more Rural Town Rural (other) capital and more material input, both in absolute terms and relative to the number of Non-Manufacturers people they employ. For example, the .3 median of the capital stock per worker for large urban manufacturing firms is more .2 than 50 times larger than the median capital stock per worker in remote rural areas. .1 Figure 19 illustrates these differences by plotting kernel densities of the capital-labor 0 -10 -5 0 5 ratio for rural and urban firms on a log scale. Log of the Capital-Labour ratio Large Urban Rural Town Rural (other) 136. These differences in factor intensity Source:2006/07 RICS-Amhara and 2006 EES are also strongly correlated with differences in scale ­ larger firms are more capital intensive and also use more inputs per worker. We find sizeable differences in factor intensity across rural and urban areas even among firms of a comparable size; the median capital intensity of urban microenterprises is approximately 15 times the median capital intensity of enterprises located in rural towns. -60- Box 11: Theoretical and Empirical Framework for the Rural-Urban Comparison Market integration can lead to aggregate efficiency gains because of economies of specialization. But what happens to the relative development of the rural and urban sectors in the economy is less clear. In a simple trade model two individuals produce and consume two goods ­ food and non-food products. If individuals living in rural areas are not able to trade, they spend most of their working hours producing food, regardless of their underlying skills. If they are able to trade, individuals whose skills are better suited to non-food production can specialize in that and buy food in the market. Without trade, the rural economy is close to a point of complete specialization in food production, and so the gains from further intensification of food production by the individual with skills biased towards food production will be modest. However, the gains from increased non-food production by the individual with these skills may be much larger, because his productive skills are now more efficiently employed. As a result, production of nonfarm goods increases relative to the production of farm goods. This suggests the nonfarm sector might gain more from market integration, in terms of positive output effects, than the farm sector. The implications of this for rural development in Ethiopia are potentially important. While it is true that at present the nonfarm sector in Ethiopia is not very large and not always very profitable it does make an important contribution. It could be that by integrating the rural market, performance in the nonfarm sector may rapidly improve. A related consideration is that the effects of productivity gains in the nonfarm sector, perhaps generated by technological improvements brought about by an improved investment climate, may be much larger if markets are well integrated compared to if they are isolated. Thus, better market integration and an improved investment climate can move in tandem to spur development and diversification of an economy that is dominated by agriculture. Technological progress may be much enhanced if accompanied by market integration and the returns to investing in the capital stock are likely to be much higher in well-integrated markets (since capital enhances the productivity of labor). There are several other mechanisms which could explain why market integration could spur asymmetric growth: increasing returns to scale in the production of nonfarm goods; technological spillovers or other forms of agglomeration economies; preferences resulting in demand less skewed towards food if incomes are higher; and backward and forward linkages (Haggblade et al. 2007). Such effects are probably important, in which case they probably enhance the result of asymmetric growth. Demand for food is likely to play an important role. If one of the effects of market integration is to raise individuals' incomes and this in turn lowers the relative importance of food consumption, then this will certainly enhance the pattern of asymmetric growth in nonfarm production documented above. Technological or pecuniary externalities may also be important. For example, better access to information, inputs and skilled labor resulting from market integration will probably benefit the nonfarm sector. It should be noted, however, that some forms of externalities, e.g. technological spillovers, are likely to be highest in technologically advanced economies, and so be of limited importance in rural Ethiopia. The rural data are from the 2007 RICS-Amhara. Basic features of the nonfarm enterprise sector in Amhara are similar to the nonfarm enterprise sector in the four major regions of Ethiopia. The urban data are drawn from the 2006 Ethiopian Enterprise Survey (EES), which covered 14 towns and cities located in 7 regions of Ethiopia, with approximately half of the data coming from Addis. The EES comprised three separate surveys; a survey of 360 manufacturing firms and a survey of 124 services enterprises, as well as a survey of 126 micro- enterprises. Enterprises in the former two surveys are referred to as "large" enterprises and were supposed to employ at least 5 employees, while firms in the microenterprise survey are referred to as "small" enterprises and were supposed to exclude firms with 5 employees or more. The sample of urban microenterprises exhibit similar characteristics to the rural enterprises: many firms were informal or unregistered family-run small enterprises, with high participation yet low profitability; many managers had low education and were young; and the market was predominantly localized. Source: Adapted from Söderbom and Rijkers, 2009; Rijkers, Söderbom and Loening, 2009. -61- C. RURAL AND URBAN ENTERPRISE PRODUCTIVITY Firms in rural towns are as productive as small firms in urban areas 137. Overall, urban firms are much more Figure 20: Distributions of Value Added productive than rural ones. The median value-added per full-time equivalent worker Manufacturers .5 in large urban manufacturing firms, US$ 1208, is almost 15 times as high as in rural .4 towns, US$ 83. Labor productivity is even .3 lower in remote rural areas. The relative .2 dispersion of productivity is much higher in .1 rural areas, indicating that there is less 0 competition, which may explain why -5 0 5 10 Log of Value Added per Worker unprofitable firms manage to survive. Large Urban Urban Microenterprises Differences in labor productivity are Rural Town Rural (other) strongly correlated with differences in the size distribution across rural and urban Non-Manufacturers .4 areas. .3 138. Regressions on pooled small urban manufacturers and rural manufacturers' data .2 reveal that although firms located in remote .1 rural areas are some 50-60 percent less productive than firms located in urban areas, 0 firms located in rural towns are as -5 0 5 Log of Value Added per Worker 10 productive as those in urban areas. The Large Urban Rural Town Rural (other) coefficient estimate on being located in a rural town is very similar to the coefficient Source: 2006/07 RICS-Amhara and 2006 EES. estimate on being located in another major urban area or even in Addis, indicating that the benefits of agglomeration are concavely related to city-size. In other words, productivity levels of firms in rural towns are not very different from those in urban areas, but firms in rural remote areas are much less productive than firms located elsewhere. D. RURAL AND URBAN ENTERPRISE GROWTH Firms in rural areas, even rural towns, very rarely grow 139. The large differences in the rural and urban size distributions suggest that the rural investment climate does not favor factor accumulation and growth. Comparing the average annual growth rate of workers in rural and urban firms indicates that this is indeed the case; whereas urban manufacturing microenterprises grow some 5 percent each year and large urban manufacturing firms grow an average 9 percent each year, the rural enterprise growth rate is less than one percent for enterprises located in rural towns and one percent for enterprises located in remote rural areas. In addition, rural enterprises are much less likely to invest, which is consistent with their lower capital intensity. -62- 140. Growth matrices of rural and urban manufacturing firms confirm that rural firms are mostly not growing, while there is a substantial movement across size categories in urban areas. In particular, the results reveal that a minority of currently medium and large-sized urban firms started as small firms, which indicates that small firms are capable of escaping their initial size category in urban areas, though the very smallest firms, 1-person enterprises, are least likely to do so. By contrast, all rural enterprises have remained small. Table 23: Ethiopia ­ Transition Matrix: Urban Manufacturing Firms, 2006 Size at Start-up Current Size (employees) (employees) 1 2-5 5-10 10-5 50-100 >100 Total 1 67% 25% 15% 0% 0% 0% 11% 2-5 17% 69% 60% 31% 11% 1% 38% 5-10 8% 4% 21% 24% 6% 1% 13% 10-50 8% 1% 4% 43% 53% 27% 23% 50-100 0% 0% 0% 2% 17% 11% 4% >100 0% 0% 0% 1% 14% 59% 12% Total 100% 100% 100% 100% 100% 100% 100% Source: 2006 EES. Table 24: Amhara ­ Transition Matrix: Rural Manufacturing Firms, 2007 Size at Start-up Current Size (employees) (employees) 1 person 2-5 persons Total 1 person 98% 30% 85% 2-5 persons 2% 70% 15% Total 100% 100% 100% Source: 2006/07 RICS-Amhara. -63- 141. Basic growth regressions using information on the age of the firm and its size at start-up found that firms in rural towns do not grow faster than firms in other rural areas, despite being more productive (Rijkers, Söderbom, and Loening, 2009). The poor growth performance of rural firms suggests that the costs of dynamic losses due to market fragmentation may be many times higher than the static losses. In addition, the fact that firms in both rural towns and remote rural areas do not grow suggests that better integration of rural towns into the economy at large--for example by fostering stronger rural-urban linkages and interconnecting rural towns with each other and with urban centers--may help to achieve dynamic gains from clustering of economic activity. E. SUMMARY 142. Comparing across rural and urban areas we find a substantial performance gap, with large differences in firm size, productivity and growth. Whereas a significant number of urban firms are very large, we find practically no firms with more than 10 workers outside urban areas. Focusing only on small firms, we find that enterprises located in rural towns record very similar levels of TFP to those of urban firm and much higher levels of TFP than enterprises located in remote rural areas--defined as any rural area that is not a rural town. Despite their similarities, however, it appears that firms in rural towns are less able to realize growth potential than urban microenterprises. Urban microenterprises display a healthy dynamism whereas very few firms in rural areas, even in rural towns, grow their workforce. In conjunction with the finding that there are only a few large firms in rural areas, this suggests that conditions in rural areas are not conducive to firm growth. 143. In sum, it seems the investment climate in rural towns can support comparable productivity performance of microenterprises, but cannot support comparable dynamic performance. This could be because the level of market integration in rural towns is not sufficient to generate incentives for firms to invest and expand, or possibly, because supply-side constraints present a more insurmountable barrier to growth in these areas. Overall, these findings suggest first, that rural towns should be a focus for development in rural areas; and second, that alleviating the barriers to growth in rural towns could potentially yield high returns if it releases the dynamic potential of small firms. -64- 7. POLICY OPTIONS FOR PROMOTING RURAL DIVERSIFICATION A. OVERVIEW 144. This chapter focuses on the policy and programmatic implications of the findings in previous chapters. Ethiopia's main strategy for rural development--as elaborated in the Plan for Accelerated and Sustainable Development to End Poverty (PASDEP)--seeks to address three over-riding challenges: (a) promoting growth within smallholder agriculture, (b) addressing food insecurity, and (c) creating centers of growth with strong linkages to the local economy. The Ethiopia Rural Investment Climate Assessment can inform the debate on alternative approaches to addressing these three key challenges. It also identifies opportunities for enhancing the potential contribution of the rural nonfarm economy. Summarizing the previous findings, the chapter first looks at the current limitations of the rural nonfarm economy in Ethiopia and explores options for enhancing the role of the sector. B. SUMMARY OF KEY FINDINGS FOR POLICY 145. Ongoing population growth and land degradation increase the need for income diversification strategies. The PASDEP considers the promotion of nonfarm enterprise activity as an additional catalyst for rural development, though in practice promoting nonfarm activities has had a limited role, partly because of limited knowledge of the sector in Ethiopia, where it is often believed that rural equals agriculture. 