Report No. 40108-TZ Tanzania Pilot Rural Investment Climate Assessment Stimulating Non-farm Microenterprise Growth June 2007 Sustainable Development Network Eastern Africa Country Cluster 1 Africa Region Document of the World Bank Tanzania Pilot Rural Investment Climate Assessment Stimulating Non-farm Microenterprise Growth June 2007 Sustainable Development Network Eastern Africa Country Cluster 1 Africa Region This volume is aproduct of the staffof the International Bank for Reconstruction and Development/ The World Bank. The findings, interpretations, and conclusions expressed inthis paper do not necessarily reflect the views o fthe Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy o f the data included inthis work. The boundaries, colors, denominations, and other informationshown on any map inthis work do not imply anyjudgment on the part o f The World Bank concerningthe legal status o f any territory or the endorsement or acceptance o f such boundaries. The material inthis publication is copyrighted. Copying andor transmitting portions or all ofthis work without permission may be a violation o f applicable law. The InternationalBank for Reconstructionand Developmend The World Bank encourages dissemination o f its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part o f this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, http:llwww.couvright.cond. All other queries onrights and licenses, including subsidiary rights, should be addressedto the Office o fthe Publisher, The World Bank, 1818 H Street NW, Washington, D C 20433, USA, fax 202-522-2422, e-mail pubrights@,worldbank.org. Acknowledgements This RuralInvestmentClimate Assessment is basedon analysis of survey data collected by Tanzania's National Bureau o f Statistics (NBS). The survey was conducted between January and March2005, The World Bank team would like to acknowledge the excellent collaboration with N B S staff. A team ledby William Lane launched the survey and rural investment climate work, Josef Loening prepared the final synthesis report, and research assistance was provided by James Keough. Inputsat various stages came from Tilahun Temesgen (survey design, sampling, training and survey supervision, background descriptive analysis), Reka Sundaram-Stukel, Klaus Deininger and Songqing Jin (productivity analysis), RamRamaswamy (rural finance), and Alexander Lotsch (maps). Robert Townsend, Henry Gordon, Cornelis van der Meer and Renate Kloeppinger-Toddprovided technical advice. Substantive comments were receivedfrom Robert Utz, Michael Wong and Hans Hoogeveen. Mary Hallward-Driemeier, Shobha Shetty and Xavier Gine peer-reviewed the report. Editingand translation assistance from Sharon Dotto Abu and Ichikaeli Maro-Mzobora The Tanzania Rural Investment Climate Assessment, which i s the first pilot for the African continent, i s part o f a global piloting exercise on the Rural Investment Climate, coordinated by Cornelis van der Meer. The team thanks the Bank Netherlands Partnership Program and the Norwegian Trust Fundfor its support to this initiative. Excellent overall guidance was received from Judy O'Connor (Country Director) and KarenMcConnell-Brooks (Sector Manager). Shukrani Tathmini hiiya Hali ya Uwekezaji Vijijini inazingatia uchambuzi wa data za utafiti mpana zilizokusanywa na Ofisi ya Takwimu ya Taifa, Tanzania. Utafiti ulifanywa kati ya Januari na Machi 2005. Timuya Benki ya Dunia inapenda kutoa shukrani za dhati kwa ushirikiano mzuri ilioupata toka kwa wafanyakazi wa Ofisi ya Takwimu ya Taifa. Utafiti wa hali ya uwekezaji vijijini ulizinduliwa na timu iliyoongozwa na William Lane; Josef Loening (aliandaa ripoti ya mwisho ya makusanyo); na James Keough (alikuwa mtafiti msaidizi). Michango minginekatika hatua mbalimbali ilitolewa na watu wafuatao (majukumu yao yameoneshwa katika mabano): Tilahun Temesgen (kubuniutafiti, kuteua sampuli, mafunzo na usimamizi wa utafiti, mchanganuo wa usuli); Reka Sundaram-Stukel, Klaus Deininger na Songqing Jin (uchambuzi wa tija); RamRamaswamy (masuala ya fedha vijijini); na Alexander Lotsch (ramani). Ushauri wa kiufundiulitolewa na Robert Townsend, Henry Gordon, Cornelis van der Meer na Renate Kloeppinger-Todd. Maoni muhimuyalitolewa na Robert Utz, Michael Wong na Hans Hoogeveen. Ripoti ilipitiwa kwa pamoja na Mary Hallward-Driemeier, Shobha Shetty na Xavier Gine. Shukrani ziende kwa Sharon Dotto Abu na Ichikaeli Maro-Mzobora kwa msaada wa tafsiri na urekebishwaji. Tathmini ya Hali ya Uwekezaji Vijijini Tanzania, ni kazi ya majaribio ya kwanza Barani Afrika ikiwani sehemu ya zoezi la majaribio la kiulimwengu kuhusuHali ya Uwekezaji Vijijini. Tathmini hii iliratibiwa na Cornelis van der Meer. Timu inatoa shukrani za dhati kwa Programu ya Ushirika ya Benki ya Uholanzi na Mfuko wa Maendeleo wa Norwei kwa msaada wake katika juhudi hizi. Timuinawashukuru pia Judy O'Connor (Mkurugenzi Mkaziwa Benkiya Dunia) na Karen McConnell-Brooks (Meneja wa Sekta) kwa mwongozo wa kazi yote. Tableof Contents ExecutiveSummary (Englishand Kiswahili) .............................................................................. 1 1 Introduction . .............................................................................................................................. What I s the Rural InvestmentClimate?..................................................................................... 13 13 15 2 Profileof RuralNon-farmEnterprises .Snapshotof Tanzania's Rural Economy ................................................................................... .................................................................................. Basic Characteristics ................................................................................................................. 20 20 Economic Activities .................................................................................................................. 23 Labor Productivity..................................................................................................................... 26 3 EnterpriseDynamics . ............................................................................................................... Business Closures...................................................................................................................... EntryInto the Non-farm Sector................................................................................................. 33 33 34 35 4 The Impactof aBetterInvestment Climate .Enterprise Growth..................................................................................................................... .......................................................................... Constraints to Enterprise Operations and Growth -Perceptions ............................................ 39 39 Finance. Infrastructure. and Governance -Objective Measurements..................................... 45 Simulating Gains from a Better Investment Climate ................................................................ 48 5 Reflectionsfor Policyand FutureAnalysis . ........................................................................... 52 52 Finance. Infrastructureand Institutions..................................................................................... Agriculture and Rural Trade...................................................................................................... FutureAnalytical Work............................................................................................................. 53 54 Appendices Appendix 1: Summary Tables ................................................................................................... .................................................................................................................................... 56 56 Appendix 2: Regression Analysis ............................................................................................. 61 -74 Appendix 4: Rural Finance........................................................................................................ Appendix 3: Survey Methodology ........................................................................................... 77 Bibliography ................................................................................................................................. 89 BackgroundDocuments .............................................................................................................. 92 ListofBoxes 3 Box 2: Design of the Tanzania Rural Investment Climate Survey ............................................... Box 1: Characterization o f Rural Non-farm Enterprises ................................................................ 14 Box 3: On Structural Transformation o f the Rural Economy inAsia .......................................... 18 B o x 4: Women's Microenterprises inRuralTanzania ................................................................. 22 29 36 Box 7: Why D o RuralNon-farm Enterprises Grow?.................................................................... Box 6: Typology of Rural Non-farm Enterprise........................................................................... Box 5: Costs and Benefits o f BeingInformal............................................................................... B o x 8: Productivity Analysis of Tanzania's RuralNon-farmEnterprise Sector .......................... 37 50 Box 9: Snapshot of Microfinance Institutions inTanzania .......................................................... 80 ListofFigures 4 Figure 2: Growth of Agricultural GDP......................................................................................... Figure 1: Top FiveRural Business Constraints .............................................................................. 16 Figure 3: Evolution of Main Sources o f Household Cash Income ............................................... 18 Figure 4: Number of Workers per Enterprise ............................................................................... 21 Figure 5: Distribution o f Enterprises by Age................................................................................ 22 24 Figure 7: Seasonality o f Non-farm Actvites inTanzania ............................................................. Figure 6: Sectoral Distribution of Enterprises .............................................................................. Figure 8: Median Sales per Day o f Labor by Location and Registration..................................... 25 27 Figure 9: Median Sales per Day o f Labor by Sector and Registration......................................... 27 Figure 10: Reasons for Not RegisteringWith Government.......................................................... 28 Figure 11: Median Sales Per Day o f Labor by Size...................................................................... 31 Figure 12: Median Sales per Day o f Labor by Sector .................................................................. 31 32 Figure 14: Sources o f Start-up Capital ......................................................................................... Figure 13: Median Sales per Day o f Labor by Region................................................................. 34 Figure 15: Perceived Reasons for Closure o f Business - and Reasons Preventing Start-up ........35 Figure 16: Distribution of Enterprise Employment Growth......................................................... 36 Figure 18: FirmGrowth, Size and Age inRural Tanzania ........................................................... Figure 17: Employment and Sales Growtho f Formal and InformalEnterprises by Region.......-37 38 Figure 19: Top Five Constraints o f All Rural Non-farm Enterprises, 2005 Figure 20: Comparison o f SelectedRural and UrbanBusiness Constraints inTanzania............-40 and Their UrbanICA Ratings .................................................................................... 39 Figure21: Comparisono f SelectedRural Business Constraints: Tanzania versus Sri Lanka and Selected African Countries ................................................................. 41 Figure 24: Confidence inConflict Resolution and Legal Environment by Communities, 2005 ..42 Figure 23: Top Five Constraints o f Rural Areas........................................................................... Figure 22: Top Five Constraints o f Rural Market Towns............................................................. 42 47 Figure25: Improvingthe Rural Investment Climate: EstimatedGains on Enterprise Employment Growth............................................................................ 49 Figure 26: Visualization of Business Constraints' Impact on Employment Growth 50 Figure 27: Distribution o f Approved Loans by Annual Interest Rate........................................... over 10 Year-horizon ................................................................................................. 83 Figure 28: Distribution o fApproved Loans by Value .................................................................. Figure 29: Collateral Requirements.............................................................................................. 84 85 86 Figure31: Sources o fLoans......................................................................................................... Figure 30: Purpose o f Loans......................................................................................................... 87 ListofMaps Map 1: Percentage o f PopulationBelow the Poverty Line........................................................... 17 Map 2: Density o f Rural Non-farm Enterprises............................................................................ 20 Map 3: Major and Severe Business Constraints by GeographicalZone....................................... 44 45 Map 5: EstimatedTravel Time to Rural Market Towns............................................................... Map 4: Mean Distance To Rural Financial Institutions................................................................ 46 Listof Tables Table 1: Snapshot o f Tanzania's RuralNon-farmEnterprise Sector.............................................. 2 Table 2: House Income Characteristics With and Without Non-farm Enterprise ........................ 19 Table 4: Average Market Shares for Main Product or Service..................................................... Table 3: Market Links................................................................................................................... 26 Table 5: Transaction Costs and Taxes for Formal Non-farmEnterprises .................................... 26 30 Table 7: RoadTypes Within and Outside Communities .............................................................. Table 6: Decomposition o f Start-up by Enterprise Size ............................................................... 34 Table 8: Enterprises ReportingMajor and Severe Constraints to Growth and Operations ..........46 56 Table 9: Top five Major or Severe Constraints PreventingHouseholds from Starting a Nonfarm Enterprise (percentages among households without non-farm enterprise) ................57 Table 10: Top five Major or Severe Constraints Causing Households to Close Table 11: Basic Enterprise and Community Characteristics by Region....................................... Their Non-farm Enterprise......................................................................................... 57 58 Table 13: Basic Infrastructure Use of Enterprises by Region....................................................... Table 12: National Real Prices for Goods and Services inRural Communities........................... 59 60 Table 14: Determinants of Employment and Sales Growth ......................................................... Table 15: Community-level Investment Climate Constraints and Employment Growth.............62 63 Table 16: Simulation Results o f Business Constraints Impact on Employment Growth .............65 Table 17: Probability o f Being Registered.................................................................................... 66 Table 19: Determinants o fNew Investment ................................................................................. 70 Table 18: Determinants for Non-farm Sector Participation.......................................................... 67 Table 20: Determinants of Total Factor Productivity ................................................................... 73 Table 22: Original Sample Sizes for Enterprises, Household and Community Survey................76 Table 21: Names o f Selected Regions and Zones and Number of Enumeration Areas................75 Table 23: Number o fRespondents for Enterprises, Household and Community Survey ............76 Table 24: Access to Formal Loans by Enterprises and Households............................................. 78 79 Table 26: Institutional Providers o f Microfinance Services ......................................................... Table 25: Access to Credit by Formal and Informal Enterprises.................................................. 81 Table 27: Distribution of Enterprises with Financial Statements by Sales................................... 82 Table 28: Access to Formal Loans by Enterprises inDifferent by Sales ..................................... 82 Table 29: Interest Rates Chargedby Different Lenders ............................................................... 84 EXECUTIVE SUMMARY Tanzania's Pilot Rural InvestmentClimate Assessment (RICA) measures the economic environment o fnon-farm enterpreneurs. The pilot assessment has three key objectives: it aims to better understand the ruralnon-farm economy inTanzania, shed light on rural enterprise dynamics and business constraints, and reflect on areas where government policies are readily directed to help promote rural non-farm enterprise activity. The RICA i s based on an analysis of a unique survey data set collectedby the National Bureau o f Statistics (NBS) during January and March 2005, covering enterprises, households, and communities in all seven geographical zones o f the country. Selected findings are summarized below. Non-farm enterprise characteristics Rural nonfarm enterprises matter.Non-farm activities are an improtant source o f income for approximately 1.4 million rural households, an increase from 1.2 million in2001. The highest enterprise densities are in the Lake region and inCentral Tanzania. Over the past decade the share o frural non-farm self-employment income has almost doubled. In2004, some 28 percent o f rural households reportedthat at least one member was working in a non-farm business. Non-farm enterprises are an essential source o f livelihood for a significant proportion o f Tanzania's rural population. Households that runa non-farm enterprise have an income that i s about 24 percent higher than that o f those without, suggesting that access to informal employment inthe rural non- farm sector could provide a path out of poverty. Tanzanian rural nonfarm enterprises differ from their urban counterparts.The capitalizationo f businesses i s extremly low. The medianvalue-added per laborer o f a ruralnon-farm business i s only US$ 83, a stark contrast to an urban micro-enterprise inTanzania that has an estimated median value-added per laborer o f US$474 (World Bank, 2004b). About one-half o f the enterprises are locatedinrural areas, while the other half i s located inrural market towns. Non- farm enterprises are very small with the majority operatedby one person duringmost o f the year. However, duringpeak seasons enterprises often employ part-time or casual labor with these being mostly family members. The level o f education among enterprise owners i s highby rural standards, with the majority at grades seven and eight. Rural trade dominates. The overall landscape o fnon-farm enterprises inTanzania i s quite diverse. However, the predominant entrepreneurial activity i s trading. About 57 percent o frural enterprises are engaged inwholesale or retail trading. Consequently, more than 75 percent o f Tanzanian enterprises are heavily affected by seasonality, which typically constrains enterprise growth. Non-farm enterprises inrural Tanzania buy and sell locally, operating inrelatively thin markets. Only 19percent o f the enterprises are formally registered. Labor productivity is low. For a typical rural business, sales are less than US$ 1.5 per day o f labor. However, there are differences. Small enterprises are relatively more productive than their larger counterparts (which could be due to the intensive use o f household family labor). The opposite i s true o f urban enterprises, which are more labor efficient at the higher end of the employment spectrum. Enterprises inTabora are more productive than enterprises inany other surveyed region. Productivity differences by sector, however, are not very pronounced. The exception i s mining,where labor productivity i s higher than in other sectors. Registration is associated with higher labor productivity. Formal enterprises have higher levels o f labor productivity than informal. The median formal enterprise generates US$ 149per laborer, versus US$ 82 for informal. Regulatorybarriers to entry are costly, estimated at about one-third -1- o f enterprise annual gross sales. The productivity differences between formal and informal enterprises are more pronouncedinrural market towns. However, additional study i s necessary to understandbenefits derived from enterprise formality. Table 1: Snapshot of Tanzania's Rural Non-farm Enterprise Sector, 2005 Total number o f enterprises ai 1.2mio Formally registered enterprises 19% Sector o f operation Trading 57% Services 21% Production 19% Two or more sectors 3yo Location Rural towns 46% Rural areas 54% Production Mediannet earnings per enterprise (US$) 113 Median value added per worker (US$) 83 Labor force Enterprisesusing only family labor 83% Enterprisesusing only hired labor 1Yo Both family and hired labor 15% Family workers (average) 1.6 Hired workers (average) 0.6 Total workers (average) 2.2 Characteristics o f entrepreneur Female 23% Primary education (1-6 years) 11% Primary education (7-8 years) 69% Secondary education 17% Tertiary education 3% Work experience (average inyears) 4.9 Source: 2005 Tanzania RICS. a/200l HBS Ruralenterprise dynamics The rate of newfirm creation appears to be lower than in other African countries. This could be a result o f investment climate constraints or a weaker rural enterprise culture. Inline with the findings from RICAs for other countries, rural entrepreneurs believe that access to finance and basic infi-astructure are among the most important constraints that impact on enterprise start-up and closure. The majority o f start-ups are small firms, and entry into the non-farm sector i s dependant on income generated from agriculture - 77 percent o f start-up capital i s from agricultural earnings. When agriculture is prospering and overall demand for non-farmproducts or services i s high, starting a business can meanprosperity. Butwhen agriculture i s languishing or population growth i s high, start-up jobs may simplyreflect the news that firms are acting as a sponge, soaking-up excess workers inmarginal activities. InTanzania, both demand-pull and supply-pushforces seem to determine entry into the ruralnon-farm enterprise sector. A minority of enterprisespropels employment growth. About one-third o f established rural enterprises (operating five or more years) are highperformers. The estimated overall high employment growth rate o f 4.5 percent annually for these established enterprises i s impressive, consideringthat the majority o f small enterprises inrural areas did not grow at all. Employment -2- growth i s regionally defined and occurs mostly inthe formal sector. Tabora employment growth proved strongest among the regions. Box 1:Characterizationof RuralNon-farm Enterprises The rural non-farm enterprise sector inTanzania i s quite heterogeneous. Nevertheless, at risk o f oversimplification, some characterization is possible. Firms are owned and solely operated by a male with five years relevant experience and often with primary education. Owners occasionally hire seasonal labor, but seldom look hrther than for householdlabor. The large majority o f enterprises are involved ininformal wholesale or retail trading or processing o f agricultural commodities. Earnings are subsistence with a net income o f some US$ 113 per year - or about one third o f annual income. Most entrepreneurs complement their agricultural income with seasonal non- farm activities. Business operations are dependent on readily available resources within the proximity o f the local community. Investment climate constraints Due to relatively rapid agricultural growth inrecent years, demand exists for more rural non-farm economic activity. However, entrepreneurs are now constrained mainly from the supply-side in their response to this increased demand. Overall, the report finds that access to rural finance and road infrastructure are among the main rural investment climate constraints. Access to finance was generally perceivedby rural entrepreneurs as the main business constraint. This i s not a new findingby ICAs or other studies o f the rural economy inAfrica. The interpretationis complex as this could simplyreflect the desire for additional financialresources. Appendix 4 therefore places the results inthe context o f what i s currently known on rural finance inTanzania. Business constraints are assessed inthree ways throughout the report: 1. Entrepreneurs are asked directly about what they believe are the major constraints to business operations and growth (Figure 1). Rural entrepreneurs generally perceive rural finance, public utilities and road infrastructure as major constraints. More than 60 percent o f entrepreneurs believe that access to finance hampers growth. Regionally, Tabora scores relatively well inthree aspects o f the investment climate -access to finance, transport, and governance. Financing constraints are perceivedas particularly severe in the Lake region, Northern Highlands and Southern zones. Access to public utilities and transport infrastructure i s perceived as a major and severe constraint inthe Western zone. Demand-side and governance business constraints show lower variability and magnitude. 2. The perceivedbusiness constraints are benchmarked with the Sri Lanka Rural and Urban ICA (World Bank, 2004c) and a comprehensive study on rural non-farm enterprises in Africa (Liedholm and Mead, 1999). Such a comparison shows that the overall level o f constraints perceptions i s relatively highinTanzania. A comparisonwith the urban and formal industrybased ICA for Tanzania (World Bank, 2004b) shows that the level o f perceived constraints for the urban and formal industry-based enterprises i s generally higher than inrural Tanzania. A common finding, however, i s the perceptionthat access to finance and transport significantly constraint growth. Interestingly, rural and urban enterprises inTanzania perceive access and costs o f finance as a problem o f almost similar magnitude. -3- Figure 1: Top Five RuralBusinessConstraints,2005 3. As part o f the analysis o f this report, the relative impact o f investments constraints on enterprise growth i s measuredwith objective data at the community level via econometric techniques. Overall, perceived business constraints generally coincide with these more objective measurements. An empirical analysis suggests that better access to roads and rural finance would have the strongest impact on enterprise employment growth. Interestingly, rural cell phone communication ranks third. Demand-side factors related to agriculture rank fourth. For those rural entrepreneurs who do use electricity, increased reliability o f the electricity grid could stimulate growth. Simulations suggest that even marginal improvements o f the rural investment climate could significantly increase non- farm enterprise employment growth. Reflectionsfor policy and future analysis The pilot approach and methodology taken inthis assessment call for a careful evaluation o fthe followingrecommendations, which are presented to stimulate dialogue and future analysis. This Rural Investment Climate Assessment i s the first o f its kindinTanzania, and only a few assessments are available from other regions o f the Bank. Acknowledging the regional dimension and heterogeneity o f rural enterprises i s important. The main issues are: 1. Most rural non-farm enterprises inTanzania are highlydependant on the performance o f agriculture. This suggests that favorable policies and investment for agriculture play a big role. Policies and investments to meet the Government's agricultural growth targets, as described inthe Agricultural Sector Development Strategy, are fundamental for the non- farm rural enterprise sector. Operationalizingthe strategy through the recently developed Agricultural Sector DevelopmentProgram therefore remains a priority. 