Report No. 44286-ZM Zambia What Are the Constraints to Inclusive Growth in Zambia? A Policy Note Poverty Reduction and Economic Management Poverty Group (PRMPR) Economic Policy and Debt Department (PRMED) Southern Africa Unit (AFTP1) Document of the World Bank Table of Contents Acknowledgements ....................................................................................................... 5 Abstract ......................................................................................................................... 6 1 IntroductionandRecentDevelopments . ............................................................... 7 2 . A Frameworkfor SharedGrowthAnalytics ...................................................... 17 3 . SharedGrowthAnalyticsfor Zambia ................................................................ 21 Are employability and productive assets mainconstraintsto productiveemployment?..........27 Education 27 Health .................................................................................................................................................. ............................................................................................................................................. 29 I s the high cost of capital an obstacleto income growth? ......................................................... 30 Dolow social returnsimply low income growth?;................................................................... 36 Mining. agriculture and natural endowments 36 Geography ........................................................................................................................................... ...................................................................................... 38 Infrastructure ....................................................................................................................................... 39 Humancapital ...................................................................................................................................... 42 Are government failures aconstraint to shared growth?........................................................... 43 Macroeconomic environment .............................................................................................................. 43 Tax code .............................................................................................................................................. 46 LandRights ......................................................................................................................................... 49 Regulatory uncertainty. government interventions and other administrative procedures.................... 48 Governance .......................................................................................................................................... 50 Are market failures reasonsfor low returns?............................................................................ 51 4. Conclusions .......................................................................................................... 57 Figures Figure 1:Zambia's GDP per capita and annual growth. 1962-2006................................... 7 Figure 2: Growth accounting decompositionfor Zambia ................................................... 8 Figure 3: Inflation1986to 2006 (percent) .......................................................................... 8 Figure4: Copper production and copper prices 9 Figure5: Internalmigration flows. 1996-2006................................................................... .................................................................. 9 10 Figure 7:Non-oilHerfindahl Index for countries inSSA. 2004....................................... Figure 6: Degree of export diversification. Zambia .......................................................... Figure 8: Composition of exports inZambia..................................................................... 11 11 12 13 Figure 11:Poverty rates inrural and urbanareas. 1991-2006........................................... Figure 10: Loans and advances to the private sector......................................................... Figure 9: Changes inreal growth rates (96) ....................................................................... 13 Figure 12: Incidence of poverty by stratum....................................................................... 14 2 Figure 13: Percentage distribution of households that perceived a change inwelfare Figure 14: Shared growth analytics ................................................................................... 18 duringthe recent year (2005-2006) ................................................................................... 14 19 Figure 16: Real value added per worker inZambia .......................................................... Figure 15: Businessenvironment analysis ........................................................................ Figure 17: Distribution of labor inZambia by sector, 2004 (9%of total labor force) ........23 24 Figure 18: Number of people employed by stratum, 1996-2006...................................... 24 Figure 19: Number of people employed by sector inthe rural and urbanareas (2004) ....25 Figure 20: Poverty maps, Zambia...................................................................................... 26 Figure21: Percentage of households with accessto education facilities within five kilometers.......................................................................................................................... 28 Figure22: Examples of self-assessedreasons for poverty, rural andurbanareas (percent) 31 Figure23: Interest rate spreads (lending minus deposit rate) ............................................ ........................................................................................................................................... 31 Figure 24: Real cost of capital (averages) ......................................................................... 32 34 Figure 26: Number of commercial banks' branches .......................................................... Figure25: Percentage of populationwith abank deposit account.................................... Figure 27: Reasons for not having a bank account............................................................ 34 35 35 Figure 29: Indirect cost to agriculture ............................................................................... Figure28: Reasons for not having a micro finance institution (MFI) loan....................... 38 Figure30: Access to air transport...................................................................................... 39 Figure 31: Cost structure: firm-level averages b y country ................................................ 41 Figure 32: Mobile rates per minute ................................................................................... 42 Figure33: Skilled migration andhumancapital inAfrica, 2000 ...................................... 43 Figure 34: The development o f the Kwacha exchange rate .............................................. 44 Figure35: Agricultural exports 2000-2007 ....................................................................... 45 Figure36: Import and export shares, Zambia 1960-2007 ................................................. 46 Figure 37: Country comparison of global rankingon tax level......................................... 47 Figure 38: World Bank Governance Indicators for Zambia.............................................. Figure39: The number of households with accessto facilities within five kilometers ....51 52 Figure 40: Constraints to tourism growth as identifiedby the hospitality sector inZambia ........................................................................................................................................... 55 Tables Table 1:Mean shares of household income by source. income quintile. rural areas ........21 Table 3: Shares intotal value added inZambia (percent) ................................................. Table 2: Mean shares of household income by source. income quintile. urbanareas.......21 22 Table 4: Industries' contribution to real growth inZambia (percent) ............................... 23 Table 5: Structure of the farming sector ............................................................................ 25 Table 6: School enrollment rates ....................................................................................... 27 Table 7: Literacy rates: cross-country comparisons .......................................................... Table 8: Mean years of schooling of householdhead in2002/03 ..................................... 28 28 30 Table 10: Investment andsaving indicators ...................................................................... Table 9: Access to sanitation andwater ............................................................................ 33 3 36 Table 12: Comparative prices of diesel fuel (US$/liter) .................................................... Table 11:Country comparison o f some natural resource endowments............................. 39 Table 13: Comparative electricity costs (US$/kwh. 2001)................................................ 42 48 Table 15: Agricultural production and yields.................................................................... Table 14: Tariffs on key categories of goods. 2003 (percent) ........................................... 53 Table 16: Spearman rank correlation between farm yields per hectare and services performance inZambian districts...................................................................................... 54 4 ACKNOWLEDGEMENTS This paper was prepared by the team leaders Elena Ianchovichina andSusanna Lundstrom and is ajoint product of PRMPR, PRMED and AFTP1. It was prepared upon request from the Zambia country team to serve as an input into the Zambia CAS. We thank Ricardo Hausmann, DaniRodrik, Lant Pritchett, Jos Verbeek, Aditya Mattoo, andother participants inthe Harvardworkshop on Zambia for useful comments onthis work, as well as participants at the seminars inLusaka with the government and NGOs, andwith donors. We are grateful to Alan Gelb, Jonathan Kydd,Arne Bigsten, Lindavan Gelder, and Louise Fox, for their written comments on this study, and to Ana Revenga, Louise Cord, Kapil Kapoor, John Panzer, Kenneth Simler, Emmanuel Skoufias, Francesco Caramazza, Verena Fritz, BrianLevy, Marie Sheppard, and Helen Mbao who sharedwith us valuable comments. We would also like to acknowledge the comments of participants inthe workshop on Shared Growth and Job Creation, held at the University of Oxford, and ina seminar organized by the Poverty Group of PREMnetwork at the World Bank. Dotilda Sidibe (program assistant, AFTP1) andLuluMilinga (team assistant, AFTP1) provided administrative and mission assistance. 5 ABSTRACT Despite positive, relatively broad-based and stable growth record inrecent years and immense untappedpotential inagriculture, miningand services, Zambia's poverty rates remain high. Income growth is limitedby coordination failures such as poor access to domestic and international markets, inputs, extension services and information. High indirect costs - most of which attributable to infrastructure service-related inputs into production including energy, transport, telecom, water, but also insurance, marketing and professional service - undermineZambia's competitiveness, limitjob creation and therefore serve as a major constraint to inclusive growth. Continuedreal appreciation i s another serious threat to the competitiveness of export-oriented and import-competing sectors andhence to job creation. For Zambia to stay competitive and sustain the growth momentum, it will be critical to improve productivity - includingthe productivity of its labor force, and to lower indirect production costs related to basic services. Carefully crafted monetary andfiscal policies will also be critical inresponding to the real appreciation pressures. Improving the quality and access to secondary andtertiary education is essential ifthe poor are to benefit from future growth of the non-farm economy. Weak governance andinparticular poor government effectiveness, are factors behind the market coordination failures andthe identified government failures, and are as such major obstacles to inclusive growth in Zambia. 6 1. INTRODUCTIONAND RECENTDEVELOPMENTS 1. Zambia i s a country which despite its mineral wealth and fertile soil i s one of the poorest countries inthe world.' Its rank inthe UNHumanDevelopment Index for 2007- 08 is 165 out of 177 countries. Its per capita income is still muchbelow the per capita income at independence in 1964 (Figure 1) and the poverty rate i s as high as 64 percent. At independence the country's income level was 75 percent above the African average andfour times that of East Asia (Bigsten andTengstam, 2007). Today the per capita income is below that of the African average and a quarter of what it i s inEast Asia. 2. After decades o f declining standards of living, Zambia's economy started growing andper capita incomes started risinginthe late 1990s (Figure 1).The recent positive and stable growth recordhas been accompanied by productivity improvement (Figure 2), and presents a distinct break with the past of highgrowth volatility. Can this highgrowth recordbe sustained and made more inclusive?This study tries to identify the key constraints to inclusive and sustained, rapid growth inZambia. An emphasis on increasing the opportunities for the poor to contribute and benefit from the growth process i s critical given the fact that the majority of Zambian people are poor and/or vulnerable. Figure1:Zambia's GDPper capita andannualgrowth, 1962-2006 15 10 5 8 0 ,I -5 -10 -15 ~ ' GDP per capita growth (annual Yo) L-JGDP growth (annual %) -m- GDP per capita (constant LCU) Source: World Bank (DDPdata). In2006per capita income inZambia was $365 (constant 2000 $). 7 Figure 2: Growth accountingdecompositionfor Zambia I Source: Authors' calculations assumingCRTS (see specifics about the calculations inAppendix) using WorldBankdata andBarro andLee (2000). 3. Before we start the analysis it i s important to understand what was behind the change inproductivity and growth that occurred around 1998 and the country-specific context of economic development inZambia. The improved macroeconomic environment duringthe 1990sis often mentionedas a major factor behindthe improvement inthe business environment andhas most probably had, with some time lag, a significant impact on growth. For example, inflation was over 180percent inthe early 1990s but has now returned to single digits (Figure 3) and the government budget deficit halved as a share of GDP between 2003 and2006. Figure 3: Inflation 1986to 2006 (percent) Source: WDI andMF (2007). 4. Another important factor that i s often assumedto have influencedthe latest positive economic developments was the sharp increase inthe price o f copper. This has certainly helped growth inZambia but it i s also important to notice that the increase in 8 productionstartedbefore the sharp increase inthe international price of copper in2005 (Figure 4). Figure 4: Copperproductionandcopper prices 800.0 700.0 600.0 500.0 400.0 300.0 200.0 100.0 0.0 1- Production (rnt'000) LHS -+International price (constant 1990USD rnt) RHSi Source: Data from Bank of Zambia. 