146. Ethiopia's nonfarm enterprise sector is sizable and significant. About 25 percent of rural households participate in nonfarm enterprise and participation rates range from only 20 percent in Amhara to 37 percent in the SNNP region. Nonfarm enterprise profits on average account for 42 percent of total income among households that run an enterprise. By implication, nonfarm enterprise income represents around 10 percent of aggregate rural household income. Nonfarm enterprise more often complements than substitutes for agriculture 147. Enterprise is predominantly part-time and complementary to agriculture. Despite high participation rates, very few households participate exclusively in nonfarm enterprise activity. Less than 3 percent rural households rely exclusively on income from nonfarm enterprises. The majority of nonfarm enterprises are run part-time, either in parallel with agriculture, or periodically as a substitute for agriculture to provide an alternative source of income in periods when the level of activity in agriculture is low. 148. Policies to promote rural income diversification in Ethiopia should take into account theses seasonal patterns. Seasonality may act as a constraint to rural enterprise growth: an ebb and flow of labor into the activity only when it is surplus to agriculture -65- may hamper continuity and ability to upgrade skills and specialize. Moreover, as it is often risky or not worthwhile to establish the business on a permanent basis, seasonality can drive entrepreneurs into informality. Finally, and here in particular in the manufacturing and construction sectors, seasonality often implies an additional need for short-term capital, which cannot be easily met. Nonfarm income is important for those lacking alternatives 149. Women are important actors in the sector and tend to rely more on nonfarm enterprise income. Female-headed households own nearly one-half of all enterprises. Yet, women head only one-fourth of households. This implies that almost every second household headed by a woman operates a nonfarm enterprise. Furthermore, nonfarm enterprise income tends to be more significant as a share of total income, or as the only source of income, for female-headed households. They are more likely to engage in nonfarm enterprise as a primary activity rather than a secondary complement to agriculture. 150. Women, predominantly single women, are more likely to be "pushed" into nonfarm enterprise because they face constraints in other domains, especially agriculture, and not necessarily because they are well positioned to exploit profitable market opportunities. Women tend to concentrate in activities with relatively lower revenues but also earn less than men do within the same sector. Although women's enterprises are smaller and less profitable than men's, they appear to offer an important opportunity for employment and income generation, especially for those in vulnerable situations such as single women and others without access to land. 151. Although the relatively high participation of women in non-farm activities indicates that they do not face disproportionately high entry-barriers, policy support to non-farm activities should take into consideration the gender-specific nature of those activities. In particular, women face certain constraints more intensely than men: access to water, low demand, access to informal credit, and fear of not repaying a loan. Some of these may relate to the sector of activity and generally small scale of activity, but overall suggest that women have greater difficulties than men in solving the basic operational problems of their enterprises. 152. Nonfarm enterprise is particularly important for poorer households. Similar to (and overlapping with) the case of female-headed households, the poorest quintile of rural households have highest participation in, and get the highest proportion of income from, enterprise activity. Analysis based on the WMS including nonfarm enterprise self employment and wage employment income found that an increase in nonfarm income has a small but negative effect on inequality. This suggest that promoting nonfarm activities, especially among disadvantaged groups, is not only a pro-poor development policy, reducing agricultural dependence, but also reduces inequality. -66- Average returns are low, but there is a lot of variation in performance 153. Overall, the profits from nonfarm Box 12: Allene, a Grain Trader from a enterprise are low. In fact, at 5.6 Birr profits Small Market Town per workday, profits are less than a dollar per Allene is a licensed trader from a small town workday and are lower, on average, than the situated at about 27 km away from the capital of a Wereda in Amhara Region, located along daily wage rate for casual agricultural an important trade route. The town has about workers. The average annual profit, averaging 10,000 residents. The town does not have the across inactive and active periods, is 340 Birr, status of municipality and infrastructure is precarious as no telephone line or electricity or approximately US$ 27. are available. Until 2000, he was a farmer and, like many 154. Of course, there is a lot of other farmers in the area, he managed to heterogeneity across firms: some perform accumulate capital from selling his agricultural much better than the average, others much products. In 2000, he had saved Birr 10,000 and decided to start-up his own business, worse. Enterprises engaging in trading on though he continues farming. He constructed average yield higher returns than enterprises a small mud house and started to trade grain. engaging in services. The high returns to He used to buy directly from farmers and sell to grain traders in the town. He became a trading activities could reflect arbitrage successful trader and expanded his business. opportunities due to limited economic He subsequently constructed a small integration. Manufacturing enterprises yield warehouse and started to buy grain from local traders and sell it to wholesalers. Using the lowest returns. Better performing sectors brokers, he used to sell grain up to Addis are those which require more significant start- Ababa. The brokers' commission is up to 3 up capital. Mobile enterprises or those that percent of the sales or two Birr per quintal. The profit margin Allene gets is on average operate close to a market are more profitable Birr 23 per quintal of grain. Allene mentioned than others. Enterprises in rural towns perform that up to 2003 he had been very satisfied better than those in remote rural areas. The with the results of his business. performance of local agriculture affects In 2003, a new tax system was enforced and productivity, probably because of an increase he was required to pay Birr 42,000 - what he called "a very unfair amount." He paid the tax in local demand. but decided to abandon grain trade. He gave up his grain trade license and started a 155. There is a lack of growth and transport business. In 2005, there was a tax dynamism from within the sector, although reform, resulting in a reduction of taxes. Local authorities allowed the traders to apply for a there is high churn and increasing recovery of previously paid taxes. Allene participation. Most enterprises are young, very applied and recovered Birr 14,000 from the small, and static. Very few firms invest and previously paid amount. Allene restarted trading grain. grow. Average capital stock is in the region of Allene mentioned that the town does not have US$ 16. Only 1 percent of all enterprises the status of municipality, which often employ more than three workers and only 8 hampers businesses to get land for building percent of firms have expanded their labor premises and use their property as collateral for loans. He expressed concerns about the force since start-up. Despite this it seems that subjective way the taxes are still calculated. nonfarm enterprises are close to their optimal He also mentioned that cooperatives are size. engaged in grain trade. He feels that it has become difficult to compete. Markets are small and fragmented Source: Bakker (2007). 156. The main constraints to growth are on the demand side. Self-reported data on the most severe constraint to running and starting-up an enterprise indicate markets, credit, -67- and to a lesser extent transportation, are the most important for all groups. Market demand is the most commonly cited constraint to running an enterprise, and is much more frequently cited in Ethiopia than in Tanzania for example, where due to a rapidly growing agricultural sector in recent years, demand-side constraints are limited and rural enterprise constraints operate mainly from the supply-side. Access to credit is the most common constraint for starting-up an enterprise. 157. The survey findings and econometric analyses support the notion that demand- side constraints are severe: · Markets are small and localized. For example, more than 90 percent of entrepreneurs walk to the market and very few firms sell to customers outside their own community. · Enterprise sales are also strongly correlated with the agricultural performance of local and adjacent communities. The reason appears to be that demand for nonfarm products is much higher when agricultural performance is strong. In addition, uncertainty regarding agricultural performance limits incentives to invest, at least in the short run. · Firms in rural towns perform better than those in remote rural areas. 158. It seems that the combination of poor infrastructure and remoteness result in high transaction costs, as a result of which markets are small and highly localized. Consequently, demand for nonfarm enterprise products is low, which limits incentives to invest and expand and helps explain why most enterprises remain small. 159. Policies facilitating the integration of markets would make nonfarm enterprises less dependent on the local rural economy, which may help these enterprises develop beyond supplying a small and volatile local market with low value-added products. Supporting market integration through the promotion and development of small market towns is a particularly promising policy option. C. POLICY OPTIONS FOR THE NONFARM ENTERPRISE SECTOR 160. The Rural ICA has not identified binding supply-side constraints which severely limit the growth of the nonfarm sector. There are some investment climate problems, in particular in access to finance, transport and infrastructure, and to a lesser extent dissemination of technology. These issues and potential interventions are considered below. However, it appears that in the market environment faced by nonfarm enterprises these constraints do not "bite" and the returns to alleviating them may be limited. 161. Rather, the Rural ICA has found that low demand--due to small and fragmented markets, and volatile demand vulnerable on the performance of the agriculture sector-- are the major constraints to nonfarm enterprise, limiting returns and incentives to investment. On this issue there are two clear conclusions: (a) the nonfarm sector cannot be seen in isolation from agriculture; and (b) the promotion and development of small market towns is a promising area for intervention. -68- The nonfarm sector cannot be seen in isolation from agriculture 162. Whilst nonfarm enterprise is secondary to agriculture for most people, it is a crucial alternative for others. This suggests that a balanced approach to rural development is needed which capitalizes on the linkages and complementarities between the sectors. 163. This analysis has found that profits from agriculture are the major source of start- up capital for nonfarm enterprises, that income from agriculture is a major source of consumption demand for nonfarm enterprises, and accordingly the performance of the nonfarm sector is affected by the performance of agriculture. This study has not looked explicitly at the impact of the nonfarm sector on agriculture but others have found that an increase in nonfarm income raises agricultural output and productivity because cash from nonfarm activities is used to buy agricultural inputs such as fertilizer (Woldehanna, 2000) 164. From an overall policy perspective, the analysis highlights the need for a more balanced approach to promoting food security in Ethiopia. Currently, the focus is on revitalizing agriculture through investments in land rehabilitation and enhancing farming opportunities (through support to livestock investments, adoption of improved farming technology, and diversification to high value crops) together with a better-managed transfer system to households facing food shortages. 165. Where possible, policymakers should capitalize on the complementarities between agriculture and the nonfarm enterprise sector. It is likely that policy reforms that benefit nonfarm enterprises also benefit the agricultural sector and vice versa. Better access to credit, upgraded transport facilities and improved insurance, for example, would benefit farmers and entrepreneurs alike. Moreover, enhanced agricultural performance is likely to stimulate the performance of nonfarm enterprises, while improved off-farm performance might stimulate agricultural growth, by acting as a "pull" factor. 166. On a more general level, to the extent that rural nonfarm enterprises are part of agricultural input and output markets or agricultural service delivery, their efficiency will support smallholder farming depending on the contribution of such services to improved agricultural performance. The analysis of the RICS provides little evidence of this linkage. Very few rural nonfarm enterprises are part of agricultural input and output markets and participation in rural service delivery is at best insignificant. Instead, most nonfarm activity is in production or trade for local consumption. This may be because production linkages are weak and there is some evidence in the literature that this is the case. For example, based on an analysis of nonfarm enterprises in Tigray, Woldehanna (2000) shows that nonfarm activities are strongly related to population density while weakly related to farm income, and argues that production linkages are weak because purchases of agricultural inputs and marketed surplus is low. 167. This is a key policy issue. Clearly, the development of the agro-food processing system and the integration of smallholder farmers into this system are important for their growth. This is dependent on the efficiency of agricultural markets and such systems may be weakened if inter alia key rural actors (nonfarm enterprises that interface with farmers in this system) are absent, inefficient or face high transaction costs. Policy on agricultural -69- market development should therefore be based on an understanding of the role of different actors, including rural nonfarm enterprises, along the agro-food marketing and processing chain, and action taken accordingly--to develop appropriate support institutions and mechanisms that encourage the contribution of all actors to a vibrant supply chain. 