2. Almost 60 percent o frural non-farm enterprises are trading enterprises. Maintaining favorable internal trade policies may therefore be of utmost importanceindetermining enterprise performance. Revenues o f these enterprises come mainly from local sales. Therefore, internal trade policies set by both local government authorities and line -4- ministries may need to be revisited. Continued enforcement o f recent changes should also be a priority. 3, Road infrastructure i s among the main constraints that impact on ruralbusiness operations. Easing bottlenecks inrural infrastructure i s therefore important. Priority areas are maintenance and rehabilitation o f the existing road network. Differing regional impacts shouldbe considered inresource allocation for rural infi-astructure, particularly if rural employment growth i s a key objective. This should be considered inbothnational level expenditure prioritization and the local government formula base allocations. Prioritization could be based on the expected rates of return to infrastructure and impacts on the agricultural and rural non-farm economy. Private sector participation may require a strengthening o fregulatory institutions and ensuring their independence. 4. Access to rural finance appears to be among the main supply-side constraint -but interpretationi s complex and requires future analysis. Microcredit may offer a tool for promoting rural non-farm activity. However, interventions should pay sufficient attention to the performance o f the agricultural economy. Inbuoyant rural markets, where ongoing agricultural income growth drives demand for non-farm goods and services, injections o f credit can play a role inenabling non-farm entrepreneurs to participate ingrowing market niches. Priorities may be the promotion o frural saving schemes, the establishment o f greater linkages between commercial banks, SACCOs and MFIs, and credit guarantee schemes to offset risks. 5. Cell phone communication reduces transaction costs. Exploring options for better telecommunications via private sector cell phone nodes may therefore be an attractive policy option. This includes the adoption o f a new Electronic Communications Bill, the implementation o f the new licensing framework, and the review o f policies and regulations to generate competition and reduce communication and operational costs. In addition, capacity buildingand the continueduse o f global experiences to enhance the efficiency o fthe telecoms sector could be important. 6. The large share o f informal rural non-farm enterprises may be explainedby the fact that being formal i s costly. Reduce direct costs o f doing business may therefore be important. Despite recent reforms, transaction costs and taxes for formal non-farm enterprises remain very high.Continuation o f business registrationreform and effective implementation at the local level remains a highpriority. Future analysis o frural non-farm enterprises should focus on three aspects: (a) assessment o f the role o f larger firms and their economic linkages, particularly in small rural market towns, (b) identifying entry or mobility barriers to high-return niches within the dynamic part o fnon-farm economy, and (c) for cost-effective interventions, analysis o f a handful o f specific subsectors, and supply chains within them, that holdthe potential for growth. -5- MuhtasariJumuishi Tanzania: Tathminiya Majaribioya Hali ya Uwekezaji Vijijini Kuchochea Ukuaji wa Shughuli Katiti Zisizo za Kilimo Tathmini ya Majaribio ya Hali ya Uwekezaji Vijijini nchini Tanzania inapima mazingira ya kiuchumi ya wajasiriamali wasio wakulima. Tathmini ya majaribio ina malengo makuu matatu: kuuelewa vyema uchumi usio wa kilimo nchini Tanzania; kubainishahali za shughuliza vijijini na vikwazo vya biashara; na kuahsi maeneo ambayo sera za Serikali huelekezwa katika kusaidia kukuza shughulizisizo za kilimo vijijini.Tathmini hiiinajikita katika uchambuzi wa vifunganishi vya data za kipekee za utafiti mpana zilizokusanywa na Ofisi ya Takwimu ya Taifa kati ya Januari na Machi, 2005, ikijumuisha shughuli,kaya, najumuiya katika kanda zote saba za kijiografia nchini Tanzania. Ufuatao nimuhtasari wa matokeo ya utafiti huo. Sifa za Shughuli Zisizo za Kilimo Suala la shughuli zisizo za kilimo za vijijini. Shughuli zisizo za kilimo ni chanzo muhimucha kipato kwa kaya takribani milioni 1.4 nchini Tanzania, ikiwa imeongezeka toka kaya milioni 1.2 mwaka 2001, Shughuli hizinyingihufanyika katika kanda ya Ziwa na katikati ya Tanzania. Katika kapindi cha muongo mmoja uliopita mchango wa shughulizisizo za kilimao katika pato litokanalo na kujiajiri vijijini limeongezeka maradufu. K w a mfano, mwaka 2004, kiasi cha asilimia 28 ya kaya za vijijini ziliripoti kuwa angalau mwanakaya mmoja alikuwa akijishughulisha na kazi isiyo ya kilimo. Shughuli zisizo za kilimo nichanzo muhimucha ustawi wa sehemu kubwa ya Watanzania waishio vijijini. Kaya zinazoendesha shughulizisizo za kilimo zina kipato cha takribani asilimia 24 zaidi ya kile ambacho kaya zisizo na shughulihizo hupata, jambo linaloashiriakuwa fursa za ajira zisizo rasmi katika sekta ya shughulizisizo za kilimo vijijini zinaweza zikasaidia katika kuondokana na umaskini. Shughuli zisizo za kilimo vijijini hutofautiana nu shughuli kama hizo mijini. Mkazo katika biashara nimdogo mno. Wastani wa pato kwa kila anayeshughulika kwenye shughuliisiyo ya lulimo vijijini ni dola za Kimarekani 83 tu, kiasi ambacho kinatofautianasana na lule cha anayejishughulisha na shughuli ndogondogo za mijini nchini Tanzania ambapo wastani wa pato lake nidola za Kimarekani 474 (Benlu ya Dunia, 2004). Kiasi cha nusuya shughulihizi ziko katika maeneo ya vijijini wakati ambapo nusunyingine iko katika maeneo ya masoko vijijini. Shughuli zisizo za kilimo nichache sana ambapo kiasi kikubwa huendeshwa na mtummoja katika kipindi chote cha mwaka. Hata hivyo, wakati wa msimu, shughulihizi mara nyingi huajiri vibarua wa muda mfupi ambao mara nyingi huwa ni wanafamilia. Kiwango cha elimu miongoni mwa wamiliki wa shughulihizi nichajuu kwa viwango vya vijijini, wengi wao wakiwa na elimu ya hwango cha darasa la saba au zaidi. Kukithiri h a biashara za vijijini. Hali yajumla ya shughulizisizo za kilimo nchini Tanzania ina sura mbalimbali. Hata hivyo, shughuliya kiujasiriamali inayotawala sana nibiashara ndogo ndogo. Kiasi cha asilimia 57 ya shughulizisizo za kilimo vijijininibiashara zajumla au rejareja. Matokeo yake, zaidi ya asilimia 75 ya shughuliza Kitanzania huathiriwa sana na misimu, ambayo kwa kiasi kikubwa huathiri ukuaji wa shughuli hizo. Shughuli zisizo za kilimo vijijini nchini Tanzania hununua na kuuza kienyeji, zikiendeshwa katika masoko madogo sana. Ni asilimia 19 tuya shughulihizo ndizo zimesajiliwa rasmi. Tijaya kazi ni ndogo.K w a shughuliya biashara halisi za vijijini, mauzo nichini ya dola 1.5 kwa siku ya nguvukazi. Hata hivyo kuna tofauti. Shughuli ndogondogo zina tija zaidi kuliko kubwa -6- Cjambo linaloweza kutokanana matumizi makubwa ya nguvukazi ya familia ya kaya). Hali hiini hnyume kwa shughuliza mijini, ambako tija ya nguvukazi nikubwa zaidi kileleni mwa mlolongo wa ajira. Shughuli mkoani Tabora zina tija zaidi kuliko sehemu zingine zozote ambazo zilihusika katika utafiti huu.Tofauti za tija kisekta ,hata hivyo, si bayana sana. Sekta ya migodi imeonesha tofauti kubwa, ambapo tija ni kubwa sana kuliko sekta zingine. Picha ya Sekta ya Shughuli Zisizo za Kilimo Vijijini NchiniTanzania, 2005 Jumla ya shughuli 1.2 mi0 Shughuli zilizosajiliwa 19% Sekta ya Uendeshaji Biashara 57% Huduma 21% Uzalishaji 19% Sekta mbili au zaidi 3% Mahali Mijiya vijijini 46% Maeneo ya vijijini 54% Uzalishaji Pato la wastani lajumla kwa kila 113 shughuli (dola za Kimarekani) Thamani ya wastani iliyopatikana kwa kila shughuli (dola za Kimarekani) 83 Nguvukazi Shughuli inayotumia nguvukazi ya familia tu 83% Shughuli inayotumia nguvukazi ya kuajiriwa 1% Familia na nguvukazi ya kuajiriwa Familia zote mbilina nguvukazi 15% ya kuajiriwa Wafanyakazi wa familia (wastani) 1.6 Wafanyakazi wa kuajiriwa (wastani) 0.6 Jumla ya wafanyakazi (wastani) 2.2 Sifa za mjasiriamali Mwanamke 23% (miaka 1-6) 11% Elimuya msingi (miaka 7-8) 69% Elimuya sekondari 17% Elimuyajuu 3yo Uzoefukazini (wastani katika miaka) 4.9 Chanzo: Utafiti wa Tathminiya Hali ya Uwekezaji Vijijini Nchini Tanzania,2005. d2001 HBS Usajili unahusishwa na tija kubwa ya nguvukazi. Shughuli rasmi zina viwango vikubwa vya tija ya nguvukazi kuliko zile zisizo rasmi. Shughuli rasmi ya wastani huzalisha kiasi cha dola za Kimarekani 149kwa mhusika, ikilinganishwana dola za Kimarekani 82 kwa shughuliisiyo rasmi. Masharti ya udhibitiwakati wa kusajili shughulini ghali sana; inakadiriwakuwa ni 1/3 ya mauzo yote ya mwaka. Tofauti za tija kati ya shughulirasmi na zisizo rasmi niza wazi zaidi -7- katika miji ya vijijini. Hata hivyo, utafiti zaidi unahitajikakubaini faida zinazopatikana na urasimi wa shughuli Haliya shughuliza vjijini Kiwango cha uanzishwaji wa shughulii mpya kinaonekana kuwa cha chini zaidi kuliko katika nchi nyingine za Kiafrika. Hiiinawezakuwa imesababishwa na hali ya vikwazo vya uwekezaji au utamaduni wa kutotilia maanani shughuli za vijijini. K w akuzingatia matokeo ya Tathmini ya Majaribio ya Hali ya Uwekezaji Vijijini katika nchi nyingine barani Afrika, wajasiriamali wa vijijini huamini kuwa tatizo la upatikanajiwa fedha namiundombinu mingine ya msingini miongoni mwa vikwazo vikuuvinavyoathiri uanzishaji na ukamilishaji wa shughulihusika. Shughuli nyingimpya nindogondogo, na uingiaji katika sekta isiyo ya kilimo hutegemea kipato kitokanacho na kilimo. Kilimo kinapostawi na mahitaji ya bidhaa au huduma zisizo za kdimo yanapoongezeka, uanzishaji wa biashara huweza kuleta mafanikio. Lakini kilimo kinapodorora au ongezeko la idadi ya watu linapokuwa kubwa, shughulimpya huweza kutafsiriwa kama njia ya kufyonza nguvukazi ya ziada katika shughulizilizopuuzwa. Nchini Tanzania, msukumo wa mahitaji na nguvuya usambazaji huonekana kuwa nimambo muhimuyanayomsukuma mtu kuingia katika sekta isiyo ya kilimo vijijini. Idadi ndogoya biashara huchochea ongezeko la ajira. Kiasi cha 113 ya shughulizilizoanzishwa vijijini (zinazoendeshwa kwamiaka mitano au zaidi) hustawi. Makadirio ya kiwango chajuu cha jumla cha ukuaji wa ajira cha asilimia 4.5 kwa mwaka cha shughulihizi nicha kufurahisha, ikitiliwa maanani kuwa shughuli nyingi katika maeneo ya vijijini hazikukua kabisa. Ukuaji wa ajira hutafsiriwa kikandana mara nyingi hutokeakatika sekta rasmi. Ukuaji wa ajira mkoani Tabora ulikuwa ni madhubuti zaidi kuliko katika mikoa mingine. Kisanduku1: Sifa za ShughuliZisizo za Kilimo Vijijini Sekta ya shughuli zisizo za kilimo vijijini nchini Tanzania hutofautiana. Hata hivyo, bila kujali kama mambo yatakuwa yamerahisishwa sana, inawezekana kuainisha baadhi ya shughuli hzo. Shughuli humilikiwa na kuendeshwa na wanaume wenye uzoefu wa miaka mitano na mara nyingi wenye elimu ya msingi. Mara chache wamiliki huajiri vibarua wa msimu, lakinikwa kawaida hutumia nguvukazi ya kaya. Kiasi kikubwa cha shughuli niza biashara ya jumla au rejereja isiyo rasmi au usindikaji wa bidhaa za kilimo. Mapato niya kujikimuikiwa na kipato chajumla cha dola za Kimarekani 113 kwa mwaka, au kiasi cha 113 ya pato la mwaka. Wajasriamali wengi hujazilia kipato chao kitokanacho na kilimo kwa kipato kitokanacho na shughuli zisizo za kilimo. Uendeshaji wa biashara hutegemea rasilimali zilizopo katika jamii husika. Vikwazo vya mazingira ya uwekezaji Kutokana na ukuaji wa haraka wa kilimo katika miaka ya hivi karibuni, kuna mahitaji makubwa ya shughuliza kiuchumi zisizo za kilimo vijijini. Hata hivyo, wajasiriamali sasa wanakabiliwa hasa na lukwazo cha usambazaji katikajitihada zao za kuitikia ongezeko hili la mahitaji. K w a ujumla, ripoti inabaini kuwa tatizo la upatikanaji wa fedha na miundombinu ya barabara ni miongoni mwa vikwazo vikuu vya hali ya uwekezaji vijijini. K w a ujumla, ukosefu wa fedha ulitajwa na wajasiriamali wa vijijini kuwa ndio kikwazo kikuucha biashara. Haya s i matokeo mapya ya tathmini ya mazingira ya uwekezaji vijijini au tafiti nyingine za uchumi wa vijijini barani Afrika. Tafsiri ningumukwa kuwa hali hiihuweza kuashiriatu shauku ya kupata -8- rasilimali fedha za ziada. Kiambatisho 4 kinabainisha matokeo katika muktadha wa kinachofahamika kuhusiana na fedha vijijini nchini Tanzania . Katikaripoti hii, vikwazo vya biashara vinatathminiwa kwa njia tatu: Wajasiriamali wanaulizwa moja kwa moja kuhusuwanachoamini kuwa nivikwazo vikuu vya uendeshaji na ukuaji wa biashara (Mchoro 1katika tafsiri ya kiingereza). Wajasiriamali vijijini kwa ujumla huonakuwa tatizo la fedha vijijini, rasilimali za kijamii, na miundombinu ya barabarani vikwazo vikuu. Zaidi ya asilimia 60 ya wajasiriamali vijijini huamini kuwa ukosefu wa fedha huathiri ukuaji. Kimkoa, Tabora inafanya vizuri katika maeneo matatu ya hali ya uwekezaji-upatikanaji wa fedha, uchukuzi, na utawala. Vikwazo vya fedha vimechukuliwa kuwa ni vibaya zaidi katika kanda ya Ziwa, kanda ya Nyanda za Juu Kaskazini na kanda za Kusini.Tatizo la upatikanaji wa rasilimali za kijamii na miundombinuya uchukuzi limeonekana kuwa kikwazo kikuuna kibaya katika kanda ya Magharibi. Upande wa vikwazo vya mahitaji na vya usimamizi wa biashara vinaonesha kutotofautianasana na kutokuwa tatizo kubwa. 2. Vikwazo vilivyotajwa vya biashara vinabainishwa pia katika tathmini ya hali ya uwekezaji vijijini na mijininchini Sri Lanka (Benki ya Dunia, 2004c) na utafiti makini kuhusushughulizisizo za kilimo vijijini baraniAfiika (Liedholm andMead, 1999). Ulinganisho huo unaonesha kuwa kiwango chajumla cha mitazamo ya vikwazo ni cha juukidogo nchini Tanzania. Ulinganishowa mijinina sekta rasmi kwakuzingatia tathmini ya hali ya uwekezaji kwa Tanzania (Benki ya Dunia, 2004b) unaonesha kiwango cha vikwazo vinavyotajwa kwa shughuliza mijinina za sekta rasmi ni chajuu kidogo nchini Tanzania. Mwelekeo kwa ujumla ,hata hivyo, nikuwa matatizo ya fedha na usafirishaji huathiri ukuaji wa biashara vijijini nchini Tanzania. Kinachofurahishani kwamba, shughuliza vijijini na mijininchini Tanzania huchukulia tatizo la upatikanaji wa fedha na gharama za fedha kama tatizo lenye uzito sawa. 3. Kama sehemu ya uchambuzi ya ripoti hii, matokeo ya vikwazo vya ukuaji wa shughuli hupimwa kwa data halisi zilizotokana na jamii kwa kutumia mbinu za kiikonometriki. K w a ujumla, vikwazo vilivyotajwa vya biashara huoana na vipimo hivi vyenye uhalisia mkubwa zaidi. Uchambuzi yakinifu unaashiria kuwa hali nzuri ya upatikanaji wa barabara na fedha vijijini ingeweza kuwa luchocheo madhubuti cha ukuaji wa ajira katika biashara. Kinachofurahisha, mawasiliano ya simu za mkononi vijijini yameshika nafasi ya tatu. Sababu za kimahitaji zinazohusiana na kilimo zimeshika nafasi ya nne. K w a wajasiriamali wa vijijini wanaotumia umeme, wanaamini kuwa kuongezwa kwa uhakika wa umeme wa gridi kutachochea ukuaji. Maigizo yanaonesha kuwa hata uimarishaji wa kiasi kidogo wa hali ya uwekezaji vijijini unaweza kuongeza kwa kiasi kikubwa ukuaji wa ajira katika shughuli zisizo za kilimo vijijini. Tafakari kuhususera na uchambuziwa baadaye Mkabala na mbinuzilizotumika katika tathmini hii unadai upimaji wa mapendekezo yafuatayo, yaliyotolewa ilikuchochea majadiliano na uchambuzi wa siku zijazo. Tathmini hiiya Hali ya Uwekezaji Vijijini niya kwanzaya aina yake nchini Tanzania, na nitathmini chache tundizo zilizopo kutoka katika kanda zingine za Benki ya Dunia. Nimuhimukutambua hali na uchangamani wa lukanda wa shughuliza vijijini .Masuala makuuni: 1. Shughuli nyingizisizo za kilimo vijijini nchini Tanzania nihutegemea mafanikio ya kilimo. Hiihumaanisha kuwa sera nzuri na uwekezaji katika kilimo nimambo ya -9- muhimu.Seranavitegauchumivyenye kutimiza malengoya Serikali ya ukuaji wa kilimo, kama ilivyoelezwa katika Mkakati wa Kukuza Sekta ya Kilimo, nimuhimusana kwa sekta ya shughulizisizo za kilimo vijijini. Kutekelezamkakati huukupitia Mpango uliobuniwa hivi karibuni wa Kukuza Sekta ya Kilimo kunabalu kuwa nikipaumbele cha kwanza. 2. Karibu asilimia 60 ya shughulizisizo za kilimo vijijini nibiashara. Kudumisha sera za ndani zinazofaa kunaweza hatimaye kuwa muhimusana katika kuamua utekelezaji wa biashara. Mapato ya shughuli hizi kwa kiasi kikubwa hutokana na mauzo ya kawaida. K w a hiyo, sera za biashara ya ndani zilizowekwa na mamlaka za serikali za mitaa na wizara zinaweza kuhitaji kutazamwa upya. Kuendeleakutekeleza mabadiliko ya sasa ni swala linalopaswa kupewa kipaumbele. 3. Tatizo la miundombinu ya barabara nimiongoni mwa vikwazo vikuu vinavyoathiri uendeshaji wa biashara vijijini. Kuondoa matatizo ya miundombinu vijijini ni muhimu sana. Maeneo ya kipaumbelenimatengenezo na ukarabati wa mtandao wa barabara uliopo. Tofauti za kikanda zizingatiwe katika uelekezaji wa rasilimali kwa ajili ya miundombinuya vijijini, hususani kama lengo kuuniukuaji wa ajira vijijini. Jambo hili lizingatiwe katika vipaumbele vya matumizi ya kitaifa na Yale ya ulekezaji wa fedha katika serikali za mitaa. Vipaumbele vinaweza kuzingatia matarajio ya viwango vya faida kwa miundombinu na matokeo yake kwa kilimo na uchumi usio wa kilimo. Ushirikishaji wa sekta binafsi unaweza kuhitaji uimarishaji wa asasi za udhibitina kuhakikisha uhuru wao. 4. Tatizo la upatikanajiwa fedha vijijini linaonekana kuwanimiongoni mwa vikwazo vya usambazaji-lakini ufafanuzi wake nimgumuna unahitaji uchambuzi zaidi. Mikopo midogo midogo inaweza kuwa nyenzo ya kukuza shughulizisizo za kilimo vijijini. Hata hivyo, uingiliajikati lazima uzingatiekwa makini uendeshaji wa uchumi wa kilimo. Katika masoko ya vijijini, ambako ukuaji wa lupato kitokanachona kilimo huongeza mahitaji ya bidhaa na huduma zisizo za kilimo, utoaji wa mikopo unaweza kuwa na jukumumuhimukatika kuwawezesha wajasiriamali wasio wakulima kushirikikatika shughuliza soko linalokua. Vipaumbele vinaweza kuwa uendelezaji wa mifuko ya akiba vijijini, uanzishaji wa uhusiano mkubwa kati ya benki za biashara na vyama vya kuweka na kukopa pamoja na asasi za kati za fedha, na mipango ya dhamana ya mikopo ili kuepuka hatari. 5. Mawasiliano ya simu za mkononi yanapunguza gharama. Kutafuta njia mbadala ya mawasiliano bora ya simu kupitia mitandao ya simu ya sekta binafsi kunaweza kuwa ni juhudinzuri kisera. HiihujumuishaMswada Mpyawa Mawasiliano, utekelezaji wa mfumo mpya wa leseni, na upitiaji upya wa sera na kanuni ilikuchochea ushindani na kupunguzagharama za mawasilianona uendeshaji. Hali kadhalika, ujenzi wa uwezo na kuendelea kutumia uzoefuwa kiulimwengu katika kuendeleza ufanisi wa sekta ya simu niswalamuhimu. 6. Sehemu kubwa ya shughuli zisizo rasmi zisizo za kilimo vijijini zinaweza kuelezwa kwa kutumia ukweli kuwa kuingia katika sekta rasmi ni ghali. Kupunguza gharama za moja kwa moja za kufanya biashara kunaweza kuwa muhimu.Pamojana marekebisho ya hivi karibuni, gharama na kodi za kuuza na kununua kwa shughulirasmi zisizo za kilimo bad0 ni zajuu sana. Kuendeleza marekebisho ya usajili wa biashara na utekelezaji madhubuti wa katika kiwango chajamii ni swala la kupewa kipaumbele. -10- Uchambuzi wa siku za usoni wa shughulizisizo za kilimo vijijini ulenge mambo matatu: (a) Tathmini ya majukumu ya shughulikubwa na uhusiano wake wa kiuchumi, hususani katika miji midogo vijijini, (b) Kubainisha vikwazo vya kuingia au kusogea kwa shughuli zenye faida kubwa ndani ya sehemu muhimu za uchumi usio wa hlimo, na (c) K w a ajili ya uchambuzi wa uingiliaji kati wenye faida, uchambuzi wa masuala ya sekta mahususi, na kuweka mikufumiongoni mwao, inayotoa fursa za ukuaji. -11- 1.INTRODUCTION This report is organized into five chapters. The first chapter lays the analytical groundwork for assessing the rural investment climate inTanzania and establishes a broader context for the empirical findings. The second chapter describes the profile o f Tanzania rural non-farm enterprises.' The third chapter analyzes enterprise dynamics: start-up, closures and growth. The fourth chapter i s dedicated to the rural investment climate that determines a large part o f this dynamics. The fifth chapter provides reflections for policy and future analysis. The following chapter argues that the rural investment climate measures the economic environment o f the poor. By assessing supply- and demand-side constraints o f the local economy, one can identify critical areas o f reformand prioritize public investments. Changes inmeasures o f poverty inTanzania are largely determined by the performance o fthe rural economy. Private entrepreneurs inthese areas are o f particular importance because they create beneficial links between the non-farm economy and agriculture. Inthis context, ruralnon-farm enterprises contribute to alleviating rural poverty, and are o f growing significance. WHAT I S THE RURAL INVESTMENT CLIMATE? Assessingthe economicenvironment of the poor Investmentclimate refers to the opportunities and incentives for firms to invest productively, createjobs, and expand (World Bank, 2004a). Among others, the investment climate includes factors that are incentives or disincentives for starting and runninga business, including financial services, infrastructure, governance, regulations, taxes, labor, and conflict resolution. The investment climate i s recognized as important to improve output, employment, and enterprise productivity (Dollar et al., ZOOS), all o f which hold the potential to stimulate employment growth and reduce poverty. Micro-entrepreneurs inrural areas create jobs neededto increase income. They provide goods and services and often pay taxes neededto fundpublic investments, but the size o f their contribution largely depends on the environment inwhich private business can operate. Bothrisks and barriers can undermine rural entrepreneurship, hence, it i s important to understand the conditions necessary to develop rural non-farm enterprises. The Tanzanian Rural Investment Climate Assessment (RICA) i s among the first to take a comprehensive look at the business environment inrural areasa2The majority o f Investment Climate Assessments (ICA) has not considered the heterogeneity o f the investment climate across rural areas and industries. The standard approach i s heavily biased toward registered (bigger) enterprises inthe manufacturing sector, which are typically located inurban areas. Rural areas have lower populationdensities, 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 i s that the rural populationtypically ' Non-farm enterprises include all rural businesses engaged innon-primary productive activities. This includes the transformation, transport, and marketing of primary products, but excludes agriculture, forestry, hunting, and fishing. Households primarily engaged in the productionofgoods and services for home consumption are excluded. Tanzania is the first pilot assessmentfor the African continent. Part o f a larger World Bank initiative, these piloting RICAs cover Sri Lanka, Nicaragua, Tanzania, Indonesia, Benin, and Ethiopia. Two related studies are also carried out in Bangladesh and Pakistan. -13- works on farms or inmicro-enterprises. InTanzania, where about 75 percent o f the population resides inrural areas, it i s thus essential to conduct comparable analyses inrural areas.3 Box 2: Design of the Tanzania R u r a l Investment Climate Survey The empirical basis o f this report is a pilot Rural Investment Climate Survey (RICS). Tanzania's National Bureau o f Statistics (NBS) conducted the survey during the months o f January andMarch 2005. Data was collected using face-to-face interviews with members o f selected rural households, community leaders and owners or managers o f non-farm enterprises. Three separate, but interrelated survey questionnaires for households, enterprises and communities were used to collect data (for more details on the methodology, see Appendix 3). The survey covers a total o f 150 communities, 1,239 enterprises and 1,610 households inselected rural areas and small market towns. Agricultural households that operate a non farm enterprise make about 40 percent o f the sample, households that do not operate an enterprise make up another 40 percent, and enterprises that are not household-based another 20 percent. The survey was focusing o n non-farm enterprises and did not cover commercial farms. A stratified multi-stage cluster sampling was used for each survey module. T o ensure representation o f all geographical and climatic zones, mainland Tanzania was initially stratified into seven zones based on agro- ecological characteristics. One region from each geographical zone was selected into the sample. Thus, Morogoro, Kilimanjaro, Tabora, Kagera, Kigoma, Mtwara and Mbeya were selected respectively from the East, Northern Highland, Central, Lake Victoria, West, Southern and Southern Highland zones. Overall, the Tanzania RICS has collected extremely detailed rural data -covering both the farm and non- farm economy. It achieved highresponse rates for all o f the three survey modules. The data also compares favorable with the 2000/2001HouseholdBudget Survey. Unfortunately, the weights that were originally prepared could not be used (see Appendix 3). Inaddition, a wealth o f the survey data could not be fully explored. Merging the three survey modules required careful but time-consuming revisions. Given the uniqueness o f the Tanzanian RICS, these issues may be taken up in firther analysis Understandingconstraintsof rural enterprises Both supply and demand constraints affect rural non-farm enterprises. InTanzania, demand constraints for rural enterprises are mainly related to agriculture. Profits from agricultural production, income earned from non-farm enterprises, and demand generated outside the rural economy can all contribute to effective demand for the goods and services producedby rural entrepreneurs. Which o f these sources o f demand i s the most important depends on the local environment and the degree o f development inwhich the enterprise operates. A virtuous cycle ofdevelopment can arise through the interactiono f farm andnon-farm activities. Agricultural and non-farm activities are linked in several ways -through consumption (demand for final products), production (backward and forward supply o f inputsamongbusinesses), finances (remittance and savings channeled through urban institutions), and labor market links.In Tanzania, agriculture has major growth links to the non-farm sector, but almost entirely through consumption. Estimated expenditure multipliers range fiom two to three -Tsh. 1,000 (US$ 0.77) o fnew household income from crop sales ina remote area can leadto a further Tsh. 2,000 inadditionallocalemployment inthe productionofgoods and services. Thisis ademonstration An urban ICA for Tanzaniahas beenconductedby the World Bank in 2004. A basic comparisonbetweenthe urbanand rural ICA findingsis developedinChapter4. -14- o f the importanceo f agricultural growth, which provides the necessary stimulus to create other economic activities (World Bank, 2000).4 Onthe supply side, a wide variety o f factors determines the ability o frural enterprises to produce goods and services. Supply constraints also affect the cost o f goods and services that may include the state o f local infrastructure, ability to access finance and the cost o f doing so, cost and quality o f labor, quality o f the local regulatory environment, and extent o f competition, knowledge o f market opportunities, and stability and security inthe area. Ifenterprises use old and highly labor- intensive technologies to deliver goods and services, unit costs can be high and productivity low. Under such circumstances, it i s only profitable for enterprises to serve a local clientele because o f hightransaction costs. What i s the role o f the investment climate inthis context? First, private entrepreneurs are needed inthe creation ofthe beneficial linksbetween the non-farm economy andlocal farmers, for example, through agricultural trade. However, unjustified risks, transaction