5. Rapid growth between 1996 and 2006 was accompanied by risingdomestic migration inresponse to the structural changes inthe economy (Figure 5). There were strong rural-to-rural and urban-to-urbanlabor movements andmuchless pronounced increases inrural-to-urban migration. The later reflects perhaps the fact that Zambia already had a highlevel of urbanization (35 percent) given its development stage, andthe excess of urban labor resultingfrom the privatization of mines. It i s important to notice that urbanto rural migration was also on the rise between 1996 and 2004. However, inall reported years since 1996 data suggest positive and risingnet rural-to-urban migration. In the period 2004-06 migration flows seem to have stabilized with the exceptionof rural- urbanflows. Figure 5: Internal migrationflows, 1996-2006 I 160000 140000 120000 100000 I 1996 I199% %OOOO 0 2004 60000 0 2006 40000 20000 0 Ruralto Rural Ruralto Urban Urbanto Rural Urbanto Urban Source: CSO (forthcoming). 9 6. I s the recent growth episode a one time event related to a rebound from the improvement inthe macroeconomic environment, the increase incopper prices, and the structural change accompanying reforms inthe early to rnid-l990s, or has Zambia embarked on a new, sustainable growth path? There are indications that recent growth i s not only a sign of these external events, but an outcome of more fundamental changes in the economy that have ledto new sources o f growth. Zambia has managedto broaden its export base. Inthe period 1980-2004 the country nearly doubledthe number of products exported (Figure6) and halved its Herfindahlindex2breakingaway from the group of least diversified economies inSub-Saharan Africa (Figure 7). While in 1980 the five largest Zambian exports accounted for 96 percent of its exports, in2004 they made up about 80 percent of exports (Figure 6).3 Figure 6: Degree of export diversification, Zambia I 0 4 c 0 00 1980 1985 1990 1995 2000 2004 -+-No. I of products exported (> 10,000 US$) +Herfindah1 index --*c Five largest (share of merchandise exports) Source: World Bank, Export diversification data, PRMED. 7. Zambia has diversified by capitalizing on its advantage inland-intensiveprimary goods (Figure 8). Miningproducts still dominate merchandise exports, but while inthe three decades after independence Zambia relied exclusively on exports of ores and metals, inthe last 17 years agricultural exports, including non-traditional farm exports, started playing a muchmore prominent role.4 Zambia's share o f food and other farm products intotal exports increased from 4 percent in 1980s to 20 percent inthe early 2000s. Evensmall scale farmers have diversified-by 2002/03 one out o f five grew cotton, 45 percent derived income from animal products and 17 percent from horticulture. The Herfindahl index measuresthe degree of export diversification. The higher the Herfindahl index, the lower the degree of diversification. Although the economy has grownrapidly and trade has expanded inabsolute terms, it i s declining in relative terms (see Figure 36). The total gross value of agricultural output has risenby over 50 percent between mid-90s and 2001-2004. Cotton and tobacco has contributed to export-led growth. Cassava, sweet potatoes, cotton and groundnuts production have increased. 10 Figure7: Non-oilHerfindahlIndexfor countries inSSA, 2004 Source: World Bank, Export diversification data, PFWED. Figure8: Compositionof exportsinZambia 0250 1200 $0200 1000 4 .- c 0800 c 3 ~ - 3 0150 0600 u0 0100 2 0 400 2 -- z50 050 0200 g 0 000 0 000 ~ Food as a share of primary goods 0RawMaterialsasashareofprimarygoods 1 Manufacturing as a share of exports Textiles as a share of Manufactig -Ores and Metals as a share of exports +-Primary Goods (non ores and mineral) as a share of exports Source:World Bank, Export diversification data, PRMED. 8. The decline inthe export share of mininghides importanttrends inthe mining sector which has diversified away from copper into other base metals and precious stones. In 1995 other base metals were fifth on the list o f Zambia's top 5 exports, and they accounted for less than2 percent of total exports. By 2002 other base metals have moved to the second spot and represented 15 percent of Zambia's merchandise exports. Precious stones, which were not on the top 5 list in 1998, represented nearly 4 percent o f total exports. 9. Inadditionto mining,inthelasttenyears, growthwas drivenbystrong expansions inservices and construction, and to a lesser degree inmanufacturing. The 11 change inaverage growth rates from the period 1991-98 to 1998-06, was 4.4 percent in total of which 1.8 percentage point came from mining, 1.1from services and 1.4 from construction, but only as little as 0.4 from manufacturing(Figure 9). Growth within services was mainly driven by growth incommunity and social services, real estate and business services, and wholesale and retail sales (Mattoo and Payton, 2007). These sectors are more often an "employer of last resort" unlike service sectors such as tourism, transport and communication, and finance and insurance, which tendto reflect broader economic dynamism. Understandingthe constraints to growth inthese latter service sectors and the manufacturing sector i s an important step toward understandinghow to sustain growth inZambia inthe long run. Figure 9: Changes inreal growth rates (%) Services, etc 1998-06 Manufacturing I1991-98 Mining and quarrying Construction Agriculture Net indirect taxes GDP Grow th -1.5 -0.5 0.5 1.5 2.5 3.5 4.5 Source:WorldBank (2007b). 10. Robust increases inforeign direct investment inflows accompanied the boom in the mining sector. In2005 FDIinflows increasedby 166 percent, compared to just 2 percent in2004. Investmentpledges increased the most inmanufacturing and mining,5 and decreasedinagriculture, tourism andtransport. Credit to the private sector increased significantly for almost all sectors indicating increased domestic private sector activity (Figure 10).The sectors with the strongest growthinprivateloans and advances were agriculture, wholesale and retail trade, manufacturing and other sectors.Loans to other sectors includedpersonal loans (usually used for investment ina sector rather than consumption), loans to miningsuppliers, law firms, audit firms, NGOs and development organizations, private hospitals, cleaning services, book publishers, and others. It hasbeenestimatedthat US$1.4 billion were injectedinthe miningsector inthe last 3 years. 12 Figure 10: Loans and advancesto the private sector 1,600,000 -Agriculture, forestry,Fishinc hunting 1,400,000 -and Mining and quarying 1,200,000 Manufacturing Electncity, gas, water and 1,000,000 energy -Construction 800,000 -Wholesale and retail trade -Restaurants and hotels 600,000 -Transport, storage and 400,000 , communications ---Financial seMces 200,000 Community, social and personal seMces Real estate 0 1 Other sectors Source:Bankof Zambia. 11. Despite robust and increasingly broad-based growth inrecent years, aggregate poverty rates inZambia have declined only slightly and remain high.According to the household survey in2006,64 percent of the people inZambia are still poor (Figure 11). Poverty rates remain highest inrural areas (80 percent) where two-thirds of Zambia's population resides. This implies that the vast majority of the poor (72 percent) live in rural areas. Other measures of well-being paint a disturbingpictureincluding increasing child malnutrition over the 199Os, highprevalence o f HIV/AIDS, and the lowest life expectancy inthe world in2007. Figure 11:Poverty rates inrural andurbanareas, 1991-2006 270 250 # fj& 230 210 40 - -* t 190 30 - "t 170 20 : : : : : : : 150 ~ 9 N 8 i Q 8 8 ~ 8 8 8 9 f l 8 8 8 8 ..-I-NationalHeadcount(LHS) -----m- -.National - +-.- - National Rural (LHS - Urban (LHd) -GDP4- Po\~rtheadcountat 1US$aday FYer Capita (RHS) ! Source: Republic of Zambia (2006a), CSO, WDI, andBank staff estimates. 13 12. Looking separately at trends inrural and urbanpoverty rates, one sees a sharp decline inurbanpoverty between 2004 and 2006, and a slight increase inrural poverty duringthe same period (Figure 11).Understanding why urbanpoverty levels seemto have responded to the new economic opportunities but not rural poverty i s crucial for understandingthe constraints to shared growth. Figure12: Incidenceof povertyby stratum 100 90 80 70 60 1996 50 0 2004, 40 El 30 20 10 0 RuralAgr Rural Agr Rural Agr Rurd Non- Urban Low Urban Urban Hgh Small Medium Large Agr Cost Medium Cost cost Source: CSO (forthcoming). Note: Small, Mediumand Large inthe rural area refer to the scale of the farm, and Low, Mediumand High costs inthe urban area refer to the cost of the residential area. Figure13: Percentagedistribution of householdsthatperceivedachangeinwelfare during the recent year (2005-2006) 70 60 50 40 Better off The Sam 30 Worse off 20 10 0 Rural Fbral Rural Rural Urban Urban Urban Agr Agr Agr Non- Low Wdium Hgh Small Wdium Large Agr Cost Cost Cost Source: CSO (forthcoming). Note: Small, Mediumand Large inthe rural area refer to the scale o f the farm, and Low, Medium and High costs inthe urban area refer to the cost o f the residential area. 14 13. Poverty rates are highest among small and medium-sizedfarmers andnon-farm rural residents (Figure 12). When asked directly how they perceive their welfare change duringthe precedingyear, households involved inlarge-scale farmingwere most likely to have felt an improvement, followed by urbanhouseholds inhighcost residential areas (Figure 13). The larger the size of the farm, or the higher the cost level inurbanareas, the stronger was the perceived improvement inthe household's living standards. Finally, across all categories, the share of households reporting improved welfare was larger than the share of those reporting deterioration. 14. Growth inZambia must accelerate inorder to reduce poverty ina sizable way. Quantitative analysis by Bigsten and Shimeles (2007), who analyze the trade-off between growth and redistribution for poverty reduction inseveral African countries inthe period 1981-2001, find that annualper capita growth inZambia would need to be 4 percent in order to halve poverty between2001 and 2015, assuming a constant Gini. Although many assumptions underlie this conclusion it indicates that growth needs to accelerate in Zambia even beyond the 3 percent average annual per capita growth recorded inthe strong growth period 2001-2006. 15. This paper investigates the nature of the growth process inZambia andthe bindingconstraints to shared growth, i.e. inclusive growth inwhich the poor contribute andbenefit from the growth process. Inour analysis we employ a method inspiredby the growth diagnostics approach of Hausmann,Rodrik and Velasco (2005) and the framework of the `Integrated Economic Analysis for Pro-Poor Growth' (Sida, 2006). The method guides us inthe process of identifyingthe most bindingconstraints to the poor's income growth through productive employment, be it wage-employment or self- employment. 16. Our approach is distinctfrom the growth diagnostics exercises a laHausmann, Rodrik and Velasco (2005) inthat we look at both demand-side and supply-side conditions that prevent the poor from takingpart inthe growth process. Onthe supply side we focus on constraints that hamper employability andaccess to labor markets. On the demand side we consider obstacles to job creation and productivity improvements. We focus onthe ability of the poor to participateinthe growth process and not on redistributive measures for poverty reduction. This implies that our analysis needs to take a longer terms perspective than the growth diagnostic analyses a la Hausmann, Rodrik andVelasco (2005).6 17. We find that the mainbindingconstraints to shared growth inZambia are coordinationfailures. These include poor accessto domestic and internationalmarkets, inputs,services and information. Highindirect costs -most of which attributable to infrastructure service-related inputs into production includingenergy, transport, telecom, water, but also insurance, marketing and professional service - undermine Zambia's competitiveness, limitjob creation and productivity, and therefore are also major constraints to pro-poor growth. Continued real appreciation i s another serious threat to Itis not only about finding the constraintthat ignitegrowth and poverty reduction,but the binding constraintsthat ignite sustained and inclusive growth. 15 the competitiveness of export-oriented and import-competingsectors andto job creation. Carefully crafted monetary and fiscal policies will be critical inresponding to the real appreciation pressures. For Zambia to stay competitive and sustain the growth momentumit will be critical to improve productivity. Improving the quality and access to secondary and tertiary education and health services i s essential if the poor are to benefit from future growth of the non-farm economy. Weakgovernance and inparticular poor government effectiveness are factors behind the coordination failures observed in Zambia, and are as such major obstacles to inclusive growth. 18. Thepaper is structured as follows. Section2 presentsthe analytic framework and data for shared growth diagnostics. Section 3 describes Zambia's economic growth record and the economic situation of the poor, and applies the framework to the case of Zambia. Section4 concludes with a summary and some caveats. 16 2. A FRAMEWORK FORSHARED GROWTHANALYTICS 19. The economic agent inthe shared growth diagnostics framework i s the individual rather thanthe firm, since we assumethat the main instrument for a sustainable and inclusive growth i s productive employment. Employment growth generates newjobs and income for the poor - from wages inall types of firms, or from self employment, usually inmicro firms - whileproductivity growthhasthe potentialto liftthe wages ofthose employed and the returns to the self-employed. 20. The ability of the poor to be productively employed depends on the one handon their employability, which inturn depends on their individual resources. An employability analysis would include analysis of (i) existing stock of humancapital, the such as education and health; (ii) the ability of the poor to acquire skills and stay healthy; and (iii) to labor marketswhere individuals can earn income by offering their access skills. On the other hand, the ability of the poor to be productively employed depends on the opportunities for the poor to makefull use of these resources as the economy evolves over time. The analysis therefore looks at ways to strengthen the productive resources and capacity of the poor on the labor supply side as well as ways to open up new opportunities for productive employment on the labor demand side. 21. Ifthe mainproblemis lack of employment for the poor dueto limitedsupplyof certaintypes of labor skills, the constraints are relatedto the productive resources and capacity of the poor as individuals rather than the environment inwhich they can use these resources. This situation calls for an in-depth employability analysis that will shed light onthe resources of the poor, e.g. the poor's labor skills and the productivity attributes that they bring to ajob. Ifthe mainproblem is low labor productivity or lack of employment opportunities for the poor due to limiteddemand for labor, an analysis of the bottlenecks inthe business environment i s necessary. 