168. The PASDEP recognizes that integration and interdependence between the agricultural and industrial sectors play a key role in the country's economic development and bringing about socio-economic transformation. However, the linkages between the two productive sectors have remained very weak, and the industrial base of the economy has continued to be very limited. The on-going ADLI strategy, designed to address the underlying structural problems, targets these critical objectives. 169. Possible actions include: Continued emphasis on agricultural development as a major pre-requisite for interventions in support of the rural nonfarm sector. Policies to promote rural entrepreneurship need to take into account the inter- relationships with agriculture and heterogeneity of the rural nonfarm sector. Interventions should aim to maximize spillover from related support (for example extension). Development of small towns and infrastructure 170. The promotion and Box 13: Small Towns, Great Significance: Institutions development of small towns as Shaping Rural Enterprise Development in China centers of marketing, commerce, and In China, a dynamic rural nonfarm enterprise sector has service delivery is an area of been a major contributor to the country's remarkable growth. In India, the growth in nonfarm enterprise output intervention which would support and employment has been rather stagnant. What can development of both the agriculture explain the observed patterns? Tracing the development and nonfarm sectors. Others have for more than 20 years, Mukherjee and Zhang argue that the differences are due to the institutional system in both argued that small towns are important countries. for rural development. Dercon and Regulations initially intended to protect small enterprises Hodinott (2005) argue that small in India may have hindered their growth compared to the towns are key to improve welfare of more spontaneous experience in China. In the planned rural Ethiopians. Woldehanna (2000) area, protection was mainly on the state-owned enterprises in China. With the success of agricultural suggests that rural towns act as a reforms in the early 1980s, agricultural productivity focal point in the development of the increased dramatically, channeling surplus to the rural economy and are essential to development of local rural enterprises. Since then, China gradually reduced protection, facilitated migration to ensure adequate economic and social small towns, and has adopted a fiscal decentralization infrastructure to develop demand for policy, providing strong incentives for local governments high value goods. to develop rural township and village enterprises. Facing tough competition, local governments must be 171. For the nonfarm and innovative and rural enterprises must be competitive to survive in the market place. As a result, and benefiting agriculture sectors alike, small towns from a policy promoting market linkages, the rural serve as centers that can link itinerant enterprise sector gradually took the share of previously and small-scale rural enterprises with state-owned enterprises. often complex and far-flung trading, Source: Mukherjee and Zhang (2007). -70- administrative and service systems. For the nonfarm sector, the evidence suggests that growth of small towns would address fragmented markets, reduce transaction costs, accelerate specialization, and increase productivity. Similarly for the agriculture sectors, small towns would increase marketing opportunities, reduce transaction costs, and improve access to inputs. 172. Since the returns to market integration seem to be highest at the lowest levels of market integration, promoting rural market towns appears to be a good way to enhance the productivity of the nonfarm sector. This would mean making small towns a focus for investment in transport, power, water, and communications infrastructure. Improved transport infrastructure connecting small towns with their rural surroundings is key to integrating markets and reducing transaction costs. Moreover international evidence shows that support activities, banks, marketing and service centers, training centers, etc, locate where infrastructure is high (Binswanger, 1989). 173. Improved transport links between rural market towns and larger towns and cities is also an important consideration. The overall slow dynamic performance of rural nonfarm enterprises suggests that rural towns themselves might need to be better integrated into the regional and national economy to foster sustained growth. On the other hand, better transport links will lower the costs of distance and open up rural towns and remote rural areas to competition from the larger urban areas where firms benefit from economies of scale. Whilst this is desirable in terms of efficiency and growth, the location and distribution of the efficiency gains will be a concern. 174. This suggests that, for the rural poor to benefit, better transport links will not be enough. In terms of commerce and service delivery, small towns will need to have a strong enough offering that they are not bypassed with easier access to a larger center. In terms of supporting the competitiveness of rural enterprises, small towns will need public investment in other infrastructure such as working premises, agro processing and storage facilities, and marketing facilities (on a cost recovery basis). These local investments and other programs to address investment climate constraints discussed below will have greater impact if firms have access to larger markets. Thus, better market integration and an improved investment climate can move in tandem to spur development and diversification. 175. The PASDEP recognizes that inadequate road network and transport services have contributed to weak spatial integration, predominance of rural settlements in isolation from one another, and low economic activity. The Ethiopian Rural Travel and Transport Sub-Program focuses on reducing the travel and transport burden of the rural population by constructing road infrastructure, providing social and economic infrastructure facilities, and enabling the people to utilize the road infrastructure effectively. Plans for a Universal Electrification Access Program, expected to bring electrification to over 6,000 rural towns and villages and some 24 million within five years, will also open new opportunities for nonfarm enterprises. 176. Whilst focusing on large towns and cities, the Government's Urban Development Strategy is relevant to rural market towns. In particular, the fourth pillar of the strategy, to -71- promote rural-urban linkages, includes a Small-Towns Development Program, which will provide support services, such as development plans, basic services, and digital mapping to 600 small towns; preparing and providing management support resources for provision of basic services; and market infrastructure development in smaller towns. 177. Possible actions include: Stakeholder consultation and consensus on a regional pilot program to stimulate small market town development, private enterprise growth, and rural-urban linkages. Prioritization exercise for investment in transport infrastructure and other public goods in small market towns based on spatial economic analysis and any local economic and business development strategies. Some basic spatial master planning to prioritize and manage investment in infrastructure within rural towns. Improving access to finance in rural areas 178. Access to finance is identified by entrepreneurs and non-entrepreneurs as a major barrier to participation. Moreover, the very low levels of capital in firms suggest it is an important issue with scope for improvement and potentially high returns, especially if improving access to capital and increase access among the poor to higher value added activities. 179. Despite significant efforts during the past years, the rural financial markets in Ethiopia are still under-developed. Similarly to other developing countries, financial institutions find it difficult to operate in rural areas due to the high transaction costs involved. Coverage is therefore low although with the expansion of microfinance institutions it is slowly expanding. It is commonly estimated that banks, micro-finance institutions and multipurpose cooperatives cover less than the total demand. Microfinance institutions (MFIs) provide only a narrow range of financing products, focusing on agricultural inputs, short maturing loans and often on group responsibility, the latter not being favored among rural clients in Ethiopia. 180. An important factor limiting access to credit is the low capital base of the MFIs. The Development Bank of Ethiopia seeks to supports MFIs through the Rural Finance Intermediation Program to address this issue and there may be scope to expand this support. Another approach to help address this gap is to build grassroots institutions to expand outreach of financial services to rural areas. In addition to micro-finance institutions, rural savings and credit cooperatives are slowly emerging as providers of financial services in rural areas within Ethiopia. International experience suggests that such financial cooperatives can be sustainable providers of financial services and that have proved to be a good conduit to increase rural outreach, including to the poor--there is a long track record of external intervention, much of it positive in its impact. -72- Box 14: Rural Finance in Ethiopia: Limited Access and Variety of Products Rural microfinance in Ethiopia has grown significantly. In 2001, some 23 microfinance institutions had a total of 460,000 clients with an estimated outstanding portfolio of about Birr 300 million in loans and Birr 240 million in savings. By the end of 2008, the number of lending institutions rose to 29 and the total number of clients over 2.2 million. Outstanding loans rose to almost Birr 4.8 billion and Birr 1.8 billion in savings. Active promotion of microcredit and some changes in the regulatory framework helped foster the development, including allowing MFIs to offer 12 specific services, the elimination of the cap on interest rates charged by MFIs, removal of the Birr 5,000 limit on loan sizes, and the extension of the loan repayment period for up to five years. Anecdotal evidence also suggests that growth of the microfinance industry led to a reduction in informal credit and moneylender interest rates. But supply does not meet demand for microfinance. In spite of enormous growth of the microfinance industry, virtually every sector in Ethiopia continues to consider access to finance as major obstacle. The demand for financial services largely exceeds supply, with the majority of the rural population not having access to them. According to the 2004 WMS, some 87 percent of rural households never used any microfinance service. Similarly, according to the 2007 RICS-Amhara, only 22 percent of the rural population and households residing in small towns report access to microfinance. Absence of competition among MFIs and high demand for financial services are the primary reasons for lack of market analysis and new product development. Few institutions are present and rural clients have limited choice. MFIs are the dominant formal providers for credit to small enterprises. Semi-formal lending institutions such as Iquib (Rotating Savings and Credit Association) are traditional institutions and popular by small entrepreneurs. Multi-purpose cooperatives and NGOs are present, but often deliver financial services in a fragmented way. The Ministry of Agriculture and Rural Development has been a major financier of input credit to farmers. Financial institutions typically lack skills and implementation capacity. MFIs mostly offer the same products with little variation. Moreover, the provision of financial services through non-financial institutions or non-specialist cooperatives only provides short-term relief, which may not be sustainable. · Market analysis, product development, and scaling-up successful experiences are important. One of the main problems is the availability of very limited variety of financial products and services, which is a particular challenge for small enterprises. MFIs in Ethiopia largely focus on agricultural clients and are often not financially viable. Although individual lending is allowed, most of the loans MFIs offer are on group guarantee methodology and for a short repayment period. But small enterprises often prefer individual loans to group loans with longer repayment periods. MFIs typically place no emphasis on the marketing their products and services. MFIs need to evaluate customer needs, conduct market analysis, and offer innovative products or services. · Rural Ethiopia lacks deeper outreach of savings mobilization. Most of the MFIs offer two types of savings products: compulsory savings for credit customers, and individual voluntary savings. But the outreach of savings services in rural Ethiopia is typically poor. In rural market towns savings mobilization could be an attractive option because the capacity for resource mobilization is typically higher than in rural areas, and with reduced administrative costs. · Gradual foreign investment in microfinance may enable the development of the industry. The existing regulatory framework does not allow foreign direct investment in financial services. As a result, MFIs do not have access to foreign microfinance expertise, management skills, and cheaper capital. But allowing foreign competition would facilitate in bringing the best out of the institutions involved. The industry would stand to gain by having access to the best financial management, operational practices existing in the rest of the world, and to the cheaper capital available in international markets. One option may be to gradually allow foreign or NGO ownership in the MFI businesses. · Human resource development is important. A focus on providing training and business development services would enable MFIs to graduate into activities and financial products that are in demand. Linking rural markets better to major urban markets would expand the opportunities for rural nonfarm businesses, and help to a great extent in their income generation. In turn, this would also help the growth of MFIs, through increased credit demand from its customers. Source: Bakker (2007) and Ramaswamy (2008). -73- 181. One must keep in mind, however, that while in buoyant rural economies injections of credit can play a valuable role in enabling the poor to participate in growing market niches, in stagnant rural markets, enhancing access to finance may yield limited results, as it would merely encourage new entrants into an already constrained environment. Neither is credit the only factor for effective participation in the nonfarm sector. Its impact is often felt in conjunction with other constraints such as access to inputs, and limited business skills. Credit initiatives to promote rural diversification must therefore be accompanied by market development through the identification and delivery of a limited number of key missing ingredients along supply chains most relevant to the rural nonfarm economy. 