costs, or other barriers to business operations can underminerural entrepreneurship. Second, the investment climate not only affects rural non-farm entrepreneurs but also farm activities. For example, poor access to rural finance and infrastructure hitsboth farm and non-farm activities. This RICA may therefore be useful ina broader context. Assessing the economic opportunities and constraints o frural firms sheds light on the general factors pertinent to poverty and rural development. By quantifying the associated costs o f a weak business environment, this assessment can helpto prioritize rural investments. SNAPSHOT OF TANZANIA'S RURALECONOMY Overalleconomicperformanceimproved Tanzania i s among the world's poorest countries, with a per capita income o f about US$330 when measured at the official exchange rate in2004. Duringmost o f its post-independence history, Tanzania pursuedsocialist policies that led to extendedperiods when economic performance was below the country's potential. Inthe mid-l980s, Tanzania embarked on economic reforms that were not sustained, and after an initialperiod o f economic growth inthe late 1980s, the early 1990s were again characterized by macroeconomic disequilibria and poor economic growth. Inthe mid-l990s, Tanzania resumed its reform course with a commitment to macroeconomic stability. Macroeconomic stabilization was accompanied by wide-ranging structural reforms, includingprivatization o f state owned enterprises, liberalization o f the agriculture sector, efforts to improve the business environment, and strengthening the management o f public expenditures. Economic performance inTanzania has improved consistently over the past decade. Inflation fell from 27 percent in 1995 to 4 percent in2004. The exchange rate i s more stable, with positive effects on agricultural trade, inparticular export crops. Annual average GDP growth increased from about 3.5 percent inthe mid-1990s to about 6.3 percent in2004. A key feature o f the Tanzanian economy is the continued large share o f informal sector activities: estimates suggest that informal activities, including agriculture, may count for up for 60 percent o f Tanzania's GDP (World Bank, 2006~). Also from the development literature, there is overwhelming evidence for the potential of agriculture to cause non-farm economic growth. For a recent overview and analysis, see Tiffin and Irz (2006). -15- Increasedgrowth occurred inall sectors, with industry-inparticular manufacturing-as the fastest growing sub-sector. Increased aid financed an expansion o f government investments and created favorable demand conditions that supported accelerated growth. Inthe manufacturing sector, productivity was a result o f accelerated entry and exit o f firms. An important result o f prudent monetary and fiscal policy, combined with ongoing financial sector reforms, i s the recovery o f credit to the private sector that grew by more than 30 percent annually inrecent years (World Bank, 2006~). Recentincreasein agricultural growth Agriculture plays a dominant role inTanzania's economy, accountingfor nearly 46 percent of GDP and employing around 75 percent o fthe labor force in2004. Agriculture providesthree- quarters o f merchandise exports. Intotal, about 5 million hectares are cultivated annually, o f which 85 percent grow food crops. For the past 10 years, the sector has grown more rapidly than inmost other Africancountries. Agricultural growthhas beenincreasing steadily andat arate higher than population growth since 1999 (Figure 2). Given the magnitude o f agnculture, improvementsinoverall economic growth rely heavily on the performance o f the sector. Figure2: Growthof Agricultural GDP, 1991-2004 (inpercent) 'I- I 6.0 6 - 5 - 4 - 3 - 2 - 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 [-+Annual growth ..... Mean growth 1991-20041 Source: TanzaniaNational Bureau of Statistics Recent increases inagricultural growth stem from overall economic and sector reforms that began inthe mid-1990s. Farmershave respondedto improved incentives and adaptedto the challenging external price environment for traditional exports by increasing the production o f exportable food crops. Starting from a low base, productivity levels have also improved for several crops. However, agriculture i s largely rainfed, and the major constraint for the agricultural and rural sector remains low labor and landproductivity. Inthe absence o fmajor technological breakthroughs or diversification into higher value crops, agricultural growth i s mainly driven by cultivation o f new land and growth o fthe labor force (World Bank, 2006~). -16- Significance of rural economic growth As most people live inrural areas, changes inthe headcount o fnationalpoverty are almost exclusively determined by the performance o f the rural economy. Simulations suggest that rural economic growth has a strong effect on overall poverty (Demombynes and Hoogeveen, 2004). M a p 1: Percentage of Population Below the Poverty Line, 2001 I Source Tanzania Bureau of Statistics (2002) Rural growth patterns observed duringthe 1990s may have led to an initial increase intotal poverty, followed eventually by a decline. According to these simulations, during the first half o f the 1990s per capita incomes actually declined, but inthe mid-l990s, economic growth started to accelerate again. The genuine change inpoverty will only be known with the next representative household survey.' Due to the increase in agricultural and rural growth, however, projections suggest rural poverty may have declined from about 39 percent in2001 to 34 percent in2004. However, according to official figures, ruralpoverty remained virtually unchanged from 1991 to 2001. A comparison o f poverty indicators calculated from the National HouseholdBudget Surveys (HBS) shows that total poverty declined only marginally from 39 to 35 percent from 1991 to 2001. Inrural areas, povertyremainedalmost unchanged (moving from 41 to 39 percent). The poverty rate inrural Tanzania i s substantially higher than inurbanareas, where the incidence o f poverty declined. Only Dar es Salaam experienced a statistically significant change inpoverty levels. Regionally, poverty rates are highinmost regions o f the country, but are highest inthe South, Singida, and along Lake Victoria (Map 1). Rural non-farm enterprises matter Non-farm enterprises are essential for a significant proportion o fTanzania's rural population, and the sub-sector i s o f growing importance. According to community data from the Tanzania Rural Investment Climate Survey (RICS), some 28 percent o f the households report that at least one An updatemaybuild on the Tanzania RICS because the household module has detailed income and asset data -17- member i s working ina non-farmbusiness. This i s still a relatively low number. For Sub-Saharan Africa, fkequently cited figures claim that on average up to 40-45 percent o f households participate inrural non-farm wage and self-employment activities (Barret et al., 2001). Figure3: EvolutionofMain SourcesofHousehold Cash Income,1992-2004 (inpercent) 1- 12% ] A = 10% / x 86% I 4% 85% 2% 85% 0% 84% Source: 2005 Tanzania RICS Box 3: On StructuralTransformationof the RuralEconomyinAsia The rural non-farm economy i s a result o f economic transformation. The process often begins with a countryside dominated largely by self-sufficient and primarilyagricultural households. These produce largely for themselves most o f whatever farm and non-farm goods and services they need. There is little trade or commerce and the prevailing agricultural technologies require few if any external inputs. Some non-farm activities can prosper inrural areas dominated by agriculture, particularly inthe larger villages and rural market centers where they canbetter capture local demand (for example, retail establishments, shops, and agricultural services). Rural towns grow inimportance and as the rural economy continues to grow, trade with larger urban centers also expand and more urban goods become available. These often displace many traditional rural products, forcing structural changes inthe composition o f the rural economy and its towns. G Gradually, as population densities and market access increase, new technologies and modern farm inputs become available, leading to increased agricultural surpluses insome commodities and increased opportunities for trade. Increasing agricultural productivity also raises income, which inturn increases the number and amount o f consumer goods and services that rural households wish to purchase. Household beginsto specialize, taking greater advantage o f their particular skills, resource endowments, and market opportunities. Some non-farm activities that were initially undertaken by farm households for their own consumption expand and are spun off as separate full- or part-time businesses. There is greater trade among rural households and insmall market centers and rural towns. The latter i s beginning to grow more rapidly. Source: Adaptedfrom Rosegrant and Hazel1 (2000) Over the past decade, however, the share o frural non-farm self-employment income has almost doubled inTanzania (Figure 3). The average share o f household self-employment income inthe non-farm sector rose from about 6 percent in 1992 to more than 12percent in2004. While the -18- sale o f food and cash crops i s the main source o f householdincome, the share o f agricultural income has declined over the past decade. A significant body of empirical evidence shows that ruralnon-farm enterprises positively affect household welfare inTanzania.6The Tanzania RICS also confirms a positive impact o f non-farm activities on household income. Table 2 illustrates the incidence o f enterprise ownership and its relation to household income. Self-employed households that runa non-farm enterprise have an income nearly 24 percent higher than that o f those without (Sundaram-Stukel, Deininger and Jin 2007). Moreover, the average earnings from enterprises account for about one-third o f the total income generatedinhouseholds currently operating enterprises. The differences inper capita income are statistically significant and suggest that non-farm activities are important for the generation o f additional income. Table 2: House Income Characteristics With and Without Non-farm Enterprise, 2005 Statistically significant Total Non-enterprise Enterprise difference (5% households households level) Income and its composition Per capita income (TTshs) 288.7 256.8 317.9 YES Share from crop production 40.7% 54.4% 30.2% YES Share from livestock 15.5% 18.5% 11.1% YES Share from non-farm self- YES employment 21.1% 4.6% 35.9% Share from wage 12.3% 9.7% 14.6% YES Share from transfer 10.4% 12.7% 8.3% NO Source: 2005 Tanzania RICS Self-employment inthe rural non-farm sector does not reduce household engagement in agriculture. When comparinghouseholdincome with and without enterprises, the average earning level from agriculture inboth groups i s almost equal, with no statistically significant difference in the level o f agricultural income between the two groups. Inaddition, a comparison o f the average area farmed, about 4.6 acres per household, reveals that both groups farm approximately the same area. One explanation for Tanzanian household-based enterprises engaging inagriculture i s the needto diversify risks across agricultural and entrepreneurial activities (Angermann, 2001). In rural Tanzania, farm and non-farm enterprises are therefore complementary. A decomposition of changes inrural consumptionsuggests that shifts from agricultureto non-agriculturalactivitieshavebeenan importantcontributor to povertyreduction(World Bank, 2006~).Also Lanjouw et al. (2001) and Ellis (2003) find that non-farm activitiesoffer an importantroute out of poverty inrural Tanzania. -19- 2. PROFILEOFRURALNON-FARMENTERPRISES This chapter profiles Tanzania's rural non-farm enterprises sector. The highest enterprise densities are inthe Lake Region and Central Tanzania. About one-half o f the enterprise populationi s located inrural market towns. Formal primary schooling i s an important qualification for entry. Enterprises are typically very small and operate inlocal markets with limited competition. About 57 percent o f rural businesses are engaged inwholesale or retail trading, predominantly inthe informal sector. Local regulatory barriers to entry into the formal sector appear to be insurmountable for most enterprises. Yet, registration has a significant impact on the scale and success o f rural business operations. A labor productivity analysis reveals that self-employment i s largely the most profitable form o f rural non-farm activity. Labor productivity i s highest inTabora, which i s associated with an apparently more favorable rural investment climate than inthe rest o f the country. BASICCHARACTERISTICS Magnitude and location Tanzania's rural non-farm sector includes about 1.2 millionrural enterprises. Regionally, there are large differences. Enterprise densities (the number o f non-farm enterprises per 1,000 households) ranges from only 13 inRuvuma to up to 126 in Shinyanga. The highest enterprise density i s around the Lake region and inCentral Tanzania (Map 2). Map 2: Densityof RuralNon-farmEnterprises,2001 CloO hh] 7-13 I 4 38 33-69 70 8Y f ambia Source 2001 HBS -20- Thispattern tends to mirror the concentration o froads or railways and associated economic activities. However, the concentration i s not related to one single factor. For example, high densities can be found inregions with highagriculturalproductivity, but also inregions that tend to perform poorly. About one-half o f the enterprises are located inrural areas, while the other half i s located in small rural market towns.' Enterprises inrural towns tend to be bigger. Rural non- farm enterprises are almost entirely sole proprietorships. About 77 percent are owned by men in contrast to other Sub-Saharan countries where a larger share are owned and operatedby women. Small-scale activities may explain the low share (Box 4). Butother aspects o fthe Tanzanian sector are very similar to other countries inSub-Saharan Afkica (Liedholm and Mead, 1999). Inparticular, the capitalization o f rural businesses i s low. The median total value for fixed assets claimed by Tanzanian rural entrepreneurs i s only US$ 120 per enterprise.' About one-half o f the enterprises own buildings and land, but only 20 percent own storage facilities and less than 6 percent own machinery or other equipment. The main means o f transport are bicycles or pack animals. Less than 1percent o f the firms own motorvehicles. Enterprisesare young and small Figure4: Number of Workers per Enterprise,2005 Non-farm enterprises inrural Tanzanian are very small. The majority are operatedby one person duringmost parts o fthe year. Self-employment is thus a crucial element inrural Tanzania. However, duringthe peak season, enterprises often employ part-time or casual labor. Figure 4 shows the estimated number o f workers per enterprise, including permanent, part-time, and casual laborer. Some 58 percent o f the enterprises are managed by the owner (self-employment). About 26 percent o f the enterprises employ up to two workers (including the owner) but mostly through household family labor. Very few enterprises are larger than this size. Most enterprises are young and new firms emerge rapidly. The median firm age i s 5 years (Figure 5). There i s little regional variation in enterprise size and age. The only exceptions are 'As defined by NBS geographical classification. Ruraltowns have higher population densities than rural areas and usually have their own markets or social service providers, such as schools and health centers. * Whenever the distribution of a variable is skewed, the median instead of the arithmetic mean is used throughout the report. -21- Kilimanjaro, where businesses tend to be bigger, and Tabora, where enterprises are more experienced. Figure5: DistributionofEnterprisesby Age, 2005 Box 4: Women's MicroenterprisesinRuralTanzania To supplement their husband' income or their own, a large share o f women engages insome cashearning activity. These activities are significant, but at the same time too small to be fully captured through the Tanzania RICS. A small survey among village women inthe Morongo andRuvuma region inthe 1990sfinds that more than 90 percent have at least one income generating activity, and almost two thirds have two. Almost half o f the women think that their mainbusiness i s a reliable source o f income. Interestingly, the most common problem i s not lack o f capital; but rather a lack o f raw materials (the economy was not yet liberalized), inadequate technology, and low market demand. Brewing and beer selling top the list o f women's business ventures: this popularity i s due to substantial income and because it does not require regular labor. This i s followed by cooking and selling food, and by selling agricultural or fishing surplus products. A variety o f other occupations i s also common, for example hair plaiting and hairdressing. Although profits can be high, demand tends to be low, and selling these services i s done on a sporadic basis. Source: Adapted from Tovo (1991) Formaleducationis importantfor entry Formal schooling appears to be an important prerequisite for entrepreneurial activities. Some 75 percent o f rural entrepreneurs have primary education. About 11percent have completedprimary schooling. Secondary education i s less common inrural areas, but it becomes more important when the enterprise i s located ina rural market town. Also an empirical analysis, undertaken for this study, revealed that the probability to start-up a non-farm enterprises raises by 2.3 percent for each additional year o f schooling o f the household head (Sundaram-Stukel, Deininger and Jin 2007). The educational profile o f hired workers i s rather similar to the entrepreneurs -almost all -22- employees working inthe rural non-farm sector have some sort o f primary education. Men and women entrepreneurs do not have significantly different education levels or work experience. Entrepreneurs have as little as 5 years average working experience intheir sector. More than three-quarters o f the entrepreneurs have learned their management skills from relatives, friends, or through self-learning. Only a minority receivedformal vocational training or relevant working experience. Training through NGOs or local associations i s not common generally. Moreover, formal schooling seems to be more rewarding than vocational training. The marginal rate o f return o f 1year o f formal education ranges between 4.8 and 17.5 percent. This i s incomparison to a marginalrate o f return to 1year o f vocational training that ranges between 1.4 and 2.8 percent. Vocational training, therefore, does not appear to be a substitute to formal education for entrepreneurs inrural Tanzania (Kahyarara and Teal, 2006). ECONOMIC ACTIVITIES Ruraltrade dominates The overall landscape o f non-farm enterprises inTanzania i s quite diverse. However, the predominant entrepreneurial activity o f rural non-farm enterprises across all regions i s trading. Figure 6 shows that 57 percent o f rural enterprises are engaged inwholesale or retail trading. Rural services also play an important role with a participation o f 21 percent. The production sector accounts for 19percent o f all enterprise activity. Activities without a clear sectoral association are not very c ~ m m o n . ~ Ofthe trading enterprises, 42 percent o frural enterprises buy and sell unprocessedagricultural commodities, while about 31percent trade processedagricultural products. Despite the dominance o f agriculture inrural Tanzania, only 2 percent o f the trading enterprises are engaged inagriculturalinputtrading. The service sector is dominated bya variety ofpersonal andbusiness services, followed by hotels, restaurants, and the transport sector. About 0.5 percent o f the enterprises are engaged inrural financial services." Seasonality constrains growth Seasonality i s a hallmark o f the Tanzanian rural non-farm sector, a variation largely due to labor supply, demand for rural products, and availability o fraw materials. More than 75 percent o f Tanzanian enterprises are heavily affectedby seasonality. Sales inall sectors usually peakbefore planting and after harvesting seasons. Not surprisingly, the seasonal variation in sales i s particularly pronounced inthe trade sector. Figure 7 illustrates the seasonal patterns from January to December. For eachmonth, entrepreneurs were asked to rate the level o f activity from very low to very high.The busy season arrives earlier for productionand trade enterprises, with a pronounced lull early inthe year among service enterprises. Enterprises that indicated activities in more than one sector experienced fewer fluctuations. The sectors are defined by the largest proportion o f annual sales producedby the establishment in a specific sector. Enterprises are classified here as multisectorial ifthey attribute equal shares of sales revenues o f a sector (for example, 50 percent in agroprocessing and 50 percent in trade). This definition underreports the multisectorial character o f rural activities. loThe detailed sectoral disaggregation for trade and services i s derived from a decomposition o f sales revenues. -23- Figure 6: Sectoral Distributionof Enterprises, 2005 Mining8, Quarrying 1% Agroprocessing Construction sectm Services 21% Financialservices Businesssei 18% Transpolt semces - 9% Restauran 9% 11% Wood and furniture pll, 8% i Unprocessed Textiles and clothing -agricultural products 7% 42% Agricultural inputs 2% Processed agricultural products Source: 2005 Tanzania RlCS 31% Seasonality negatively affects enterprise performance inrural Tanzania inthe following ways. First,with worker participationinbothagriculture and the non-farm economy, many firms experience an ebb and flow o f workers that hampers continuity and ability to upgrading skills. Second, seasonal demand fluctuations can also drive entrepreneurs into informality with a variety o f implications. Third, seasonal rains deteriorate ruralroads and increase transaction costs or make mudroads impassable. Seasonality often implies an additionalneed for short-term capital that cannot be met. For example, manufacturingor construction enterprises cannot buynecessary inputs, even when raw materials are available duringpeak periods. Inconstruction, which i s concentrated duringcertain periods o f the year, raw materials such as cement can not be purchased any time so producers often try to buildtheir inventories o f finishedproducts inanticipation o f seasonal demand peaks, but are constrained bytheir limitedsupply o fworkmg capital (Angermann, 2001). -24- A Production B: Services : c o o 3 I C O S O 839" H.eq ngn 60Lo 6390 H r g r OAierage L C O , 4 C i J 0-0v. 200. 2 C P " O L e q O A C: Trade I D: Mixed 103?0 100~. 80'.0 800. m\eq i ~ n 60% 60% Bn.gr BA.e.agc 4090 4GCa 0 - o n Source: WorldBank (2006a) based on 2005 TanzaniaRICS Localmarkets with thin competition Non-farm enterprises inrural Tanzania buy and sell locally with little access outside markets. Marketing i s often perceived as a key factor for enterprise success, thus one important goal i s to builda distributionnetwork that increases sales and operates at a low cost. Suchdistribution networks include intermediaries such as brokers, wholesalers, or retailers, or simply selling direct to the customer. The latter predominates inTanzania. Only manufacturing enterprises have a more diversified client structure (Table 3). Moreover, most rural enterprises activities are locally defined. Revenue i s generated almost entirely within the enterprise's own ward or district. The degree o f competition inthe rural enterprise sector depends mainly on market attractiveness and industry structure. The relationship between competitors can be a continuum, ranging from conflict to collusion, passing through competition, coexistence, and cooperation along the way. Data from the rural enterprise survey about these relationships i s somewhat scarce. Fromthe available information, however, it may be inferred that market competition inrural Tanzania i s low. On average, a rural enterprise has only five competitors." Similarly, the average market share i s 20 percent (Table 3). One interesting findingi s that informal enterprises inrural market towns face a higher degree o f competition, which could indicate additional barriers to enter the formal sector. Further analysis may clarify whether substantial entry or mobility barriers "AlsoAngermann (2001)suggestsinherempiricalandqualitativeassessmentthatruralmanufacturingenterpriseshavealowdegree o f market competition inrural Tanzania. Future analysis could deepen this assessment by verifying regional price differentials. If markets were integrated and there is competition, there should be little price fluctuation. -25- to "high return niches" within the non-farm economy limit the access to a sub-population o f relatively well-endowed households. Indicator (inpercent) Trade Services Production Buyers' structure I Government 19 19 I 14 I Traders d a d a 20 Agricultural producers & cooperatives 12 15 15 Consumers 73 66 41 Other 6 10 10 Location o f sales revenue Withinward 71 78 75 Within district 18 16 17 Withinregion 8 4 4 Withincountry 3 1 4 Average market share for mainproducthervice 26 20 17 Table 4: Average Market Shares for M a i n Product or Service, 2005 Total Formal Informal Total 20 21 19 Rural area 22 26 22 Rural town 16 18 16 Source: 2005 Tanzania RICS LABOR PRODUCTIVITY Low productivityof informalenterprises One o f the most interesting finding i s the difference between the relative productivity o f enterprises based on registration, size, sector, and region. Survey data show that about 20 percent o f the sampled enterprises claim to be unprofitable because total annual costs exceed sales revenues.'* Standardmeasures such as total value addedper worker are therefore unreliable. Total annual sales per average working day are usedas an approximate indicator for labor productivity, a measure that also considers seasonal employment patterns and can serve as an approximate welfare indicator. l2 An economic interpretation of this number is difficult given the fact that most entrepreneurs simply estimatetheir operatingcosts, which is more difficult that estimatingsales revenues. -26- Tanzania's rural economy i s dominated by informal sector a~tivities.'