22. We start by identifying the poor as productive actors who earn income either as self- or wage-employed, and further distinguishemployment by sector, size of firm, rurauurban, formaUinforma1, and other relevant characteristics. Inthe case of the self- employed, we undertake business environment analysis through the lenses of the small andmicro enterprises of the poor (Figure 14). Inthe caseof the wage employed, we undertake an employability analysis as well as a business environment analysis through the lenses of a representative firm, potentially employing the poor.7 7 Note that the analysis of labor skills as a potential constraint for the self-employed i s captured in the business environment analysis where it is analyzed as a constraint to growth of the small firm. 17 r e 14: Shared growth analytics Economic Poverty Reduction Growth 23. The business environment analysis follows, but is not limitedto, the aggregate- type of growth diagnostics suggested by Hausmann, Rodrik andVelasco (2005) as presented inFigure 15.8Investments and entrepreneurial activities are determined by the relationship betweenprivate returns to economic activities and cost offinance. Private returns are determined by social returns, which depend on factors such as geography, technology, infrastructureandhumancapital, and the private appropriability o f these returns. Private a~propriabilityreflects the extent to which social returns are translated into private returns andi s negatively affected by government failures or market failures. Government faizures, or bad policy and poor institutional environment, include macroeconomic risks such as financial, monetary and fiscal instability, and microeconomic risks, such as insecure property rights, corruption, inefficient tax collection systems, cumbersome regulations andbusiness registration procedures. Market failures include limited information and coordination externalities affecting negatively the country's ability to expand private sector development and adopt new technologies. These failures are acute when markets are fragmented either horizontally, e.g. due to geographical isolation, or vertically, e.g. inputs are limited locally inquality and quantity. * Hausmann, Rodrik and Velasco (2005). For an overview see Rodrik (2006). 18 Figure 15: Businessenvironmentanalysis Source: Adapted from Hausmann, RodrikandVelasco (2005). 24. Inadditionto the employability analysis, akey differencebetweenthe growth diagnostics framework proposed by Hausmann, Rodrik and Velasco (2005) and the shared growth analytics framework proposed inthis paper i s that instead of going through the diagnostic treefrom theperspective of a representativefirm or investor, the shared growthframework looks at the treefrom theperspective of thepoor as a productive actor - a worker and/or a self-employed. 25. To find out what this perspective on shared growth implies for a specific country, the results from an employment diagnostic analysis will be crucial. This analysis will be able to shed light on a number of important questions includingthe extent to which the poor are self-employed or wage employed; whether they live and work inthe rural or urbanarea; inmanufacturing, mining, agriculture or services; informal or informal firms; insmall, medium or large firms; exporters or non-exporters; foreign or domestic. 26. Another important question is to what extent the current employment status of the poor has a potential for future income growth from productivity improvements, or if moving out of poverty would meanfinding another type of employment or employment inanother sector. Thebusiness environmentanalysis shouldtherefore beprecededby an analysis of externalfactors explaining the country's growth and poverty reduction pattern, the overall productivity dynamics in the country, the major challenges and opportunitiesfaced, and thepossibilities for economic transformation and diversification. 19 27. The analysis of the constraints inthe business environment that the poor face in their current employment should then be complemented with ananalysis of the constraints inthe sectors with opportunities for productive employment and constraints affecting their ability to gain employment inthese sectors. Such an analysis will needto review constraints affecting labor mobility across sectors and regions, labor market constraints, access to education, finance and infrastructure. 28. Another key difference betweenthe growth diagnostics framework of Hausmann, Rodrik andVelasco (2005) and this shared growth analytics framework i s that we go beyondthe aggregate level when analyzing the business environment and rely on industry and firm level data, as well as household level data. We do so becausethe aggregate picture typically hides important details at the industrylevel. Most importantly, however, our use o f industry,firm andhouseholdlevel data i s necessitated by our main objective to identify the incidence of growth across the income distribution and the bottlenecks to the productive employment of the poor. 29. While inthe growth diagnostics approach the emphasis i s on identifyingthe reforms neededto ignite growth inthe short runthat can hopefully create a "virtuous circle" o f business activities, the shared growth approach takes a longer runperspective. This is necessarybecauseour emphasis is on improving the productivecapacity and creating conducive environment for employment of the poor, rather thanon income redistribution as a means of poverty reduction. Due to this longer term perspective, there will be a more explicit focus on structural dynamics inthe shared growth analytics framework than inthe growth diagnostics framework. This i s highly relevant inthe poorest countries where a significant part of shared growth will come from reallocation of labor from low-productivity to high-productivity sectors. Our goal i s to identify a bundle of bindingconstraints rather thanthe bindingconstraint, andthen sequence these constraints to maximize inclusive growth ina country. 30. Finally, it i s important to recognize the time lag betweenreforms and outcomes andthe rapidchanges of the economy due to external and internal factors. This implies that the analysis must identify future constraints to growth that may not be bindingtoday, butthat may needto be addressedtoday inorder to ensure sustainable andshared long- rungrowth. Shared growth analytics is about policiesthat shouldbeimplementedinthe short run, for sustainable shared growth. 31. Insummary,the focus insharedgrowth analytics is onthe poor andtheir constraints to productive self or wage employment, rather thanthe constraints to investment of a representative firm. This is why growth diagnostics i s limited to an analysis of the mainbottlenecks to private sector growth ingeneral while shared growth analytics identifies the main bottlenecks to private-sector led growth taking into account the distributional aspects of different constraints and the employability of the labor force. 20 3. SHARED GROWTHANALYTICS FOR ZAMBIA 32. To understand why poverty rates inZambia have remainedhighdespite strong growth inthe past decade this section starts by exploring the dynamics o f different sectors andthe extent to which the poor profit from the growing sectors or are dependent on stagnating sectors. We discuss the employment profile of the poor, and estimate labor productivity andjob creation at the industry level. 33. A closer look at the sources of income of the rural, mostly self-employed, household heads, suggests that farmers only get 10 percent of their income from farm sales, andthe majority of their farm output is for subsistence purposes (Table 1).Few rural households have sufficient resources to hirepoorer neighbors or provide them with loans. Richrural households tend to rely more on wage employment and less on subsistence farming than poor rural households, but there are no wide differences inrural households' wealth and education levels (Table 1).Rural inequality i s very low and better-off households are also viewed as vulnerable and their future well-being less than certain. Table 1:Mean shares of household income by source, income quintile, rural areas QuintileofNationalDistribution All Poorest 2 3 4 Richest 20% 20% Foodcrop sales 6 7 6 6 6 5 Nonfood crop sales 2 1 2 3 2 2 Non-farmbusiness 10 11 10 10 13 11 Livestock and other farm income 2 2 2 3 2 3 Salary 6 3 5 6 7 11 Remittances 6 7 6 6 7 5 Pension 0 0 0 0 0 0 Nonagriculturalrent 0 0 0 0 0 0 Other income 11 13 12 12 11 9 Consumption of own production 55 57 55 54 55 52 Source: World Bank (2007a). Table 2: Mean shares of household income by source, income quintile, urban areas Quintile of NationalDistribution All Poorest20% 2 3 4 Richest20% Non-farmbusiness 24 25 27 27 26 19 Salary 50 38 40 46 50 60 Remittances 6 8 6 5 6 5 Other income 15 20 19 16 14 13 Consumption of own production 5 9 7 7 4 3 100 100 100 100 100 100 Source: World Bank(2007a). 21 34. Inurbanareas, thepoorresideininformalsettlements which accordingto estimates host 50-80 percent of the urbanpopulation. The urban oor are typically self- employed, but they may also be unemployedor underemployed. The self-employed in B urbancenters arepart of the largeurbaninformal sector which inZambia provided livelihood to 56 percent of the urban workers in2002-03. They are employed ina variety of informal economic activities ranging from producing and selling buildingmaterials, to trading petty commodities, farming, and rentingreal estate. While the urbanpoor derive a muchlarger share of income from wages compared to the rural poor (Table 2), the urban poor rely much less on income from wages andhave muchfewer years of schooling than the urbanrich (Table 8 insection 3.3.1). The urban poor are typically involved inseveral different activities, including the cultivation of undeveloped urbanor peri-urban landto supplement their incomes and food intake. 35. Inthe last few years urbanpoverty declined, whereas ruralpoverty slightly increased. These trends are a result of an expansion inindustrialand service activities, which have been drivers of growth inZambia (Table 4), represented 86 percent of GDP in2006 (Table 3), andtook place mostly inurbancenters. Table3: Shares intotal value added inZambia (percent) 2000 2001 2002 2003 2004 2005 2006 Average Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Agriculture 18.3 17.1 16.1 15.9 15.7 15.2 14.6 16.1 Industry 23.3 24.2 25.9 26.8 28.3 28.9 29.5 26.7 - Construction 5.3 5.6 6.3 7.3 8.2 9.3 10.0 7.4 - Mining 6.8 7.5 8.3 8.1 8.7 8.5 8.5 8.I - Manufacturing 11.2 11.1 11.3 11.5 11.3 11.1 11.0 11.2 Services 58.4 58.7 58.0 57.2 56.1 55.9 55.9 57.2 - Banking 8.7 8.4 8.3 8.1 7.9 7.7 7.5 8.1 - Utilities 3.1 3.3 3.0 2.9 2.7 2.7 2.8 2.9 - Other services 2.1 2.4 2.5 2.5 2.5 2.6 2.9 2.5 - Real estate 10.1 10.0 10.0 9.9 9.7 9.5 9.2 9.8 -Public administration 8.2 8.3 8.1 7.8 7.4 7.2 7.4 7.8 - Transport 6.7 6.6 6.4 6.4 6.4 6.6 7.5 6.7 - Trade 19.4 19.6 19.7 19.8 19.6 19.6 18.7 19.5 Source: World Bank, DDP. Among individuals age 20 and over, 18percent of those inthe bottom quintile were unemployed according to CSO (2002-03). 22 Table 4: Industries'contributionto real growthinZambia (percent) 2001 2002 2003 2004 2005 2006 Average Total 4.5 4.5 5.8 6.1 5.8 6.7 5.6 Agriculture -0.5 -0.3 0.8 0.7 0.4 0.3 0.3 Industry 2.0 2.8 2.5 3.2 2.3 2.6 2.6 - Construction 0.6 1.o 1.4 1.5 1.6 1.3 1.2 - Mining 1.o 1.2 0.3 1.1 0.2 0.6 0.7 - Manufacturing 0.5 0.6 0.9 0.5 0.4 0.6 0.6 Services 3.0 1.9 2.5 2.2 3.1 3.8 2.8 - Banking 0.0 0.3 0.3 0.3 0.3 0.3 0.2 - Utilities 0.4 -0.2 0.0 0.0 0.1 0.3 0.I - Other services 0.5 0.1 0.2 0.2 0.3 0.4 0.3 - Real estate 0.4 0.4 0.4 0.4 0.4 0.3 0.4 - Public administration 0.5 0.1 0.1 0.0 0.3 0.6 0.3 - Transport 0.2 0.1 0.3 0.4 0.5 1.4 0.5 - Trade 1.o 1.o 1.2 1.o 1.2 0.4 1.0 Source: Authors' calculations basedon data from World Bank, DDP. 36. Labor productivityinindustry and services is also muchhigher than inagriculture (Figure 16). The huge gap betweenfarm and non-farm labor productivityimpliesthat Zambia, which has 70 percent of its labor force employed inagriculture (Figure 17), uses its labor extremely inefficiently. Figure 16: Realvalue addedper worker inZambia 9 P u) g.- 3 Y 2000 2001 2002 2003 2004 2005 2006 Source: Authors' estimates basedon data from Government o f Zambia and World Bank. 23 Figure17:Distribution of laborinZambiaby sector, 2004 (% oftotal labor force) Source: Government of Zambia. 37. The growing sectors created few job duringthe dynamic period 1998-2006 (Figure 18). A possible positive development i s the decline inthe number ofjobs inthe segment where the poorest are concentrated, i.e. small scale agriculture accompanied by an increase inthe number ofjobs inurbanlow cost and mediumcost areas. However, in the period 2004-06 there has been a decline inthe number of urbanjobs inmediumand highcost areas, andthe numberofjobs inlarge-scale agricultural firms remained negligible. Figure18: Numberof peopleemployedby stratum, 1996-2006 2500000 2000000 1500000 I19961 I19981 o 2004 1000000 0 20061 500000 0 Rural Rural Rural Rural Urban Urban Urban Agr Agr Agr Non- Low Medium High Smll Medium Large Agr Cost Cost Cost Source: CSO (forthcoming). Note: Small, Medium and Large inthe rural arearefer to the scale of the farm, and Low, Mediumand High costs inthe urban arearefer to the cost of the residential area. 38. These trends suggest that inthe past decade the path out of poverty was through migration to urbanlow cost andmedium cost areas where the number ofjobs increased andproductivity was higher than inthe rural economy. Would this bethe path of out poverty inthe future? What would it take to increase the speed of non-farm employment creation? What would it take to raise labor productivity? 24 39. The distribution of employment by industry (Figure 19) implies that even a sizable percentage increase inemployment inthe urban industries will result inrelatively few urbanjobs inthe near term. Furthermore, the capital intensive nature o f mining impliesthat increases ininvestmentinflows to this sector will not generate the number of jobs requiredto lift a significant share o f people from poverty, and may have adverse effects on competitiveness through real exchange rate appreciation effects. This suggests that Zambia needs to focus on lifting constraints to productivity and employment creation insectors other thanminingthat will bethe source of incomefor the majority of Zambians. Figure 19: Number of peopleemployedby sector inthe rural andurbanareas, 2004 2500000 __I- 2000000 --r. 1500000 -- 1 0bra1 , Urban 1000000 -- ~ 500000 -- Agric. Mning hnuf. Hect. Constr. Trade Hotels Transp. Finance Corrrrun. ~ Source: CSO (forthcoming). 40. Agriculture i s a sector of special interest as most people inZambia, and especially the poor, are employed and derive their income from farming. But does this sector have a potential for productivity improvements inthe future?Can it be a potentialpath out of poverty? This sector has a dual structure representingon one handa small number of export-oriented commercial farmers that boast productivity levels similar to developed countries, andon the other hand, a large number o f small-scale, subsistence farms that have productivity levels typical of Sub-SaharanAfrica. There are also hybrid operations by medium-sized and emergent farms that produce for both commercial and subsistence purposes (Table 5). Table 5: Structure of the farming sector Small scale Emergent Mediumscale Large scale No of farmers 459,000 119,200 25,230 >40 Hectares per holding 0.5-0.9 10-20 20-60 >60 Crops grown Food crops Foodcash crops Foodcash crops Cash crops Production focus Subsistence Commercial/ Commercial/ Commercial Subsistence Subsistence Source: Republic of Zambia (2000). 