182. Possible actions include: Review current efforts to improve access to credit in rural areas focusing on the need to increase coverage and to promote more flexible product lines. Invest in grassroots financial institutions and supply chains relevant to the rural nonfarm enterprise. Feasibility analysis for market potential of urban and semi-urban/rural mobile- banking taking into consideration infrastructure and regulatory constraints. Pilot for mobile-banking schemes in urban and semi-rural areas. Box 15: Rural Enterprise Support in Ethiopia: A Crowded Landscape The policy and institutional environment for nonfarm enterprises includes many actors at the federal, regional and local levels including the Ministry of Agriculture and Rural Development, the Ministry of Trade and Industry, the Federal Micro and Small Enterprises Development Agency and their local and regional offices. The non-commercial support system provides services on a no-fee basis to encourage and enable micro- enterprises. The main actors are the Micro and Small Enterprise (MSE) support centers and to a lesser extent NGOs. NGOs are predominantly active in food insecure Weredas. In an exploratory analysis undertaken by Bakker (2007) in two Weredas, Meket in North Wello and Burie in West Gojjam, one MSE support center focused exclusively on the Wereda capital while the other provided support to rural entrepreneurs because it was supported by Food Security Program funding. In both Weredas, the major reason for promoting MSEs is their capacity to create jobs through self-employment, especially for young people and women. Indeed the number of jobs created is a key performance indicator. There are no indicators relating to the growth and sustainability of the MSEs which seems to indicate that there is a lower emphasis placed on MSE's performance and sustainability. This imbalance in priorities may undermine identification and establishment of MSEs with real growth potential. NGOs are predominantly active in the food insecure areas. Besides the positive role in providing direct support to MSEs, the NGOs make an important contribution in building the capacity of the government staff to provide business development services and to foster the consolidation of the private sector. In food secure Weredas, where very few NGOs operate, the government staff more often lack the financial and technical skills necessary to support MSE development. The MSE support service is well represented at all levels and the coordination and communication between offices, and with other departments and institutions such as NGOs, is good. The Wereda and zonal offices have motivated young staff with good technical knowledge but weaker business development skills. The kebele extension agents, however, are overstretched and hindered by poor transport. Overall, the activities and services provided seem to be supply driven, reflecting policy objectives. A demand driven offering would probably include more training in business and management skills for new and established entrepreneurs. Source: Bakker (2007) and Mulugeta (2007). -74- Providing support to entrepreneurs 183. This section looks briefly at the institutional support arrangements for nonfarm entrepreneurs in rural Ethiopia and considers three mains areas of support to entrepreneurs: promotion of improved technologies; skills development; and support to clusters of similar businesses with growth potential. 184. The major actors providing business support in rural areas are government agencies, mainly the Offices of Agriculture and Rural Development and Regional Micro and Small Enterprises Development Agencies (MSE support centers). Some NGOs are active in providing business development and technical training (carpentry, masonry) but their scope is very limited. Of these institutions, it is the MSE support centers located at the Wereda and in some cases Kebele levels that have primary responsibility for nonfarm enterprise development. Agricultural Technical Vocational Education and Training centers and Farmers Training Centers also play a role in nonfarm skills development and creating opportunities for skilled labor to participate in rural nonfarm enterprise. 185. The impact of support services on the rural nonfarm sector has been minimal; perhaps because support is uneven and these institutions are new, under-funded, and focus on urban areas. Service delivery for both skills development and introduction of new technology is likely to remain in the public domain for the near future. Internationally, there are significant successes in public provision of services related to rural nonfarm enterprise, especially in the area of technology development and dissemination. However, less successful examples also abound. On balance, experience suggests that such efforts must: (a) focus on key widely produced products/services; (b) link with local input suppliers (importers, manufacturers, repair services) to ensure sustained and affordable access to the necessary inputs; and (c) provide short-term assistance in facilitating the transition of small firms to new technologies and possibly also to new marketing channels (Haggblade et al, 2007). 186. The type of technology applied in nonfarm enterprises and opportunities for innovation affect the costs of production and service delivery, competition, access to lucrative markets, and adherence to quality standards. Advances in technology within the nonfarm economy may take place through private innovation and adaptation of external technologies, or through promotion by external actors such as the Government or NGOs. A review of a large number of case studies by Haggblade and others (2007) document fewer instances of technological advance through private actors in countries where agriculture is at a low level as such regions offer fewer economic incentives for technological advancement. Nevertheless, there are some notable successes in promoting innovation in the nonfarm sector by NGOs and government technology institutes in such contexts, resulting in significantly increasing revenues to rural households in resource poor areas. 187. Support to groups of similar businesses affected by the same supply-side constraints is efficient and seems to be a promising area of intervention, especially for local NGOs (Haggblade et al, 2007). Support would probably need to focus on activities with market potential outside the immediate area and promotional efforts focus on -75- matching local resources to external, even international, consumers. Support would include supply chain reviews and problem solving on an activity by activity basis. In particular solutions may be found to collective action problems and facilitate group solutions such as machinery and equipment leasing or bulk buying of inputs. A comprehensive support program may require additional skills, resources and capacity building for the local MSE support centers and extension services, building on the experience of the current cluster development pilot project within the Federal Micro and Small Enterprises Development Agency. 188. The Rural ICA has not looked explicitly at the relationship between business skills, or indeed vocational skills, and the rural nonfarm economy but the overall impression created is that entrepreneurial skills are underdeveloped. Whilst the study has looked at the impact of education on the sector and found that additional years of schooling is positively associated with enterprise start-up and participation, the relationship between education and enterprise performance is more ambiguous. What can be said with confidence is that formal education remains at a very low level in rural Ethiopia and that generally the returns to education in rural areas are high, and are perceived to help relieve the pressure on agriculture to absorb all of the rural workforce by opening up other options, especially to young people. 189. Skills development for the rural labor force remains within the public domain with virtually no private training institutions targeting rural areas. Traditional apprenticeships in the nonfarm sector may constitute an important contribution but this is un-researched. Technical and Vocational Training Colleges, and various public training institutions for specialized services such as agricultural extension, veterinary services, and human health services. Of these institutions, it is the MSE support centers (mandated to serve both rural and urban areas) that have primary responsibility for delivering training to rural entrepreneurs among a host of other responsibilities. 190. Strengthening small and micro enterprises is explicit in the ADLI and in addition, there is a National Micro and Small Enterprise Development Strategy. In particular the strategy recognizes that MSEs are important in the context of Ethiopia's poverty reduction strategy as they are seedbeds for the development of medium and large enterprises (vertical integration), and because they absorb agriculturally under-employed labor, and diversify the sources of income for farming families (horizontal integration). 191. Possible actions include: Review of strengths and weaknesses and measures implemented by line ministries and regional governments. Establish a monitoring team to supervise agreed implementation arrangements by line ministries and regional governments. Consider extending the scope of extension services to include nonfarm enterprise. Consider developing local economic and business development strategies. -76- 192. Possible actions on the support provided include: Review of experiences by NGOs and public service delivery systems including cost-benefit analysis of interventions. Take successful experiences in delivery of services (skills development and advisory services, technology dissemination) to scale as appropriate. General market development efforts through the identification and delivery of a limited number of key missing ingredients along supply chains most relevant to the rural nonfarm economy. Considering gender implications in the provision of support 193. Investment climate and enterprise development policies should be mindful of the different needs and constraints experienced by women entrepreneurs. However, if targeted appropriately, the some of the highlighted program areas--access to finance, supply chain reviews, and skills development--appear to be particularly relevant. Targeting female entrepreneurs would be in particular of interest at the project level, considering government or donor supported investments that aim to enhance rural entrepreneurship. Addressing food insecurity through nonfarm enterprise 194. Setting up a nonfarm enterprise is a critical and effective household livelihood strategy, important in optimizing labor use among households unable to apply available labor and/or skills optimally in farming--due to lack of complementary resources such as productive assets or land, because of an inadequate mix in adult labor, or simply because of excess labor. This is especially relevant for food insecure households that tend to have small and often degraded land holdings, insufficient livestock (oxen, sheep and goats-- that have been disposed off in response to shocks) necessary for the mixed farming systems carried out in most Ethiopian highlands, and in some cases inadequate adult labor. The main limitations are the low opportunities for nonfarm enterprises in food insecure areas (even as a secondary activity) and the low level of profits generated. 195. The presence and income from nonfarm activities can help households cope better with shocks and be more food secure. This suggests that even low-return nonfarm activities may prove to be important from a welfare point of view, although not necessarily a substitute for higher-return activities, such as wage labor. In food insecure rural areas, the nonfarm sector could potentially play a very important role in ensuring rural livelihoods. 