~ Only 19percent o f the enterprises are formally registered. Enterprises inrural areas are more likely to be informal than their counterparts inrural towns. About 27 percent ofruralenterprisesintowns are registered compared to 14percent inrural areas. There i s little variation by enterprise size, economic sector, or region. Asked about reasons for not registering a business, 54 percent o f rural enterprises claim that there i s no needto register. However, about 30 percent o f businesses perceive registration and license fees as too high. Also, empirical analysis undertaken for this study reveals that enterprise size, registrationcosts, and education are correlated with the decision to enter into formal sector (Appendix 2). Figure 8: Median Sales per Day of Labor by Location and Registration, 2004 I RuralToms Rum1 Areas 2 3 0 1 3 USUDay of Labor Figure 9: Median Sales per Day of Labor by Sector and Registration, 2004 lnfwmd Formd Trading Trading Services Services Manufaduring Manufaduri ng Agrop rocessing Agmprncessing Mining E', Quarrying Mining &Quarrying Construction Construction Twoor More Sectors Two or More Sectors 0 2 4 6 8 0 2 4 6 8 l3Formality o f rural non-farm enterprises is defined as notbeing registered with any government agency and not complying with any legal obligation concerning taxes, safety, or labor laws. This definition somewhat oversimplifies the Tanzanian reality. Many small enterprises operate under various degrees of semi-formal legal status. For example, they do not register butpay taxes to local authorities. -27- Registrationi s therefore associated with the scale o f rural business operations. Figure 8 illustrates that informal enterprises have lower sales than their formal counterparts. Informal enterprises in the manufacturing and miningsectors, however, report higher sales levels than their formal counterparts (Figure 9). Manufacturing activities that take place at home are usually on a very small scale. Being informal inthe mining sector appears to be advantage when exploiting precious metals. Figure 10: Reasonsfor NotRegisteringWith Government,2005 Excessive proceedures Excessivefees 0 20 40 60 (%) ~~ Source: 2005 Tanzania RICS Median sales per labor day for formal enterprises inrural towns are more than double than for their counterparts locatedinrural areas. The difference i s statistically significant and suggests that most informal enterprises can expand their operations only up to a certain threshold. The difference between formal and informal businesses i s more pronouncedfor businesses located in rural towns -inrural areas, productivity gains would be marginal iftransitioning to formal. It therefore appears rational for many enterprises inrural areas to stay informal to avoid associated cost increases. Inrural market towns, however, becoming formal appears an attractive option. What are the reasonsfor low productivity of informal enterprises? Many small and informal firms inrural Tanzania that are profitable prefer to reinvest their revenue into agriculture or eventually set up an additional small enterprise (Angermann, 2001). The reason i s that expanding the rural business beyond a certain threshold would require an entry into the formal sector. Expansionwould mean moving beyond the local market and incurring higher transaction costs, registration, managing a more complex organizational structure, and improved (and more expensive) service, production, or trading methods. -28- IBox 5: Costs and Benefitsof BeingInformal I Informality offers benefits to the rural entrepreneur. Informal entrepreneurs can avoid taxes and fees. Seasonal demand fluctuations make it easier for an informal fmto adjust because o f its simple and flexible technology, and hence it can avoid some costs associated with idle capacity. The ease with which an informal firm can vary its employment level saves labor costs. Inaddition, entrepreneur's skill requirements are less demanding. Government policies and regulations, to the extent that they apply, can be circumvented. There are also other regulations, such as laws pertainingto property rights, which informal firms may avoid. These advantages mustbe weighted against the costs andrisksassociated with operating informally. Rural entrepreneurs may receive fewer services from the state, such as access to electricity and water. Informality also means that it i s difficult to access financial and other commercial services. Informal firms maybeunable to use formal channels o fdispute resolution andhave to rely onlocal networks, confining them to local markets. Source: Adaptedfrom Bigsten and Soderbom (2005) Inparticular, dealing with government agencies involves highcosts. As such, entry into the formal sector under the existing regulatory environment i s a challenge for most rural enterprises. Average one-time official registration costs are approximately 6 percent o f annual gross sales. Formal enterprises have to deal with more than three government agencies, which frequently increase unofficial registrationcosts. These are estimated about 4 percent o f sales revenue. In addition, rural enterprises have to pay annual fees for operating permits or licenses, which can be upto about 4 percent o f annual sales. Furthermore, depending onthe sector, the estimated annual tax rate on enterprise profits i s approximately 20 percent. Formal enterprises may have to pay in total up to one-third o f their sales revenue -a strong disincentive to enter the formal sector. These findings shouldbe placed into a context o f broader local government tax reforms, implemented since 2003. The main elements o f the reform were the abolition o f the flat rate development levy in2003 along with "nuisance taxes," and the abolition o f business license fees for enterprises below a certain size-and capping o f those fees for larger enterprises-in 2004. The pre-reform situation inTanzania had variable market fees, dues distorted relative prices, small start-up businesses were taxed arbitrarily, collection costs were highrelative to amounts collected, taxes were patently not fair (the flat rate development levy was self-evidently regressive), there was little transparency regarding amounts collected and disbursed, and citizens were unable to perceive links between the public services they received (or failed to receive) and the majority o f taxes that they paid. A preliminaryrapid assessment o fthese ongoingreforms suggests that the impacts o fthe reforms varied between groups, but were broadly progressive (World Bank 2006~).Businesses recorded a 14percent decrease intax burdenoverall. Withinthis, medium businesses recorded 11percent less tax, small businesses 36 percent less tax, and microbusinesses (under Tsh. 54,000 turnover) 11percent more tax. The increased paymentby micro-businesses (an exceptionto the otherwise progressive trend) probably results from their nonpayment o f previous business license fees, coupledwith the wider use o f other taxes (such as billboard fees) by councils after the reform: these were all imposed on micro businesses with greater vigor than were the defunct license fees.14 l4Futurework is currentlyundertakenwith World Bank support to identify viable sources of revenueand to modelthe impact of various scenarios.This work includesmodelinglocal taxation options. -29- Table 5: TransactionCosts and Taxes for FormalNon-farmEnterprises,2005 '' Indicator Value One-time registration Numberof days to completeregistration 21 Numberofgovernment agenciesinvolvedinregistration process 3.5 Official registration cost (US$) 24.8 Unofficial registration cost (US$) 25.7 Average official and unofficial registration costs (as % o f gross sales) 5.6 Operating permit Number o f days to obtain operating permit 22 Numberofgovernment agenciesinvolvedinobtaining operating permit 3.4 Official operatingpermit cost (US$) 19.5 Unofficial operation permit cost (US$) 12.5 Average official and unofficial permit costs (as % o f gross sales) 3.6 Operating license Numberofdays to obtain operating license 23 Numberofgovernment agencies involvedinobtaining operating license 2.3 Official operating license cost (US$) 22.5 Average official operating license cost (as % o f gross sales) 2.5 Taxes Average income tax rate for manufacturing enterprises 21 Average income tax rate for trading enterprises 19 Source: 2005 Tanzania RICS. a/ The survey data and technical documentationdoes not provide sufficient details on the type of taxes or how the data was collected. Numbersare approximatedue to small sample size and omittedresponses. Self-employedentrepreneursare mostproductive One-person enterprises are relatively more productive than their larger counterparts (Figure 11). Onaverage, these enterprises generate about US$ 1.5 on sales revenue per working day.15 Overall, rural labor productivity tends to decline with enterprise size. The exception i s enterprises that employ more than four workers. Self-employment appears to be more attractive than wage employment inthe non-farm sector. However, due to the seasonality o f non-farm activities, self- employed entrepreneurs needto substitute part o f their income trough agriculture. Productivity differencesby sector are small Productivity differences by sector are less pronounced. Figure 12shows that the median sales per day range from about US$0.9 to US$ 1.5, depending on the sector. Median sales generated through the services, trade, or construction sectors are almost identical. The only exception i s the miningsector, which generates three to four times more enterprise revenue than any other sector. Miningactivities are mainlyrelated to the discovery o fgold aroundthe country. Tanzania is becoming an emerging gold producer with major gold miningactivities located inthe Biharamulo District (Kagera Region) in the Lake Victoria goldfields. IsThis finding runs againstmainstreamevidence. A possibleexplanation couldbe the use of family labor.That is, largerrural firms coulduse ahigher amount of relatively less productivefamily labor. -30- - Figure 11: Median Sales Per Day of Labor by Size, 2004 Four Person Enretpnse Two Person Enterprise Three Person Enterprise Five or More Person Enterprise , I I 0 .5US$/Day 1 1.5 of Labor Source: 2005 Tanzania RlCS Figure 12: Median Sales per Day of Labor by Sector, 2004 Manufacturing Two or More Sectors Trading Construction Mining & Quarrying Services Ag10 processing I I I I I 0 1 2 3 4 US$/Dayof Labor Source: 2005Tanzania RlCS Regionaldifferences -doesthe ruralinvestmentclimatematter? Enterprises inTabora are more productive than their counterparts inother surveyed regions (Figure 13). This findingi s important because statistical analysis o fthe RICS data revealed this productivity could be associated with a better investment climate inthe region. Inparticular, objective measurements o froad infrastructure and financial constraints at the community level are significantly correlated with higher levels of labor productivity inTabora. Moreover, entrepreneurs inTaboraperceive lower levels o f major and severe business constraints inthree key areas -finance, transport, and governance. The perceived business constraints are significantly lower from the constraints reportedinother regions o f the country. -31- Figure 13: Median Sales per Day of Labor by Region, 2004 Kagera Mtwara Kilimanjaro Mbeya Morogoro Tabora t I I I I 0 1 2 3 4 US$/Day of Labor Source: 2005Tanzania RIGS The remainder o f this report expands on these findings.Chapter 3 discusses factors that affect establishment and growth o f rural non-farm enterprises, while Chapter 4 describes the magnitude and regional dimension o f the rural investment climate. It shows that Tabora i s the only region that has significant employment growth o f rural non-farm enterprises inthe informal sector, the sector in which the majority of enterprises operate. Chapter 5 argues that a combination o f factors that determine the rural investment climate seem to matter. Tabora i s rarely the region with the lowest level o fperceived or objective infrastructure, finance, or governance constraints. Butit i s the only region that scores relatively better inall o f these areas. -32- 3. ENTERPRISEDYNAMICS This Chapter analyzes factors that affect entry and growth o fruralnon-farm enterprises. Enterprise start-up i s closely relatedto income generated from agriculture. These enterprise "birth rates" are inthe order o f 11percent, which i s lower than inmany other countries inSub-Saharan Africa. The majority o f start-ups are small firms. The single most important factor that determines start-up and closure i s lack o f access to formal credit. Employment growth i s regionally defined, occurs inthe formal sector, and i s systematically higher among small and young firms, a powerful finding for those concerned withjob creationinrural Tanzania. ENTRY INTOTHE NON-FARM SECTOR Moderate"birthrate"amongsmall enterprises The non-farm enterprise sector inrural Tanzania i s less dynamic than in comparable countries. The annual rate o f new start-ups was about 11percent in2004,16a rate higher than the 6 to 7 percent rate often reported for industrializedcountries, but substantially lower than the approximate 20 percent reportedfor urban and rural enterprises in other Sub-Saharan African countries (Liedholm, 2OO2)." The comparatively low rate could be a result o f highinvestment constraints, or possibly due to weaker entrepreneurship inTanzania than in other countries. The majority o f new enterprises are small firms -more than 60 percent are created as one- person establishments, mostly inthe informal sector (Table 6). Formal enterprises are more likely to start as relatively big enterprise. A sectoral breakdownreveals that inthe construction, manufacturing, and agro-processing sectors comparatively more enterprises are created inthe category o f having five or more workers. Drivingforcesbehindstart-up enterprises Start-ups o f new rural non-farm enterprises can indicate "good" or "bad" news. When agriculture i s prosperingand overall demand for non-farm products or services i s high, starting a business can mean prosperity. But when agriculture i s languishing or population growth i s high, start-up jobs may simplyreflect the news that firms are acting as a sponge, soaking-up excess workers in marginal activities. Unfortunately, the survey data do not reveal the driving forces behind creation because entrepreneurs were not asked their motivation for starting a new business. Limited empirical evidence from other enterprise surveys suggests that inrural Tanzania both factors may play a role. About one-half o f Tanzania's rural enterprise creation i s due to demand- pull factors, while the other halfis due to supply-pushforces (Angermann, 2001). Enterprise start-up i s closely related to agriculture (Figure 14), with about 55 percent o f start-up capital fiom agricultural production. The survey data provide some support for the finding that both supply and demand drive the creation o frural non-farm enterprises inTanzania. Seventeen percent i s from non-agricultural income sources and more than 13 percent from local friends or relatives. I6Thisnumber i s likely to provide a lowerbound estimatebecause the estimatedoes not include firms that opened and closedduring the survey period.The calculations are based on cross-sectionaldataand follow the methodologyadvocatedby Liedholm and Mead (1999). "SurveyswereundertakenforBotswana,Kenya,MalawiSwaziland,Zimbabwe,Lesotho,Niger,Nigeria,andSouthAfrica,covering more than 50,000 rural enterprises(Liedholm and Mead, 1999). -33- Table 6: Decomposition of Start-up by Enterprise Size, 2005 Percentage distributionby enterprise size Category Number o f workers 1 2 3 4 5+ Overall 63 22 6 4 5 Formal 54 25 5 6 10 Informal 66 21 6 3 4 Sectoral breakdowna/ Trading 65 22 6 4 3 Services 63 23 6 3 5 Manufacturing 62 20 2 1 15 Agroprocessing 58 18 8 3 13 Construction 64 16 2 2 16 Mining and quarrying 80 20 0 0 0 Two or more sectors 47 29 12 6 6 Source: 2005 TanzaniaRICS. a/ Thebreakdown is approximatedue the small number of observations in theproduction sertor, and the small number of observationsfor larger enterprises Figure 14: Sources of Start-up Capital, 2005 I The most important factors that constrain rural entrepreneurs are capital and basic infrastructure. A regression analysis on the determinants o f entry undertaken for this study reveals that credit, along with access to roads, i s significantly correlated with new enterprises (Appendix 2). BUSINESS CLOSURES Why do enterprisesinrural Tanzania close? Tanzanian entrepreneurs perceive lack of access to formal credit as their main reason for closure. A surprisingfindingis that only a minority o frural entrepreneurs attribute "traditional" business -34- failure, such as the lack o f market demand, as an important reason for closures. Lack o f market demand i s often cited amongst the most important causes o f business failure in Sub-Saharan Afiica (Liedholm and Mead, 1999). Another surprisingfindingi s that electricity access ranks second even though a large majority o f rural entrepreneurs are traders without immediate need for electricity. One reason mightbe a difficulty in separating household and enterprise needs. It i s also remarkable that the reasons for closure and preventing start-up are almost identical (Figure 15), which could suggest that those who have closed their enterprises were able to immediately set-up a new business, and were as such not able to separate the constraints. Figure15: PerceivedReasonsfor Closure of Business- and ReasonsPreventingStart-up,2005 Source: 2005 TanzaniaRICS ENTERPRISE GROWTH One-thirdof ruralenterprises are highperformers Employment growth generated by rural non-farm enterprises has been low. The mean annual growth rate o f labor days for the period 2000 to 2004 i s about 4.5 percent. However, employment growth i s being propelled by a minority o f enterprises (Figure 16). The distribution o f average annual employment growth shows that about 60 percent o frural non-farm enterprises have been stagnant, about 5 percent have contracted over the past years, with the remaining 35 percent growing, some quite substantially.'8 The differences are more pronouncedbetween the formal and informal sector. Formal enterprises grew faster. A decomposition o f the relative contribution o f start-up and existing enterprises for 2004 suggests that most ruraljobs (94 percent) were created from the growth o frelatively high-performing firms. Employment generation through new start-ups had a relatively limitedrole (6 percent). The cross-sectional and recall character of the data implies that the growth numbers are approximations. Employment growth could be over- or underestimated depending on firm survival and new entries. -35- Figure 16: DistributionofEnterpriseEmploymentGrowth, 2000-2004 (inpercent) I I Box 6: Typology of RuralNon-farmEnterprise Survivalists. These enterprises have survived the perils o f start-up. Enterprises are often runby those who have no choice but to generate non-farm income. The level o f income may be at the poverty line or below. The enterprise will not grow and eventually collapse. Trundles.These enterprises have beeninexistence for some time. Enterprise turnover i s roughly static and entrepreneurs show no great desire to expand. Income is at the poverty line. Enterprises have added to their workforce since starting but only insmall amounts. Flyers. Enterprisesrunby entrepreneurs who see opportunities for growth. Income levels may meet more than basic needs. Enterprises will hire new labor and may graduate to the small enterprise spectrum. Source: Adaptedfrom Liedholm and Mead (1999); Duncombe and Heeks (2002) Enterprise growth is regionally defined Ingeneral, when mediangrowthrate ofsales and employment are compared, annual sales growth i s always higher than employment growth (with the exception o f formal enterprises inKigoma). One possible explanationi s that only 50 percent o f entrepreneurs are investing intheir businesses: additional income could be used for non-business purposes. The data also show that employment growth i s regionally defined. Significant employment generation over 2000-2004 only took place inKigoma, Kagera, and Tabora (Figure 17) but employment generation was almost entirely due to jobs inthe formal sector. The exception i s Tabora, the only region that also showed significant employment growth inthe informal sector. -36- Figure 17: Employmentand Sales Growthof Formal and InformalEnterprisesby Region,2000- 2004 (Upper bars show medianemployment growth) Source: 2005 Tanzania RICS Box 7: Why Do RuralNon-farmEnterprisesGrow? Growth o f rural non-farm enterprises canbe measured inseveral ways, including sales growth, profits, and number o f working days. Ifmeasurement error were not a problem, defining growth interms o f sales or profits mightbe preferable to a labor-basedmeasure from an accuracy standpoint. However, the Tanzania RICS data rely on a retrospective technique. Since most proprietors do not keep records, they can only estimate their sales or profits, even at the present time. Expecting that guesses from five years ago would be accurate might be folly. As a result, the key measurement o f growth used inthe Tanzanian RICA i s number o f working days. There is no specific growth theory for rural non-farm enterprises, but by combining theoretical insights withempirical evidence, it is possible to identify potentialvariables (Jovanovic 1982; McPherson 1996; Evans 1987; and Sleuwaegen and Goedhuys, 2002). Besides the factors than determine the rural investment climate, the two key determinants o f enterprise growth are age and initial size. "Learning models" o f enterprise growth along with empirical evidence from the United States and developing countries support an inverse relationship between these two variables and enterprise growth. Once firms are established they learn about their efficiency, and competition forces the least efficient ones to exit. Managers learn about ,their efficiency and adjust their scale o f operations accordingly. Young and small firms that are at the initial stage o f uncovering their own efficiency level grow faster. It is thus the youngest along with the smallest firms at start-up that are more likely to createjobs -a powerful finding for those concerned withjob creation inrural Tanzania. Determinantsof enterprise growth Among those firms that did grow between 2000 and 2004, employment growth i s systematically higher among smaller and younger firms. The inverse relationshipbetween size and age on growth suggests an important role for these firms inrural Tanzania. Figure 18 predicts enterprise growth as a function o f size and age to facilitate interpretation o f an empirical analysis -37- undertaken for this study.lgThe estimate i s based on coefficients obtained from a regression analysis o f enterprise employment growth. The analysis shows that after start-up, one-person rural enterprises inTanzania will only grow duringthe first four years andthen remain stagnant. The average enterprise size is about 1.4 employees, a number that coincides with descriptive survey data for one-person start-ups (40 percent growth). By contrast, a bigger enterprise with an initial start-up size o f five employees contracts slightly duringthe first year, but grows relatively fast for five subsequent years (20 percent growth). Thereafter, employment growth declines and the firm eventually start to contract. Figure 18: FirmGrowth,Size and Age inRuralTanzania Ir----- ____ I 2 ................................................................... I- Initialsize = 1-Initialsize = 5 j This "stylized" growth process shed light on the distribution patterns o femployment growth in Figure 18. Employment generatedby rural enterprises i s rather low and occurs mostly for a minority o f small and young enterprises. However, after a certainperiod small enterprises appear to never grow substantially -unless other growth obstacles are considered. The following chapter analyzes to what extent the rural investment climate aligns with this growth process. l9See Appendix 2 for the analysis -38- 4. THE IMPACT OF A BETTER INVESTMENT CLIMATE This chapter assesses the impact o f the rural investment climate on growth o fnon-farm employment. Entrepreneurs generally believe that they are mainly affected by supply-side constraints and that access to rural financial services and roads are the main constraints to rural business operations. More than 60 percent o f entrepreneurs believe that access to finance hampers growth. Regionally, Tabora scores better inthree aspects o f the investment climate -finance, transport, and governance. Perceived business constraints generally coincide with measurements that are more objective. The only exception i s electricity, where reliability rather then access matters. An empirical analysis suggests that better access to markets, finance, and cell phone communication would have the strongest impact on growth. Demand-side factors related to agriculture rank fourth. Even marginal improvements inthe investment climate would affect growth. CONSTRAINTS TO ENTERPRISE OPERATIONS AND GROWTH PERCEPTIONS - Financeand infrastructureas main constraints One o fthe main goals o f a rural investment climate assessment is to identify the leading factors that constrain enterprise productivity and growth. The survey asked entrepreneurs whether they perceived various problems as an obstacle. Although these subjective rankings are not a definitive priority-setting tool, they can be a useful starting point. Additional and more objective data from the community and household survey and quantitative analysis, which are presented inthe next section, can add weight to the survey results. Figure19: Top Five Constraintsof All Rural Non-farmEnterprises, 2005 and Their UrbanICA Ratings,2003 Source: 2005 TanzaniaRICS InruralTanzania, non-farm enterprises are most concerned about access andcosts ofrural finance (Figure 19). About 61 percent o frural entrepreneurs rate financing as a major or severe constraint to business operations. Other important perceivedconstraints are access to public utilities (mainly electricity, and water) and transport (roads). A surprising finding i s that only 29 percent see demand (marketing) for rural non-farm services and goods as a major or severe -3 9- constraint. Since the large majority o f businesses operate inthe informal sector, less than one- thirdo frural entrepreneurs perceive that governance negatively affects ruralbusiness operations. The claim that limited access to public utilities i s the second most important constraint i s difficult to interpretbecause 57 percent o f rural entrepreneurs are traders who may not need electricity or water access for their rural businesses, but instead may reflect their household's desire for better access to services. Benchmarkingnationalandinternationaldata A comparisono fthe ranking o fperceivedconstraints with the urbanor formal industrybased ICA (World Bank, 2004b) reveals several interestingfindings(Figure 19; Figure 20).20Inurban areas enterprises are mainly concerned with taxation (73 percent ratedtax and 65 percent rated tax administration as a major or severe obstacle). Corruption and economic policy are also mentioned as important constraints. By contrast, taxation, corruption, or the overall policy environment are rarely mentioned as a problem inrural areas. The fact that rural entrepreneurs do not perceive these factors as a severe constraint to business operations reflects the highlevel o f informality inrural areas. A finding common among rural and urban enterprises inTanzania i s the perceptionthat access to finance, electricity, and transport constrains business operations. Figure20: Comparisonof SelectedRural andUrbanBusinessConstraintsinTanzania, 2003 and 2005 at Cost of finance., Economic policy BO ,, , .,. ".....l........-"" Taxes `N. Corruption '.. : `. Crime Telecom access Road access Road quality [&Tanzania Rural 2005 .....Taman$ Urban 2003 j Source: 2005 Tanzania RICS and 2003 Urban ICs a/ On a scale from zero to 60: percentage of enterprises reporting major and severe constraints Interestingly, there are differences and similarities inthe level o frural and urban constraints. The level o f perceived constraints inurban Tanzania i s generally higher than inrural areas. Rural and urban entrepreneurs perceive access and cost o f finance as a problem o f almost similar magnitude. This observation points to structural factors inthe financial sector that constrain both rural and urban enterprises. It i s not surprisingto note that smaller, informal enterprises perceive 20Comparisons are based on the full sample of the urban ICs. Comparing constraints for urban and informal microenterprises would show less pronounced differences but are omitted due to small sample size. -40- governance as a smaller constraint due to the weak presence o f governmental institutions inrural Tanzania. Also a comparison with other countries confirms that finance is the main investment climate's bottleneck inTanzania. The comparator countries are Sri Lanka and selected Eastern and Western Africa rural economies. Figure 21 reveals that the overall level constraints perception i s greater in rural Tanzania than inany other country.21The exception i s market demand for which rural Tanzania scores slightly lower. Tanzania scores particularly highon all aspects o f rural finance: access, costs and tedious loan procedures. Internationalcomparison o f rural data shouldbe taken with prudence inthe light o f different concepts o frural space and non-farm activities. Nevertheless, the comparison does confirm the earlier analysis. Figure21: Comparison of SelectedRural BusinessConstraints: Tanzania versus Sri Lanka and Selected African Countries a/ Corruption 50 Market demand Crime Taxes Cost of finance Electricityaccess Access to finance Electricityquality Telecorn access ' Roadquality Road access I1..a,-Sri Lanka 2M)3 +Tanzania 2005 Botswana,Kenya, Malawi, Swaziland,Zimbabwe 1990s ~ ~ Source 2005 TanzaniaRICS, 2003 Sn LankaRICS, Liedholmand Mead (1999) a/ On a scale from zero to 50 percentage of enterpnsesreporting major and severe constraints, or pnncipal problems Are rural enterprisessupply- or demand-sideconstrained? Rural entrepreneurs generally believe that they are mainly affected by supply-side constraints. Demand-side constraints, such as marketingproblems, seem to play a much less significant role. Rural enterprises perceive markets as a lower priority than their urban counterparts (Figure 19). Also an empirical analysis with objective investment climate data at the community-level reveals that demand-side constraints are relatively less important than other supply-side constraints (see Figure 25Error!Referencesource not found.). 