41. Opportunities for farrnjob creation incommercial agriculture exist inZambia. A World Bank (2007~)study projects the employment and income implications of 25 commercial agriculture expansion and estimates that expanding irrigated commercial agriculture have the potential to generate, for example, 2 full timejobs per hectare in coffee production, 0.5 inwheathoya, 1infodder crops, 2 inlocal horticulture and as many as 25 infloriculture. Depending on the types of benefits the expected wage would be around $3.6-4.5 a day," which i s much higher than the average return a day of $1.2 for a small-scale cotton farmer, $0.3 for a small-scale maize farmer without fertilizer subsidies, and $1.3 for small scale commercial maize farmers with 50% fertilizer subsidies. The wage differential reflects productivity differentials," andthe potential to fillthisgap andincrease average agriculture labor productivity will becovered inthe subsequent analysis. 42. Another dimension that is crucial to this analysis i s economic geography. There are limited financial resources and capacity to make public investments and the ones that get done need to be targeted to areas where the net benefit to inclusive economic growth will be largest. Zambia's population distribution is highly uneven with 65 percent living inareaswith less than 150personsper squarekilometer. Adding apoverty perspective to this makes the trade-off inpublic investments evenmore difficult. The poverty maps on Figure 20 show the inverse relationship between the poverty headcount and the poverty density and help us distinguishthree groups of poor: (i) urbanpoor; (ii) poor in rural serviceable areas, and (iii) poor inremote areas. This study focuses on the first two rural groups, while policies to increase mobility combinedwith some kind of social safety nets or subsidies (not analyzed inthis study) need to be considered inremote areas. Figure20: Poverty maps, Zambia Source: Simler (2007). loThesejobs would pay ZMK 11,000 a day, plus benefits, holidays and sometimes access to health clinics and schools. Commercially-oriented mediumand small firms pay about ZMK 5,000-6,000 a day. For example, the average yield of Zambian smallholder cotton growers is 0.8 tons per hectare and needs to expand with at least 2 tons per hectare to reach the same returns as wage workers incommercial cotton farms. 26 43. This analysis suggests that the mainreasonsfor low income growth of the poor in Zambia are low returns to self employment -not least inagriculture, andlimited growth of and/or access to wage employment. So what are the main factors limitingreturns to labor andjob creation inZambia? Followingthe tree inFigure 14 we first turnto the questionof employability. AREEMPLOYABILITY AND PRODUCTIVEASSETS MAINCONSTRAINTS TO PRODUCTIVE EMPLOYMENT? 44. This section examines the factors determining the qualitative supplyof labor and whether these factors are a major constraint to income of the private household/individual. These factors including education and health determine the prospects of the poor to seize opportunities inthe business environment inthe longer term. The question of whether the supply of labor i s a constraint to investment o f firms i s analyzed as part of the business environment analysis under social returns. Education 45. Zambia has highprimary school enrollment rates and literacy rates above the average for SSA and LICs (Table 7), but gross secondary school enrollment rates for the period 2002-04 were lower thanthe average for SSA and much lower than the average for LICs (Table 6). In2004, only 24 percent of Zambians attended secondary school and only 18 percent completed the full 12 years of schooling. There is no major difference between the rural and urban areas inaccess to primary schools but insecondary school access the difference is apparent (Figure 21). Table 6: School enrollmentrates Primary Primary Secondary Secondary school school school school enrollment, enrollment, enrollment, enrollment, net (%) . . aross (%) net (%Iaross (%) - . , I , , . I Zambia 80 99 24 26 Uganda 118 13 16 Tanzania 91 106 South Africa 89 105 90 Namibia 74 101 37 50 Malawi 95 125 25 29 Kenya 76 111 40 48 China 118 73 India 90 116 54 Sub-SaharanAfrica 93 30 Low income countries 80 104 45 OECD 96 102 92 101 Data are most recentlyavailable from 2002-2004period Source: WorldBank (DDP data). 27 Table 7: Literacy rates: cross-countrycomparisons CountryName 1990 2005 Zambia 65 68 Sub-SaharanAfrica 54 59 Low income 51 61 Lower middle income 80 89 Middleincome 82 90 Uppermiddleincome 92 93 Highincome 98 99 Source: WorldBank (DDP data). Figure 21: Percentageof householdswith access to educationfacilities within five kilometers. 100 90 80 70 60 50 8 Rural 40 Urban 30 20 10 0 ,I Middle Basic School Secondary School i (1-7) I Source: CSO (forthcoming). 46. While worker skills are not perceived as a major obstacle to private sector growth inZambia, for the poor, education beyondthe primarylevel is amajor constraint to successful self-employment'2 and employment inthe formal sector. While there i s no major difference inmean years of education betweenthe rich andthe poor household heads inthe rural areas (Table 8), school attendance inurbancenters i s differentiated by consumptiodincome levels. In2002/03 the householdheads inthe richest deciles had nearly twice the mean years of schooling thanthose of the householdheads inthe poorest deciles (Table 8).13 At all ages, children from households inthe top quintiles are more likely to be inschool than those inthe poorest quintile. Because the poor typically start school later, the greatest difference is at young ages. Table 8: Meanyears of schoolingof householdhead in2002/03 All Poorest 20% Richest20% Rural 5.3 4.4 6.2 Urban 9.3 6.6 11.1 Source: WorldBank (2007a). l2 Firmstudies indicatethat educationof firmproprietor is most importantfor growthof small indigenous firms. l3 Tertiary education,which is essentialfor jobs inlargecompanies andthe public sector, reachesjust a little over 2 percentof the population. 28 47. Finally, despite improvements inschool infrastructure, and the introduction in 2002 of free education up to seventh grade, the quality o f education remains poor and there i s a dire lack of capacity to meet needs. Fees are still charged for education after grade seven and there are fewer slots inmiddle and highschool. Health 48. Health i s another important dimension of employability and its poor status in Zambia i s a constraint to productive employment for many poor. Estimates of the costs related to HIV/AIDS inZambia are almost 1percent inGDP growth per year according to Zambia C E M (World Bank, 2004b). The highprevalence of HIV/AIDS affects income growth negatively because it underminesthe stock o f available labor, its productivity and limitsincentives for investments for future consumption (physical as well as human capital investments). 49. The HIV/AIDS epidemic i s more devastating inZambia than inmany other SSA countries. In 2005 HIV infection rates inthe working-age population stood at 17 percent inZambiacomparedto the SSA average of6 percent. The infectionratewas muchhigher inurbanareaswhere it affected22 percent ofthe working-age population, butHIV/AIDS underminedthe capacity to supply labor inrural areas as well. 50. Although access to healthfacilities has improved since 1998, it i s still a problem limitingaccessto 12percent of the ruralhousehold that reportedto be more than 15 kilometers away from a health facility. Because of poor health care provision, many of the infected individuals without access to healthcare and medications are unable to continue working productively, if at all. The 2002-03 LCMS examined reasons for urban- rural migration, and found that most migration was infact not inresponse to economic stress, instead many of them were people with HIV/AIDSwho returned to their villages duringtheir final months of life. Shortage of labor is mentionedby many as a problem dividing the self-sufficient and the food deficient andhouseholds with highdependency ratio, i.e. low labor supply per householdmember, are muchmore likely to be poor. 51. Inconclusion, poor quality and low levels of secondary andhighereducation and highprevalence of HIV/AIDS underminethe ability of the poor inZambia to seize economic opportunities. The conclusion comes from our focus on the poor as individuals rather than a representative firm.l4 Next se turnto the business environment analysis in Figure 14. We start with anevaluation of the cost o f capital as a potential binding constraint to shared growth (Figure 15). l4See section 3.3.5 where human capital i s discussed from the perspective of the firm. 29 Table 9: Access to sanitation and water 1990 1995 2000 2004 Improved sanitation facilities, rural Zambia 31 39 46 52 SSA 24 24 27 28 SouthAfrica 53 51 48 46 Improved sanitation facilities, urban Zambia 63 62 60 59 SSA 52 53 53 53 SouthAfrica 85 83 81 79 Improved water source, rural Zambia 27 32 36 40 SSA 36 39 41 43 SouthAfrica 69 69 71 73 Improved water source, urban Zambia 86 88 89 90 SSA 82 81 81 80 SouthAfrica 98 98 99 99 Source: World Bank (DDP data). I S THE HIGHCOST OF CAPITAL AN OBSTACLE TO INCOMEGROWTH? 52. The Zambia GrowthDiagnostic study (World Bank, 2007b) identifiedthe high cost of capital and poor access to finance as bindingconstraints to growth inZambia. According to Zambia's InvestmentClimate Assessment (ICA) report (World Bank, 2004a),15 in2003, the highcost of capital was perceived as the top-most constraint to business operations of Zambianfirms.16 More than 80 percent o f firms inZambia thought that the cost o f financing i s the most bindingconstraint to their operation compared to 60 percent inUganda, and 73 percent inKenya. Lack o f capital to expand or start a business was perceived as the mainreason for their poverty status by 30 percent of ruralresidents and42 percent of urbanresidents in2002/03 (World Bank, 2007a). 53. Recent data however suggest a more nuanced picture. In2006, access to agricultural inputs and low wages, not high cost of capital, were cited by the largest share of poor people as reasons for poverty inrural and urbanareas, respectively (Figure 22). Lack of capitavcredit to extend or start a business was perceived as the mainreason for poverty by 14 percent inrural areas and 19 percent inurbanareas. l5Itis importantto keepinmindthat Zambia's 2007 InvestmentClimate Survey data was collectedin 2003 and many aspects of the businessenvironment have changed since then. Moreover, over 70 percent of the work force is employed inthe informal sector, and the I C A survey covers only formal companies. The large size of the informal sector indicates serious problems with the investment climate. l6Thehighcost of capitalandthe limitedaccessto finance were perceivedas amajor or severeobstacleto growth by 82 percent and 54 percent of business owners, respectively. The average interest rate spreadwas 17percent in2005 which was substantially higher than the one inOECD (3 percent) and SSA (11percent). Domestic credit to the private sector was on average only 7 percent o f GDP inthe period from 1999 to 2005, compared to 176percent inOECD and 62 percent inSSA. 30 Figure 22: Examplesof self-assessedreasonsfor poverty, rural andurbanareas (percent) 0 5 10 15 20 25 30 Lackof inputs Cannot afford agricuitural input Inadequateland Low agricuituralproduction Low prices for agricukuralproduce 7 Rura Lackof market for agricutluralproduce w Urba Lackof credit or capital Lack of errployment opportunities Salaries/w ages too low Rices of comdities too high Hard econon-ic times/econon-ic decline Source: Summary from CSO (forthcoming). 54. While risk premiums on lending to firms have beenhighby international standards (Figure23), they are now approaching averages inother-African countries and LICs. Moreover, on average the real cost of capital inZambia has been below that of Uganda and other large, booming, copper exporting countries such as Mongolia, and very close to the real cost of capital inSouth Africa (Figure 24). If any, it i s poor access to and highcost of capital onloans to small, domestic companies that has constrained the growth of the private sector. Figure 23: Interest rate spreads (lending minus deposit rate) Source: Authors' own calculations using World Bank data. 31 Figure 24: Realcost of capital (averages) 1000 900 80 0 70 0 I 600 :500 fi 400 300 20 0 100 0 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 +Zambia -Mongolia +South Africa Uganda Source: World Bank,DDP. 55. According to Zambia's I C A (World Bank, 2004a), in2003, nearly 50 percent of largerfirms (over 100employees) had a loan, while only 19percent of small firms (10-49 employees) had a loan. The cost of these loans to small firms was also more than 10 percentage points higher than those to large firms. Similar differentials existed between the cost of capital of exporters and non-exporters, as well as domestic and foreign companies. Informal firms, not represented inthe survey, typically face even steeper constraints when it comes to cost and access to capital. These firms have to finance their operations largely from own funds. Poor access to capital for small formal firms gives informal companies little incentives to become formal. 56. The reasonfor the poor access to andhighcost of finance for small andmicro firms appears to be poor financial intermediation rather than low domestic savings or bad internationalfinance. Domestic saving as a share of GDP climbed up fromjust 6 percent inthe 1990sto 16.5percent in2000s (Table 10).In2006 Zambia's domestic saving as a share of GDP surpassedthe corresponding average for SSA. Foreign direct investment and aid were higherthanthe averages for SSA and LICs both inthe 1990sand 2000s (Table 10). 57. Financial intermediation i s limited bothby the small size of the bankingsector, its volatility, and an inadequate supporting financial infrastructure (e.g. public registries). At less than 5 percent, the percentage of people with a bank account i s much lower than that inother African countries, suggesting that the depthof Zambia's financial sector is very low (Figure 25). Furthermore, according FinScope data in2005 two thirds of Zambians were not served by formal or informal financial institutions and only 5 percent of adults and 8 percent of business owners usedmicrofinance (FinTrust,2007). 