196. Policies seeking to address food insecurity in rural Ethiopia should consider the potential contribution of the rural nonfarm enterprise sector. Current support programs for food insecure rural households such as the PSNP provide an alternative livelihood-- essentially an additional income source to farming resulting from wage labor. The share of total income derived from nonfarm enterprise is relatively low in food insecure areas, particularly in the purely rural areas. But promoting nonfarm enterprise may offer a sustainable alternative. Consideration should be given to understanding why participation is currently lower in insecure areas, particularly among women; and access to external -77- markets not vulnerable on local agricultural performance. The recommendations above, particularly on the development of small towns and infrastructure, are relevant. 197. A topic for further study is labor-based safety nets and engagement in nonfarm enterprises. The share of social benefits in total income, which is predominantly cash/food-for-work transfers, is much higher than enterprise income in food insecure areas. This suggests that participation in the PSNP--which targets the poor households and has flexible demand for household labor --may well offer a better livelihood strategy than rural nonfarm employment. Nevertheless, the PSNP is a temporary mechanism and it is evident that nonfarm enterprises do provide complementary income source for poor households. 198. It is therefore important that policies seeking to address food insecurity in rural Ethiopia also consider the potential contribution of rural nonfarm enterprise. The PASDEP recognizes focusing on crop and livestock production alone may not entirely solve the problem of food insecurity in some areas. For such areas, income diversification through promoting nonagricultural activities is of paramount importance. Policy makers should explore further the role of the nonfarm economy in promoting improved welfare of poor, food insecure households, the interaction between labor-based safety nets and engagement in nonfarm enterprises. -78- ANNEX ANNEX 1: SELECTED SUMMARY TABLES Table 25: Ethiopia ­ Participation Rates, Industry Type, and Mean Age of Enterprises, 2007 Nonfarm Mean Industry type participation Age Manufacturing Trade Services Total % % % % % N Rural Ethiopia 24.6 36.4 52.1 11.5 100 6.1 Region Tigray 22.4 30.9 56.6 12.5 100 6.3 Amhara 18.2 45 42.6 12.5 100 7.3 Oromia 22.9 36.2 51.9 11.9 100 5.8 SNNP 36.6 31.9 57.8 10.2 100 5.6 Zones in Amhara North Gonder 15 53.4 35.5 11.1 100 8.3 South Gonder 10.6 52.5 40.7 6.9 100 7.7 North Wello 10.6 51.8 40.1 8.1 100 10.9 West Gojjam 16.2 53.8 33.9 12.4 100 7.5 Gender of Household Head Male 15.1 23.5 64.3 12.2 100 5.5 Female 40.8 50 37.8 12.3 100 6.4 Source: 2006/7 RICS-AgSS. -79- Table 26: Ethiopia ­ Percentage Distribution of Enterprises by Constraints that Prevent Operations and Growth, 2007 Electri- Tele- Water Trans- Financial Market Govern- Safety Tech- Registration Taxa- Labor city communications portation services s ment nology & Permits tion issues % % % % % % % % % % % % Rural Ethiopia 1.4 0.2 1.1 12.9 36.4 38.7 2 0.8 2.2 0.2 0.5 3.6 Region Tigray 2.8 0 7.4 10.7 22.2 41.5 1.1 3.8 2.9 0.5 0.3 6.9 Amhara 3.4 0.1 1.3 12.4 28 42.5 3.4 0.7 3.5 0.3 1.3 3.1 Oromia 0.5 0 0.2 15.3 35.7 40.6 2 0.5 3 0.2 0.1 2 SNNP 0.6 0.7 0.8 10.9 45.6 33.4 1.3 0.6 0.3 0.1 0.3 5.3 Zones in Amhara North Gonder 0.7 0.6 0 9.7 28.6 52.7 0 0.8 1.2 0.4 0 5.3 South Gonder 2.3 0 0 5.7 36.1 38.1 3.4 1.3 5.8 1.3 0.7 5.3 North Wello 5.6 0 1.2 7.6 20.1 60.8 0 0.6 2.9 0.6 0 0.6 West Gojjam 4.7 0 0 12 31.4 46.2 1.1 1.2 1.3 0 1.2 0.8 Industry type Manufacturing 2.4 0.4 2.4 9.6 28.8 46.9 0.7 0.4 4.1 0.2 0.4 3.9 Trade 0.2 0.2 0.1 15.8 43.5 31 2.9 1 0.8 0.2 0.6 3.6 Services 3 0 1.1 10.4 29.4 46.9 2.5 0.9 2.5 0.1 0.4 2.9 Gender of Household Head Male 1.6 0 0.3 15.6 34 36.9 3 0.8 2.4 0.5 0.7 4.2 Female 0.9 0.4 2.5 10.2 35.8 42.8 0.6 0.8 2 0 0.1 4 Number of employees 1 employee 1 0.2 0.8 11.8 36.8 40.5 1.6 0.7 2.2 0.1 0.5 3.6 2-3 employees 2.1 0.3 1.9 15.9 34.1 34.8 3.5 0.8 2.5 0.3 0.5 3.4 4-9 employees 5.7 3.9 0 21 23.7 42.6 0 0 0 3.2 0 0 10+ employees 24.4 0 0 21.7 11.9 42.1 0 0 0 0 0 0 Source: 2006/7 RICS-AgSS. -80- Table 27: Ethiopia ­ Percentage Distribution of Households by Constraints that Prevent Opening a Nonfarm Business, 2007 All households Any household Electrici Tele- Water Postal Trans- Financial Markets Govern- Safety Tech- Regis- Taxa- Labor Total member plan to ty commun service portation services ment nology ration & tion issues open a nonfarm ications Permits enterprise Yes No % % % % % % % % % % % % % % % % Rural Ethiopia 22.9 77.1 0.6 0.1 0.4 0 9.3 40.4 24.2 0.9 0.5 6.7 0.3 0.2 16.5 100 Region Tigray 17.7 82.3 0.9 0.1 1.3 0.1 8.3 45.3 20.9 0.8 0.5 5.6 1 0.2 15 100 Amhara 15.4 84.6 1.5 0.1 0.3 0 8.3 39.6 28.8 0.8 0.3 8.5 0.2 0.2 11.3 100 Oromia 22.6 77.4 0.1 0.1 0.1 0 12 39.8 23.3 0.8 0.6 7.8 0.2 0.1 15.3 100 SNNP 34.9 65.1 0.3 0.2 0.8 0 6.4 41 20.9 1 0.5 2.9 0.4 0.2 25.5 100 Zones in Amhara North 20 80 0.1 0 0.1 0 11.7 40.3 24.7 0.4 0.3 7.7 0.1 0 14.6 100 Gonder South 14.8 85.2 0.1 0.2 0 0 2.9 47.1 18.2 0.8 0.8 3.9 0.1 0.1 25.9 100 Gonder North Wello 17.1 82.9 1.5 0.1 0.1 0.1 5.7 43.6 24 0.5 0.2 5.3 0.3 0.2 18.6 100 West 11.6 88.5 0.7 0.3 0.3 0 10.5 49.5 28.6 0.8 0.3 5.2 0.2 0.1 3.6 100 Gojjam Gender of Household Head Male 22.3 77.7 0.7 0.1 0.3 0 10 40.2 24 1 0.4 7.4 0.3 0.1 15.5 100 Female 19.6 80.4 0.5 0.1 0.9 0 7.6 40 25.9 0.3 0.7 5.2 0.2 0.2 18.5 100 Source: 2006/7 RICS-AgSS. -81- Table 28: Ethiopia ­ Percentage Distribution of Enterprises by Main Reason for Starting an Enterprise, 2007 Obtain Househol No income to Support Advice d lost access support from from Social & wage to agric Low/volatile agricultural Market NGO/co- relatives/ economic earnings land agric income work opportunity operative friends independence Other % % % % % % % % % Rural Ethiopia 1.9 11.3 28.7 47 3.3 0.1 2.6 1.7 3.4 Region Tigray 3.7 17.9 26.2 41.1 7.5 0 0.8 0 2.8 Amhara 2.9 17.1 28.7 41.1 4.2 0.3 2.4 0.9 2.4 Oromia 1.2 10.1 23.8 51.9 2.3 0.2 3.4 1.1 5.9 SNNP 1.6 7.7 34.6 46.6 3 0 2 3.1 1.4 Zones in Amhara North Gonder 5.5 8 28.2 45.3 10.2 0.5 1.7 0 0.6 South Gonder 5.9 16.6 37.5 34.5 3.6 0 1.2 0 0.7 North Wello 1.7 9.4 30.6 49.1 1.9 0 0 6.7 0.6 West Gojjam 0.5 11.5 27.3 48.3 4.8 0 0.7 3.9 3.1 Industry type Manufacturing 2.3 12.2 26.5 49.1 3.1 0.2 2 1.5 3.3 Trade 1.6 10.4 30.7 46.4 3.2 0 2.9 1.7 3.1 Services 1.9 12.8 26.6 43.4 4.1 0.7 3 2.3 5.2 Gender of Household Head Male 1.1 9.3 29 50.3 3.4 0 3.5 1.9 1.4 Female 2.7 13.4 27.3 43.6 3.1 0.3 2 1.3 6.3 Number of employees 1 employee 1.8 11.9 29.5 45.5 2.8 0.1 2.4 1.9 4.1 2-3 employees 2.1 11.8 26.8 48.6 4.8 0.3 2.9 1 1.7 4-9 employees 0 5.5 20.7 62.9 0 0 0 0.9 10.1 10+ employees 0 0 0 87 0 0 13 0 0 Source: 2006/7 RICS-AgSS. -82- Table 29: Ethiopia ­ Percentage Distribution of Enterprises by Main Source of Start-up Capital, 2007 Nonfarm Bank or self- Wage or co- Family Private Agricultural employment salary Sale of operative or money income income income Remittances assets loan friends lenders Other % % % % % % % % % Rural Ethiopia 59.2 8.5 1.2 0.3 0.6 1.8 11.5 9.9 6.9 Region Tigray 47.2 15.7 2.8 0.5 2.6 10 8.4 8.4 4.5 Amhara 59.2 9.2 1.3 0.2 0.7 3.6 10.6 7.3 8 Oromia 60.9 8.1 1.1 0.5 0.5 0.8 10.4 8.6 9.1 SNNP 59.6 7.2 1 0.2 0.4 0.3 13.9 13.3 4.2 Zones in Amhara North Gonder 51.3 15.7 0 0.1 0.3 2.5 14.6 6.6 9.1 South Gonder 55.4 7.7 1.9 0 0 5.5 9.5 12.7 7.5 North Wello 66.3 3 1 0 0.6 6.3 11.6 2.5 8.7 West Gojjam 56.5 5.9 5.2 0 0.4 2.8 12.4 8.9 7.9 Industry type Manufacturing 61.3 9.4 0.9 0.3 0.8 0.6 10.4 8.9 7.5 Trade 58.9 7.5 1.1 0.1 0.5 2.9 13.4 11.3 4.4 Services 54.1 10.2 2.8 0.9 0.6 1 6.8 7 16.6 Gender of Household Head Male 65 7 1.6 0.7 0.7 2.5 9.7 7.3 5.6 Female 55.1 8.4 1.2 0 0.8 1.5 11.5 12 9.6 Number of employees 1 employee 58.5 8.2 0.8 0.1 0.5 1.8 12.6 9.8 7.8 2-3 employees 58.3 10.9 2.3 0.5 1.2 2.4 10.2 9.8 4.4 4-9 employees 77.9 1.3 0 0 0.8 1.3 0 0 18.7 10+ employees 88.6 11.4 0 0 0 0 0 0 0 Source: 2006/7 RICS-AgSS. -83- Table 30 : Ethiopia ­ Percentage Distribution of Enterprises Closure, 2007 Operating today Duration of Plan to reopen closed enterprises Yes No Yes No % % Years % % Rural Ethiopia 74.8 25.2 4.6 62.5 37.5 Region Tigray 69.4 30.6 4.9 42.7 57.3 Amhara 75.6 24.4 5.1 54.9 45.1 Oromia 73.2 26.8 4.6 63.5 36.5 SNNP 77.1 22.9 4.3 72.2 27.8 Zones in Amhara North Gonder 68.1 31.9 6.1 62.3 37.7 South Gonder 77.7 22.3 5.2 50.3 49.7 North Wello 76.6 23.4 8.2 38.1 61.9 West Gojjam 71.8 28.2 3.7 40.3 59.7 Industry type Manufacturing 79.5 20.5 7.2 57.2 42.8 Trade 69.8 30.2 3.3 67.7 32.3 Services 82.7 17.3 5.2 46.5 53.5 Gender of Household Head Male 75.5 24.6 3.9 62 38 Female 74.1 25.9 4.9 57.2 42.8 Number of employees 1 employee 74 26 4.3 64.3 35.7 2-3 employees 80.1 19.9 5.6 60.3 39.7 4-9 employees 88.2 11.9 17.3 37.9 62.1 10+ employees 58.5 41.5 5.3 28.5 71.5 Source: 2006/7 RICS-AgSS. -84- Table 31: Ethiopia ­ Enterprises by Number of Employees, Sales Growth, and Share of Profits in Household Income, 2007 Estimated share of Share of Perceived change in sales in past year household income enterprise sales Workers at Current Average from enterprise going to Category start-up workers sales Increase No change Decrease profits operating cost No No Birr % % % % % Rural Ethiopia 1.3 1.4 393 50.8 20.2 29.1 37.4 51.5 Region Tigray 1.5 1.7 447 45.5 23.8 30.8 38.6 37.3 Amhara 1.2 1.3 297 40.8 23.6 35.7 36.5 46.6 Oromia 1.3 1.4 365 55 16.9 28 41.1 49 SNNP 1.4 1.5 478 53.6 20.8 25.6 33.8 60.1 Zones in Amhara North Gonder 1.2 1.4 444 42.2 27.2 30.6 44.4 49.4 South Gonder 1.1 1.2 357 43 23.9 33.2 36 44.5 North Wello 1.3 1.3 348 40.3 18.9 40.8 40.8 40.6 West Gojjam 1.3 1.6 330 45 21.5 33.5 44.2 45.4 Industry type Manufacturing 1.3 1.4 169 48 25.2 26.8 37.8 45.5 Trade 1.3 1.4 579 54 17.5 28.6 35.6 57 Services 1.4 1.5 269 45.2 16 38.9 44.3 45.8 Gender of Household Head Male 1.4 1.5 567 58 17.5 24.5 38.3 53.1 Female 1.2 1.2 154 43 22.3 34.8 37.1 48.7 Number of employees 1 employee 1 1 321 50.4 20.8 28.7 37.5 49.9 2-3 employees 1.9 2.2 557 53.2 17.5 29.4 37.4 53 4-9 employees 4.2 5.9 1,267 47.1 5.9 47.1 45.8 65.4 10+ employees 15 30 1,473 53.9 0 46.1 19 51.7 Source: 2006/7 RICS-AgSS. -85- Table 32: Ethiopia ­ Average Distance to Agriculture Input and Output Markets and All-weather Roads, 2007 All households With enterprise Without enterprise Distance Distance Distance Distance Distance Distance to to Distance Distance to to Distance Distance to to Distance Distance markets markets to road to road markets markets to road to road markets markets to road to road in km in mins in km in mins in km in mins in km in mins in km in mins in km in mins Rural Ethiopia 8.1 78 11.1 103 7.1 66 10 90 8.4 82 11.5 107 Region Tigray 10 91 8.4 108 8.8 68 7 114 10.2 97 8.7 107 Amhara 9.1 88 13 126 8.2 75 12.5 110 9.3 91 13.1 129 Oromia 8.3 82 9.6 83 7.8 76 7.8 71 8.4 84 10.1 87 SNNP 6 54 11.5 104 5.4 48 10.5 94 6.3 58 12 110 Zones in Amhara North Gonder 9.9 94 20.5 218 7.4 72 15.7 154 10.3 98 21.4 229 South Gonder 9.4 94 13.8 137 8.9 88 10.5 105 9.5 95 14.2 140 North Wello 8.6 85 13.8 130 9.3 91 15.9 146 8.6 84 13.6 128 West Gojjam 8.1 85 12.2 127 6.9 73 12 119 8.4 88 12.2 128 Gender of Household Head Male 8.3 80 11.4 106 7.3 69 9.6 88 8.4 82 11.7 109 Female 8 75 10.9 100 7.2 66 11.1 100 8.5 82 10.7 100 Source: 2006/7 RICS-AgSS and 2006/7 RICS-Amhara. -86- Table 33: Amhara ­ Number and percentage distribution of Socio-economic Characteristics of Enterprise Owners Gender of household Zones head Amhara Urban Rural North South North West Gonder Gonder Wello Gojjam Male Female % % % % % % % % % Gender Male 43.6 36 46 38.7 48.7 53.7 42.7 70.6 4.2 Female 56.4 64 54 61.4 51.3 46.3 57.3 29.4 95.8 Age category Less than 25 12.7 16.7 11.4 13.1 8.6 5.1 18.1 12.7 12.6 25-34 26.3 24.8 26.8 23.8 28.6 28.8 27.2 31 19.4 35-44 29.7 25.1 31.2 35.9 28.3 21.6 25.7 32.2 26.1 45-54 17.1 20.9 15.8 14.4 16.9 20.3 19.3 11.3 25.4 Over 55 14.3 12.6 14.8 12.7 17.5 24.3 9.9 12.8 16.5 Education level No schooling 70.9 58.2 74.9 73.8 71.7 67.2 68 62.4 83.2 primary school 24.2 29.3 22.5 21.9 23 26.6 26.9 31.5 13.4 Above primary school 5 12.5 2.6 4.4 5.3 6.2 5.1 6.1 3.3 Source: 2006/7 RICS-Amhara. -87- Table 34: Amhara ­ Percentage Distribution of Enterprises by Start-up Capital Category Amount of start-up capital (Birr) Less More Mean Start- than 1,000- 5,000- than up capital 1,000 5,000 10,000 10,000 Total % % % % % No Amhara 87.9 9.8 0.8 1.6 100 596 Urban 85.4 10.3 2 2.3 100 808 Rural 88.7 9.6 0.4 1.3 100 524 Zones in Amhara North Gonder 89.8 9.2 0.4 0.7 100 402 South Gonder 93.7 6.3 0 0 100 233 North Wello 85.3 13.2 0.7 0.7 100 438 West Gojjam 82.7 11.4 1.8 4.2 100 1,175 Sector Manufacturing 95.5 2.5 0 2.1 100 470 Trade 77.7 20.1 1.3 1 100 775 Services 75.7 20.4 3.4 0.4 100 785 Gender of Household Head Male 81.2 15.1 1.3 2.4 100 884 Female 97.8 1.9 0 0.3 100 167 Source: 2006/7 RICS-Amhara. -88- Table 35: Amhara ­ Number and Percentage Distribution of Households by Source of 100 Birr in Case of Emergency, All Households All Households Source of 100 Birr in case of emergency Ability to raise 100 Loan/gifts Birr Sale of Sale of Sale of Bank Loan from From animal Sale of forest Household Own savings from family/ nonfarm product crops product assets cash account Equb Edir Bank friends enterprise Other No % % % % % % % % % % % % % Amhara 1,144,010 63.6 36.2 22.1 0.3 0.3 7.6 0.4 0.6 0.2 0.2 29.2 1.2 1.6 Urban 71,164 60.2 4.6 5.6 0.4 2.1 32.7 1.6 1.1 0 0.2 46 4.2 1.4 Rural 1,072,846 63.9 38.3 23.2 0.3 0.2 5.9 0.3 0.6 0.3 0.2 28.1 1 1.6 Zones in Amhara North Gonder 388,713 70.5 39.1 21.2 0 0.1 6.7 0.4 1.3 0 0.3 27.9 1.9 1.3 South Gonder 245,778 56.1 27.1 15.5 0 0.2 7.6 0.3 0.5 0.3 0 44.2 0.5 3.8 North Wello 201,957 56.6 52.1 13.3 1.1 0.8 6.3 0.8 0 0 0.3 24.2 0.8 0.4 West Gojjam 307,562 68.1 29.4 34.5 0.6 0.2 9.5 0.2 0.4 0.7 0.1 22.2 1.3 0.9 Gender of Household Head Male 960,205 69.5 37.9 23.1 0.4 0.3 7.7 0.4 0.2 0.2 0.2 27.1 1.3 1.2 Female 183,805 44.1 27.8 17.7 0 0.5 7 0.4 2.7 0.4 0.2 40.4 0.9 2.4 Source: 2006/7 RICS-Amhara. -89- Table 36: Amhara ­ Number and Percentage Distribution of Households by Type of Shock during the Last 12 months All Households Without enterprise With enterprise Rural Urban Zone Rural Urban Zone Rural Urban Zone South Gonder South Gonder South Gonder North Gonder North Gonder North Gonder West