2'Benchmarkingperceived constraints with regional and non-regional comparator countries i s a widely used approach. Inthe case o f Tanzania, the non-regional comparison is based on the availability o frural data. Sri Lanka i s chosenbecause it i s the only pilot study that has been completed. However, also preliminary data from Nicaragua and Indonesia suggests that finance i s among the top three constraints. -41- This is an important difference betweenthe rural and urbanICA.InurbanTanzania, enterprises appear to be more driven by demand-side factors. Differences inthe structure of enterprises could be one explanation (Daniels, 2003). Ruralenterprises have low costs of entry, particularly when operating inthe informal sector. Limited capital, skills, or experience do not prevent entrepreneurs from entering the non-farm sector. By contrast, the more capital-intensive industries inthe urban sector require higher skill levels and are therefore more vulnerable to fluctuations inmarket demand. Figure 22: Top Five Constraints of Rural Market Towns Source: 2005 TanzaniaNCS Figure 23: Top Five Constraints of Rural Areas, 2005 Source: 2005 TanzaniaRICS -42- Rural areas versus market towns A comparisonbetweenrural areas and small market towns reveals that constraints related to governance and taxation increasewith the level o f urbanization and market access ( Figure 22 and Figure 23). This finding i s consistent when the rural and urban ICAs are compared. The level o f perceivedbusiness constraints i s generally higher inrural areas than in small rural market towns. Not surprisingly, governance constraints score higher inmarket towns than inrural areas where government presence i s limited. However, the level o fperceived tax constraints does not differ between rural areas and market towns. Independent by the type o f location, the accessibility and cost o f rural finance are perceivedas the main constraints. The perception that finance i s the main constraint to entry and growth o f existing rural enterprises i s therefore robust throughout this report. It echoes a large body o f similar analyses for countries in Sub-Saharan Africa (Liedholm, 2002; Bigstenand Soderbom, 2005). Regional differences Factors that constrain enterprise productivity and growth differ by geographic zone. Map 3. plots the top five business constraints identifiedby rural entrepreneurs -finance, public utilities, transport, marketing, and governance. Three key findings emerge from the visualization. First,finance, utilities, andtransport infrastructure clearly emerge as the main factors that impede business operations and growth, but there are large regional differences. Financingconstraints are perceived as particularly severe inthe Lake region, Northern Highlands and Southern zones. Access to public utilities and transport infrastructure i s perceived as a major and severe constraint inthe Western zone. Finally, the map clearly indicates that Tabora i s the only zone that scores better inthree aspects o f the rural investment climate (finance, transport infrastructure, and governance). With the exception o f rural finance, it i s rarely the region with the lowest level o f business constraints. However, it i s the only zone that scores relatively better inall o f these areas. -43- Map 3: Major and Severe BusinessConstraintsby GeographicalZone, 2005 Finline f 21 4-i 130 61180 mlm81.1uo Source: 2005 TanzaniaRICS Note: Business constraints for geographical zones are approximate. Morogoro, Kilimanjaro, Tabora, Kagera, Kigoma, Mtwara and Mbeya represent the East, Northern Highland, Central, Lake Victoria, West, Southern and Southern Highlandzones, respectively. -44- FINANCE, INFRASTRUCTURE,AND GOVERNANCEOBJECTIVEMEASUREMENTS - Limited access to financial services Access toformal financial services for individual enterprises i s extremely limited. The average distance to the nearest money-lending institution i s 30 kilometers. About 58 percent o f the surveyedcommunities claim to have access to financial services, predominantly through informal channels. More than one-half o f the financial institutions are either cooperatives or other community-based establishments, one-third are government-owned institutions or private banks, and the remaining sources o frural finance are private moneylenders or other sources. Inabout two-thirds o f these communities, however, households can access loans for non-farm investment purposes. Community-level data therefore strongly support the claim from entrepreneurs that access to rural finance i s insufficient. Regionally, access to rural financial institutions i s particularly poor inthe northern and southern parts of the country, but better inTabora (Map 3).'* M a p 4: M e a n Distance To Rural Financial Institutions, 2001 Source 2000/2001 HBS Road and transport infrastructure Community-level data supports perceived constraints from entrepreneurs that business activities suffer from poor road infrastructure. About 17percent o f the surveyed communities do not have a mainroad connection. Of those communities that have road access, about 40 percent are isolated duringthe rainy seasonbecause the roads are seasonal ( Table 7). The available means o f transportation are also limited. Only 28 percent o f communities have public transport services. Bicycles or pack animals are the main means o ftransportation for about 8 percent of ruralhouseholds. 22The Central zone encompasses Tabora, Dodoma, Singidaregions. It i s the driest zone in the country with an annual rainfall of less than 500 mm.The major crops are millet and sorghum. -45- As a consequence o fpoor road infrastructure, the time to travel to markets is high.For rural households, it takes on average more than 80 minutes to travel to the next city, and more than 40 minutes to travel to the next market Map 5 displays spatial patterns of access to rural market towns). Travel time i s slightly lower for enterprise households than for non-enterprise households. The difference i s statistically significant and underlines the importance of infrastructure for rural enterprises. Transportation costs for rural non-farm enterprises to the next market are high-the travel costs to the next market are about Tsh. 90 per kilometer. This suggests that, on average, a rural non-farm enterprise pays approximately US$ 3 to travel to the next market. Map 5: EstimatedTravel Time to RuralMarketTowns Source: Minot et al. (2006) Table 7: RoadTypes Within and OutsideCommunities,2005 Type ofroad Within community (%) Outside community (%) Mud 73 52 Concrete 19 30 Asphalt 3 13 Gravel 3 4 Other 2 2 Source: 2005 TanzaniaRICS -46- Electricity and telecommunications Only 40 percent o f communities are electrified. Not only do most o f the surveyed communities lack access to electricity, but even inelectrified communities most households do not have access to power. As few as 30 percent o f households inelectrified communities use ele~tricityIn~those . ~ communities that have access to electricity, respondingcommunity leaders report that getting a power connection for new businesses took more than 140 days (three times longer than inurban areas as measuredby the urban ICA). The public electricity supply i s not very reliable. It was interruptedon average 71times during2004. Consequently, 73 percent o frural non-farm enterprises couldnot use national gridpower for productivepurposes (Temesgen, 2005b). Development o f electricity and telecommunicationinfrastructure often goes hand inhand, so most entrepreneurs do not have access to basic means o f communication. Only 13 percent ofrural entrepreneurs own a fixed line or cell phone. These number change slightly when disaggregatedby rural areas (8 percent) andrural towns (19 percent). Poor telecommunicationalso impliesthat many entrepreneurs have limitedtimely access to market information. Figure24: Confidencein Conflict ResolutionandLegalEnvironmentby Communities,2005 Source: 2005 Tanzania RICS Local governance and conflict resolution The evidence describing local governance i s somewhat uneven. About two-thirds o f the surveyed communities do not report conflicts with local authorities that negatively affect the business environment, but the other one-third does. A large majority o f communities report confidence in local dispute resolution and contract enforcement mechanisms (Figure 24). When conflict occurs, it is mainlybecause o f disputes over landholdings. About 60 percent o fthese conflicts were resolved through local networks (Temesgen, 2005a). Withinthe past five years, more than 75 percent o f communities claim to have taken action to improve the local business environment. According to community leaders, a large majority o f households participate ina variety o f small projects that aim to improve physical or social infrastructure. Local initiatives that have helped to develop non-farm businesses conditions include improved market facilities, telecommunication, or electrification. 23ThisIS still higher than in the2001 HBS where only some 11 percent of the communitieswere collected to the gnd -47- SIMULATINGGAINSFROMA BETTER INVESTMENTCLIMATE Enterprises are mainly supply-sideconstrained Empirical analysis undertaken as part o f this report suggests that non-farm enterprises could benefit substantially from an improved rural investment climate. Among the main supply-side constraints are infrastructure, finance, and telecommunications. Demand-side constraints that are linked to the performance o fthe agricultural economy rank fourth. The analysis confirms much o f the earlier descriptive evidence. Moreover, business constraints perceivedby rural entrepreneurs are broadly consistent with objective measurements at the community level. They also have a quantifiable effect on enterprise growth. Simulations were conducted on the determinants o f objective investment climate constraints on enterprise employment growth. The simulations are helpful to visualizethe impact o fpotential gains if such improvements could be made, but shouldbe read with caution. They rely on empirical data and methods that are subject to measurement error, do not fully consider some o f the interactions that encompass the rural investment climate, do not address causality issues, and provide little guidance on how to achieve the selected improvements. Figure 25 also illustrates that the estimated impact on employment growth sometimes has a large margin o f error. The simulations are based on a regression analysis o f the determinants o f enterprise employment growth. Key determinants were enterprise size and age, and a number o f objectively measurable investment climate constraints at the community level. Parameters that significantly affect employment growth include transport infrastructure, access to finance, access to cell phone communication, registration with a government office, a reduction inregistrationdays, and reductions inviolent social conflicts. Interestingly, and contrary to the perceptions o f entrepreneurs, access to electricity does not turn out to significantly affect employment growth. Butfor those rural entrepreneurs who do use electricity, reliability matters. A decrease in interruptions could stimulate growth. Because most entrepreneurs are traders, these findings appear plausible. Infrastructure,finance, andcell phonecommunicationare key Removing the constraints o f inadequate road infrastructure and finance would have the strongest effect on employment growth. The simulations assumed a 50 percent improvement o f selected investment climate indicat01-s.~~The ranlung o f a constraint's impact on growth does not change with different assumptions. Figure 25 shows that improved access to markets would have the strongest effect on employment growth, followed by access to rural finance. Interestingly, rural cell phone communication ranks third.Demand-side factors such as higher rural wages due to productivity increase inagriculture or other factors, rank fourth. For those rural entrepreneurs who do use electricity, a decrease in interruptions could stimulate growth. Also legal registration and lower registrationcosts could boost growth. Finally, reduced conflicts could potentially benefit growth. 24For example, mean distance to the next market was assumed to decrease from 17 to 11kilometers. -48- Figure 25: Improvingthe RuralInvestmentClimate: EstimatedGainson EnterpriseEmployment Growth Roads 5010 reauction in aierage market a stance i Finance 50% increasedaccess to lending i Communications 50% increased access to cell phones Demand 50% I increaseof I I 8 agr cultLral hage rate Electricitysupply 50% decrease in interruptions Registration 50% increaseof formal registration Social cohesion 50% reductionof violent conflict Business environment 50% Upper and lowerthin bars show reductionof standarderror of simulation registrationtime 0.0% 0.1% 0.2% 0.3% 0.4% Source: 2005 Tanzania RlCS Marginalimprovementsinthe investmentclimate matter for growth The simulations show that the estimated effect of selected measures o fthe investment climate would range from less than 0.1 up to about 0.3 percent on annual employment growth. H o w big i s this for a typical enterprise at start-up? Over the mediumterm, even a marginal improvement in the rural investment climate could be significant and lift the rural economy out o f stagnation. Figure 26 builds on the simulations andplots the stylized enterprise employment growth process. The scattered lines assume that a broad improvement o f the rural investment climate would result ina 0.1percent increase inemployment growth (much lower than the estimated impact o f individual constraints ranging from 0.04 up to almost 0.3 percent, respectively). Even a marginal improvement o f the investment climate could provide quite substantial gains for the rural economy. Over a 10-year period, a one-person enterprise would reachthe two-worker category and experience continued growth. After an initial period o f stagnation, a five-person enterprise would generate on average up to four additionalworkers. Overall, this i s inline with the findingspresented inthe previous chapter. Inrelative terms, smaller rural enterprises would benefit most from an improved investment climate. Over a 10-year horizon, a one-person start-up firmcould double while a five-person start-up enterprise could grow by 80 percent. -49- Figure 26: Visualization of BusinessConstraints' Impact on Employment Growth over 10 Year- horizon 10 - ... ... .................. ......................................... ~ , - Initial size = 1 Initial size = 5 Initial size = 1 and 0.1% annual growth increase 8 --- Initial size = 5 and 0.1% annual growth increase ........ ........................................................................... / / 0 0 1 2 3 4 5 6 7 8 9 10 Firm age (years) Source: 2005 Tanzania RICS Box 8: Productivity Analysis of Tanzania's Rural Non-farm Enterprise Sector Also a detailed analysis o f the determinants o f entry, investment, andproductivity o f non-farm enterprises suggests significant potential gains from improvements inthe rural investment climate. The study, which was undertaken for this report, assesses the impact o f entrepreneur perceptions and objective investment climate constraints at the community level via multiple regression analysis. Itfinds that elimination o fmajor business constraints could: Increase participationinnon-farm entrepreneurial activity by about 8- 12 percent, Expand new enterprise investments by about 20 percent, and Boost total factor productivity by about 28 percent. Access to markets and roads mainly affects total factor productivity. By contrast, financial constraints impact more on entry into entrepreneurial activity, and o n new investments. Small rural enterprises suffer more from poor investment climate constraints than bigger firms, which are often able to overcome constraints. Source: Adaptedfrom Sundaram-Stukel, Deininger and Jin (2007). For details see the Appendix -50- Finally, careful collection o f information on inputs, outputs, and inventories for different enterprise types allows analysis o fthe effect o f specific constraints on total factor productivity. Doing so, this study find strong evidence o f infrastructure-related constraints beingcritical for the ruralnon-farm sector to expand and be most productive. With few exceptions, the productivity o f small enterprises i s more severely affectedby investmentclimate constraints than that o f large ones who are ina muchbetter position to take action to avert such constraints. This suggests that, inthe case ofTanzania, policiesto try andremoveconstraints ofthis nature wouldbe avery important strategy to facilitate pro-poor growth. -51- 5. REFLECTIONSFOR POLICY AND FUTUREANALYSIS The rural non-farm economy inTanzania has grown too big for policymakers to ignore. Tanzania encompasses more than one millionrural microenterprises. Results o f this pilot survey suggest that some 20 percent o fruralhouseholds have at least one family member working ina rural non- farm enterprise. Evidenceregularly suggests that rural nonfarm enterprise activity i s a key source for income growth and diversification for the rural poor inTanzania (World Bank 2006c, Lanjouw et al. 2001, and Ellis 2003). This pilot assessment describes aruralmicroenterprise sector struggling to compete ina difficult business environment. About one thirdo frural enterprises are growing. A number o f factors need to be addressed ifthe full potential o f private sector-led growth inrural areas i s to be unleashed. A central findingo fthe report is that evenmarginalimprovementso fthe ruralinvestment climate matter. Perceived constraints and constraints measured with objective data at the community- level are similar, suggesting some robustness o f the empirical results. Moreover, major findings o f this assessment also compare favorably with earlier empirical work on rural microenterprises for nine African counties inthe 1990s (Liedholm and Mead, 1999). However, it i s important to emphasize that the assessment and recommendations are based on a pilot approach and data collection exercise. This Rural InvestmentClimate Assessment i s the first o f its kmdinTanzania, and only a few o f these assessmentshave been completedelsewhere by the Bank." Acknowledging the regional dimension and heterogeneity o frural enterprises i s important. Overall, this calls for a careful evaluation o f the following reflections. These are thought to stimulate dialogue and future analysis. Muchremains to be learned about the rural investment climate and its impact onnon-farm enterprises. AGRICULTURE AND RURAL TRADE Policies and investment for agriculture Policies and investments to meet the Government's agricultural growth targets, as described in the Agricultural Sector Development Strategy, are fundamental for the non-farm rural enterprise sector. Most rural enterprises inTanzania are highly dependant on the performance o f agriculture. Increases inagricultural incomes generate local demand for goods and services, and agricultural savings to invest inthe start-up and expansion o f non-farm rural enterprises. The improved performance o f agriculture since the mid 1990s has induced an increase innon-farm enterprise growth. Operationalizingthe strategy through the recently developed Agricultural Sector Development Programtherefore remainspriority. The emphasis on agriculture aligns with the finding that, inthe survey year 2005, supply-side constraints are more important than demand-side constraints. Demand exists for more rural non- farm economic activity due to the relatively rapid agricultural growth inTanzania inrecent years. Potentialentrepreneurs are now constrained intheir response to this increased demand. However, over the longrun, sustained agricultural growth i s the basis for the development o f the rural non- farm sector inTanzania. Similarly, inresource-poor areas and inregions with unexploited potential, restarting agricultural growth will remain a priority. However, where a more buoyant economic base exists, efforts are neededto promote non-farm activities. 25Only the Sri Lanka Rural and Urban InvestmentClimate Assessment has been disseminated. Draft analysis of the RICS has also been undertaken for Nicaragua and Indonesia but still has to be completed. -52- Public investments and policies could benefit both agricultural and non-agricultural growth. While an agricultural-ledrural growth strategy does require specific investments, for example agricultural extension and research, many other investments or interventions actually lie outside agriculture. Both agricultural as well as rural non-farm activities would benefit from these investments or interventions. Internaltrade policies As almost 60 percent o fruralnon-farm enterprises are trading enterprises, trade policies are of utmost importance indetermining enterprise performance. Revenues o f these enterprises come mainly from local sales. Therefore, internal trade policies set by both local government authorities and line ministries should be revisited. Inparticular, local taxation o f trade across district boundaries should be avoided. There have been recent improvements inthese policies and associated regulations. The Government re-issued a notice in 2003 to remove physical controls on crop movements within and across Tanzania's borders. The number o f taxes has beenreduced, including removal o f the double tax (at point o ftransit and original sale) for crops that moved through formal market channels. Continued enforcement o fthese recent changes should be a priority, particularly local level tax compliance with the Public Finance Act. FINANCE, INFRASTRUCTUREAND INSTITUTIONS Access to rural finance Access to finance i s perceivedto be the biggest constraint to business start-up and expansion, more so than interest rates. Butinterpretation o f this finding i s complex. Microcredit could offer a tool for promoting rural non-farm activity. However, interventions should pay sufficient attention to the performance o f the agricultural economy. Instagnant rural markets, injections o f microcreditmay increase the number o f start-ups -but not increase enterprise growth. Microcredit in stagnant rural areas could therefore merely "redistribute poverty" as new entrants divide a fixed pie into ever-smaller increments. Inbuoyant ruralmarkets, where ongoing agricultural income growth drives demand for non-farm goods and services, injections o f credit can play a role inenablingnon-farm entrepreneurs to participate ingrowing market niches. Promotingrural saving schemes could be a priority. Over seventy percent o f start-up capital for rural enterprises comes from own savings with about 25 percent from friends or family and informal sources. Only 1percent i s from private moneylenders and 1percent from Bank Loans. Greater linkages between commercial banks, SACCOs, and MFIs could be made to improve access to credit. Each has their own advantages, the deeper outreach and low cost structure o f SACCOs, more rigorous credit assessments, monitoring and enforcement mechanisms o f MFIs, and more financial resources o f commercial banks. Fiscal incentives for rural banlung facilities could be established. Private sector guarantee schemes to offset risks could be promoted. Also, enhancing the capacity o f rural institutionsthrough training may equally be important. These options could form part o f the activities under the Second Generation Finanical Sector DeepeningProgram. Bottlenecks in roadinfrastructure The roadnetwork is important to reduce transactions costs. Priority areas are maintenance and rehabilitation o f the existing roadnetwork. Differing regional impacts shouldbe considered in -53- resource allocation for rural infrastructure, particularly ifrural employment growth i s a key objective. This should be considered inboth national level expenditure prioritization and the local government formula base allocations. Priontization should be based on the expected rates o f return to infrastructure and poverty impacts. Private sector participation would require a strengthening o f regulatory institutions and ensuring their independence. Options for better cell phonecommunication Cell phone telecommunications reduce transaction costs, by improving information flows. The analysis shows that this contributes significantly to the development o frural non-farm enterprises. Advances intechnology, as well as card phones and mobile phones, are contributing to rapidly expanding networks, lower costs and more affordable telephone systems. Phones themselves often create small businesses with landlines and mobile phones `rented' to occasional callers. However, inTanzania, tariffs remain highand teledensity i s one o f the lowest inthe region. Poor telecommunications access has been the norm for most rural communities in Tanzania. Explore options for better telecommunications via private sector cell phone nodes. This includes the adoption o f a new Electronic Communications Bill,the implementation o f the new licensing framework, and the review o fpolicies and regulations to generate fair competition and reduce communication and operational costs. Inaddition, capacity buildingand the continueduse o f global experiences to enhance the efficiency o fthe telecoms sector would be important. Costsof doingbusiness The large share o f informal rural non-farm enterprises can be explainedby the fact that being formal i s costly. Transaction costs and taxes for formal non-farm enterprises remain very high. These are estimated at about 30 percent o f gross sales at the time o f the 2005 survey. While local government `nuisance' taxes were abolished in2004, the overall tax rate remains high.However, the abolition o f licensing, registration and permit costs could increase enterprise revenues, and reduce welfare losses that stem from the lack o f access to formal credit. Continuation o f business registrationreformand effective implementation at the local level remains a highpriority. There has been progress inreducingbusiness registrationcosts since 2004 with the abolition o f licensing fees for small enterprises and the removal o f annual licensing requirements. However, it will be important for the Bill on Business Activities Registrationto address adequately all fees on business registration. The Bill- submitted to Parliament in2005 - simplifies start-up procedures for businesses and eliminates the multiplicity o f regional and national licenses by introducinga single registrationcertificate. It also eliminates the necessity to renew licenses on an annual basis as well as activity specific fee schedules. FUTURE ANALYTICALWORK Role of larger firms andtheir economiclinkages This assessment shows that assistance aimed at small and younger firms maybe worthwhile. The identificationo f this enterprise segment i s a powerful finding for those concerned withjob creation inrural Tanzania. However, further validation may be worthwhile for two aspects. First,the smallest firms incertain sectors maynot be the best places to start given that there is ample evidence that small firms are often engaged in survival activities and are thus less likely to -54- graduate into higher size categories (Liedholm and Mead, 1999). Second, larger firms frequently shape opportunities for smaller enterprises. Because o fthese economic linkages, assisting larger rural enterprise development insmall rural market towns may be important to unleashing growth opportunities. Entry barriers into non-farm sector Future work could identify entry or mobility barriers to high-return niches withinthe dynamic part of the non-fam economy. Tanzania's heterogeneous rural non-farm sector offers opportunities for the rural poor as well as the rich. Poor rural households could seek economic refuge through distress diversification into low-skill nonfarm activities. Simultaneously, the more affluent households could