32 Table 10: Investmentandsaving indicators 1991- 2000- Latest year* 1999 2005 Gross capital formation (% of GDP) Zambia 13.7 23.8 25.8 (2006) Sub-Saharan Africa 17.8 18.9 19.2 (2006) Low income countries 21.4 25.0 28.8 (2005) OECD 21.3 20.5 20.3 (2004) Gross capital formation (annual % growth) Zambia 11.1 7.9 13.1 (2006) Sub-Saharan Africa 3.1 7.1 13.0 (2006) L o w income countries 5.6 8.4 15.9 (2005) OECD 2.1 2.6 5.0 (2004) Aid (% of gross capital formation) Zambia 193.4 72.8 50.3 (2005) Sub-Saharan Africa 30.0 25.6 26.7 (2005) L o w income countries 16.2 10.5 9.9 (2005) OECD na na na FDI,net inflows (% of grosscapital formation) Zambia 27.6 14.3 13.8 (2005) Sub-Saharan Africa 8.4 16.1 13.9 (2005) L o w income countries 5.0 5.8 5.1 (2005) OECD 6.0 12.2 7.1 (2004) FDI,net inflows (% of GDP) Zambia 4.0 3.3 3.6 (2005) Sub-Saharan Africa 1.5 3.0 2.7 (2005) L o w income countries 1.1 1.4 1.5 (2005) OECD 1.3 2.4 1.9 (2005) Gross domestic savings (% of GDP) Zambia 6.0 16.5 18.1 (2006) Sub-Saharan Africa 16.2 18.0 16.2 (2006) L o w income countries 18.6 22.5 25.1 (2005) OECD 21.6 19.8 19.4 (2004) Source:World Bank, DDP. *Note: Latest year shown inparenthesis. 33 Figure 25: Percentage of populationwith a bank deposit account I Source: Martinez (2006). 58. Another problem is lack o f public registries or private bureaus with information onborrowers' creditworthiness, Despite a good legal framework that protects the rights o f borrowers and lenders, inpractice Zambia's recovery rate on bank loans i sjust 22 cents -higher than inLICs but significantly lower than inSouth Africa and HICs. 59. Theproblemof poor financial intermediation is especially severe inrural areas. Before liberalizationthere were government-runinstitutions providing agricultural credits, but in 1997 this ended and was combined with a period of prohibitively high interest rates. Informal borrowing i s normally not an option as the rate can go up to several hundredpercent annually. 60. However, there i s evidence that financial intermediation has improved lately. The number of commercialbank brancheshas grown rapidly inthe period 2006-07 (Figure 26). Some microfinance institutions operated by NGOs and outgrower schemes (mainly incotton, paprikaandtobacco) havebeensuccessful inproviding credit to fanners, but the micro finance and outgrower credit channel remains limited. Figure 26: Number of commercial banks'branches 185 180 175 170 165 169 155 150 145 140 2W1 2002 2W3 2004 2005 2M)6 2W7 Source: Bank of Zambia (2007). 34 61. A question that needsto beposedbeforemakinga conclusion about the cost of or access to capital as a binding constraint to growth is why people did not use financial services. Among the top reasons for not having bank accounts was low and irregular income, rather than physical access or highfixed costs (Figure 27). This indicates that constraints to income growth rather than access to saving institutions are crucial for shared growth at present. A similar picture emerges when one looks at the reasons for not having micro credit. Figure 27: Reasons for not having a bank account We cannot populatetireaccessfrontier usingdata onreasons peopledo nothavea bankaccountalone,althoughit canprovide an insightintocriticalaccessbarnen Reasonno bank account 1doilothawmney CQ JU?istod bank Ido T lyIthew* rq&crym ~ * t u r n r n Us60lbdmW rUMn)-Q*Lam$mF + 9-z:- ipl(aou*rtmb Vmm- 'bWNi8aNbarY W W d 0% 20% 4Q% 80% Source:FinTrust(2007). FinScopedata from 2005. Figure 28: Reasons for not having a micro finance institution (MFI)loan Introducingthe accessfrontier for a loanfmnia MFIin m i a i.mP Source: FinTrust(2007). FinScopedata from 2005. 35 62. Figure 28 shows that only 1percent of business owners have micro credit loans. O f the ones without micro credit only 17 percent say they have access to credit but do not need it because they have own funds. The majority of business owners with no access to micro credit cannot afford it (63 percent) or are too poor (23 percent). However, a significant share of business owners is unaware of micro-credit opportunities (32 percent). 63. Insummary, access to creditmay stillbe aconstraint for some groups of poorbut it is difficult to conclude whether it is abindingconstraint for shared growth. Signsof improvement can be found inboth "objective" and perception data, indicating that the financial infrastructure seems to respond to new economic opportunities. Positive effects of access to credit are mainly seen when credits are combined with access to other types of inputs, and output markets, such as inoutgrower schemes. This suggests the importance of coordination when supplyingproducers with key services. We continue with an evaluationof social returns as a potential reason for low returns to investment, which inturn limit shared growth (Figure 15). D OLOW SOCIAL RETURNSIMPLY LOW INCOME GROWTH? Mining, agriculture andnatural endowments 64. Relativeto other Southern African countries (Table 1l), is well endowed Zambia with natural resources, includingarable land, favorable climate and pattern of rainfall over large sections o f the country, ample water, forestry and miningresources (World Bank2003). These endowments offer opportunities for income growth from a diverse set of economic activities includingmining, agriculture, forestry, fisheries, tourism and hydro energy, but also food processing, and manufacturing related to Zambia's endowments of mining, forestry and others. Table 11:Country comparisonof some naturalresource endowments AnnualRenewable Forest Coverage Nationally Protected Water Resources (1000ha) 1995 Areas (1000ha) 1999 (km3), 2000 Zambia 80 31355 6366 Zimbabwe 14 8626 3071 Malawi 18 3213 1059 Mozambique 100 16834 4779 SouthAfrica 45 7204 6619 Namibia 6 12374 10616 Tanzania 80 1224 13817 Source: African Development Indicators (2000). 65. Miningis Zambia's main source of export revenues, and inthe face of rising demand for natural resources from emerging Asia and elsewhere, will continue to be a key source of growth o f export revenue. Dominated by copper, mining inrecent years has 36 diversified into other metals andminerals. Possibilities remain open for Zambia to capitalize on its deposits of nickel, zinc, lead, coal, emeralds, gold, silver and uranium. 66. Another, natural resource that remains largely untapped i s land. Only 10percent of total land and40 percent of arable land is currently cultivated. However, as shown in Figure 16, average returns to agriculturehave remained low despite possibilities for output expansion through both land expansion and intensification of non-commercial farming. 67. Low productivity inrural areas i s partly a result of many years o f development neglect, when Zambia's agriculture was mainly seen as a source of cheap foods sold in urbanarea." Duringthe 90s major reforms were carried out within Zambia's agriculture sector, including market determined prices, reduction of producer subsidies, and other measures. The nature of these reforms impliedthat mostly large scale firms and firms with market accessbenefited, while many small-scale farmers struggledto access markets, inputs and information. The government however continuedto intervene in agricultural markets. ** These interventions created misallocation of resources (Jayne et al. 2007) andraised indirect costs. The indirect costs to agriculture, which include direct taxes and indirect costs through macroeconomic and other distortions, have diminished substantially inmost countries, except inZambia (Figure 29). These costs created by the economic structure o f Zambia are highcompared to other African countries, andhave even increasedsince the beginningof the 80s. 68. The abundance of fertile landand the possibility to remove these efficiency limitingdistortions, implies that there is a potentialto expand farm output inZambia. According to a World Bank study (2007c), Zambia's commercial agricultural expansion will be mainly along the extensive margins as their yields are already comparable to those indeveloped countries. Formediumand small scale farms however there willbe opportunities to increase output both at the intensive and extensive margins. 69. Zambia has untapped social returns inhydro power and tourism. Based on statistical analysis, Mattoo andPayton (2007) find that tourist flows are much lower than expected, given Zambia's natural endowments. The main reasons for the lower than expected tourist flows are the condition of road infrastructure and the cost o f doing business as an indirect measure o f the level o f "tourism establishment" serving and accommodating tourists. Their estimates suggest that if the road infrastructure andthe business climate were as inSouth Africa Zambia's tourist flows would be 51 percent higher than existing levels. l7Revenuesfrom the copper mines were formerly usedto finance government interventionsinagricultural markets. These included guaranteedprices for maize and subsidized inputs and credits. The incomes of the rural households were further ensured by remittances from workers in, at that time, public employment in the mines. Hence, the agricultural sector was mainly encouraged through public and private transfers and there were few incentives to diversify due to the lack of investment ingeneral productivity enhancing infrastructure. l8Section 3.4.3 discussesthis issue. 37 Figure 29: Indirectcost to agriculture Kenya Clgands M a degashser Camer o o m Sudan. Ehsna Wigpm-i a Ethiopia Banzaniet Zambia Cbte d'dvo ire --am -60 -49 -233 a 20 40 Norninml cat= af assistance. % Source: World Bank (2008a). Note: The nominal rate of assistance, the "net taxation", takes into account distortions onboththe output and input side. It adjusts for direct output subsidies, the difference inoutput price at the farm gate and at the border, input subsidies, differences betweenthe domestic and the internationalprices of inputs, distortions inthe market for foreign currency, and others. Geography 70. Zambia i s a landlocked country which makes it potentially harder to reach export marketsandrealize economies of scale, as well as access cheap import^.'^ Onthe other hand, being landlockedmay also serve as an import tax protectingdomestic import- competingfirms. 71. One concern posed by Zambia's landlocked position i s its ability to export bulky low-value products (e.g. some agricultural products). The extra cost o f getting such products to the coast needs to be compensated b y more efficient production compared to coastal countries. Indeed, increasingly more of Zambia's agricultural products are exported by air - a shift that requireda focus on highvalue and low weight and volume products, but also improved access to air transport. However, Zambia's access to air transport is still well below the access level expected given its GDPper capita level (Figure 30). There is evidence that as some firms have suspendedhorticultural exports to Europe, the frequency of flights has decreasedraisingair-freight costs for remaining firms. The average GDPper capita inlandlocked SSA countries was about half of the GDP per capita for the coastal SSA countries in 1999-2005, and Zambia's GDP per capita was only 60 percent of the average of the landlockedcountries during the same period. However, interms of GDPper capita growth during the same period, the order was reversed implying that Zambia is catching up with other landlocked countries. 38 Figure 30: Access to air transport - - OIUP p0r capita (logs) Fittedvalues -95% CI Source: Mattoo andPayton (2007). 72. When it comes to regional trade, Zambia's landlockedposition has proved to be an advantage as Zambia borders eight other countries andis the beginning, destination or transit country for five of the 18 major transit corridors inSub-SaharanAfrica. Regional and internationaltransit infrastructure costs are relatively low inZambia. All five transit routes have unitroad transport costs that are below the regional average. Low transport costs on the maintransit corridors have facilitated regional trade and shifted Zambia's exports to SADC and COMESA countries and away from the EUand the UnitedStates (World Bank, 2007b). Still Zambian exporters face significantly higher obstacles to trade thanexporters inother countries interms of time to import and export goods.20 Infrastructure 73. Zambia's status as a landlocked country cannot be viewed as a major problem, but highdomestic transport costs are a constraint to growth. The domestic transport cost for one ton per one kilometer was US$0.07 inZambia, which was higher than incountries like South Africa (US$0.02), Malawi (US$0.07) andEthiopia (US$O.O4-0.06) (World Bank,2007b). The highprice of transport seems to derive frompoor domestic road infrastructure outside of the main transit corridors, highfuel costs (Table 12) due to high taxes on diesel, and highcost of tires.21 Zambia RSA Tanzania Kenya Zimbabwe Pre-budget Post-budget Reduction in Excise Tax 0.86 0.73 0.32 0.65 0.59 0.55 74. Poor quality andunevenly distributeddomestic roadnetwork i s a constraint to growth of rural economic activities, including commercial agriculture, new mining activities, tourism, among others. The domestic road quality has improved inrecent years Source: DoingBusinessIndicators,World Bank. 21Diesel and tires contribute over 50 percentof the costs of transportation(World Bank 2003). 39 and the percentage of roads that are paved i s higher inZambia than the average for SSA. However, only a few districts inthe Lusaka province have roads with fairly good quality. The railroad has not been an alternative so far but a recent privatization of Zambia Railways may change that inthe future. 75. Poor market access is a constraint to the farmers' ability to sell their output and their ability to access inputs needed for an efficient production as large partsof Zambia's rural area are sparsely populated with long distances between villages and low road quality. World Bank (2007a) reports that in2002/03 half of the rural households were more than 9 kilometers from the nearest food market, and over 25 kilometers from the nearest agricultural input marketsfor fertilizers and seeds. Poor market access has a negative effect on returns to farming.More remote households have less landunder cultivation, lower returns per household member and lower returns to land.Alwang and Siege1(2003) show that net returns are roughly 10percent lower for remote households. 76. Cotton production, a success of the post-reformperiod for smallholders, offers an example of the disadvantages faced by remote households. The possibility o f taking on cotton outgrower schemes i s only available to farmers who live inareas that are situated close to cotton ginners. Outgrower agents only operate insuch areas because traveling over wider areas with low populationdensity i s not cost effective. As a result, cotton i s not an option for smallholders inmost parts of the country. Another example i s the government fertilizer subsidyprogramthat does not reach farmers inremote areas of the country.22 77. Inadequate infrastructure and poor quality and expensive basic services are perceived as a major constraint by business owners. Connectivity services are crucial for economic integration within the domestic economy and with the rest of the world. A recent study of 42 developing countries, based on Investment Climate Surveys, found that Zambianfirms havethe second-highest share of "indirect" costs, most of which are attributable to services-related inputs used inproduction - energy, transport, telecom, water, insurance, marketing,travel, independentprofessionals and accounting (Figure 31).InZambia the share of indirect costs i s on average 22 percent of gross value added, which i s twice the share of labor costs. The study found that the highlevel o f indirect costs attributable to the highprices of services i s likely to have underminedthe competitiveness of Zambian firms inexport markets and therefore slowed downjob creation. A specific example of high-cost connectivity services i s the mobile phone business which sells services at muchhigher cost inZambia compared to other countries (Figure32). '*See section 3.4.3 for further discussion. 