Gojjam West Gojjam West Gojjam North Wello North Wello North Wello Types of Shocks No % No % % % % % % % % % % % % % % % % % Food shortage due to Flood 122878 7.3 2290 1.9 4.4 11.2 6.7 6.2 7.3 2.0 4.6 11.4 6.4 6.0 7.8 1.9 3.9 9.8 8.8 7.1 Food shortage due to drought 227574 13.6 4672 4.0 18.4 16.5 12.6 3.0 14.2 3.9 20.8 17.4 12.8 3.0 9.2 4.1 9.5 10.0 11.3 2.9 Flood 130837 7.8 2979 2.5 5.2 10.0 4.2 10.2 7.8 2.2 5.3 10.3 3.4 11.0 7.6 2.8 5.0 7.6 10.3 6.2 Crop damage 304328 18.1 8963 7.6 15.5 21.9 17.1 15.8 18.5 9.5 16.9 22.2 16.4 16.9 15.7 5.8 10.1 19.1 22.8 10.2 Loss/death of livestock 281828 16.8 4566 3.9 17.4 16.2 13.6 15.7 17.4 3.4 19.7 16.3 13.6 17.1 12.7 4.3 9.1 16.1 13.9 8.7 Price shock 29667 1.8 6915 5.9 2.5 2.7 2.2 0.7 1.4 4.7 2.1 2.4 1.3 0.4 3.9 6.9 4.1 5.0 9.1 2.4 Loss of job Household member 18387 1.1 3244 2.7 0.8 1.0 1.0 2.1 1.2 2.3 0.8 1.0 1.0 2.2 0.6 3.1 0.9 1.1 0.7 1.9 Illness of Household member 324417 19.3 22017 18.6 19.3 19.8 17.9 19.8 19.3 13.0 19.0 19.9 17.0 20.0 19.5 23.8 20.3 19.6 24.6 18.8 Death of Household member 57432 3.4 4568 3.9 3.2 5.2 2.5 2.7 3.4 2.8 2.9 5.2 2.8 2.6 3.5 4.9 4.3 5.6 0.7 3.5 Other 47705 2.8 9303 7.9 4.6 4.5 1.6 1.4 2.6 7.4 3.8 4.5 1.6 0.8 4.5 8.3 7.3 4.9 2.1 4.3 Source: 2006/7 RICS-Amhara. -90- Table 37: Amhara ­ Number and Percentage of Households that Suffered from Food Shortages during the Last 12 months Total Without enterprise With enterprise Zone Zone Zone South Wello South Wello South North North North North North North Wello West Gojjam West Gojjam West Gojjam Gonder Gonder Gonder Gonder Gonder Gonder Months Urban Rural Urban Rural Urban Rural No % No % % % % % % % % % % % No % % % % % No Shortage 90,501 76.6 1,138,457 67.8 61.9 65.2 57.6 87.8 39.1 59.8 51.1 58.1 50.2 74.1 37.5 8 10.8 7.1 7.4 13.7 Experienced Shortage 27,714 23.4 541,217 32.2 38.1 34.8 42.4 12.2 60.9 40.2 48.9 41.9 49.8 25.9 62.5 92 89.2 92.9 92.6 86.3 Months of Shortage 02-Jan 10,194 36.8 217,105 40.2 41.6 36.1 43.7 34.3 43.9 39 39.3 34.5 43.5 38.7 32.6 46.1 47.5 48.2 45.5 22.3 04-Mar 10,510 37.9 244,643 45.3 41.1 51.7 44.7 41 27.2 47.3 43.5 54.4 45.4 39.1 44.2 34.9 34.9 31.8 38.1 46.1 06-May 4,131 14.9 46,464 8.6 7.3 9 9.6 13 17.3 8.7 7.5 8.2 9.1 16.2 13.5 7.9 6.7 15.3 13.6 4.3 08-Jul 1,179 4.3 5,766 1.1 2.2 0 1.3 0.6 5.1 0.9 1.8 0 1.1 0.8 3.8 2 3.4 0 2.8 0 10-Sep 464 1.7 8,514 1.6 2.8 0.3 0.4 3.7 0 1.1 2.8 0.1 0.4 0 2.7 4 2.9 1.9 0 13.5 12-Nov 1,236 4.5 18,068 3.3 5 2.8 0.3 7.5 6.5 3 5.1 2.8 0.4 5.2 3.3 5.2 4.6 2.9 0 13.7 Source: 2006/7 RICS-Amhara. -91- ANNEX 2: SELECTED RESULTS FROM REGRESSION ANALYSIS Table 38: Ethiopia ­ Probability of Rural Nonfarm Enterprise Ownership, 2007 (1) (2) (3) Any enterprise Any enterprise Any enterprise Explanatory variables last 3 years today today Food insecure Wereda^ -0.037*** -0.046*** (0.008) (0.007) Tigray^ (base=Amhara) 0.061*** 0.042*** 0.007 (0.015) (0.014) (0.014) Oromia^ 0.030*** 0.021*** 0.031*** (0.009) (0.008) (0.010) SNNP^ 0.161*** 0.140*** 0.130*** (0.010) (0.009) (0.011) Km to major agricultural market -0.006*** -0.004*** -0.004*** (0.001) (0.001) (0.001) Km to all-weather road -0.001*** -0.001** -0.001*** (0.000) (0.000) (0.000) Water Resource Satisfaction Index -0.002*** (WRSI)2005 (0.000) WRSI 2004 0.000 (0.000) WRSI 12-year average 0.002*** (0.001) Constant 0.266*** 0.207*** 0.206*** (0.010) (0.009) (0.033) Number of observations 14,095 14,072 12,515 Source: 2006/7 RICS-AgSS. Probit estimates, marginal effects, and standard errors are in parentheses. ^ indicates binary variables (=1 if yes, else 0). Statistical significance: ***=1%, **=5%, *=10%. WRSI is the water requirement satisfaction index of the Wereda for the crop season of that year; higher values indicate better rainfall levels and patterns. -92- Table 39: Ethiopia ­ Probability of Rural Nonfarm Enterprise Closures, 2007 (1) (2) (3) (4) (5) All areas: Food secure Food insecure enterprises Explanatory variables All areas All areas Weredas Weredas opened as of 2005 Food insecure Wereda^ 0.079*** (0.016) Tigray^ (base=Amhara) 0.020 -0.004 0.087** 0.041 0.064*** (0.035) (0.087) (0.043) (0.037) (0.021) Oromia^ 0.020 -0.049** 0.156*** -0.009 -0.031 (0.020) (0.024) (0.036) (0.022) (0.021) SNNP^ -0.034* -0.062** 0.029 -0.051** -0.060*** (0.021) (0.028) (0.034) (0.023) (0.019) Km to major agricultural market 0.001 0.001 0.003 0.003 -0.001 (0.002) (0.002) (0.002) (0.002) (0.001) Km to all-weather road -0.002** -0.002** -0.002 -0.003*** -0.001** (0.000) (0.000) (0.002) (0.001) (0.001) WRSI 12-year average -0.001** -0.006*** (0.001) (0.002) WRSI 2005 0.000 (0.001) WRSI 2004 0.004*** (0.001) Constant 0.229*** 0.267*** 0.230*** -0.001** 0.319*** (0.022) (0.026) (0.035) (0.001) (0.061) Number of observations 3,424 1,769 1,655 2,925 2,380 Source: 2006/7 RICS-AgSS. Probit estimates, marginal effects, and standard errors are in parentheses. ^ indicates binary variables (=1 if yes, else 0). Statistical significance: ***=1%, **=5%, *=10%. WRSI is the water requirement satisfaction index of the Wereda for the crop season of that year; higher values indicate better rainfall levels and patterns. -93- Table 40: Ethiopia ­ Determinants of Enterprise Profits, 2007 Explanatory variables Coefficient Standard error Household characteristics Household size -0.005 0.009 Age of household head 0.033*** 0.007 Age of household head squared /1000 -0.396*** 0.080 Household head is a male^ 0.483*** 0.051 Schooling of household head (years) 0.050* 0.026 Schooling of household head squared /1000 -3.671 2.424 Location Rural town^ -0.072 0.088 Distances Distance to all weather road (km) 0.041** 0.019 Distance to the food market (km) -0.041 0.025 Seasonality Activities seasonal^ -0.037 0.042 Activities Hotels and restaurants^ (base = food) 0.281*** 0.090 Retail trade via stalls and markets^ 0.075 0.085 Services^ -0.229*** 0.086 Whole sale trade^ 0.596*** 0.124 Transport services^ 0.180 0.195 Manufacturing^ -0.180*** 0.059 Grain milling^ 0.110 0.254 Other specialized services^ 0.210 0.861 Retail not stalls and market^ 0.177** 0.080 Region Tigray^ (base=Oromia) 0.359*** 0.081 Amhara^ -0.069 0.062 SNNP^ -0.347*** 0.059 Base of operation Inside residence^ (base=outside residence) -0.065 0.076 Market^ 0.528*** 0.089 Shop^ 0.521*** 0.098 Road^ 0.150 0.142 Mobile^ -0.208 0.166 Other^ 0.101 0.170 Number of observations 2,474 Source: 2006/7 RICS-AgSS. OLS estimates with robust standard errors clustered by enumeration area. ^ indicates binary variables (=1 if yes, else 0). Statistical significance: ***=1%, **=5%, *=10%. -94- Table 41: Amhara ­ Enterprise Cobb-Douglas Production Function, 2007 Explanatory variables (1) (2) (3) (4) Production factors Labor (log of days worked) 0.546*** 0.519*** 0.542*** 0.514*** (0.116) (0.116) (0.116) (0.116) Capital (log) 0.125** 0.123** 0.123** 0.121** (0.055) (0.054) (0.055) (0.053) Material inputs (log) 0.295*** 0.293*** 0.296*** 0.295*** (0.056) (0.052) (0.056) (0.051) Share of paid labor -0.317 -0.380 -0.276 -0.347 (0.314) (0.338) (0.321) (0.347) Sector Manufacturing^ (base=other) -0.706** -0.666** -0.700** -0.652** (0.320) (0.317) (0.316) (0.313) Food and beverages^ -0.463* -0.508* -0.463* -0.501* (0.271) (0.281) (0.270) (0.279) Grain milling^ -1.325 -1.117 -1.317 -1.071 (1.120) (1.202) (1.115) (1.194) Hotels and restaurants^ -0.215 -0.321 -0.216 -0.317 (0.360) (0.342) (0.358) (0.340) Retail trade via stalls and 0.483 0.350 0.477 0.339 (0.401) (0.405) (0.399) (0.401) Services^ -0.333 -0.342 -0.334 -0.338 (0.425) (0.437) (0.424) (0.435) Whole sale trade^ 0.375 0.355 0.368 0.356 (0.379) (0.366) (0.379) (0.367) Transport services^ 0.164 0.303 0.164 0.333 (0.689) (0.859) (0.694) (0.855) Characteristics of manager Manager's age -0.049* -0.044* -0.051** -0.046* (0.025) (0.025) (0.025) (0.024) Manager's age squared/1000 0.418 0.356 0.432 0.372 (0.267) (0.264) (0.264) (0.259) Manager's is male^ 0.489** 0.560*** 0.482** 0.551*** (0.197) (0.202) (0.197) (0.202) Manager's schooling (years) -0.170** -0.176** -0.170** -0.176** (0.085) (0.085) (0.085) (0.085) Manager's schooling 19.266** 17.550* 19.073** 17.113* (9.204) (9.013) (9.243) (9.027) Local Demand Mean WRSI 2006 0.106*** 0.101*** 0.106*** 0.101*** (0.032) (0.033) (0.032) (0.033) Geography Remote rural^ 0.009 -0.044 (0.392) (0.393) Location in rural town^ 0.565* 0.566* (0.290) (0.296) Distance financial institution -0.113 -0.123 (0.126) (0.127) Distance all weather road (log 0.183* 0.183* (0.100) (0.101) Distance market (log km) 0.037 0.062 (0.212) (0.214) Financial institution in 0.119 0.116 (0.216) (0.216) Continued on next page. -95- Proportion of firms in -0.306 -0.279 community not using electricity (unavailable) (0.307) (0.303) Opportunity cost of 0.202 0.219 labor Daily wage male casual (0.229) (0.230) worker in agriculture (log) Competition Between 1 and 5 -0.270 -0.276 competitors^ (base=no (0.253) (0.253) More than 5 competitors^ 0.205 0.200 (0.205) (0.205) Selection Correction Inverse Mills Ratio -0.007* -0.008** (0.004) (0.003) Constant -6.812* -7.071* -6.684* -7.056* (3.503) (3.836) (3.485) (3.828) Number of observations 384 384 384 384 Source: 2006/7 RICS-AgSS. OLS estimates with robust standard errors clustered by enumeration area. ^ indicates binary variables (=1 if yes, else 0). Statistical significance: ***=1%, **=5%, *=10%. -96- Table 42: Rural-Urban Comparison Production Functions, OLS Regressions on separate samples Large urban Small Urban Rural Large urban Small Urban Rural Sample Sector Specification Manufacturin Manufacturing Manufacturing Manufacturing Manufacturing Manufacturing g Baseline Baseline Baseline Baseline + IC Baseline + IC Baseline + IC coef/sd coef/sd coef/sd coef/sd coef/sd coef/sd Factors Log K 0.149*** 0.096** 0.213*** 0.156*** 0.087* 0.223*** (0.047) (0.048) (0.054) (0.047) (0.051) (0.053) Log L 0.761*** 0.850*** 0.854*** 0.749*** 0.795*** 0.857*** (0.087) (0.173) (0.177) (0.086) (0.196) (0.172) Activity Food and beverages 0.026 -1.038* 0.056 -1.034* (0.295) (0.553) (0.304) (0.532) Garments and -0.486 0.026 -1.006* -0.481 0.047 -1.012* (0.310) (0.253) (0.542) (0.312) (0.271) (0.522) Leather 0.021 0.980 0.002 1.117 (0.414) (0.859) (0.425) (0.856) Wood, furniture & -0.274 -2.144*** -0.281 -2.144*** metal (0.283) (0.667) (0.293) (0.639) Management Female management -0.008 0.017 -0.588** 0.017 -0.004 -0.466* (0.179) (0.308) (0.265) (0.180) (0.331) (0.270) Manager's schooling -0.089 0.565*** -0.146* -0.103 0.458* -0.125 (0.098) (0.217) (0.084) (0.097) (0.249) (0.082) Manager's schooling2 0.006 -0.032** 0.013 0.006 -0.025* 0.011 (0.005) (0.013) (0.010) (0.005) (0.015) (0.009) Constraints Credit -1.161** -0.494 -1.253*** (0.571) (0.944) (0.449) Transport 0.397 2.525 0.349 (1.090) (2.123) (0.480) Utilities -1.112 0.388 -0.261 (0.912) (2.113) (0.380) Geography Rural town 0.544*** 0.459** (0.211) (0.213) Constant 3.166*** -0.527 1.758*** 3.836*** -0.350 2.459*** (0.627) (0.763) (0.573) (0.744) (0.888) (0.586) N 301 53 294 301 53 294 R2 0.732 0.458 0.261 0.743 0.479 0.291 Adjusted R2 0.724 0.388 0.235 0.732 0.370 0.259 Median Solow shares Log K 0.10 0.24 na 0.10 0.24 na Log L 0.90 0.76 na 0.90 0.76 na Mean Solow shares Log K 0.15 0.31 na 0.15 0.31 na Log L 0.85 0.69 na 0.85 0.69 na Note: - .01 - ***; .05 - **; .1 - *; Robust standard errors in parentheses Source: 2006/07 RICS-Amhara and 2006 EES. -97- Table 43: Rural Urban Comparison Production Functions, OLS Regressions on pooled small manufacturing firms sample Factors (1) (2) (3) (4) (5) (6) coef/sd coef/sd coef/sd coef/sd coef/sd coef/sd Log K 0.183*** 0.226*** 0.217*** 0.215*** 0.215*** 0.223*** (0.058) (0.044) (0.047) (0.043) (0.049) (0.048) Log L 0.528** 0.890*** 0.785*** 0.795*** 0.743*** 0.748*** (0.212) (0.149) (0.157) (0.149) (0.162) (0.159) Activities Food and beverages -0.549 -0.494* -0.634* -0.591** -0.604 -0.496 (0.348) (0.276) (0.357) (0.281) (0.373) (0.371) Garments and textiles -0.520* -0.455* -0.540** -0.509** -0.493* -0.408 (0.270) (0.248) (0.270) (0.249) (0.285) (0.288) Leather 1.302* 1.338* 1.417* 1.448* 1.424* 1.638** (0.745) (0.749) (0.766) (0.764) (0.785) (0.791) Wood, furniture & metal -1.672*** - -1.710*** -1.670*** -1.709*** -1.633*** (0.523) (0.519) (0.507) (0.502) (0.511) (0.492) Management Female management -0.414* -0.416* -0.496** -0.493** -0.535** -0.437* (0.227) (0.221) (0.217) (0.215) (0.222) (0.225) Manager's schooling -0.084 -0.072 -0.105 -0.101 -0.103 -0.086 (0.066) (0.065) (0.066) (0.065) (0.066) (0.068) Manager's schooling2 0.007 0.006 0.008* 0.008 0.007 0.006 (0.005) (0.005) (0.005) (0.005) (0.005) (0.005) Rural town Rural Area -0.475 (0.334) Rural Area*Log L 0.414 (0.277) Rural Area*Log K 0.043 (0.079) Location Dummies Addis 0.306 0.197 0.292 (0.267) (0.266) (0.272) Other city of over 200,000 people 0.313 0.106 -0.120 (0.310) (0.314) (0.332) Rural town 0.275 0.495 0.461 (0.421) (0.463) (0.496) Other rural area -0.257 -0.524** 0.010 0.047 (0.428) (0.205) (0.477) (0.502) Utilities usage Electricity usage 0.691** 0.675** (0.279) (0.287) Power outages -0.349* -0.420* (0.208) (0.218) Owns a landline 0.403 0.442 (0.320) (0.322) Owns a cell phone -0.064 -0.056 (0.262) (0.262) Constraints Credit -1.157*** (0.437) Transport 0.299 (0.469) Utilities -0.244 (0.380) Constant 1.942*** 1.398*** 1.513*** 1.743*** 1.218*** 1.760*** (0.361) (0.282) (0.367) (0.297) (0.380) (0.409) N 347 347 347 347 347 347 R2 0.424 0.422 0.434 0.434 0.440 0.460 Adjusted R2 0.403 0.407 0.412 0.417 0.411 0.427 Note: - .01 - ***; .05 - **; .1 - *; Robust standard errors in parentheses Source: 2006/07 RICS-Amhara and 2006 EES. -98- ANNEX 3: SURVEY METHODOLOGY 1. Definition of Nonfarm Activities and Nonfarm Enterprises Nonfarm activities include all economic activities in rural areas except agriculture, livestock, fishing and hunting. Nonfarm enterprises are defined as all activities performed as self-employed, employers or unpaid family workers in sectors other than agriculture excluding wage and salary employment. More details on the definitions and survey methodology can be found in a Basic Information Document (CSA, 2008b). 