participate inmore sophisticated, high-productivity activities. These entry barriers may have the potential to limit the access for a subpopulation o f relatively well- endowedhouseholds. Subsector and supply chain analysis Future work could help identify a handful o f specific subsectors, and supply chains within them, that hold the potential for growth and participation by the rural poor. With more detailed analysis, identification o f a limitednumber o f key missing ingredients offers prospects for cost-effective intervention. Concentration on a single trade or industrygroup likewise serves to focus strategic injections inways that can open up growth opportunities. Available diagnostic tools used elsewhere in Sub-Saharan Africa provide techniques for evaluating current supply chain structure, dynamics and opportunities for expanding output and income for many like firms at once. This leverage, focused on supply chains where the poor participate, will be instrumentalinforging cost-effective, equity-enhancing interventionsto promote non-farm enterprise activities inrural Tanzania. -55- APPENDICES APPENDIX1:SUMMARY TABLES Table 8: Enterprises ReportingMajor and Severe Constraints to Growth and Operations, 2005 (in percent) Constraints Finance Utilities Transportation Marketing Governance Business Taxation Land Labor Other Registration Policy Policy Policy - Region Kilimanjaro 63 47 30 34 26 17 15 14 7 30 Morogoro 53 23 14 23 13 14 10 3 2 I O Mtwara 76 66 47 44 30 27 26 12 12 34 Mbeya 60 40 32 25 40 18 22 19 4 21 Tabora 35 44 20 23 18 12 10 4 4 14 Kigoma 77 85 51 36 28 29 24 19 5 20 Kagera 62 61 14 25 25 14 17 20 4 27 Industries Production 65 49 35 30 21 19 16 11 5 22 Service 67 52 35 28 28 17 14 12 4 20 Trade 59 50 28 31 27 19 20 12 5 21 Location Rural towns 56 35 18 23 28 16 19 12 4 19 Rural areas 65 61 39 34 25 21 18 13 6 24 Enterprise Age Less than 3 yrs 58 50 26 28 28 17 18 10 7 20 3-5 yrs 60 46 30 30 24 21 17 14 6 22 6-10 y r ~ 60 49 33 31 30 18 24 13 6 23 More than I O yrs 64 50 29 28 25 18 15 14 4 21 Size 1laborer 65 47 30 29 28 17 17 12 5 23 2 laborers 62 51 35 37 23 24 21 15 7 24 3 laborers 72 67 38 39 38 24 21 19 4 37 4 laborers 59 50 20 18 23 14 16 23 7 14 5+ laborers 70 51 21 18 28 16 22 18 4 18 Soiirce: 2005 TanzaniaRICS -56- Table 9: Top five Major or Severe Constraints Preventing Households from Starting a Non-farm Enterprise (percentages among households without non-farm enterprise) Region Finance Utilities Transport Marketing Governance Other Kilimanjaro 49 29 7 4 6 5 Morogoro 48 24 11 8 4 6 Mtwara 46 26 10 10 2 8 Mbeya 73 8 10 2 2 5 Tabora 39 19 14 8 13 7 Kigoma 57 10 17 5 1 11 Kagera 51 22 17 6 4 1 Source: 2005 Tanzania RICS Table 10: Top five Major or Severe Constraints Causing Households to Close Their Non-farm Enterprise (percentages among households with closed non-farm enterprise) Region Finance Utilities Transport Marketing Governance Other Kilimanjaro 14 43 7 18 7 11 Morogoro 37 36 7 7 4 7 Mtwara 53 18 7 15 2 5 Mbeya 61 12 5 13 2 6 Tabora 28 24 22 13 9 3 Kigoma 45 19 17 11 0 8 Kagera 28 31 25 9 3 3 Source: 2005 Tanzania RlCS -57- Table 11:BasicEnterpriseand Community Characteristicsby Region,2005 a Characteristics Total Kilimanjaro Morogoro Mtwara Mbeya Tabora Kigoma Kagera Staffing Average number o f laborers (including 2.2 2.7 3.1 1.5 1.6 2.7 1.9 2.2 1laborer (%) 47 66 65 65 22 45 63 47 2 laborers (%) 28 12 26 26 49 34 16 28 3 laborers (%) 14 4 6 5 7 12 4 14 4 laborers (%) 4 3 1 2 14 4 10 4 5+ laborers (%) 6 15 2 3 7 6 7 6 Average number o f household laborers 1.6 1.4 1.3 1.3 1.2 1.6 1.2 1.3 Average number o f hired laborers 0.7 1.2 1.1 0.1 0.3 1 .0 0.3 1.0 Average owner's ewerience (vrs'l 4 9 5 7 5 3 4 7 3 X 4 x 6 4 5 4 ManagersIOwners wl primary education 80 76 84 84 80 67 94 70 Managerdowners w/ secondary 17 22 12 14 17 26 6 27 Managerdowners w/ tertiary education 3 3 3 2 3 7 0 3 Male manager (%) 77 78 64 84 79 87 87 62 Age and Sector Age <3years (%) 20 17 21 27 16 15 13 29 Age 3-5 years (%) 26 25 31 22 31 19 25 25 Age 6-10 years (%) 23 19 17 26 26 22 28 24 Age >10 years (%) 31 39 31 25 27 44 35 22 Industry(%) 22 41 44 30 7 33 57 14 Services (%) 21 23 25 22 32 30 38 24 Trade (%) 57 58 58 61 72 61 70 80 Ownership andFormality Sole proprietorship (%) 92 89 86 95 95 92 89 96 Registered (%) 19 21 14 14 19 28 18 26 Medianregistration fee (US$) 30 18 64 17 47 17 28 28 Median license fee (US$) 23 30 30 16 23 23 18 46 Median federal & local tax & levy fee 46 29 46 39 46 134 30 54 Sales and Assets Median value added (US$) 113 119 46 64 188 321 188 20 Median value addedper worker (US$) 83 73 28 46 138 115 135 18 Seasonal sales (%) 75 74 73 68 76 55 98 84 Average local market share (%) 20 37 25 29 12 26 35 34 Mediannet assets (total assets -total 230 387 157 134 189 272 420 786 Medianvalue o f all fixed assets (US$) 193 367 152 73 165 230 384 55 1 Median investment infixed assets (US$) 9 9 51 7 0 0 142 6 Infrastructure (community level) Average time to nearest city (minutes) 84 59 87 93 61 110 122 76 Average distance to nearest city (km) 20 14 21 19 15 36 14 18 Mainroad connecting community to city 83 85 96 80 77 70 93 85 Distance to nearest market (km) 9 8 11 10 9 13 4 6 Average distance to nearest financial 16 13 21 15 15 14 19 19 Access to financial services in 83 95 83 60 100 74 73 81 Educationo f government official (yrs) 7 8 7 6 8 7 8 8 Time current government inpower 30 34 29 16 22 39 24 48 Electricitv within the communitv (%) 40 60 42 35 20 d , I . _ 20 67 13 Source: 2005 Tanzania RICS a/ Median values used inplace of mean to correct for outliers in select indicators -58- Table 12: National Real Prices for Goods and Services in RuralCommunities, 2002-2004 (averages) Statistically Annual Significant Growth Rate Change from 2002-2004 2002-2004 Goods and Services 2004 (in%) (10% level) Petroleum(US$/Liter) 0.62 2.0 No Fertilizer (US$/20kg bag) 6.52 3.3 No Cement (US$/SOkg bag) 8.57 9.7 Yes Galvanized steel sheet for roofing (US$/3 meters) 5.34 3.6 Yes Electricity - less than 100 Kwhconsumption (US$/Kwh) 0.06 -11.1 No Electricity - more than 100 Kwhconsumption (US$/Kwh) 0.11 -2.8 No Telephone call to nearby region (US$/Minute) 0.35 1.o No Cell phone call to nearby region (US$/Minute) 0.37 -2.8 Yes Commodity transport to nearby district (US$/Mt) 18.00 -2.6 N o Male daily casual laborer wage rate inagriculture (US$/Acre) ai 19.31 6.5 Yes Male daily casual laborer wage rate inagriculture (US$/Day) 1.07 5.5 N o Male daily casual laborer wage rate inconstruction (US$/Day) 1.82 14.7 Yes Male daily casual laborer wage rate inpublic works (US$/Day) 1.40 4.1 No Female daily casual laborer wage rate inagriculture (US$/Acre) 14.53 5.1 No Female daily casual laborer wage rate inagriculture (US$/Day) 1.27 7.0 No Female daily casual laborer wage rate inconstruction (US$/Day) 1.70 7.4 Yes Female daily casual laborer wage rate inpublic works (US$/Day) 1.11 7.5 Yes a/ Male wages are significantly different than female wages (10 percent level) -59- 0 3 0 0 0 0 0 1 0 3 0 1 . 2 - r - ' br v i m - 0 0 3 0 0 0 0 0 S W w N N 1 m w - f - - N 8x Y t? rd 0 a s n - i a d - h paumc APPENDIX 2: REGRESSIONANALYSIS Determinantsof ruralnon-farmenterprise employment growth This appendix presents results from an analysis o fthe impact o f investment climate constraints on rural non-farm enterprise employment growth.26Following Evans (1987), the basic empirical model i s a general growth function g in size and age: where S,,and &are the size o f a firm for the period t' and inperiod t, respectively, andA,i s the age o fthe firm inperiod t. Inaccordance with the main arguments o f this report, this functional relationshipcan be moderated through a set o f investment climate variables IC: G =g(St,At)eb" The equation thus suggests the following regression framework: ln(S'r) -ln(") a,ln(S,) =const + +a, ln(A,) +a3 ln(S,) x ln(A,) + 2biIC +E, d i=l where the dependent variable corresponds to the average annual growth rate, d stands for the number o fyears over which the growth rate i s measured, and a and b are the coefficient vectors. The partial derivates o f growth with respect to size and age allow testing for alternative theories o f firm growth. Learning models o f firm growth such as Jovanovic (1982) suggest that these shouldbe negative. Inline with Evans (1987), higher order expansions o f the logarithmic expression for firm size and age, and an interactionterm between size and age are included inthe regression. The basic framework also incorporates six regional dummies and a dummy for enterprise participation inthe formal sector. A basic regression is IWI without investment climate constraints on average real sales and employment growth as a first step. Ifmeasurement error were not a problem, defining growth in terms o f sales or profits might be preferable to a labor-based measure. However, the Tanzania RICS data rely on a retrospective technique. Since most proprietors do not keep records, they can only estimate their sales or profits, even at the present time. It i s likely that measurement errors o f sales growth make the regression to perform poorly (Table 14). The key basis for the following growth estimate i s therefore the number o f working days. Changes inworking days are a more robust measure o f enterprise growth inrural areas (McPherson, 1996). For rural entrepreneurs that do no keep books or records, a measurement i s easy to remember. 26Preparedby Josef Loening. Approaches that analyze microenterprise growthinAfrica, using size and age as main explanatory variables for employment growth, are Sleuwaegen and Goedhuys (2002) and McPherson (1996). -61- Table 14: Determinants of Employment and Sales Growth, 2000-2004 Dependent growth vanable: Annual growth Annual growth of of labor days sales 2002 Explanatory vanables 2000-2004 2004 (1) (2) Inage 0.400* 0.057 (2.22) (0.62) Inage squared -0.156* -0.047 (-2*18) (-1.07) Inage cubic 0.020* 0.008 (2 19) (1.03) Insize -0.329** -0.990** (-19.6) (-3.65) Insize squared 0.228** 0.179** (19.5) (3.54) Insize cubic -0.037** -0.011** (-16.3) (-3.51) Insize x Inage -0.021 ** 0.005 (-3.42) (0.42) Formallyregstred 0.009 0.009 (1.70) (0.47) Constant -0.1 19 1.871** (-0.82) (3.93) Regional dummies YES YES Observations 722 828 AdjustedR-squared 0.65 0.18 Robust t statistics inparentheses *significant at 5%; ** significant at 1% Source: 2005 Tanzania RICS Table 14 shows that the regression on employment growth performs relatively well. The regressionreports robust t-statistics to correct for heteroskedasticity. The relationshipbetween size and age on growth i s nonlinear. The results are stable inthe sense that usingaverage instead o f initial size inthe regressions to address the problem o f transitory fluctuations o f enterprises (Mazumdar and Mazaheri, 2003) does not significantly change the sign or significance o f the coefficients. Inaddition, sample censoring does not seem to bias the results significantly. The functional relationship i s therefore considered robust. Figure 18 (main text) predicts enterprise growth as a function o f size and age, which facilitates interpretation o f the coefficients. The results suggest an important role for small and young firms. The analysis shows that after start-up, an average one-person rural enterprise inTanzania will only grow duringthe first four years and then remain stagnant. The average enterprise size i s about 1.4employees, a number that coincides with descriptive survey data for one-person enterprises, -62- Table 15: Community-levelInvestmentClimate Constraintsand EmploymentGrowth, 2000-2005 Coefficients Explanatory variables Statistically Statistically N Adj. R' significant insignificant Finance Access any non-farmfinancial service -0.003 589 0.61 (-0.53) Access to rural pnvate bank a/ -0.009 537 0.66 (-0.86) Access to urban pnvate bank a1 0.014 537 0.66 (1.39) Access to cooperative bank a/ -0.005 537 0.66 (-0.83) Access to community group bank a/ -0.012 537 0.66 (-1.10) Access to money lender a/ 0.016"" 537 0.66 (2.62) Access to other financial sources a/ -0.004 537 0.66 (-0.55) Access to government bank a1 -0.001 537 0.66 (-0.12) Inpastructure Roadside location 0.029* 627 0.64 (2.49) Distance to next market or city (x10 rn km) -0.005** 560 0.66 (-4.22) Access to cellular phone service 0.008* 560 0.66 (2.08) Access to electncity 0.003 627 0.63 (0.56) Hectncity mteruptions (numberlmonth) -0.001** 257 0.52 (-3.35) Average duration o f mteruptions (hours) -0.002" 257 0.52 (-2.71) Market demand Agnculturalwage rate (~1000m TSh/day) 0.003" 604 0.62 (2.06) Construction wage rate (~1000TSh/day) 0.001 594 0.61 (1.01) Public works wage rate (~1000Tshlday) 0.002 563 0.61 (0.35) Business environment Number o fdays to register (x100) -0.005" 578 0.60 (-2.75) Social violence m community -0.011" 602 0.61 (-2.48) Number ofthefts rn community (x100) -0.013 510 0.61 (-1.86) Robust t statistics inparentheses. a/ Specific finance constraints are regressedjointly. * significant at 5%; ** significant at 1%. Source: 2005 Tanzania RICS -63- By contrast, a bigger enterprise with an initial start-up size o f five employees contracts slightly duringthe first year, but grows relatively fast for the five subsequent years. Thereafter, employment growth declines and the firm eventually contracts. This "stylized" growth process also sheds light on the distribution patterns o f employment growth inFigure 16 (main text). Employment generatedby rural enterprises i s low and occurs in a minority o f small and relatively young enterprises. However, employment generationby these small enterprises will never grow substantially unless other growth obstacles are considered. Investmentclimate constraints are included into the employment growth regression as a second step." The results are displayedinTable 16. Objective measurements (community constraints) are preferred to subjective measurements (perceived business constraints). Inthe case o f the Tanzania RICS, subjective measurements either have an insignificant impact on employment growth, or the wrong sign. Potential constraints are regressedindividually on growth because o f multicollinearity, unclear causalities and the complicated interactionprocess among business constraints (Ayyagari et al., 2006; Bigsten and Soderbom 2005). For example, some constraints may affect firm growth only indirectly through their influence on other obstacles. Inaddition, if multiple investment climate variables were included simultaneously, many observations are being lost.28 Finally, an econometric simulation i s conducted to facilitate the interpretationo f the investment climate coefficients. The simulations should be taken with some caution. They rely on empirical data from two RICS modules that proved challengingto merge, use econometric methods that are subject to measurement error, and do not address causality issues. Finally, it i s also evident from the table that the estimated investment constraints have a large margin o f error. Nevertheless, the simulations are useful incomparing the magnitude o f individual investment climate variables with respect to their impact on growth. The simulation i s done with a macro for the Stata statistics package (King et al., 2000). It uses a Monte Carlo simulation technique that can produce standard errors o f the parameters. The simulations assume a 50 percent reduction or improvement o f those variables that are statistically significant inthe regressions (for instance, mean distance to the next market was assumed to decrease from currently 17.1 to 11.4 kilometers). It i s important to note that the mainpurpose o f the simulations i s to visualize the magnitude and then rank the respective impact o f constraints on enterprise growth. Assuming an improvement of, for instance, 10percent would change the magnitude o f the coefficientsbut does not affect the respective ranking o f the investment climate variable. Improved access to road infrastructure and rural finance impact significantly on employment growth. Figure 25 (main text) shows that improved access to markets would have the strongest impact on employment growth, followed by access to rural finance. Interestingly, rural cell phone communicationranks third. Demand-side factors such as higher rural wages due to productivity increases inagriculture or other factors, ranks fourth. For those rural entrepreneurs who do use electricity, an increase ininterruptions could stimulate growth. Inaddition, legal registrationand 27 The RICS contains numerous investment climate variables that could impact on rural enterprise growth. To facilitate selection, business constraints were first correlated with sales and employment growth, and only those variables that showed a sufficient degree o f correlation were selected for the regressions. A similar approach has been done in the Tanzania Urban ICA (World Bank, 2004b). -64- lower registration costs could boost growth. Finally, a reduction inviolent conflicts could potentially benefit growth. 29 Table 16: SimulationResults of BusinessConstraintsImpact onEmploymentGrowth Mean impact on Community-levelconstraint annual employment Standard errors growth Business environment: 50% reduction o fregistration time 0.041% 0.014 Social cohesion: 50% reduction o fviolent conflict 0.109% 0.044 Registration: 50% increase o fformalregistration 0.138% 0.060 Electricity supply: 50% decrease ininterruptions 0.195% 0.059 Demand: 50% increase o fagricultural wage rate 0.215% 0.105 Communications: 50% increased access to cellphones 0.236% 0.141 Finance: 50% increased access to lending 0.239% 0.091 Roads: 50% reduction in average market distance 0.279% 0.063 Source: 2005 Tanzania RICS Determinants of formal registration The standard approach to study the determinants o f formality (firm i s registered by any government office) i s a probit regression framework (Bigsten et al., 2004). The parameters o fthe coefficients can be estimated using maximum likelihood procedures. The results o f the analysis are presented inTable 17. Firmsize has the strongest impact on registration. Increasing annual sales revenue by only 1,000 Tsh (US$0.77) increases the probability o f beingregistered by 2.6 percent. Other factors that strongly affect registrationare secondary and tertiary education, and the location of the enterprise. Female entrepreneurs are less likely to register. Itmay be that the opportunity costs are higher for women given their householdresponsibilities. Registrationcosts have a negative impact on firmregistration. For example, a 5 percent reduction ino f the share o f registrationcosts insales could boost registrationby 11percent3' 29The ranking of business constraints identificd through the regressions i s considered robust. Usingspatial econometrics to assess the determinants of rural wage labor in Tanzania, also Mduma and Wobst (2005) identify similar constraints. 30The usual caveat of causality issues apply. For example, registration could lead to higher sales but also higher sales (more productiveenterprises) to higher productivity. -65- Table 17: Probability of BeingRegistered,2005 Dependent variable: Explanatory variables Enterprisei s formally registred (1) (2) Age of enterprise (years) 0.009* 0.015** (-2.30) (-2.84) Age squared -0.001** -0.001** (-2.94) (-2.89) Sales (x1000 inTsh) 0.026** 0.018* (-4.65) (-2.28) Managers work experience (years) 0.003 0.002 (-1.75) (-0.95) Manager has secondary education (base = primary) 0.097** 0.114* (-3.01) (-2.50) Manager has tertiary education 0.193* 0.212* (-2.55) (-2.33) Household owns firm 0.100* 0.124* (-2.50) (-2.51) Male manager 0.1 19** 0.130** (-4.29) (-3.68) Rural area (base = rural town) -0.083** ... (-3.49) Location on main road (base = other) ... 0.197** (-3.06) Share of average registration cost insales ... -0.022* (-2.14) Regional and sectoral dummies YES YES Observations 1094 590 Pseudo R-squared 0.13 0.18 Reports marginal changes; robust z statistics inparentheses *significant at 5%; ** significant at 1% Source: 2005 Tanzania RICS Determinantsof enterprise participation Inclusion o f a sample o fhouseholds without enterprises inthe data allows estimating determinants for participation inthe rural non-farm sector.31Given the positive welfare impact o f enterprise ownership, determiningwhether entry barriers exist, and how they may be overcome, i s o f great interest. To do so, households were indexedby i and communities (GNs) byj to estimate a probit equation for operation o f an enterprise that i s o f the form where Zi i s a dummy variable equaling one ifhousehold i operated a non-farm enterprise and zero otherwise, Hi,cj, lcj are vectors o f households' physical and human capital endowment; access to infrastructure and the regulatory environment governing enterprise operation, respectively, Dj i s a set o f provincial dummies, al to a4are coefficient vectors to be estimated, and qi s an iid error term. Variables includedinHiare household size, land endowments, the household head's age 31This section draws from Sundaram-Stukel, Deininger and Jin (2007). -66- and education, a dummy for whether the head's parents operated a non-farm enterprise. C, includes dummy for electrification, distance to city, dummy for existence o fpublic transportation to market, dummy for mudroad, and distance to the nearest bank, ICj includes the number of days requiredto register an enterprise and average tax rates inthe community. Table 18: Deternlinatits for Non-farm Sector Participation Specification (1) (2) (3) Householdcharacteristics Household Size 0.024" ** 0.024*** 0.025*** (3.48) (3.43) (3.52) Head's age (log) 2.058** 2.049** 2.093** (2.17) (2.16) (2.21) Head's age squared -0.294** -0.293** -0.299** (2.31) (2.30) (2.35) Head's years o f education 0.023*** 0.023*** 0.023 *** (5.23) (5.22) (5.27) Years o f education o f head's father 0.008* 0.008* 0.008* (1.88) (1.88) (1.81) Head's parents operatedbusiness 0.114** 0.111** 0.116*** (2.55) (2.46) (2.58) Dummyfor female head -0.161*** -0.162" ** -0.165*** (3.84) (3.85) (3.92) Land endowment -0.045 ** -0.046** -0.045** (2.05) (2.08) (2.05) Investment climate variables Dummyfor Electrification 0.121*** 0.126*** 0.104*** (3.43) (3.55) (2.77) Distance to city -0.001** -0.001** -0.001*** (2.48) (2.57) (2.70) Public transport to market available 0.083* 0.083* 0.089* (1.79) (1.79) (1.89) Dummyfor mudroadonly 0.031 0.030 0.035 Distance to the nearest bank 0.001 0.052 (1.04) (1.43) Days required to complete a registration -0.000 process (0.03) Average tax rate inthe community 0.023 (0.28) No. o f observations 1593 1593 1593 Robust z statistics in brackets.*significant at 10%;**significant at 5%;***significant at 1% -67- Results fromregressions for household's participation innon-farm employment (Table 18 ), highlightthat, inaddition to household characteristics, access to infrastructure and services are key to facilitate participation inthe rural non-farm sector.32Households with higher levels o f education, more family labor, a male head, and parents who had experience inthe non-farm sector, are more likely to do so with estimates suggesting that an additional year o f schooling by the headincreases the probability o f participation by 2.3 percentage points, that this likelihood peaks at an age o f 34 years and reduced by 16percentage points by having a female head. Parental education and involvement inthe non-farm sector both increases the probability o f participation, by 11points, consistent with what was found inChina (Mohapatra et al. 2004). A second set o f findingsrelates to the importance of infrastructure access and investment climate. Livinginan electrified village is estimated to increase the probability o fnon-farmparticipation by 12points, an effect that is equivalent to more than the estimated difference between households with and without parents inthe non-farm sector or an increase inthe head's level of education by almost 5 years. Though only marginally significant, a similarly large impact i s found for availability o f public transport, estimated to increase the probability o f enterprise startup by 8.3 percentagepoints. Itis o finterest to compare this to the coefficient on distance to the next city inwhich, while highlysignificant, is small, implyingthat for everybody locatedupto about 80 kmfrom a town, public transport would more than compensate for the impact o f distance. The coefficient on the distance to the next bank remains insignificant, thus providing little support to the hypothesis that improving financial services would provide the basis for a significant increase inenterprise startups. This i s contrary to what i s expected giventhe overriding importance o f financial constraints in subjective assessments and suggests that use o f subjective constraints inthis way may indeedmix different concepts. Finally, tax and other regulatory policies which have emerged as key constraints inurban surveys emerge as having little relevance for operation o f rural enterprises, presumably because the concerned enterprises are small and informal anyway. Determinants of newinvestments Restricting the sample to only existing enterprises only allows exploring factors affecting enterprise expansion and productivity. As investment i s a different measure o f firm growth than the size o f the labor force, firms firms were indexedby k and estimate a Probit or obit regression o f the form where Zk i s a dummy that equals 1if firmk invested within a given period for Probit regressions or the value o f such investment inTobit regressions, Ek i s a vector o f enterprise characteristics including dummies for size, sector, and age o f the enterprise, the value o f fixed assets and number o f workers, education and experience o f the top manager, the magnitude o f the firm's informal credit line as explainedearlier, qi s a vector o f investment climate constraints (access to infrastructure variables) as discussed above, &. i s an indicator o f enterprise size that equals one for enterprises with more than 2 full-time workers,33Dj denotes regional dummies, a. through aj are scalars or vectors o f coefficients to be estimated and &k i s an iid error term. For any constraint inthe vector q,the corresponding element o f a2or az a4then denote the estimated impact on + 32Note that what is reported in the table are the marginal effects from the Probit regression. 