40 Figure 31: Cost structure: firm-level averagesby country Mozambique Zambia Eritrea Tanzania Kenya Ethiopia Nigeria Uganda Bolivia Morocco India Senegal Bangladesh Nicaragua China 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 I indirect labor capital IinputsI Source: Eifert, Gelb, and Ramachandran(2005). 78. While electricity costs are not highcompared to countries within and outside the region (Table 13), there are big variations in quality of and access to electricity supply in different parts of the country. Insome areas there i s no supply of electricity while firms around Ndola and Kabwe suffer frequent power outages. Only about 20 percent of the country's populationhas access to electricity. 79. The problem is not limited supply of electricity. Zambia has installedcapacity which i s well inexcess of current domestic demand and exports electricity within the region. The inconsistency between excess supply and poor access i s due to the dominance of the miningindustry inthe energy sector. Such monopsony power has discouraged the expansion of electricity infrastructureto other productive sectors. Poor access to basic services is more acute for small firms which have to wait longer than large firms to gain access to electricity, water and phone connections (Figure 39). 80. Informal businesses are significantly more disadvantaged by the lack o f affordable basic services as they lack resources to supply own water and generate own electricity. Much of the infrastructure ininformal urbanareas i s outdated, poorly maintainedand overcrowded. While access to markets and quality o f housing are not constraints for urbanresidents, market infrastructureininformal settlements i s inadequate. 81. Itis also extremely expensive to buildwarehouses -anessential type of infrastructure inurbanareas. The cost of obtaining the necessary licenses and permits for a warehouse construction i s 1766 percent of per capita income compared to 1048 percent inSSA, 996 percent inLICs and72 percent indeveloped countries. 41 Figure32: Mobilerates per minute Mauritius Zimbabwe Mozambique Namibia Swaziland Tanzania South Africa Malawi Botswana 32 Za.?o,s 4 Source: Mattoo andPayton(2007). Table 13: Comparativeelectricity costs (US$/kwh, 2001) Zambia Kenya Uganda Madagascar India Bangladesh Sri Nepal South Lanka Africa 0.03 0.05- 0.04 0.06 0.07 0.06 0.06 0.08 0.03- 0.07 0.07 Source:World Bank (2003). 82. An extension of the rural roadnetwork and other public utilities inrural and urban, informal settlement areas is anecessary complement to all other investments and reforms to foster shared growth inZambia. However, these investments need to be managed carefully and positioned strategically to foster positive externalities for as many households and industries as possible.23 Humancapital labor is not a constraint to business growth. Worker skills and education are ranked lo* 83. Zambia's growth diagnostics study (World Bank 2007b) concludedthat skilled ina list of 17obstacles to business operations inZambia's ICA conducted in2003 (World Bank, 2004). This suggests that, for an average firm, there i s no mismatch between skills demanded by companies and skills workers provide. Moreover, a decline inemployment inmedium-cost andhigh-costurbanareas inrecent years signals limited demand for formal employment (Figure 18). Therefore, inthe near term supply of different types o f efficient labor i s unlikely to constrain formal job creation. 84. Another way to analyze excess demand or excess supply o f labor i s to look at international migrationpatterns. Brain drain canhave a hugenegative effect on growth in a country with limited human capital and limited education infrastructure and resources. Overall migration rates are generally lower inAfrica relative to the rest o f the world, and 23See further the discussion on coordination failures inSection 3.5. 42 Zambia's rates are even low by African standards. Only 0.1 percent of those with primary school education migrated from Zambia in2000, compared to regional averages o f 0.3 percent inSouthern Africa, 2.8 percent inNorthern Africa and 0.5 percent in South- Central Asia. For those with secondary education the migration rate is 0.3 percent which i s also very low compared to other developing countries. At the tertiary level the rate increases to 10percent but it is low compared to other countries and low given its low level o f tertiary educated (Figure 33). Figure33: SkilledmigrationandhumancapitalinAfrica, 2000 70 , 000 0.05 0.10 0.75 0.20 0.25 0.30 tetiay educated (migrants and non-migrants) per 100papulationof source country Source:Mattoo and Payton (2007). 85. Hence, at the aggregate level there are few signs of major imbalances between demand and supply of skilledlabor. However, this does not meanthat there are no imbalances for specific skills, such as for managers, technical engineers, etc. Moreover, given the time lag of changes inthe level, quality or focus of education andthe actual supply of skills, there mustbe a strategy on how to meet potential skill constraints inthe future. 86. We conclude that within social returns, infrastructure and basic services are bindingconstraints for growth as well as inclusivegrowth. The discussion on infrastructure as a constraint will be picked up also when discussing coordination failures insection 3.5. Wenow turnto private appropriability issuesinsearchof government and market failures (Figure 15). ARE GOVERNMENTFAILURESA CONSTRAINTTO SHAREDGROWTH? Macroeconomic environment 87. In2003 macroeconomic instability rankedsecond inthelist ofbusiness constraints publishedinthe I C A (World Bank, 2004a), with 74 percent o f the firms 43 naming it as a major or severe obstacle to businessoperations inZambia. The macroeconomic situation has improved substantially since 2003. Inflationfell downto single digits in2006, and despite rising energy prices the terms of trade improved inthe period 2003-2007 due to rising commodity prices, especially prices of copper. The overall government budgetdeficit (including grants) fell as a share o f GNP from 13.4% in2003 to 7.6% in2006.24 88. The main macroeconomic concern surrounds the appreciation of the Kwacha. The path of the exchange rate is illustrated inFigure34, and was fairly stable inthe range 4400-4800 K/$from 2002 to early 2005. Appreciation commenced in2005 and there was a rapid rise inNovember 2005, taking the rate to 3600W$. InMay 2006 the rate fell below 3000K/$, and has since returnedto around 3400W$. The real exchange rate has also appreciated substantially since 2004. After staying relatively flat between 1998 and 2004, ithas appreciated by 25% in2004-05 andby 31% in2005-06. 89. Several factors have contributed to the appreciation. These include debt relief,25 scaling up of aid, foreign direct investment flows into mining, strong export performance andtight monetary policy. Debt relief has reduced Zambia's external debt from over $7 billion to less than $520 million which translates to annual foreign exchange savings of $50 million in2007, $80 million in2008, andpeakingat around $130 million in2010. Scaling up mightdouble aid to Zambia over the next 10 years and result inaid-to-GDP ratio o f 14 percent. The increase incopper revenues dwarfs debt relief and increases in aid, but it i s unclear how much of this additional copper revenue will be a net supply of foreign exchange to the Zambian domestic economy. Figure34: The developmentof the Kwacha exchange rate 150 T 140 130 120 110 100 90 80 70 Source: World Bank, DDP. 24The share of government revenue inGDPhas fallen slightly from 17.5% in2005 to a projected 16.5% of GDP in2006. Thesebudgetary issuesare partly due to the appreciation of the Kwacha, reducing the local currency value of customs duties and trade taxes, and also aid inflows. 25 Zambia passed HIPC completion in2005. 44 90. The kwacha appreciationmay already have had severe impacts on some sectors. There have been some highprofile factory closures, and the impact has been estimated by several industrysources. Recent data on agricultural exports show sharp reductions in exports of major commodities since 2005, likely linked to the appreciation of the kwacha (Figure 35). Reduction inagricultural employment of this magnitudewill translate directly into rising poverty rates as farm workers laid off from employment inlarge agricultural firms will lose their wage income and rely mainly on subsistence agriculture. Figure 35: Agricultural exports 2000-2007 I 1 20,000 A\ II I /' 15,000 I 2OW 2001 2002 2003 20M 2005 2W6 2007 Source: DatafromBank of Zambia. 91. Appreciation has fiscal effects which typically include a decline inexternal debt payments and ingovernment expenditure on imported goods, assuming unchanged level o f government imports, and on the revenue side a decline inexternal public sector grants and loans, and revenue from trade taxes assuming imports remainunchanged. Weeks (2008) estimates the total effect of the appreciation on import taxes, ODA, and debt service for Zambia to be -0.1 percentages of GDP in2005, -1.7 in2006, and -1.1 in2007. Itis difficult to come up with precise numbers for the fiscal effects of exchange rate changes, but they point to another reason for careful management of the exchange rate and its effects. 92. Increased integration into the global economy i s a priority for Zambia - a small economy for which the links to internationalmarkets allow it to realize economies of scale, access state-of-the-art technologies and improve the quality and variety of its exports. However, the real appreciation of the kwacha threatens the progress made in diversifying the export base andjumpstarting non-traditional exports. And although exports and imports have started to increase inabsolute terms since the early 2000s, in relative terms they are still much below historical data from earlier decades. Indeed, the data suggest that Zambia has not been able to scale up its export operations sufficiently to reverse the negative trend (Figure 36). 93. An appreciated currency will makeit more difficult for domestic enterprises to compete with cheap imports and domestic exporters to sell their products ininternational 45 markets. For Zambia to stay competitive and sustain its growth momentum it will be critical to improve productivity - including the productivity of its labor force, and to lower indirect production costs related to basic services. However, carefully crafted monetary and fiscal policies responding to the real appreciation pressures will also be critical. Figure 36: Importandexport shares, Zambia 1960-2007 70 60 50 40 30 20 10 0 96' \=Q `969`9 2`\9 `9 q' 2' 92` `9q9 0` 0' 0' 9 `9 `!`9 8`90' `9 `9 49 `9 9` 9' 9' 92 99` 6 8 8 q? q? q? i --cExports of goods and sefwces (% of GDP) +Imports of goods and senices (% of GDP), Source: Data from the World Bank (DDP). Tax code 94. Taxes, although on the high side, are not excessively highnor are they out of line compared to other countries (Figure 37).26The tax base however i s narrow and large firms, particularly inthe financial service industry, face much higher tax rates than small andmedium sized companies. A number of problems with the tax system are well documented: the highfrequency of unexplainedor unjustified changes intax policy, corrupt practices, and a tax administrationperceivedas arbitrary and punitive. 95. The tax system has created disincentives for enteringthe formal sector andhas disadvantaged small enterprises. The sales level for VAT registration for instance i s very high.Itstands at $40000at atime when per capita income isjust $375. This highlevel discourages even medium-sizedfirms from entering the formal sector. Specific taxes may also impede the growth of specific sectors. A 2004 study finds that VAT exempt status hurtsfarmers becausetheir effective burdenrises sharplywithout the ability to reclaim VAT on inputs.Another example is the 3 percent turnover tax levied on small agricultural producers, becauseits threshold of 20 percent profit rate is unrealistically highfor small-scale farmers. 26See World Bank (2007b). 46 Figure 37: Country comparisonof global ranking on tax level Paying Taxes -Global Rank 96. The tax systemis also problematic inthe sense that it distorts competition between and/or within some sectors. The tourism sector i s one example. This sector is seldom part o f tax incentive schemes provided to many other non-traditionalmerchandise exports. Moreover, certain tourism service providers - such as Sun International, a foreign owned firm-have resources to negotiate special deals puttingthem inan advantageous position compared to small, domestically-ownedfirms.27 97. The direct fiscal impact of the recent boom inmininghas been negligible because of the current tax exempt status of mining.This arrangement was a result of negotiations that took place under unfavorable circumstances when the sector was at the bottom of the crises. The government i s currently renegotiatingthis arrangement with the mining companies. 98. Zambia has one of the most open trade regimes inAfrica with a rating of 2 ("open") on the IMF's restrictiveness index ranging from 0 to 10 ("most restrictive") and according to the Africa Competitiveness Report, which measures openness based on levels of import and export restrictions, licensing requirements and exchange controls. Average tariff rates were also lower inZambia than inother SSA countries. The simple average MFNtariff inZambia i s 13.4 percent with a coefficient of variation of 0.7, indicatingmodest dispersion of tariff rates. 99. Ad valorem tariffs comprise four bands: 0, 5, 15, and 25 percent (Table 14). Specific rates apply to a few items. The industriesfacing the highest duties are fishing and light manufacturing such as wood products, manufactured foods, beverages and tobacco, textiles and leather. Hightariffs on agricultural and processed foods raise the prices of food with adverse consequencesfor poverty alleviation. 27See Mattoo andPayton (2007). 47 Table 14: Tariffs onkey categories of goods, 2003 (percent) Tariff bands Share of Share of Share of tariff lines imports customs revenue Raw materials 0-5 21 30 0 Capital goods 0-5 14 24 15 Intermediate goods 15 33 26 36 Finishedgoods 25 32 21 48 Source: CSO, tariff data submittedto UNCTAD. 