2. The Rural Investment Climate Survey (RICS) The Rural investment climate survey (RICS) was conducted in Ethiopia to support and provide statistics for the Ethiopia Rural Investment Climate Assessment (RICA). The data was collected by the Central Statistical Agency from December 2006 to January 2007 with technical assistance from the World Bank. The Ethiopia RICS consists of two surveys: the Ethiopia RICS-AgSS and the RICS-Amhara survey. The RICS-AgSS was conducted in the four major regions of Ethiopia - Tigray, Amhara, SNNP, and Oromia which together account for about 90 percent of the population. The RICS-Amhara covered the Amhara region in more detail. 3. RICS-AgSS Survey The RICS-AgSS survey questionnaire includes a short set of questions on nonfarm enterprises operated by households (Table A). For all those households who do not operate a nonfarm business a small sub-set of questions, including investment constraints to open and/or operate a nonfarm enterprise, are asked. Table A: Contents of the RICS-AgSS Enterprise Questionnaire Section Description Owner Particulars This section collects information on location and demographics of the enterprises owner/manager such as region, zone, gender, age, and education Nonfarm Enterprise The section collects detailed information on the enterprise operations including information · Type of enterprise · Base and geographical location of enterprise operation · Ownership status of enterprises · Sources of start-up capital and motive for enterprise start-up · Customer of enterprise goods · Seasonality of enterprise activities · Age of enterprise · Number of workers employed by the enterprise · Average sales and growth of sales · Enterprise contribution to household income · Enterprise constraints · Access to markets and roads for enterprises -99- Table B: Contents of the RICS-Amhara Household Questionnaire Section Description 1 Area Identification This section collects information on the location of the households within the survey area. It also collects information on the individuals (enumerators, supervisors, coordinators) who were involved with the collection and verification of the information. 2 Household The household demographics section collects information on the individuals who Demographics are resident in the household. It collects basic demographic information such as relationship to the household head, sex, age ethnicity and marital status among other items. It also collects literacy and education information for the household members. 3 Employment The employment section is administered to all household members 10 years old and older and collects information on: · Engagement in productive work · Primary and secondary jobs · Wages · Allowances and gratuities · Average daily wage for casual labor · Industry · Occupation · Days worked per month · Work in nonfarm household enterprise in the household 4 Living Conditions This section is administered to the household head and collects information on: · Ownership of dwelling · Size of dwelling · Sources of lighting · Sources of cooking fuel · Shocks experienced by the household 5 Household and Farm This section collects the market value of items or services consumed during the Consumption past 7 days for food items, other household and farm goods and services Expenditures 6 Sources of This section collects information on the amounts of agricultural and non- Household Income agricultural income received by the household during the last month and the last 12 months. It also collects information on the amount of gifts received by the household and gifts given by members of the household 7 Assistance from This section collects information on the aid received by members of the Government or Aid household during the last 3 years (2004, 2005 and 2006) from the government Organizations or private aid organizations 8 Credit This section collects information about loans received by members of the household during the last 5 years. It includes loans received in cash or in-kind. 9 Household and Farm This section collects information on the durable goods owned by the household Asset Ownership or farm. 10 Access to Basic This section collects information on the availability of infrastructure to the Infrastructure and household. This includes such items as telecommunications, schools, health Institutions facilities, agricultural services, police and financial institutions. 4. RICS-Amhara Survey The RICS-Amhara comprises a more detailed effort to collect information on nonfarm enterprises and their households from the Amhara region. The RICS-Amhara survey consists of three questionnaires to collect information: a household questionnaire, an enterprise questionnaire, and a community questionnaire. The RICS-Amhara household questionnaire collects information from all sample households, regardless of whether the household has any nonfarm enterprise. Table B -100- provides an overview of the modules included in the RICS-Amhara household questionnaire. The RICS-Amhara community questionnaire was designed to collect information that is common to all households in a given geographic area. During the survey a "community" was defined as a farmers' association in rural areas or a Kebele in urban areas. These are the smallest administrative units in rural and urban areas respectively. The questionnaire was administered to a group of several knowledgeable residents such as the village headman, headmaster of the local school, agricultural field assistant, religious leaders, local merchants, health workers and long-term knowledgeable residents. Table C provides an overview of the modules included in the RICS-Amhara community questionnaire. Table C: Contents of the RICS-Amhara Community Questionnaire Section Description 1 Area Identification This section collects information on the location of the community so that it can be linked to the households within the survey area. It also collects information on the individuals (enumerators, supervisors, coordinators) who were involved with the collection and verification of the information. 2 Access to Credit This section collects information on the financial services in the area. It includes information on banks, micro-finance and community groups. It asks how far the institution is from the community, and the types of services offered. 3 Income and Economic This section collects information on the important sources of Activities employment for individuals in the community. 4 Land and Agricultural This section collects information on the agricultural services available Production in the community 5 Prices of Agricultural Information on the prices received by farm producers, the costs which Products and Costs of local producers pay, the costs of infrastructure and financial services, Inputs, Infrastructure and costs of consumer goods, and costs of wages and equipment rentals. Consumer Goods 6 List of Major Enterprises List of enterprises located in the community. Available in the Community 7 Investment Climate List of possible constraints to investment and the level of constraint in Constraints the community. 8 Major Constraints The four main constraints to starting nonfarm enterprises in the community. The RICS-Amhara enterprise questionnaire was designed to collect information on all nonfarm enterprises currently owned by any member of the sampled households. The questionnaire was administered to the individual in the household who owned, either solely or with someone else, the enterprise. Table D provides an overview of the modules included in the RICS-Amhara enterprise questionnaire. -101- Table D: Contents of the RICS-Amhara Enterprise Questionnaire Section Description 1 Area Identification This section collects information on the location of the households with enterprises within the survey area. It also collects information on the individuals (enumerators, supervisors, coordinators) who were involved with the collection and verification of the information. 2 Manager/Owner Demographic characteristics of the household member that owns or manages Characteristics the nonfarm enterprise. 3 Investment Climate List of possible constraints to investment and the level of constraint. Constraints 4 Major Constraints The four main constraints facing the enterprise, how those constraints have changed over the past 12 months, and the increase in sales that would result from the lifting of the constraint. 5 Association and Start- Membership in trade or business Associations, and source of start-up capital. up 6 Labor Number of permanent and seasonal laborers in the last 12 months and the start-up year. These numbers are also divided by household and non- household members. 7 Products/ Services and Information on the most important products and services sold over the past 12 Sales months. Sales information collects units and prices, production information includes units, costs, and labor inputs. Information is collected for the past 12 months. 8 Expenditures Expenditures for the last 12 months is collected for wages, transportation, fuel, electricity, water, telecommunication, rent/leasing, and other items. 9 Investments Information on investments made since its start-up and in the last year. 10 Assets Assets owned or used by the nonfarm enterprise in terms of land, buildings, storage facilities, vehicles, and other equipment. 11 Competition Competition to the nonfarm enterprise within the community and in the country. 12 Market Information Locations in which the nonfarm enterprise markets its products and services. 13 Infrastructure Use of electricity and telephones for the nonfarm enterprise. 14 Nonfarm Enterprise Use of credit for the enterprise. Applications for loans, success in receiving Credit loans, and repayment information. 15 Enterprise Registration Information on registry of the nonfarm enterprise with any government and Permits agencies. 5. Sampling Approach The RICS in Ethiopia is largely centered on the fieldwork conducted for the Agricultural Sample Survey (AgSS). The AgSS is a long-standing effort, conducted annually by the Central Statistics Agency. It is designed to collect information from agricultural households about agricultural production and costs. Most of the RICS-Amhara households are a subset of the RICS-AgSS households, the majority of which are in turn a subset of the AgSS. The RICS-AgSS was conducted in 490 enumeration areas (EAs) in the four major regions (Tigray, Amhara, Oromia and SNNP) of Ethiopia. The RICS-AgSS visited all of the EAs visited by the AgSS in four specific zones in the Amhara region: North Gonder (44 EAs), South Gonder (44 EAs), North Wello (46 EAs), and West Gojjam (48 EAs). In the rest of the Amhara region and in the other three major regions, the RICS-AGSS visited a subset of the EAs visited by the AgSS. The number of EAs in each of the subsets is: Rest of Amhara region 50 EAs of 224, Tigray 60 EAs of 165, Oromia 79 EAs of 573, SNNP 81 EAs of 612. The total nominal sample size of the RICS-AgSS is thus 14,464 households EAs (32 households in each of the 452 rural EAs). Thereof, 13,560 -102- households are agricultural households (30 households in 452 EAs) and 904 are non- agricultural households (2 households in 452 EAs). The RICS-Amhara was conducted in four zones of the Amhara region (North Gonder, South Gonder, North Wello, and West Gojjam). In these zones, the survey visited two kinds of EAs: A subset of the EAs visited by the AgSS in the zone (which are all rural by design) and a random sample of non-AgSS EAs in small towns (operationally defined as towns with less than 10,000 habitants, often rural rural market towns). The total number of EAs visited by the RICS-Amhara in each of the special zones (North Gonder (44 EAs), South Gonder (44 EAs), North Wello (46 EAs), and West Gojjam (48 EAs) was 182 EAs. The total nominal sample size of the RICS-Amhara is thus 2,912 households (16 in each of the 182 EAs) As mentioned previously, most of the RICS-Amhara households are a subset of the RICS-AgSS households. Thus, where the same household is interviewed for the two surveys, the data can be merged and analyzed in conjunction with a few exceptions. Actual sample size differs slightly from survey design sample size due to fewer EAs in some areas and non-replacement of households that were not available in some instances. Thus, the RICS AgSS has 14,063 households instead of 14,464 and the RICS-Amhara has 2,909 households instead of 2,912 households in the survey design. The survey is representative at the zonal level. Thus, to obtain unbiased estimates from the survey data, the results should be expanded by the sampling weights provided in the data. 6. Data Quality The RICS survey was conducted in an efficient manner to ensure a high level of accuracy. In general, data quality is good as quality control was ensured through a number of procedures: (i) pilot testing of the questionnaires; (ii) two week training of interviewers including mock interviews; (iii) intensive supervision during the data collection process; (iv) the data entry and cleaning process with CSPro Census and Survey Processing System software used careful data checks to verify the skips, ranges and intra record consistency of the data; and (v) the data cleaning process with STATA intensively checked for outliers, duplicates, missing observations and variables, coding, labeling of variables and codes, consistency within files, across files and across surveys. As a further check of data quality, selected descriptive statistics of rural households in the RICS-Amhara sample were compared with the descriptive statistics of rural households in the welfare monitoring survey (WMS) for years 2000 and 2004. The WMS sample was limited to the four zones in rural Amhara covered by the RICS-Amhara survey. The resulting statistics are presented in Table 44. The statistics from the three surveys are quite close except in very few instances. This further supports the quality of the RICS- Amhara data. In addition, the interviewers were asked to