33Splitting the sample (1085 existing enterprises) along this dimension yields 942 small (enterprises with 1 or 2 full time workers) and 143 large enterprises (those with more than two full time workers). -68- investment by small and large firms, respectively so that significance o f a4highlightswhether this constraint affects large firms more or less than small ones and a t-test o f a2+ a4=O allows to determine whether large firms are affected by a given constraint. Results from Probit and Tobit regressions for new investment are reportedinTable 19 with and without the interaction o f investment climate variables with enterprise size. Inboth specifications, there i s convergence o f asset stocks as enterprise assets are predictedto increase investment at a decreasing rate with a peak at 54,598 Tsh. for the Probit and 57,957 for the Tobit. Enterprises with more workers are more likely to invest and to have higher levels o finvestment. The high elasticity (>1) inthe Tobit specification points towards disproportionate increases o f capital intensity, Le. a doubling o f workers would more than double o f investment. At the same time, for existing firms, the owner's experience i s more important for investment than formal education. Enterprise age i s insignificant or even negative. Surprisingly, sector dummies are insignificant, suggesting that, with these factors accounted for, small manufacturing enterprises do not invest more than those in other sectors. The large magnitude and highlevel o f significance o f most o f the objective investment climate variables allows three main conclusions. First,higher levels o fpublic infrastructure provision have considerable potential to leadto complementary investment by the private sector; electrification at the community level i s predicted to increase the propensity o f investment by 10 percent and almost double investment by existing enterprises; having public transport to the nearest market has an even bigger impact with an estimated 20 percent increase inthe propensity o f investment and 60 percent increase inthe amount o f new investment for those who invested. A large impact o f public infrastructure on rural small business' investment i s also implied by the negative and highly significant coefficient on dirt roads which suggest that small non-farm enterprises invillages that are accessible only by dirt road will be 10 percent less likely to invest and, even ifthey invest, have significantly lower amounts o f investment (by 88-99 percent). Furthermore, and consistent with findings from the participation regression, access to finance i s o f greater relevance for expansion o f existing enterprises than the establishment o f new ones; while the estimated impact o f bothinformal borrowing capacity and distance to banks on the probability o f investment i s very small and barely significant, bothhave a major impact inthe Tobit equation. This can to some extent help reconcile the seeming contradictionbetween the frequent mention o f finance as a key constraint by existing firms and its lack o f significance in the startup regression. Inclusion o f an interactionbetween firm size and infrastructure variables in columns 2 and 4 suggests that small enterprises suffer disproportionately from infrastructure- related constraints. Infact, conducting 2tests to assess whether infrastructure-related constraints have a significant impact on new investment or the size o f such investment by large enterprises, results for which are reportedinthe bottom o f table 5, suggest that, while all o f them are highly significant for small enterprises, none o f them i s significant for large ones. This suggests that expansion o f infrastructure investment could lead to a significant increase in startup and expansion o f small enterprises inthe rural non-farm sector. Of course, infrastructure-related constraints could still reduce productivity o f different types o f enterprises. -69- Table 19: Determinantsof New Investment Occurrence o f investment Size o f investment Probit Tobit Enterprise characteristics Total assets in2003 (log) 0.096*** 0.097*** 0.544*** 0.553*** (4.86) (4.87) (3.29) (3.35) Log o f total assets in 2003 squared -0.012*** -0.012** * -0.067* ** -0.068*** (4.49) (4.50) (3.10) (3.12) Number o f workers (log) 0.097*** 0.120** 1.002*** 1.390*** (2.64) (2.19) (3.86) (3.48) Enterprise age -0.029 -0.028 -0.335* -0.312* (1.33) (1.28) (1.86) (1.73) Service sector dummy 0.022 0.024 0.623 0.651 (0.34) (0.38) (1.27) (1.31) Trade sector dummy 0.007 0.002 0.403 0.340 (0.12) (0.03) (0.96) (0.80) Manager's Education 0.025 0.024 0.155 0.140 (0.65) (0.61) (0.46) (0.42) Owner's prior experience (years) 0.006** 0.006** 0.059** 0.063 *** (1.96) (2.06) (2.50) (2.63) Investment climate variables Dummy for electrification 0.104** 0.101** 0.974 ** 1.066** (2.24) (2.08) (2.48) (2.57) Public transportation to market 0.192*** 0.207*** 1.621*** 1.752*** (3.18) (3.18) (3.44) (3.37) Log of Informal borrowing capacity 0.010* 0.011* 0.133** 0.156*** (1.68) (1.93) (2.56) (2.90) Distance to bank -0.021 -0.028* -0.316*** -0.352*** (1.53) (1.85) (2.75) (2.87) Mudroad only -0.095** -0.105** -0.882** -0.986*** (2.20) (2.29) (2.55) (2.60) Electrification*size dummy 0.071 -0.268 (0.54) (0.28) Public transport. *size dummy -0.072 -0.326 (0.50) (0.31) Borrowing capacity*size dummy -0.021 -0.157 (1.54) (1.49) Distance to bank*size dummy 0.055 0.315 (1.45) (1.06) Mudroad*size dummy 0.077 0.837 (0.65), . (0.92) ~, Test for size effects: C!&oll=O 0.172 0.798 (1.90) (0.74) p+p1=o 0.135 1.426 (1.15) (2.27) -pyl =o -0.010 -0.001 (0.49) (0.00) 6+6 1=0 -0.027 -0.037 (0.59) (0.02) v+q 1=o -0.028 -0.149 (0.06) (0.03) Observations Robust z statistics inbrackcts. * significant at 10%; ** significant at 5%; *** significant at 1% 1085 1085 1085 1085 -70- Determinants of total factor productivity The most important issue from a policy perspective i s to obtain the impact o f exogenous constraints on total factor productivity (TFP). The approach taken inmost of the literature (Soderbom and Teal 2004, Lee et al. 2005, Guasch and Escribano 2005, Dollar et al. 2006) i s to regress the residual from a standardvalue-added production function (,u$on a vector o f such characteristics C,. With technology representedby a Cobb-Douglas production function with sector-specific coefficients, this would imply estimating lnYk=yO?l *Tk' 9 2 (InLk)*Tkf93(I&k)*Tkf?4 (Ek)?S(D,)+pk where Yk i s value added, L k i s the number o f workers, Kk the value o f fixed assets, Ek a vector o f enterprise characteristics such as type and age, D, a set o fprovincial dummies, and TL (f =1,2) i s a dummy for trade and service sectors, respectively. Assuming that observable inputs are properly accounted for, the residual,uk can be interpreted as a measure o f total factor productivity such that regressing it on the vector o f investment climate variables C,will provide an estimate o f the impact o f these on TFP. Alternatively, direct inclusion o f C, in (3) will allow estimation ina single equation which will be more efficient.34As discussed above, interact coefficients on C, with an indicator o ffirm size to allow for the impact of exogenous constraints to differ across firms o f different size. Results for determinants o f total factor productivity are reported inTable 20 with labor and capital variables interacted with sector dummies to allow elasticities to differ across sectors.3sIn line with expectations, the marginal return to labor i s higher for trade than for services (with an elasticity o f 0.55-0.69 and 0.39-0.55 depending on the specifications), with opposite patterns for capital (0.14-0.15 and 0.32-0.33 respectively). Although only marginally significant, the estimated coefficientspoint towards lower productivity inservices as compared to trade sector and that most other enterprise characteristics or not do not appear to have much effect on total factor productivity. Consistent with what was the case for investment, enterprises TFP i s significantly affected by the level and quality o f local infrastructure access. Availability o fpublic transport, a variable which, at least to the extent that such transport i s provided by the public sector, will not be independent from the estimated total factor productivity; providing such transport for firms that are currently constrained would be expected to increase TFP by 70 percent. Interestingly, once this i s accounted for, having a linkto a dirt road only does no longer have any significant impact. The second most important constraint, according to the estimates, i s availability o f electricity; providing access to the approximately 50 percent o f enterprises located invillages without electricity connection could increase their productivity by 44-49 percent. Comparedto these, doubling formal borrowing capacity would imply 10percent increase inTFP. Talung these two factors together could have a large impact; eliminating electricity and public transport constraints, which currently affect 15 and 38 percent o f the sample, would be predicted to enhance productivity by around 28 percent. Exploring whether the impact o f investment climate variables differs by enterprise size reveals a patternthat i s more differentiated, suggestingthat public infrastructure investment will be more 34Not surprisingly, results obtained by the two approaches are very similar. 35Due to negative value-added by manyproduction enterprises, the analysis focuses on the trade and service sectors only which compiise about 80 percent of the total enterprise sample. -71- important for small enterprises inalmost all the categories. Differentiating by enterprise size (col. 2) suggests that providing electricity and public transport will be more critical for small enterprises compared to big enterprises. Infact, the coefficient o f access to electricity i s not significant for large firms any more. Although the coefficient for availability o f public transport has similar magnitude o f impact on both the small and large enterprises, it i s much more significant for small enterprises than for small ones (at 1percent significant level for small ones and only 10percent for large ones). It i s also interesting that, while the informal borrowing capacity only affects the TFP o f small enterprises, the distance to commercial banks i s more significant for large than for small ones. -72- Table 20: Determinants of Total Factor Productivity (1) (2) Log o f number o f workers*service sector 0.391* 0.547* (1.69) (1.92) Log o f number o f workers*trade sector 0.552*** 0.692*** (3.60) (3.33) Log o f total assets*service sector 0.322* ** 0.332*** (3.18) (3.25) Log o f total assets*trade sector 0.139** 0.146*** (2.52) (2.61) Dummyfor zero assets*service sector 2.129*** 2.218*** (3.25) (3.37) Dummyfor zero assets*trade sector 0.987** 1.037*** (2.56) (2.64) Dummyfor home-based enterprises -0.087 -0.086 (0.57) (0.57) Dummy for service sector -1.042* -1.064' (1.65) (1.67) Dummyfor age 2-5 years -0.061 -0.055 (0.29) (0.26) Dummy for age 5-10 years 0.198 0.213 (0.94) (1 .OO) Dummyfor age > 10years 0.083 0.092 (0.38) (0.42) Manager's experience (years) 0.013 0.018 (0.56) (0.80) Owner's prior experience (years) -0.006 -0.012 (0.28) (0.54) Electrification dummy 0.444*** 0.493*** (2.59) (2.74) Public transport dummy 0.702*** 0.734*** (3.17) (2.93) Informal borrowing capacity (log) 0.096* 0.102** (1.92) (2.03) Distance to formal bank (km) -0.084* -0.057 (1.70) (1.08) Mudexternal road 0.133 0.063 (0.84) (0.36) Electrification dummy *size -0.309 (0.68) Transport dummy *size 0.000 (0.00) Inf.borrowingcapacity (log) *size 0.003 (0.06) Distance to bank*size -0.170 (1.21) Mudroad*size 0.595 (1.63) Observations 917 917 R2 0.15 0.15 Tests for size effects a+a, 0.184 (0.42) P+P1=0 0.734* (1.88) Y+ YI=o 0.105 (1.64) 6 +6,=0 -0.227 (1.73) 17+171=0 0.685** (1.96) Robustt statistics inparentheses.*significant at 10%;** significantat 5%;*** significantat 1% -73- APPENDIX 3: SURVEY METHODOLOGY Definitionof non-farmenterprises For the purposes o f the survey, a rural non-farm enterprise was defined as any self-employment (or standalone) income generating activity (trade, production or services) located inrural areas not related to primary production o f crops, livestock or fisheries undertaken either within the householdor inany non-housingunits.Any value addition (processing) to primary productioni s considered a ruralnon-farm activity. Households primarily engaged inthe production o f goods and services for home consumptionare excluded. Surveyinstruments The final survey instrument for the Tanzania RICS consisted o fthree modules: (i) household, (ii) enterprise, and (iii)community. The data were collected duringthe months o fJanuary and March o f 2005, by face-to-face interviews o f members o f selected households, owners/managers o f rural non-farm enterprises and community leaders. The household module collected information on householddemographics, sources o f income, and levels o f education. The questionnaire for this module was administered to select non-farm enterprises that were physically located within households (home based) and physically located outside households (stand-alone) as well as selected households that did not engage inrural non- farm enterprises, For households non-engaged innon-farm enterprise activities this module also collected data on factors preventing participation innon-farm enterprises. The enterprise module collected basic information on enterprise sector o f operation, start-up, income and employment generation, formality, seasonality, competition, and constraints to growth. The questionnaire was completed for each rural non-farm enterprise selected for the survey. The manager or most knowledgeable person about the firm was interviewed. The community module was usedto develop community profiles and identify community level characteristics that are important in determining the rural investment climate. This questionnaire was completedby interviewing various community leaders such as village head, local government officials, principal o f a school, etc. A community questionnaire was administered in each of the selected communities. A price component to the module gatheredprice data on key consumer commodities and services prevailing inthe main local market ineach community. Samplingapproach To ensure that the different geographic and climatic zones are well represented inthe sample and to provide highefficiency inthe estimators, Mainland Tanzania was stratified into seven zones (East, Northern Highland, Southern Highland, Central, Lake, West, and Southern zones). The zones were created based on climatic and ago-ecological characteristics, as well as cropping patterns and other geographic characteristics. Each zone has three or four regions (for a total o f 26 regions), each o f which are made up o f several districts, which inturn group are comprised o f towns and villages. The Tanzanian National Bureau o f Statistics defines an EnumerationArea (EA) as a geographical area or community with apopulation size o f 300 to 900 individuals. The survey distinguished between rural and urban EAs.Urban EAs are located within a predominantly rural area and usually contain 300-500 individuals, and usually have their own markets and social service providers (schools, health centers) that serve the surrounding vicinity. Rural EAs lack these amenities. -74- Separate sample frames were used for households, businesses and communities. Based on experience and information gathered from the RICS inother countries and the specifics of the geographical distribution o f households and non-farm enterprises inTanzania that was available from the 2002 Population and Housing Census, the National Bureau o f Statistics set sampling targets o f 1620 households, 1500non-fann enterprises and 150 communities. Since the National Bureau o f Statistics considers the regions within each zone to be highly similar, this stratification i s often used when drawing a representative sample. As a result, one region from each zone was selectedusing stratifiedrandom sampling. Then, out o f the 26 regions of Tanzania, seven were includedinthe final survey, one from each agro-ecological region. After selecting these seven regions, simple random cumulative selection was used to choose the appropriate number o f EAs from the regions. The probability o f selection depends on the size of the population in each district, even when attempts were made to ensure that all districts inthe selected regions were covered. A total o f 150 EAs were selected. Table 21 gives EA population and sample numbers by region. Table 21: Names of Selected Regions and Zones and Number of Enumeration Areas Zones Regions Districts Regional Total Total rural Selected population enumeration enumeration enumeration areas areas areas East zone Morogoro 6 1,753,362 3,086 1,629 24 Northern Kilimanjaro 6 1,376,702 2,309 1,753 20 Highland Central Tabora 6 1,710,465 2,190 1,407 21 Lake Kagera 6 2,028,157 2,387 2,104 20 West Kigoma 4 1,674,047 1,762 1,730 15 Southern Mtwara 5 1,124,481 2,078 1,273 20 Southern Mbeya 8 2,063,328 3,048 2,289 30 Highland Total 7 41 11,730,542 16,860 12,185 150 Source: 2002 Population and Housing Census Once the communities to be included into the sample were selected, listing o f the households and all non-farm establishments ineach selected enumeration area was undertaken. Listing o f the households inthose communities included information about whether any householdmember owns or operates a non-farm business. Samples o f about 10households -both with and with out non-fann businesses -were then drawn from the list prepared in each selected enumeration area. Non-farm enterprises ineach selected enumeration areas were listed by major economic activity. At least 11non-farm enterprises were then randomly selectedfrom each enumeration area depending on the availability o f such enterprises. Because the low probability o f selection, manufacturing enterprises were over-sampled to ensure sufficient observations. Table 22 and Table 23 summarize the distribution o f planned and actual sample size for the household, enterprise and community surveys by region and zone. As evidenced inthe above tables, the TRICS data collection process achieved highresponserates for all the three modules. The non-response rates, though relatively low given the informality o f these non-farm activities, were mainly caused by enterprise owner absenteeism duringvisit times. Itis also true that some o fnon-fannenterprises couldnot be locatedandthat a few o f the enterprises enumerated did not qualify as non-farm. Survey weights were found to overestimate -75- significantly and were consequently unusable for this report. Efforts are underway to adjust the weights for use infuture analysis. Table 22: Original Sample Sizes for Enterprises,Householdand Community Survey Zones Regions Enterprises Households Communities East zone Morogoro 240 240 24 Northern Highland Kilimanjaro 200 201 20 Central Tabora 210 291 21 Lake Kagera 210 240 20 West Kigoma 150 150 15 Southern Mtwara 200 200 20 Southern Highland Mbeya 290 298 30 Total 7 1,500 1,620 150 Table 23: Number of Respondentsfor Enterprises,Householdand Community Survey Zones Regions Enterprises Households Communities East zone Morogoro 238 236 24 Northern Highland Kilimanjaro 114 201 20 Central Tabora 142 291 21 Lake Kagera 123 239 20 West Kigoma 138 149 15 Southern Mtwara 199 200 20 Southern Highland Mbeya 285 294 30 Total 7 1,239 1,610 150 Comparisonofthe TanzaniaRICS HouseholdModulewithHBS Means for standard indicators are compared between the 2005 RICS and 2001 Household Budget Survey (HBS). Comparisons are made in attempt to crudely evaluate sample populationvalidity. As land ownership, household size and age are fairly static variables over a three-year time horizon, these are the comparators chosen. Average householdland ownership across households shows a slight reduction from the HBS estimated 5.8 acres per householdto the RICS estimate o f 5.3. Average household size measures at 4.87 inthe HBS and 4.97 for the RICS. Finally, average household age i s 22.5 years as reportedfor the HBS and found at 23.4 years o f age inthe RICS. Because o f spatial and temporal differences between surveys, differences are expected. Concern would be validated if these indicatorsproved significantly misalignedwith trend expectations. No evidence of sampling or survey error i s found with the chosen indicators. -76- APPENDIX4: RURALFINANCE Rural finance i s the main supply-side constraint o f Tanzania's non-farm enterprise sector. Despite financial sector reforms set inmotion a decade ago, access to rural financial services by large segments o f rural enterprises remains stunted. Most Microfinance institutions are locatedinDar es Salaam, and only few have a countrywide network that services rural areas. The principal providers o f rural microfinance are Savings and Credit Cooperatives (SACCOs) and foreign- assisted NGOs. Inrural areas, there i s a large unmet demand for credit. Rural enterprises typically obtain small amount o f loans and pay highinterest rates when they do access credit. Enterprises use their own funds to meet their start-up need, which shows evidence o f a savings culture. Rural entrepreneurs are concerned about access and costs o f credit, and often a lack o f collateral. Microfinance institutions do not meet the demand for rural credit because o f hightransactions costs and risks, infiastructure and lack o f labor to manage rural loan portfolio^.^^ Historical and InstitutionalBackground A long history of reform of the ruralfinancial system The attempt to foster rural finance inTanzania i s not a recent phenomenon. The importance o f finance was recognizedas early as the late 1960s. Specialized investment and development banking institutions developed to channel finance into neglectedsectors o f the economy, including the rural sector. The Tanzania Rural Development Bank (TRDB) was established to specialize inthe financing o f the rural sector inFebruary 1971.The Tanzania Housing Bank (THB) started in 1973 and specialized inthe financing o frural andurbanresidential, offices and commercial buildings. Through the Central Bank o f Tanzania (BOT) established the Rural Finance Fundto finance rural development. However, these institutions failed to deliver (Economic and Social Research Foundation, 2004). Most o f the banks and non-banking financial institutions geared towards financing the rural sector were restructured duringthe financial sector reforms o f the 1990s as part o f broader market oriented reforms. The financial sector reforms started in 1991 aimed to create an effective and efficient financial system. The restructuringincluded liberalization o f interest rates, elimination o f administrative credit allocations, privatizationo f state owned banks, strengthening the BOT'S regulatory and supervisory role, and allowing the entry o f privately owned financial institutions. In 1996,public awareness initiatives about microfinance startedandhelpeddevelop financial institutions with wider outreach. Recognizingthe importance o f microfinance inthe national economy, the National Microfinance Policy (NMP) was launched inFebruary 2001. The policy was intended to integrate microfinance into the broader financial sector. Reforms did not reach local communities In2005, the Government approved the Microfinance Companies andMicrocreditActivities Regulations and Financial Cooperative Societies (FICOS) and regulations to ensure a level playing field for bothregulated and unregulated microfinance service providers. The regulations stipulate that all Microfinance Institutions (MFIs) follow a best practice regulatory framework. They are also intended to help MFIs and Savings and Credit Cooperative Societies (SACCOS) 36A limitation o f the empirical analysis is small sample size. The data are based on 111enterprises that applied for a formal loan inthe past five years. To overcome this constraint, the chapter selectively includes information on rural lending from the household module of the Tanzania RICS. -77- transition into licensed financial institutions and attract private capital to support their operations (Rubambey, 2005). While the reforms improved efficiency and competition, they didnot improve access to financial services by the low-income segment o f the population, especially inrural areas. Credit as a share o f GDP has declined dramatically from about 35 percent to GDP in 1993 to only nine percent in 2004. Credit to the private sector contracted from about 15 percent o f GDP to only three percent in 1996.However, since then ithas steadily recoveredand stands now at ninepercent ofGDP (World Bank, 2006~). Commercialbanks continuedto focus on corporate clients and high-income households inurban areas, thus widening the gap between urban and rural populations and their access to financial services. Itremains to be seen whether Tanzanian financial reforms will help to overcome the inherent imperfections inrural credit markets, as well as meetingthe financial service needs o f rural enterprises. Imperfections inrural credit markets result from shortage o f realizable collateral, lack o f ancillary institutions, highcovariant risk among borrowers, and severe problems o f enforcing repayments o f loan contracts (Economic and Social Research Foundation, 2004). LimitedAccess to RuralFinance Largeunmet demandfor ruralcredit There i s a large unmet demand for formal rural credit inTanzania. According to the TRICS, 61 percent o frural enterprises believe that access to credit i s the major constraint to enterprise startup and growth. This i s very similar to the estimates made among enterprises in Sub-Saharan Africa (Liedholm, 2002; Bigsten and Soderbom, 2005). Tanzanian enterprises that list access to finance as one o f the top constraints claim they could, on an average, increase their sales revenue by 43 percent ifthis constraint was removed. Only 19 percent o f all the enterprises indicated that they wanted to apply for a formal loan for workmg capital or investment ina non-farm enterprise inthe preceding five years. Only 45 percent o f these end up applying for a formal loan. Of the enterprises that actually applied for a loan, few applicants are successful, and only 6 percent o f the all enterprises have access to formal credit. This suggests a large unmet demand for credit inrural Tanzania (Table 24). Table 24: Access to Formal Loans by Enterprises and Households, 2005 (in percent) Category Total enterprises Total households Enterpriseslhouseholds that 19.4 9.6 Appliedfor a loan 8.7 7.2 Got the loan approved 5.8 5.6 Inaddition, the household survey supports the findingthat accessto creditrather than costs is the major issue. About 12 percent o f households identify lack o f access to formal credit as the major obstacle for their non-farm businesses, which was also the top issue that prevents households from starting a non-farm business. Among households who apply only 6 percent are successful. This highlightsthe extremely limited access to formal credit ofruralhouseholds and the likely demand (Table 25). -78- Table 25: Access to Credit by Formaland Informal Enterprises, 2005 (in percent) Registered? Total Category Yes No Didnot apply 84 93 91 Appliedbut rejected 5 3 3 Appliedandapproved 12 5 6 Total 100 100 100 Source: 2005 Tanzania RICS A significantproportion o fthe initial capital for rural enterprises inTanzania comes from personal savings. Expansion o f enterprises i s mainly financed from internally generated funds. This situation has frequently ledto the argument that rural enterprises do not exhibit a high demand for external sources of finance. The initial capital required inestablishing a small enterprise may appear meager, but these amounts may account for a substantial proportion of the gross annual family income. This implies that personal savings alone i s unlikely to meet the demand for finance by the enterprises. The situation i s exacerbated inremote rural areas. -79- Box 9: Snapshot of Microfinance Institutions in Tanzania Institutions that provide financial services to low-income businesses and rural households inTanzania include licensed providers, savings and credit cooperative societies, and NGOs. Most bank branches are located inDar es Salaam, and only few have a countrywide network that services rural areas. The principalproviders o f microfinance are therefore Savings and Credit Cooperatives (SACCOs) and foreign-assisted NGOs. CommercialBanks Three commercial bankshave products and services targeted to low-income businesses: the National Microfinance Bank, the Cooperative and Rural Development Bank, and Akiba Commercial Bank. National Microfinance Bank (IVMB) was created in 1997 as part o f the restructuring o f the NationalBank of Commerce (NBC). The Government divested part of the NMB and Rabobank o f the Netherlands acquired a 49 percent stake in 2005. It is the leading bank inTanzania, with a countrywide network o f 104 branches and agencies and apresence in almost every district and regional center. Rabobank currently provides management and technical assistance. The bank started offering rural credit products mainly inthe form o f micro-loans. It i s expected that Rabobank will provide its expertise inrural lending and play a long-term role inrural development inTanzania. The Cooperative and Rural Development Bank (CRDB) was restructured and re-capitalized out o f the former government owned Cooperative and Rural Development Bank. It a private bank andhas a network o f 30 branches, o f which eight are in Dar es Salaam. Although no longer a cooperative institution, the cooperative i s still a significant stakeholder. Through its newly formed subsidiary Microfinance Company (MFC) Limited, the bank offers loans to intermediary microfinance institutions formed by individuals such as Savings and Credit Cooperative Societies (SACCOS), Savings and Credit Associations (SACAS), financial NGOs, and community banks. The beneficiary MFIs inturnprovide financial services to their customers.The strategic missionofthe subsidiary is to identify and develop banking relationships with a wide range o f MFIsand provide intermediary microfinance services. With this indirect approach, CRDB expects to reduce transaction costs and reduce the credit risk o f offering loans directly to individuals. Akiba Commercial Bank began operations in 1997 as an initiative o f more than 300 Tanzanian entrepreneurs who were inspiredto move into microfinance. Akiba's operations are predominantly focused