100. Zambianexporters also have access to inputs at world prices through the duty drawback systemwhich allows for rebates on tariffs paid on imports used inthe production of exports. The main problems with the duty drawback system include the requirementof detailed and comprehensive input-output coefficients andthe longtime taken to get the rebates. There is also the concern that the system benefits mainly large export-oriented companies, and not small and medium-sized firms 101. Although market access does not constrain Zambian exports, the preferences enjoyedby export-oriented firms will decrease as tradingpartners reduce their MFN tariffs as part of the Doha negotiations and offer similar preferences to other developing countries. It is therefore important for Zambia to focus on improving the non-tariff aspects of these arrangements (trade facilitation, rules of origin, aid for trade), strengtheningbehindthe border policies (administrative and regulatory environment, macroeconomic policies), and removing supply-side constraints (physical andhuman infrastructure) to increase its competitive strength. It i s also important for Zambia to actively participateinregional and global trade negotiations to ensure that its longer-term trade interests are adequately reflected inthese agreements. Regulatory uncertainty, government interventions and other administrative procedures 102. Regulatory uncertainty is cited as the fourth most constraining factor inthe general survey of business owners (World Bank, 2004a). Most firms (70 percent) think that officials' interpretationof regulations affecting their businesses is inconsistent and unpredictable.28The problem i s especially acute inagriculture and i s related to the contentious government's Fertilizer Subsidy Program. The government has issued confusing and haphazard policies, repeatedly promising to withdraw from the fertilizer marketbutthenre-enteringthe market underpopular pressure to assist the ruralpoor. 103. The current Fertilizer Subsidy Programhas many problems and is one example o f government failure. The government issues contracts for fertilizer purchases at the last minute and allows little time for successful bidders to import and deliver fertilizer. Because of this, importers are forced to use higher cost import routes, which has nearly 28The numberof proceduresfor startingabusiness,dealing with licensesandpermits,registering property, procedures for filing alawsuit onpaymentdisputes, as well as the costs of starting and closing a business andregisteringproperty are all below regional andLIC averages. 48 doubled the price the government pays (Mwape, 2004). Moreover, the program i s not well targeted and manyrecipients are not farmers (only 20 percent of small farmers use fertilizers) but traders, who resell fertilizers at large markups, to the well-connected and high-income groups located close to tarmac roads and district centers (World Bank, 2008a). Thus, the system opens the door to rent-seeking and corruption, distorts the market, depresses the supplyof fertilizer on the commercialmarket, and crowds out private operators. The annual uncertainty about the timing and level of government purchases i s particularly damaging. 104. According to World Bank (2008a), 5 percent of Zambia's national budget goes to agriculture of which more thanhalf i s earmarked for the Fertilizer Subsidy Programs (37 percent) and maize marketing (15 percent). Only 3 percent of the agricultural budget goes to muchneeded irrigation and other rural infrastructure, and 11percent to operating costs including agriculture extension and research. 105. Zambia fares poorly on another procedural itemthat hampers growth, especially shared growth. Firingcost are unusually highinZambia, as measuredby the weeks o f wages employers are requiredto pay - 178 weeks inZambia vs. 71 weeks inSSA, 65 weeks inLICs and 72 weeks indeveloped countries. This unusuallyhighfiring costs, coupled with highHIV/AIDS prevalence implies that the Zambian labor market i s a lot less flexible than suggestedby the rigidity of employment index, which measures difficulty of hiring and the rigidity o f hours o f employment. The regulation clearly discourages companies from hiringemployees and creates disincentives for companies to move from the informal to the formal sector. LandRights 106. The landtenure system inZambia i s dual and the majority of landi s held under customary landarrangements with limitedtransfer possibilities. Only 6 to 15 percent2' o f total landallows for ownership rights and registration underthe so called statutory tenure.30The system governing the rights over this land i s administered by the state under the Englishstatutory law, and distributed in99-year leases. 107. Despite this dominance of customary landholding, the land system i s not perceived as a bindingconstraint to shared growth inthe short to mediumterm by most stakeholders inZambia. First of all, land i s abundant inZambia. The area o f available cropland per 1,000 people is more than twice that available inSSA and other parts of the world (Table 15). Only 40 percent o f arable landi s used suggestingeconomic potential that still i s not fully exploited. 29Statistics vary by source. 30While individuals can use land and pass it on to family members through inheritance, no exclusive rights can be claimed by individual users, nor can they sell or mortgage the land. There are no formal documents of land ownership or use, and no taxes paid on this land. Unlike customary lands which are administered by traditional leaders, statutory lands are administered by the central government, and are mortgageable and subject to taxation. They are concentrated inand near Zambia's cities, along the railway line between Livingstone and the Congo border, inthe miningareas o f the Copperbelt, and incertainproductive farming areas. These areas are the most valuable and productive land inZambia. 49 108. Access to land i s not abindingconstraint for small scale farming, as of today. Most smallholders who demandmore landfrom their village chief do get land. Lack of landwas mentionedby only 4 percent of the respondents to the survey on self-assessed reasons for poverty (see section 3.2). Despite the small plot sizes under cultivation in small scale farms (Table 5), a typical household does not have the capability of cultivating more land (Alwang and Siegel, 2003), suggesting there are other constraining factors to farm output and income growth. Jorgensen and Loudjeva (2005) also conclude inthat a landreformshould notbe apriority untilcomplementary reforms haveensured improved road network, access to fertilizers at competitive prices, and functioning extension services. 109. The customary landtenure systemdoes not appear to limit the use of landas collateral. Ina survey by Smith (2001) the majority of those expressing a desire for formal landtitles did so becausethey wanted to avoid dispossession (78 percent), protect fixed investments (55 percent), and ensure transfer to heirs (50 percent). Although multiple answers were permitted, only 7 percent of respondents indicatedthat they wanted titles inorder to use land as collateral for credit. 110. However, from a longer runperspective, the landtenure systeminZambia may be perceived as a bindingconstraint to growth due to the risk it creates for the future returns on inve~tments.~~ Itmay also become anobstacle to expansion of small farms into commercial operations and the formation of more efficient farms that realize economies of scale.32Studies show that formal land titling has only been pursued by farmers already with links to commercial agriculture. Many argue that the real constraint i s lack of "serviced and accessible land" implyingthat rural infrastructure services are the binding constraint to farm operations. There i s also a backlog of land registration, which indicates that not lack of landbut inefficiencies inthe current administrative system i s a bottleneck to the commercialization of Zambia's agriculture. Governance33 111. According to the World Bank's governance indicators Zambia scores highon political stability, but only fair on voice and accountability, regulatory quality, and rule of law, while on control of corruption and government efectiveness Zambia scores very low (Figure38). Countries like South Africa, Botswana, Malawi, Tanzania, Uganda and Rwanda are all ranked higher than Zambia interms of control of corruption and only Malawi i s worse thanZambia on government effectiveness. Corruption penalizes disproportionately the poor inZambia, according to surveys conducted in2003 by the University of Zambia. Citizens inthe lowest income deciles have to pay bribes that 31 More than half of the respondents to the survey by Smith (2001) expresseddesire for formal land titles becausethey wanted to protect fixed investments. 32The current land tenure system may potentially lead to land concentration, and therefore increases in '' inequality, as those with easier access to formal landrights may overtake the more productive areas. However, others argue that a formal land titling system would leadto increased land concentration if the oor are forced to sell their land leaving them without productive assets. Inthis section we look at overall measuresof governance butother sections discuss sub-sections of governance, such as distortionary policies. 50 represent a higher share of their income than their counterparts inthe middle and high income brackets. 112. Interestingly though, ina cross-country comparison looking at the correlation between these indicators and GDP per capita, Zambia has, given its GDP per capita level, a higher than estimated score on voice and accountability, rule of law, regulatory quality, and political stability, and close to its estimated score on control of corruption and government effectivenes~.~~Moreover, the fact that government effectiveness has been improving since the late 90s along with improvements inthe economic conditions, i s a sign of effective reforms. These improvements must continue especially interms of budgetexecution, transparency and accountability (World Bank, 2004b). Figure38: World Bank Governance Indicators for Zambia Comparison between 2006, 2002, 1998 (top-bottom order) Voice and Rccountability I P o l i t i c a l Stability Government Effectiveness Regulatory Quality Rule of Lau Control o f Corruption e 25 50 75 1 Country's Percentile Rank (0-100) Source: Kaufmann, et a1(2007). 113. Despite improvements inrecent years, many governance indicators are still low implyingthat it would be difficult to address effectively the bindingconstraints to inclusive growth. This is especially evident when studyingthe coordination failures in Zambia discussed inthe next section on market failures. Inthis section we conclude that governance failures are indeed constraininginclusive growth. AREMARKET FAILURES REASONS FORLOWRETURNS? 114. Firmsneed services inorder to innovate, market their products successfully, and makea profit. These services link the supply chain between producers and consumers, andrequire simultaneous, large-scale investments invarious sectors of the economy 34Presentationby Ricardo Hausmann at Harvard, Cambridge, November 19,2007. 51 (Rodrik, 2004). These services include the provision of infrastructure and institutions linking the different steps inthe productionchain, access to markets, andbasic services such as irrigation, electricity, water. Other services not previously mentionedinclude marketing, research andproduct quality information. 115. However, the incentive to establish these kindsof services i s limited for an individual entrepreneur due to small market size inthe case of private services, and due to non-exclusiveness inthe case of public services. Hence, coordination externalities i s not an umbrella concept for arguing that government should be involved inall aspects of economic life, but should be described as the failure of the market to respond to potential investors' demands for a diverse set of services. This potential problem is especially common insparsely populated countries such as Zambia, and implies that the government needs to focus their public investments on pockets of growth rather than country-wide investments, and may need to make initial investment incertain private services that would have been provided by the market if economic activity had reached a certain level. 116. Arnold et al. (2006) analyze the relationship between the performance o f local service providers andthe productivity of firms indownstream industries,using Investment Climate Surveys with panel data o f 1,185 firms in 10 SSA countries. After controlling for systematic differences at the country, industryand firm level, the study for example predicts that Zambian firms would be 13 percent more productive if they enjoy the same access to telecommunications as South African firms, and 6 percent more productive inthe case of access to banking services. 117. Rural areas are at a disadvantage to urbanareas when it comes to very basic facilities for a functioning market (Figure 39). The limited resources of the government and the sparselypopulatedrural areas will make it impossibleto substantially improve these basic services inall areas. However, a cluster strategy to provide basic services to multiple industries and create positive externalities i s worth exploring. We present below some cases of coordination failures and specific industry cases where pockets of private sector growth have been achieved when supported by crucial services. Figure39: The number of householdswith access to facilities within five kilometers Food Input Post Office Public Public Internet Market Market Transport Phone Cafe Source: CSO (forthcoming). 52 118. Inagriculture, smallholder farmers lack accessto andinformationonproper use of fertilizer, chemicals, irrigation, seed selection which leads to low yields per hectare and low productivity. Although higher than the yields inSSA, average cereal yields in Zambia are much lower compared to those inthe rest of the world and yields inthe early 1980s (Table 15). Food security remains an issue as the volatility indomestic cereal production i s much higher thanthe average for SSA and the world. These outcomes are not surprising when one considers the use of irrigation and fertilizers inZambia. The percent of irrigated cropland inZambia inthe period 1999-2001was less than one percent, compared to nearly 4 percent inSSA and 18 percent inthe rest o f the world. In 1999, the average annual fertilizer use was not only below the use inSSA, but was nearly 10times less thanthe use of fertilizers inthe rest of the world. Table 15: Agricultural production and yields Zambia SSA World Cereal, 1999-2001 Average crop yield (kg per hectare) 1437 1221 3096 Percentagechange since 1979-81 -14% 9% 41% Agricultural land Hectares of cropland per 1,000 population, 1999 518 274 251 Percent of cropland that is irrigated 0.9% 3.8% 18.3% Agricultural inputs Average annual fertilizer use, 1999 (kg per ha) 10 12 94 Foodsecurity Volatility indomestic cereal production, 1992-2001 24.6% 6.5% 3.5% (average percent variation from mean) Net cereal imports and food aid as a percent of total 17.7% 13.5% n.a. Consumption, 1998-2000 Source: World Resource Institute (http:Nearthtrends.wri.org). 119. The market access problems due to poor roadnetworks faced by farmers are augmented by the fact that only few intermediaries serve as channels betweenmultiple smallholder farmers and/or delivery locations to food processing firms. As a result, agro- processing firms operate at low rates of capacity utilization. This makes achieving economies of scale difficult and i s an obstacle to the expansion of the agro-processing industry.The food processing industry inturn lacks access to information on modem food processing, packaging and labeling facilities. Modern food research, testing and product development facilities are absent or inadequate, limiting the ability of firms to expand exports, especially to developed country markets. 