give their opinion of the validity of responses pertaining to perceptions and opinions provided by household members and community leaders. In both cases, about 95 percent of the responses were deemed accurate by the interviewers. -103- Table 44: Amhara ­ Household Characteristics for the Four Specific Zones in Rural Amhara (mean/percent) RICS-Amhara WMS 2000 WMS 2004 Amhara Amhara Age of household head 44.2 44.4 43.5 Education of household head (years completed) 0.8 0.3 0.7 Household head has some education 21.9 6.9 19.5 Female headed household 21.8 22.0 22.0 Household size 4.6 4.6 4.5 Household members age 60 0.2 0.2 0.2 Farm income source in past 12 months 89.9 88.1 86.9 Nonfarm enterprise income source in past 12 months 12.6 11.7 7.3 Wage/salary income source in past 12 months 7.5 4.7 6.4 Rental/pension income source in past 12 months 8.1 1.2 0.8 Received transfers in past 12 months 12.9 9.1 2.2 Received social benefits in past 12 months 34.9 n/a n/a Experienced food shortage in past 12 months 32.2 n/a 37.3 Months experienced food shortage in past 12 months 1.1 n/a 1.3 Distance to nearest market (kms) 9.4 8.4 5.7 Distance to nearest post office (kms) 29.3 24.7 18.0 Distance to nearest primary school (kms) 3.5 4.1 3.2 Distance to nearest secondary school (kms) n/a 26.0 19.1 Distance to nearest health center (kms) 10.1 9.0 19.1 Distance to nearest bus stop (kms) 19.8 20.2 18.0 Distance to nearest road (kms) 17.3 16.8 12.9 Distance to nearest phone booth (kms) 17.1 28.1 19.6 Sample size 2,335 1,440 1,968 Note: Statistics are weighted. (1) RICS-Amhara covers rural areas and rural market towns in four zones in Amhara: North Gonder, South Gonder, North Wello, and West Gojjam. RICS Amhara survey period- November 2006 to January 2007. The sample here is restricted to rural households. (2) The WMS sample here is restricted to rural areas in four zones in Amhara: North Gonder, South Gonder, North Wello, and West Gojjam. WMS 2000 survey period: June 1999-February 2000. WMS 2004 survey period-June 2004 to July 2004. -104- BIBLIOGRAPHY Baker, J. 1986. The Rural-Urban Dichotomy in the Developing World: A Case Study from Northern Ethiopia. Oslo: Norwegian University Press. Baker, J. 1990. "The Growth and Functions of Small Urban Centers in Ethiopia." In J. Baker and C.F. Claeson (eds). Small Town Africa: Studies in Rural- Urban Interaction. Uppsala: The Scandinavian Institute of African Studies. Diao, X. et al. 2007. "Agricultural Growth Linkages in Ethiopia: Estimates using Fixed and Flexible Price Models" IFPRI Discussion Paper No. 00695. Washington DC: International Food Policy Research Institute. http://www.ifpri.org/sites/default/files/publications/ifpridp00695.pdf Demeke, M. 2001. "Off-farm income generation opportunities in Ethiopia: Opportunities and constraints in food-insecure Weredas of Oromia and Amhara regional states." Ethiopian Development Forum 2(1): 20-45. Dercon, S. and J. Hoddinott. 2005. "Livelihoods, Growth, and Links to Market Towns in 15 Ethiopian Villages." IFPRI Discussion Paper 194. http://www.ifpri.org/ divs/fcnd/dp/papers/fcnbr194.pdf Dollar, D., M. Hallward-Driemeier and T. Mengistae. 2005. "Investment Climate and Firm Performance in Developing Economies. Economic Development and Cultural Change 54(1): 1-30. Evans, H. 1992. "A Virtuous Circle Model of Rural-Urban Development: Evidence from Kenya." Journal of Development Studies 28(4):640-667. Fafchamps, M. and A. Quisumbing. 2002. "Control and Ownership of Assets Within Rural Ethiopian Households." Journal of Development Studies 38(6): 47­82. Fafchamps, M. and A. Quisumbing. 2005. "Assets at marriage in rural Ethiopia." Journal of Development Economics 77(1): 1-25. Feleke, S.T. 2005. "Determinants of food security in Southern Ethiopia at the household level." Agricultural Economics 33: 351-363. Haggblade, S., P. Hazell and T. Reardon. 2007. Transforming the Rural Nonfarm Economy: Opportunities and Threats in the Developing World. Baltimore: John Hopkins University press published for IFPRI. Kidane, T. and J. Loening. 2008. "Small Enterprise Growth and the Rural Investment Climate: Evidence from Tanzania." World Bank Policy Research Working Paper 4675. http://www-wds.worldbank.org/ -105- Lerman, R. I. , and S. Yitzhaki. 1985. "Income Inequality Effects by Income Source: A New Approach and Applications to the United States." Review of Economics and Statistics 67(2): 151-56. Loening, J. and G. Oseni. 2008. "Approximating Rural and Urban Welfare Effects of Food Price Inflation in Ethiopia." Background paper for the Ethiopia Food Price Inflation Policy Note. Processed. MOLSA. 1997a. Agricultural Wage Employment and Rural Nonfarm Employment in Ethiopia: Survey Results. Addis Ababa: Ministry of Labor and Social Affairs Processed. MOLSA.1997b. Survey on Agricultural Wage Employment and Rural Nonfarm Activities with Particular Reference to Appropriate Technology for Rural Employment. Addis Ababa: Ministry of Labor and Social Affairs. Processed. Mukherjee, A. and X. Zhang. 2007. "Rural Industrialization in China and India: Role of Policies and Institutions." World Development 35(10): 1621-1634. Sørensen, P. 2003. "The Need for Rural Diversification: Food-Security Policies and Nonfarm Development in the Amhara Region, Ethiopia." Danish Institute for International Studies. DIIS/GI. Kongevej Working Paper 03.11. Processed. van den Berg, M. and G. Kumbib. 2006. "Poverty and the rural nonfarm economy in Oromia, Ethiopia." Agricultural Economics 35 (supplement): 469­475. Woldehanna, T. 2000. Farm and Off-farm Employment in the Tigray Region, Northern Ethiopia: An Economic Analysis and Policy Implications. Ph.D. Thesis. Wageningen University, Netherlands. Processed. Woldehanna, T. and F. Ellis. 2005. Ethiopia Participatory Poverty Assessment. Addis Ababa: Ministry of Finance and Economic Development. Processed. World Bank. 2004. A Better Investment Climate for Everyone. World Development Report 2005. Washington DC. World Bank. 2009. Towards the Competitive Frontier: Strategies for Improving Ethiopia's Investment Climate. Finance and Private Sector Development Unit, Africa Region. Washington D.C. -106- BACKGROUND DOCUMENTS Bardasi, E. and A. Gethahun. 2008. "Gender and Nonfarm Activities in Rural Ethiopia." Background paper for the Ethiopia Rural ICA. Processed. Bakker, S. 2007. "Qualitative Sectoral Analysis of Nonfarm Enterprises in Ethiopia and Amhara: On the Role of MSE in Rural Areas." Consultant report submitted to the Royal Netherlands Embassy and the World Bank. Processed. Beegle, K. and G. Oseni. 2008. "Nonfarm Enterprises, Food Security, and Economic Shocks in Ethiopia." Background paper for the Ethiopia Rural ICA. Processed. CSA. 2008a. "Rural Investment Climate Survey: Report on Rural Nonfarm Enterprises, Investment Climate Constraints, Infrastructure Accessibility, and Household Employment." Statistical Bulletin. Addis Ababa: Central Statistical Agency. CSA. 2008b. Rural Investment Climate Survey: Basic Information Document. Addis Ababa: Central Statistical Agency. Processed. Günther, I. and M. Olapade. 2007. "A Review of the Nonfarm Sector in Rural Ethiopia: Characteristics and Dynamics." Background paper for the Ethiopia Rural ICA. Processed. Loening, J., B. Rijkers and M. Söderbom. 2008. "Nonfarm Microenterprise Performance and the Investment Climate: Evidence from Rural Ethiopia." World Bank Policy Research Working Paper 4577. http://www-wds.worldbank.org/ Muir, A., M. Wodino, B. Yenes and others. 2007. "Qualitative Livelihood Analysis of Households and Communities in Amhara: Case Study Evidence." Consultant report submitted to the Royal Netherlands Embassy and World Bank. Processed. Mulugeta, H. 2008. "Institutional and Policy Analysis on Rural Nonfarm Economies in Ethiopia." Consultant report submitted to World Bank. Processed. Ramaswamy, R. 2008. "Rural Finance in Ethiopia: A Review of Demand and Supply." Consultant report for the Ethiopia Rural ICA. Processed. Rijkers, B., M. Söderbom, and J. Loening. 2009. "Mind the Gap?A Rural-Urban Comparison of Manufacturing Firms." World Bank Policy Research Working Paper 4946. http://www-wds.worldbank.org/ Söderbom, M. and B. Rijkers. 2009. "Market Integration and Structural Transformation in a Poor Rural Economy." World Bank Policy Research Working Paper 4856. http://www-wds.worldbank.org/ Olapade, M. 2007. Distributional Effects of Nonfarm Activities in Ethiopia. Master thesis, Economie Quantitative. Université Paris I, Sorbonne. Processed. -107- 32°E 34°E TIGRAY 36°E 38°E 40°E 42°E ETHIOPIA ERIT REA REP. PARTICIPATION IN NONFARM OF ENTERPRISES, 2007 YEMEN % OF HOUSEHOLDS SECTOR SHARE OF Central WITH ENTERPRISES: ENTERPRISES BY REGION 2006 14°N West East 14°N AMHARA T I G R AY <10 Services Manufacturing Two 10 ­ 19 20 ­ 29 South S U DAN North Gonder 30 ­ 39 Wag Trade Hemira 40 ­ 50 Four >50 A FA R 12°N North 12°N Wello ZONE BOUNDARIES South Gonder One AMHARA DJIBOUTI REGION BOUNDARIES West INTERNATIONAL BOUNDARIES Metekel Gojam South Agew Wello OROMIYA Awi Five O Oromiya East 44°E 46°E 48°E BENSHANGUL Gojam Shinile a Asosa 10°N North Three DIRE DAWA 10°N Kamashi Shewa East ETHIOPIA Tongo Sp. Wellega North ADDIS Jijiga West ABABA Shewa Wellega HARARI West Shewa East S OM AL I A West Harerghe a Southwest Shewa Harerghe Illubabor ew t Sh Three One Degehabur O R O M I YA Eas Gurage 8°N 8°N GAMBELA Jimma Yem Arsi Two Selti Fik Sheka Hadiya Alaba Four K.A.T. Warder Keffa Hadiya Dawro SOMALI Wolayita Bale Korahe SOUTHERN NATIONS, Bench Sidama Maji Gamo NATIONALITES Gofa Gedeo Gode 6°N AND PEOPLES Basketo Amaro 6°N SNNP South Dirashe Omo Burji Guji Konso Afder 0 50 100 150 200 Kilometers Borena Liben 0 50 100 150 Miles This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information 4°N 4°N IBRD 36152 shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any MAY 2008 endorsement or acceptance of such boundaries. UGANDA KENYA 32°E 34°E 36°E 38°E 40°E 42°E 44°E 46°E 48°E 32°E 34°E 36°E 38°E 40°E 42°E ETHIOPIA ERIT REA REP. PERCEIVED CONSTRAINTS OF TO NONFARM BUSINESS YEMEN OPERATIONS AND GROWTH BY ETHIOPIA Central GEOGRAPHICAL ZONE, 2007 14°N West East 14°N T I G R AY TRANSPORTATION Two <10 South 10 ­ 19 North Gonder Wag 20 ­ 29 Hemira 30 ­ 39 Four A FA R 12°N North 12°N ZONE BOUNDARIES South Gonder Wello One S U DAN AMHARA DJIBOUTI REGION BOUNDARIES West INTERNATIONAL BOUNDARIES Metekel Gojam South Agew Wello Awi Five Oromiya O East 44°E 46°E 48°E BENSHANGUL Gojam Shinile a Asosa 10°N North Three DIRE DAWA 10°N Kamashi Shewa East Tongo Sp. Wellega North ADDIS Jijiga West ABABA Shewa Wellega HARARI West Shewa East S OM AL I A West Harerghe a Southwest Shewa Harerghe Illubabor ew t Sh Three One Degehabur O R O M I YA Eas Gurage 8°N 8°N GAMBELA Jimma Yem Arsi Two Selti Fik Sheka Hadiya Alaba Four K.A.T. Warder Keffa Hadiya Dawro SOMALI Wolayita Bale Korahe SOUTHERN NATIONS, Bench Sidama Maji Gamo NATIONALITES Gofa Gedeo Gode 6°N AND PEOPLES Basketo Amaro 6°N South Dirashe Omo Burji Guji Konso Afder 0 50 100 150 200 Kilometers Borena Liben 0 50 100 150 Miles This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information 4°N 4°N IBRD 36154 shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any MAY 2008 endorsement or acceptance of such boundaries. UGANDA KENYA 32°E 34°E 36°E 38°E 40°E 42°E 44°E 46°E 48°E 32°E 34°E 36°E 38°E 40°E 42°E ETHIOPIA ERIT REA REP. PERCEIVED CONSTRAINTS OF TO NONFARM BUSINESS YEMEN OPERATIONS AND GROWTH BY ETHIOPIA Central GEOGRAPHICAL ZONE, 2007 14°N West East 14°N T I G R AY CREDIT Two <10 South 10 ­ 19 North Gonder Wag 20 ­ 29 Hemira 30 ­ 39 Four A FA R 40 ­ 49 12°N North 12°N South Gonder Wello >50 One S U DAN AMHARA DJIBOUTI West ZONE BOUNDARIES Metekel Gojam South Agew REGION BOUNDARIES Wello Awi INTERNATIONAL BOUNDARIES Five Oromiya O East BENSHANGUL Gojam Shinile a Asosa North 44°E 46°E 48°E 10°N Three DIRE DAWA 10°N Kamashi Shewa East Tongo Sp. Wellega North ADDIS Jijiga West ABABA Shewa Wellega HARARI West Shewa East S OM AL I A West Harerghe a Southwest Shewa Harerghe Illubabor ew t Sh Three One Degehabur O R O M I YA Eas Gurage 8°N 8°N GAMBELA Jimma Yem Arsi Two Selti Fik Sheka Hadiya Alaba Four K.A.T. Warder Keffa Hadiya Dawro SOMALI Wolayita Bale Korahe SOUTHERN NATIONS, Bench Sidama Maji Gamo NATIONALITES Gofa Gedeo Gode 6°N AND PEOPLES Basketo Amaro 6°N South Dirashe Omo Burji Guji Konso Afder 0 50 100 150 200 Kilometers Borena Liben 0 50 100 150 Miles This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information 4°N 4°N IBRD 36155 shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any MAY 2008 endorsement or acceptance of such boundaries. UGANDA KENYA 32°E 34°E 36°E 38°E 40°E 42°E 44°E 46°E 48°E 32°E 34°E 36°E 38°E 40°E 42°E ETHIOPIA ERIT REA REP. PERCEIVED CONSTRAINTS OF TO NONFARM BUSINESS YEMEN OPERATIONS AND GROWTH BY ETHIOPIA Central GEOGRAPHICAL ZONE, 2007 14°N West East 14°N T I G R AY MARKETS Two <10 South 10 ­ 19 North Gonder 20 ­ 29 Wag Hemira 30 ­ 39 Four A FA R 40 ­ 49 12°N North 12°N 50 ­ 60 South Gonder Wello One S U DAN AMHARA DJIBOUTI >60 West ZONE BOUNDARIES Metekel Gojam South Agew REGION BOUNDARIES Wello Awi INTERNATIONAL BOUNDARIES Five Oromiya O East BENSHANGUL Gojam Shinile a Asosa North 44°E 46°E 48°E 10°N Three DIRE DAWA 10°N Kamashi Shewa East Tongo Sp. Wellega North ADDIS Jijiga West ABABA Shewa Wellega HARARI West Shewa East S OM AL I A West Harerghe a Southwest Shewa Harerghe Illubabor ew t Sh Three One Degehabur O R O M I YA Eas Gurage 8°N 8°N GAMBELA Jimma Yem Arsi Two Selti Fik Sheka Hadiya Alaba Four K.A.T. Warder Keffa Hadiya Dawro SOMALI Wolayita Bale Korahe SOUTHERN NATIONS, Bench Sidama Maji Gamo NATIONALITES Gofa Gedeo Gode 6°N AND PEOPLES Basketo Amaro 6°N South Dirashe Omo Burji Guji Konso Afder 0 50 100 150 200 Kilometers Borena Liben 0 50 100 150 Miles This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information 4°N 4°N IBRD 36172 shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any MAY 2008 endorsement or acceptance of such boundaries. UGANDA KENYA 32°E 34°E 36°E 38°E 40°E 42°E 44°E 46°E 48°E