inthe capital city with five branches inDar es Salaam, 1branch inArusha andmarketing offices inMoshi, Tanga, Mbeya, Zanzibar, and Pemba. Akiba currently offers microfinance loan products under both the traditional group and individual loan methodologies. Other products offered by Akiba are consumer loans and corporate loans and overdrafts. Financial Cooperative Societies The cooperative sector has a four-tier structure: i)Primary cooperatives at the community level, for instance the Savings and Credit Cooperative Societies (SACCOS); ii)Cooperative unions at the district or regional levels, for instance the Kilimanjaro Native Cooperative Union; iii)Apex organizations based on activity specialization, e.g. the Savings and Credit Cooperative League o f Tanzania (SCCULT); iv) The Tanzania Federation o f Cooperatives (TFC), which i s the national-level umbrella organization for all kinds and tiers of cooperative societies. A recent survey revealed that individuals operating at all levels tend not to have managerial and financial expertise, a problemthat needs to address if these institutions are to be effectively runand services expanded. The Cooperative Societies Act o f 1991 providedthe basis for the development o f SACCOS as privately owned and organized equity-based institutions. Their outreach, resources from members in terms o f share capital and savings, and the volume o f loans to members far exceed those o fmicrofinance NGOs (Randhawa and Gallardo, 2003). There are about 1,870 SACCOS inTanzania whose members constitute 0.7 percent o f the Tanzania population. A number o f reforms were undertaken inother areas, including the insurance sector where deregulation has been introduced, and inthe provident government pension fund where several restructurings have occurred. Financial NGOs (MFIs)play an important role inprovidingsmall loans to ruralbusinesses. The leadingMFIsinTanzania are PRIDE, MEDA, and FINCA. PRIDE uses a "modified Grameen" methodology and provides loans to groups of five. It has operations in 17 o f the 21 regions o f mainland Tanzania. FINCA provides group-based loans to poor businesses and households inseven regions. MEDA operates micro-credit programs inMbeya and Dar es Salaam. In addition, MEDA manages an umbrella credit programto assist other micro-credit organizations like SACCOS. The mission o f these MFIsis to provide financial services to the poor so that they can create newjobs, raise household income, and improve their standard o f living. Finally, credit by traders and marketing organization also plays an important role in agriculture, with interest rates significantly higher than commercial lending rates. -80- Table 26: Institutional Providers of MicrofinanceServices Typeiname o f institution Microfinance Marketiarea o f Main source o f products offered operation Legal status funds FinancialNGOs microfinance loans, such as Solidarityigroup-based Mandatory savings Urban and peri-urban Presidential Trust Fund, (except a few) & areas: selected rural Societies Act: Trust Donor funds Povertv Africa. YOSEFO. groupbased loans areas Individual Microfinance loans, such as: SIDO, Mandatory savings Tanzania Gatsby Trust, (except those Urban and pen-urban Mennonite Economic marked) & areas Societies Act: Trust Donor funds Development Association, individual loans Poverty Africa Village savings and credit Individual savings Societies Act, Ministry associations (SACAs) & group-based Rural villages Donor grants loans o f Home Affairs Savings and Credit Cooperative Societies Urban SACCOS Member loans only Urban areas Cooperative Societies Share capital, Act loans, grants Rural SACCOS Member savings Cooperative Societies Share capital, deposits and loans Rural areas Act loans, grants Other (savings-based) Voluntary savings Rural and Urban Cooperative Societies Share capital, SACCOS and withdrawals only areas Act loans, grants Regulated and licensed providers of microfinance services Commercial banks NationalMicrofinance Bank Savings deposits & micro-loans Nationwide Act o f parliament Depositsicapital Savings deposits: Akiba commercialBank Group and individual micro- Nationwide Companies' Act: Depositskapital enterprise loans CRDB bank New1y-organized microfinance dept. Nationwide Companies' Act: DANIDA Tanzania Postal Bank Savingsifixed (licensed as NBFI) deposits Nationwide Act o f Parliament Depositsicapital Regional banks Deoositsicauital Kilimanjaro Cooperative Savings deposits Kilimanjaro Region Companies' Act: BOT from regional SACCO union Bank and micro-loans and SACCOS Communitybanks Mufindi CommunityBank Savings deposits Mufindi District, Companies' Act: BOT Depositskapital and micro-loans Iringa Region Mwanga Rural Community Savings deposits Pare District, Companies' Act: B O T Depositsicapital Bank and micro loans Kilimanjaro Source: Adaptedfroin Randhawa and Gallardo (2003) -81- Access to credit for ruralmicroenterprises Credit constraints differ among enterprises inrural areas compared to rural towns. Butinrural areas, no particular type o fbusiness i s more successful than any other at obtaining credit, and there are no major variations when revenue level o frural enterprises was examined. Enterprises inTabora, however, are less constrainedbyrural finance. Informal enterprisesmay notbe ableto expand for a variety o freasons. One o f the most important reasons i s their inability to obtain formal credit because commercial banks prefer to provide loans to formal enterprises (Ramaswamy, 2006). Table 27: Distribution of Enterpriseswith Financial Statementsby Sales, 2005 Sales category in2004 Enterprises that assemble Enterprises that prepare audited (`000 Tsh.) a financial statement (%) financial statements (%) 0-500 16.1 0.1 >500- 1,000 17.0 0.8 >100-2,000 21.2 0.9 >2,000 23.4 3.5 All enterprises 17.6 0.8 Source: 2005 TanzaniaHCS Among smaller enterprises, loan applications are less common and loan approval rates lower than for larger enterprises. Lenders tend to be biased towards bigger enterprise customers making access to credit that much more difficult for smaller enterprises (Table 27 and Table 28). In almost all regions, over 90 percent o f enterprises have no access to formal credit. Urban enterprises inTanzania are better o f f compared to rural enterprises, with 19.9 percent obtaining loans from financial institutions (World Bank, 2004b). Table 28: Access to Formal Loans by Enterprises in Different by Sales, 2005 Category Percent of enterprises insales categories based o n 2004 sales (`000Tsh.) Category >500- 0-500 1,000 'l,ooo- 2,000 >2,000 Total (%) Didnot apply 93.8 91.2 86.7 83.6 91.4 Applied and didnot receive 2.4 2.9 2.7 5.0 2.8 Applied and received 3.8 5.9 10.6 11.4 5.8 Source: 2005 TanzaniaRICS Reasonsfor not applyingfor credit More than 80 percent o f the enterprises didnot apply for a loan for variety o f reasons. About 21 percent reported, "non-availability o f a nearby bank" followed by "the interest rate would be too high" (20 percent) as the major reasons for not applying. Disaggregatingthis information by area, rural area enterprises cited "non-availability o f a nearby bank" (27 percent) and "duration would be too short" (16 percent) as the major reasons for not applying, while enterprises inrural towns said "interest rate would be too high" (28 percent) and "insufficient collateral" (20 percent) were the major constraints to access o f formal credit. These findings show that there i s a large unmet demand inrural areas for credit. -82- RuralLending Practices and ObstaclesFor Lenders Financial providers have historically had a limitedrole inrural finance. Only 2 percent o f enterprises purchase inputsor goods for resale on credit from suppliers (Economic and Social Research Foundation, 2004). An estimated one percent o f enterprises that receivedformal loans and were therefore seen as "credit-worthy'' reportedthat they purchased inputs on credit. Only four percent o f the enterprises inrural Tanzania have an overdraft or line o f credit with banks. The median value o f such overdraft facilities i s Tsh. 500,000. Figure27: Distributionof ApprovedLoans by Annual Interest Rate, 2004 . . .., 28% 25% 20% & E c 25 15% g 10% 5% I 0% 0-25 ,2530 >50-100 >loo-500 >500 Interestrate in percentp.a. Source: 2005 Tanzania iUCS Similarly, previousresearchhighlightthat fiom 1968 to 1976, about 95 percent o f small businesses inTanzania used capital from personal savings (Satta, 2002). The amount o f rural loans approved by the two major banks inTanzania, National Bank o f Commerce (NBC) and Cooperative and Rural Development Bank (CKDB) was less than 4 percent of their total credit volume between 1986 and 1991. The rural lending situation inthe country has not much changed inthe last three decades. The following sections describe the existinglending system. Loans are small, interest rates are high Rural microenterprises obtain loans at very highinterest rates mainly from informal sources. For example, 8 percent o f the enterprises inthe survey region obtained loans from moneylenders at least once during the preceding five years. The median annual interest rate paid by such enterprises to lenders i s 125 percent, much higher than the interest rates charged by formal financial institutions inTanzania. PRIDE, the largest microfinance NGO inTanzania, provides loans with annual interest rates between 24 and 30 percent per annum. Rural enterprises prefer moneylenders because o f the flexibility they offer, shorter processing time, and no insistence on savings. PRIDElends only to those entrepreneurs who are willing to save 25 percent o fthe loan value before the loan i s granted. -83- Figure 28: Distributionof Approved Loans by Value, 2004 25% g 20% - 0 ..pE 0 15% :10% 5% 0% 0-50 s50-100 >100-200 >200-500 ~500-1,000 ~1,000- >2.000 2.000 Value of loans (in'000 ish) Source: 2005 Tanzania RICS Rural businesses pay much higher interest rates for formal loans than the average rates charged by commercial banks to urbanenterprises. The medianannual interestrate paidbyrural enterprises i s 69 percent, much higher than the average short-tern (up to 1year) interest rate of 16percent charged by commercial banks in2004. Similarly, households inrural Tanzania pay very highinterest rates (median o f 80 percent). Formal loans granted to rural enterprises are small, with about 60 percent below Tsh. 200,000 (US$ 184). However, a significant number of loans are much greater, with an average loan size o f Tsh. 487,066 (US$447). The median loan value for households i s Tsh. 200,000 (US$ 154). Table 29: Interest Rates Charged by Different Lenders, 2004 '' Category Average (%) Median (%) Moneylenders 125 Bydifferent institutions for enterprise loans 69 By different financial institutions for household loans 80 MFIs 36 Commercial banks 16 Source: 2005 Tanzania RICS and Ramaswamy (2006). a/ Interest rates charged for short term up loans (up to one year) Enterprises use their own hnds to meet their start-up capitalneeds, which show evidence o f a savings culture. Money from the entrepreneur's agriculturalproduction contributes 55 percent (Figure 14). Bank loans contribute only one percent and moneylenders one percent o f start-up capitalneeds. Commercial banks are not attracted to rural lending, a situation that lowers the amount o f capital available directly to enterprises or to microfinance institutions that lendto enterprises. Restructuring and privatization o f state banks that came with Tanzanian financial sector reforms ledto a closure o f 78 branches inthe country, mostly inrural areas (Satta, 1999). In2001, only 4 percent ofrural households participated insavings or banking activities. This went down from 13 percent in 1991 (Tanzania National Bureau of Statistics, 2002). -84- Figure29: CollateralRequirements,2004 Other, -Building, 17.7 Valuaible Durable 16.5 27.8 Machinery, 1.3 Source: 2005 Tanzania RICS Most loans require collateral Most formal loans require collateral, a requirement that i s problematic inrural areas. Inthe survey area, 81 percent o f formal loans required collateral. This i s because rural entrepreneurs are viewed as high-riskborrowers thus increasing the importanceo f collateral security. Yet MFIs approve a significant proportion (33 percent) o f the loans without collateral. PRIDE also provides loans up to Tsh. 1,000,000 on group guarantee without collateral, but the 25 percent savings deposit made by borrowers before obtaining a loan serves as partial collateral. The value o f collateral obtained by formal financial institutions for rural lending is usually very highcomparedto the loanvalue. The medianvalue is 2.5 times the loanvalue while the average i s 6 times the loan value. Different forms o f collateral are used by businesses. The survey finds that households mostly use houses as collateral to obtain loans. With about 13 percent landplays a relatively minor role. This could be explained by the fact that only 9 percent o f the total land in the survey region i s titled. Interestingly, this contrasts with frequent claims o f the role o f land as important collateral (Economic and Social Research Foundation, 2004). Landtitling may therefore not solve the entire problem because difficulty infinding markets for rural land may discourage banks from providing loans with land as collateral. High repayment rates and short-term financing model Formal loans have an average duration o f 11months. This is longer than the average duration o f 5 months for loans providedby moneylenders. PRIDEand FINCAprovide loans that are usually paid back in six weeks. This short repayment period suggests that clients are more likely to be small borrowers. However, it could also be attributed to the tendency of moneylenders to restrict loans to customers that can repay quickly and whose financial viability has been established over the years. Short-term loans limit long-term investments but can also be useful to businesses that -85- have seasonal cash flows. PRIDEand FINCA claim that their business model i s successful, with more than 90 percent repayment rate and over 90 percent ofpayments being made on time (Ramaswamy, 2006). This could demonstrate the l o w risklevel associated with the joint liability, rigidrepayment schedules, and short-term finance models. Figure 30: Purpose of Loans, 2004 Working Capital 29% Startup a Non-farm Enterprise 52% New Investment Source: 2005 Tanzania RICS Most formal loans were approved for start-up o f non-farm enterprises, both interms o f numbers and value.37Fifty-twopercent o fthe loans were for start-up o f a new enterprise, 29 percent for new investment in existing enterprises, and 19 percent for workmg capital (Ramaswamy, 2006). Interms ofvalue, 43 percent ofthe loans were for start-ups, 33 percent for new investment in existing enterprises, and 24 percent for workmg capital. The pattern i s similar inbothrural areas and towns with most loans approved for start-up o f enterprises (Ramaswamy, 2006). Most formal loans are approved by local commercial banks and MFIs (Figure 3 1).However, in terms o f total value, local commercial banks approved 46 percent o fthe loans followed by MFIs (11percent). The average size o f a loan approved by a local commercial bank is US$710 and that approved by an MFIi s US$ 177. MFIsplay a vital role inmeeting the smaller credit demands o f enterprises inrural Tanzania. The median value o f formal loans inrural Tanzania i s Tsh. 200,000 (US$ 154). MFIs approved 39 percent o f the loans below this amount. Local commercial banks approved only 4 percent. About 59 percent o f the enterprises were not provided with an explanation, 19percent were denied for insufficient collateral, and 11percent for not having a co- signer. Financialinstitutions usually do not provide an explanationwhen loans are not approved. Financial institutions disburse loans that are substantially smaller than what the enterprises request. On average, financial institutions approve two thirds o f the applied loan amount. 37This suggests that improved access to finance in Tanzania would stimulate mainly the entry of new enterprises. -86- Figure 31: Sources of Loans, 2005 Others investment Funds NGO or Trade Organization 29% Internal Funds or Retained Earnings Foreign Owned Commercial Bank i Other Domestic 15% Commercial Banks Local Commerc al Banks 46'b 0% 25% 50% Source: 2005 Tanzania RICS Perceived constraints by MFIs A combination o f factors leads to increasedtransaction costs andrisks for commercial banks wanting to serve rural clients. Eventhough commercial banks have nation-wide operations, their rural loan portfolio i s only about 5 percent o f their total business. For commercial banks operating inrural areas, most ofthe collectedsavings are transferredto Dar es Salaam to be invested in assets that have a more attractive returnriskprofile than rural investment opportunities (World Bank, 2006~).Some o fthe constraints as perceived by the financial institutions are discussed below (Economic and Social Research Foundation, 2004). Overall structural weaknesses of theJinancia1system. These include the environment for contract enforcement, and the efficiency o f the legal, judicial, and information framework. Also a low concentration inthe banking system (the average loan to deposit ratio i s low at only 34 percent) and high spreads between bank lending and deposit rates remain a problem. A large part o f the spread can be explainedby high-riskpremium chargedby the banks for rural credit risk, weak market infrastructure, and difficulties inenforcement o f creditor rights. High transaction costs and risks. Financial institutions avoid rural loans because of transaction costs and repayment risk.The top three external constraints according to financial institutions when attempting to expand their rural lending are (i) unreliability o f land and property title deeds, (ii) aspectsoflandactprovision, and(iii) andunreliablelegalsystemforloan key long enforcement. Hightransaction costs are often attributedto the low loan sizes andvalues andfragmented nature o frural financial markets. Likewise, the low household savings, geographical dispersion o f potential clients, seasonality o f agricultural production and its susceptibility to natural disasters increase transaction costs. Insurance markets and hedging instruments are virtually non-existent, -87- resulting inexposure o f lenders to highdefault risks. A weak legal system and the lack o f a developed credit information system make it difficult for financial institutions to satisfy rural credit demands and operate on a commercially viable basis. Inaddition, only 18percent ofthe enterprises inrural Tanzania create any financial statement and less than one percent prepares audited financial statements o f their operations. Unreliable financial records make credit risk assessment difficult. Even inthe case o f a successful record o f accomplishment, the lack o f any financial statements puts enterprises at a disadvantage when presenting a business propositionto financial institutions inorder to obtain credit. Financial institutions often need to have a basic understanding o f these enterprises. However, it appears though those financial institutionsthat serve rural Tanzanian businesses seem to acknowledge these conditions. About 70 percent of formal loans are to businesses that do not have a financial statement. Inadequate infrastructure. Poor physical and communication infkastructure (includingrural roads, electricity, and telecommunication) appears to be another major reason for the inaccessibility o f rural areas and the lack o f information on credit worthiness o fpotential borrowers. Inadequate infrastructure i s also reportedby enterprises as the second major constraint (next to finance) affecting their growth. This i s likely to depress boththe demand for financial services and the development o f efficient rural financial markets. Lack of expertise in rural lending. Most rural lending institutions do not have the skilled labor to manage rural loanportfolios efficiently. The major financial institutions cite their (i) strategic focus on corporate and urban customers, (ii) lack o f expertise inmicrofinance, and (iii) o f lack trained labor to ensure needed credit assessment as their top three institutional constraints. -88- BIBLIOGRAPHY Angermann, Ingrid. 2001.Gewerbliches Unternehmertum inperipheren Regionen Tansanias. Hamburger Beitrage zur Afrika-Kunde 63. Hamburg: Institut fur Afrika-Kunde. Ayyagari, Meghanna, Asli Demirguc-Kunt, and Vojislav Maksimovic. 2006. "How Important are Financing Constraints?" WorldBank Policy Research WorkingPaper 3820. Barret, C.B., T. Reardon, and P. Webb. 2001. "Nonfarm income diversification and household livelihood strategies inrural Africa: concepts, dynamics, and policy implications." Food Policy 26: 315-331. Bigsten, Arne and MAns Soderbom. 2005. "What Have We LearnedFromaDecade of Manufacturing Enterprise Surveys inAfrica?" WorldBank Policy Research WorkingPaper 3798. Bigsten, Arne, Peter Kimuyu,and KarlLundvall. 2004. "What to do with the Informal Sector?" Development Policy Review 22(6): 701-715. Daniels, Lisa. 2003. "Factors That Influence the Expansionof the Microenterprise Sector: Results from Three National Surveys inZimbabwe." Journal of International Development 15: 675-692. Demombynes, G. and J. G.Hoogeveen. 2004. "Growth, Inequality and Simulated Poverty Path for Tanzania, 1992-2002." WorldBank Policy Research WorkingPaper 3432. Dollar, D.,M.Hallward-Driemeierand T. Mengistae. 2005. "Investment Climate and Firm Performance inDeveloping Economies." Economic Development and Cultural Change 54(1): 1- 30. Dollar, D., M.Hallward-Driemeier and T. Mengistae. 2006. "Investment Climate and International Integration." WorldDevelopment 34 (9): 1498-516. Duncombe, Richard and RichardHeeks. 2002. "Enterprise Across The DigitalDivide: Information Systems and Rural Microenterprises inBotswana." Journal ofInternationa1 Development 14: 61-74. Ellis, Frank andNtenguaMdoe. 2003. "Livelihoods and Rural Poverty ReductioninTanzania." WorldDevelopment 31(8): 1367-1384. Evans,David S. 1987. "Test ofAlternative Theories ofFirmGrowth." Journal of Political Economy 95(4): 657-674. Guasch, J. L.and A. Escribano. 2005. "Assessing the impact of the investment climate on productivity usingfirm-level data: methodology and the cases of Guatemala, Honduras, and Nicaragua." Policy Research WorkingPaper 3621.Washington: The World Bank. Jovanovic, Boyan. 1982. "Selection and Evolution of Industry."Econometrica 50: 649-670. Kahyarara, Godius and Francis Teal. 2006. "General or Vocational Education? Evidences from the Returnsto EducationinTanzanian Firms."Paperpresented at the CSAE Conference. Reducing poverty and inequality -How can Africa be included?University of Oxford, March 19- 21, 2006. Processed. -89- King, Gary, MichaelT o m , and Jason Wittenberg. 2000. "Making the Most o f Statistical Analyses: Improving Interpretationand Presentation." American Journal of Political Science 44(2): 347-61. Lanjouw, Peter, Jaime Quizon, and Robert Sparrow. 2001. "Non-agricultural wage earnings in peri-urban areas o f Tanzania: evidence from household survey data." Food Policy 26: 385-403. Lee, K.,W. P. Anderson, and U.Subramanian. 2005. "Measuring the impact o fthe investment climate on total factor productivity: the cases o f China and Brazil." Policy Research Working Paper 3792. Washington DC: The World Bank. Liedholm, Carl and Donald C. Mead. 1999. Small enterprises and economic development: the dynamics o f micro and small enterprises. London: Routledge. Liedholm, Carl, MichaelA. McPherson, and Enyinna Chuta. 1994. "Small Enterprise Growth in Rural Africa." American Journal of Agricultural Economics 76(5): 1177-1182. Liedholm, Carl. 2002. "Small FirmDynamics: Evidence from Africa and Latin America." Small Business Economics 18: 227-242. Mazumdar, Dipak and Ata Mazaheri. 2003. The African manufacturing firm: an analysis based on firm surveys in seven countries. London: Routledge. McPherson, Michael A. 1996. "Growth o f micro and small enterprises insouthern Africa." Journal of Development Economics 48: 253-277, Mduma, John K.and Peter Wobst. 2005. "Village Level Labor Market Development inTanzania: Evidence from Spatial Econometrics." ZEF Discussion Paper 96. University o f Bonn. Minot, Nicholas, Ken Simler, Todd Benson et al. 2006. Poverty and malnutrition in Tanzania: New approachesfor examining trends and spatialpatterns. Washington DC: InternationalFood Policy Research Institute. Mohapatra, S., R. Goodhue, and S. Rozelle. 2004. "The Emergence o f Self Employment inRural China." WorkingPaper. Davis, California: University o f California, Davis. Randhawa, Bikkiand Joselito Gallardo. 2003. "Micro finance Regulation inTanzania: Implications for Development and Performance o f the Industry." WorldBank Africa Region WorkingPaper 51. Rosegrant, M.W.and P. Hazell. 2000. Transforming the rural Asian economy: the unfinished revolution. Hong Kong: Oxford University Press. Rumbabey, G. C. 2005. "Policy, Regulatory and Supervisory Environment for Microfinance in Tanzania." Draft paper, Bank o f Tanzania. Processed. Satta, Tadeo A. 2002. "A Multidimensional Strategy Approach to Improving Small Business Access to Finance inTanzania." Draft paper, University of Manchester. Processed. Satta, Tadeo A. 2004. "Microfinance regulation influence on small firms' financing inTanzania." Journal of Financial Regulation and Compliance 12: 64-74. -90- Sleuwaegen, Leo and Micheline Goedhuys. 2002. "Growth o f firms indeveloping countries, evidence fiom CBte d'Ivoire." Journal of Development Economics 68: 117-135. Soderbom, M.and F.Teal. 2004. "Size and Efficiency inAfrican Manufacturing Firms:Evidence from Firm-Level Panel Data." Journal of Development Economics 73 (1): 369-94. Tanzania National Bureau o f Statistics. 2002. Household Budget Survey 2000/01: Final Report. Dar es Salaam. Tiffin, kchard andXavier Irz. 2006. "Is agriculture the engine o f growth?" Agricultural Economics 35: 79-89. TOVO, Maurizia. 1991. "Microenterprises among village woman inTanzania." Small Enterprise Development 2(1): 20-3 1. World Bank. 2000. Agriculture in Tanzania Since 1986:Follower or Leader of Growth? Washington D.C.: World Bank, IFPRIand United Republic o f Tanzania. World Bank. 2004a. A Better Investment Climatefor Everyone. World Development Report 2005. Washington D.C. World Bank. 2004b. Improving Enterprise Performance and Growth in Tanzania. Investment Climate Assessment. Washington D.C. World Bank. 2004c. Sri Lanka: Improving the Rural and Urban Investment Climate. Investment Climate Assessment. Washington D.C. World Bank. 2006a. TheRural Investment Climate: It Differs and It Matters. Agriculture and Rural Development Department. Washington D.C. Draft report. World Bank. 2006b. Local Government TaxationReform in Tanzania: A Poverty and Social Impact Analysis. Report 34900-TZ. Social Development Department. Washington D.C. World Bank. 2006c. Tanzania: Sustaining and Sharing Economic Growth. Country Economic Memorandum and Poverty Assessment. Poverty Reductionand Economic Management Unit 2, Africa Region. Washington D.C. Draft report. -91- BACKGROUNDDOCUMENTS Economic and Social ResearchFoundation. 2004. "Supply Factors Affecting Rural Credit in Tanzania." Backgroundpaperfor Tanzania's Rural Investment ClimateAssessment. Processed. Keough, James A. 2006. "Selected Summary Statistics o f Tanzania's Rural InvestmentClimate Survey." Exploratoy Analysisfor Tanzanias Rural Investment ClimateAssessment. Processed. Ramaswamy, Ram. 2006. "Tanzania's Rural Investment Climate Survey -Addressing Tanzania's Key Bottleneck: RuralFinance." Backgroundpaperfor Tanzanias Rural Investment Climate Assessment. Processed. Sundaram-Stukel, Reka, Klaus Deininger and Songqing Jin. 2007. "Fostering Growth o f the Rural Non-farm Sector inAfrica: The Case o f Tanzania." Backgroundpaperfor Tanzanias Rural Investment Climate Assessment. Processed. Temesgen, Tilahun. 2005a. "Main Observations and Some Basic Statistics from Tanzania RICS Data." Preliminav draftfor Tanzania's Rural Investment ClimateAssessment.Processed. Temesgen, Tilahun. 2005b. "Tanzania: InvestmentClimate for Non-farmActivities." Preliminay draft for Tanzania's Rural Investment ClimateAssessment. Processed. Temesgen, Tilahun. 2006. "Technical Appendix on Tanzania RICS Sampling Procedure and Weights. Background notefor Tanzania's Rural Investment ClimateAssessment.Processed. -92-