120. Bramilla and Port0 (2006) collected farm level productivity data from several districts inZambia, makingit possible to show how maize and cotton productivity are negatively correlated with weak service performance as measured by the Investment Climate Surveys (Table 16). Inthe cotton sector, the firm productivity rank of Zambian districts is 90 percent correlated to the rank of availability of phone lines inthe same district, 90 percent correlated to the rank of reliability of transport services, and 87 percent correlated to the rank o f cost of finance. Inthe maize sector, the firm productivity 53 rankof Zambian districts is 77 percent correlatedto the rankof availability of phone lines, while the correlation to reliability of transport services and cost of finance could not be statistically confirmed. Table 16: Spearman rank correlationbetweenfarm yields perhectare andservices performance inZambiandistricts Source: Mattoo andPayton (2007); Bramilla andPorto (2006); World Bank (2002). Note:The hypothesis of independencebetween the two rank orderings canbe rejected at the 90 percent (*) or 95 percent (**) confidence levels. 121. The lack of middlemenand information i s especially severe insparsely populated, remote rural areas. Agricultural extension services provided by the government, NGOs, donor-fundedprojects or churches may fill that needto some extent. However, extension services provided by the government were more common before the reforms starting in 1991. The purpose of these were to help farmers inidentifying markets, adopting new techniques, reducingfertilizer costs, reducing livestock diseases and others. Ina study o f Zambia it was foundthat households with access to extension services hadhigher productivity thanthose who did not (Balat and Porto, 2004). The small scale farmers, i.e. the poor were the ones most affected by the decline inextension services as the large scale farmers can absorb more easily the fixed cost of findingthe knowledge or access alternative information channels. 122. The deterioration inextension services has encouraged some alternative methods of informationtransmission. One i s the technical assistance combined with credits under outgrower schemes. Under these schemes an entrepreneur contracts a smallholder to produce a commercial crop later marketedby the entrepreneur. The entrepreneur provides necessary technical assistanceto reach the agreed production levels and guarantees a certain level of market outlet. 123. These schemes have been very successful and indicate that overcoming coordination failures may ignite growth inthe agricultural sector. Nationalproduction of cotton tripled between 2000 and 2003, and credit repayment improved from 65 percent to more than 90 percent, with the introduction of a refinedoutgrower arrangement (World Bank, 2008a). Currently about one-third of Zambia's smallholders participateinsome form of outgrower scheme arrangement, of which 85 percent are engaged incotton production. Other crops producedby smallholders are tobacco, paprikdchili, honey, and to some extent sugar, coffee, and dairy products. 54 124. Zambia's cotton sector offers another success story, documented inEllis and Freeman (2005). Cotton production grew rapidly inthe mid-l990s, but growth was interruptedby problems with credit recovery, as new entry into the sector encouraged increased side-sellingby producers. The world's largest cotton trader, Dunavant, which operated one o f the two major cotton operations inZambia, responded to this challenge by implementing the so-called `distributor' system, whereby extension agents are transformed into self-employedcontractors, who on-lend and provide extension support to producers. The `distributors' are paid by the cotton companies on the basis of seed cotton volume delivered and the level of loan recovery achieved. Although the system is still inits infancy, yields have been gradually increasing inrecent years, production has surpassed its mid-1990s peak, and credit recovery has improved substantially. 125. Examples of coordination externalities are abundant inmanufacturing and services. Intourism, supply chains and distribution channels are dominated by international firms, who have access to global reservation systems. Inaddition, the consolidation of transport, hotels and tour operators makes it hardfor small operators to compete (World Bank 2001). The coordination failure problems within the tourism sector are also confirmed through perceptionsurvey data for which businessmen were asked to identify if a certain potentialproblem was "one of the constraints" as well as if it was the "main constraint" (Figure40). Figure 40: Constraints to tourism growth as identified by the hospitality sector inZambia poor condition of railiroad infrastructure 72 lack of awareness of the destination 63 poor quality of Livingstone airport infrastructure perceived lack of security for tourists cost of flights are not competitive no proper tourismcircuit an expensive destinationoverall 33 poor access to health/medical facilities 33 Source: Republic of Zambia (2006). 126. The gemstone miningsector is another industrywith untapped potentialthat offers an example of the obstacles posed by inadequate infrastructure and basic service provision. A preconditionfor investment inthe gemstone sector i s access to mining areas, which are usually remote. According to a World Bank survey the large gemstone firms cite lack of infrastructure as the main constraint to investment inthe sector. The most important mines are 70 kmfrom an all weather road, 51kmfrom electricity and up to 55 55 kmfrom water. Of the 891 enterprises with licenses, only one company minesyear round. Other firms work only half a year. 127. Lack of innovation, or the ability to identify profitable products for new investments is another form of market failure. Potentialreasons for poor innovation could be information externalities, which may arise when information about economic opportunities has the potential to benefit many investors, but i s costly to gather. As a result, no singlepotential investor gathers the necessary information. InZambia and other developing countries, innovation i s seen less as the actual "invention of new products", butrather as the successful diversificationof the economy includingby imitating existing products andproducing them at lower cost, developing new varieties, increasing the numberof exports andthe export destinations. 128. As mentionedearlier Zambia has successfully diversified its economy inthe past 15 years (Figures 6,7 and 8). The number o f exports increased from 501 in 1998 to 704 in2005.35A large shareof these exports -between 74 and 80percent -was exported again the following year. The numberof countries buyingZambian products expanded from 68 in 1997 to 105 in 1999, before falling down to 82 in2003, and rising again to 95 in2005. 129. Non-traditionalexports have increased, although they continue to make up only a small share of exports. Since the mainnon-traditionalexports are farm products, scaling up of these activities is boundto have apositive effect onthe income of the ruralpoor, andhas already been identified as a govenunent priority inthe ZambianPRSP. 35Source: World Bank (2007b). 56 4. CONCLUSIONS 130. Despite positive, broad-based and stable growth record inrecent years and immense untapped potential inagriculture, miningand services, Zambia's poverty rates have not declined significantly and remain high.The mainreason for limited income growth of the poor inZambia are low returns to self employment - most notably in agriculture, and limitedgrowth of and/or access to wage employment. 131. The main factors limitingreturns to labor andjob creation are market coordinationfailures such as poor access to domestic and international markets, inputs, extension services andinformation. High indirect costs - most o f which attributable to infrastructure service-related inputs into production including energy, transport, telecom, water, but also insurance, marketingand professional service -undermine Zambia's competitiveness, limitjob creation and therefore serve as a major constraint to inclusive growth. Coordinationfailures are especially severe for the poor who cannot afford the fixed cost associated with finding alternative sources for inputs, marketing and other types of services. The creation of positive coordination externalities will demanda delicate cluster strategy given the limited public resources inthis sparsely-populated, large, low-income country. 132. Continued real appreciation of the exchange rate i s another serious threat to the competitiveness of export-orientedand import competing sectors and to job creation. For Zambia to stay competitive and sustain the growth momentum it will be critical to improve productivity - including the productivity of its labor force, and to lower indirect production costs related to basic services. Carefully crafted monetary andfiscal policies will be critical inrespondingto the real appreciation pressures. Improving the quality and access to secondary and tertiary education as well as continuing the fight against HW/AIDSare essential ifthe poor are to benefit from future growth of the non-farm economy. Finally, weak governance, inparticular poor government effectiveness, are factors behindthe market coordination failures and are as such major obstacles to inclusive growth. Weak governance i s also reflected indistortionary policies, especially within agriculture. 133. This work does not propose policies to deal with the bindingconstraints to inclusive growth. It is an initial step inthe process o f developing a strategy for relaxing the constraints. The policy designi s a second step that requires an in-depth analysis of each constraint and an ex-ante analysis of the effects of policy reforms to remove this constraint. 57 References Alwang and Siege1(2003). Poverty Reducing Potential of Smallholder Agriculture in Zambia: Opportunities and Constraints. Africa Region Working Paper Series No 85, The World Bank.Washington, DC. July 2005. Arnold, J, A. Mattoo and G. Narciso (2006). ServicesMatter for Productivity: Firm- Level Evidencefrom Sub-SaharanAfrica. World BankPolicy ResearchPaper 4048. World Bank, Washington, D.C. Barro, R andJ-W. Lee (2000). International Data on Educational Attainment: Updates and Implications. Center for International Development Working PaperNo 42, Center for InternationalDevelopment at Harvard University, Cambridge, MA. Balat, J. and G. Porto (2005). Globalization and ComplementaryPolicies: Poverty Impacts in Rural Zambia. NBERWorking Paper 11175, National Bureauof Economic Research. Cambridge, MA. March2005. Bigsten A. and S. Tengstam(forthcoming). Prospectfor Poverty Reduction in Zambia, Draftfor Sida Country Economic Report. GothenburgUniversity, Gothenburg. September 2007. Bigsten, A. andA. Shimeles (2007). CanAfrica Reduce Poverty by halfby 2015.7 Development PolicyReview 25(2). Bramilla,B. andG.Porto (2006). Farm Productivity and Market Structure: Evidence from Cotton Reforms in Zambia. World BankResearchPaper 3904, World Bank, Washington, D.C. CSO (1996). Living Conditions Monitoring Survey -I 1996,Central Statistical Office, Republicof Zambia, Luska. CSO (1998). Living Conditions in Zambia (1998), Central Statistical Office, Republic of Zambia, Luska. CSO (2002/03). Living ConditionsMonitoring Survey Report 2002-03, Central Statistical Office, Republicof Zambia, Luska. CSO (2004). Living ConditionsMonitoring Survey Report 2004, Central Statistical Office, Republic of Zambia, Luska. CSO (forthcoming). Living Conditions Monitoring Survey Report 2006, Draft, Central Statistical Office, Republic of Zambia, Luska. Eifert,B., A. Gelb and V. Ramachandran(2005). BusinessEnvironment and Comparative Advantage in Africa: Evidencefrom the Investment Climate Data. Working Paper no 56, Center for Global Development, Washington, DC. Ellis,F.andH.Freeman(eds.) (2005). Rural Livelihoods and Poverty Reduction Policies, Routledge. FinMarkTrust (2006). Access to Financial Servicesin Zambia -analysis based on FinScope datafrom 2005. Datafrom http://www.finscope.co.za/zambia.html. 58 Hausmann,R., D.Rodrik andA. Velasco (2005). GrowthDiagnostics. Cambridge, MAL John F.Kennedy School of Government. March 2005. Jayne, T. S., J. Govereh, P. Chilonda, N.Mason, A. Chapoto and H.Haantuba(2007). Trends in agricultural and rural development indicators in Zambia. Working paper24, Food Security ResearchProject. Jorgensenand Loudjeva (2005). A Poverty and Social Impact Analysis of Three Reforms in Zambia: Land, Fertilizar and Infrastructure. Social Development Papers, Social Analysis, No 49, The World bank, Washington DC. January 2005. Kaufmann, D., A. Kraay, and M.Mastruzzi (2007). GovernanceMatter VI: Governance Indicatorsfor 1996-2006.The World Bank, Washington, D.C. Martinez, J. de L.(2006). Access to Financial Services in Zambia. World BankPolicy Research Working Paper 4061,November 2006. Mattoo andPayton, ed. (2007). Services Trade and Development: The Experience of Zambia. PalgraveMcMillan and TheWorld Bank:Washington, DC. Mwape, F(2004). Institutional Analysis of Fertilizer in Zambia. Background Report for ZambiaPSIA, Preparedfor Social Development. TheWorld bank, Washington, D.C., mimeo. Republic of Zambia (2000). Agriculture Bulletin 2000. Ministry of Agriculture and Cooperatives. Republicof Zambia (2006a). Fifh National DevelopmentPlan 2006-2010. Lusaka. Republic of Zambia (2006b). Livingstone Tourism Survey,Ministry of Tourism, Environment and NaturalResources,Lusaka March2006. Rodrik (2004). "Industrial Policy for the Twenty-first Century," Manuscript, Harvard University. Rodrik (2006). `Goodbye Washington Consensus, Hello Washington Confusion', mimeo preparedfor the Journal of Economic Literature. Simler, K.(2007). Micro-Level Estimates of Poverty inZambia. Working Paper, International FoodPolicy ResearchInstitute, April 2007. Washington, D.C. Sida (2006). Integrated Economic Analysisfor Pro-Poor Growth: A Methodological Approach. Method Document, Stockholm: Sida. Smith,R. (2001). Land Tenure, TitleDeeds, and Farm Productivity in the Southern Province of Zambia: Preliminary ResearchFindings (Outline). Presentedat the University of Zambia, 19 September 2001. Venables, T. (2006). Notes on TheZambian Economic Situation. DFID, London, mimeo. Weeks, J. (2008). TheReduction of Fiscal Space in Zambia - Dutch Disease and Tight- Money Conditionalities. Country Study No 14, International Poverty Centre. Brasilia. January 2008. World Bank (2003). Zambia: The Challenge of Competitivenessand Diversification. Report No. 25388-ZA, Washington, DC: World Bank, January 2003. 59 World Bank (2004a).Zambia: An Assessment of the Investment Climate. ReportNo. 29741-ZM.PrivateSector Unit,Africa Region.Washington, D.C.: World Bank. May, 2004. World Bank (2004b).Zambia Country Economic Memorandum: Policiesfor Growthand Diversification. Report No. 28069-ZA.PovertyReductionand Economic ManagementI,SouthernAfrica, Africa Region.Washington, D.C.: World Bank. October2004 World Bank (2007a).Zambia Poverty and Vulnerability Assessment.ReportNo. 32573- ZM. HumanDevelopmentI,PovertyReductionandEconomic ManagementI, Africa Region.Washington, D.C.: World Bank. World Bank (2007b).Zambia Growth Diagnostics Study,mimeo. World Bank (2007~).Market Opportunities and Constraints. Backgroundpaper prepared for the Zambia CommercialAgricultureDevelopmentProject(CADP), Identificationmission, Annex 3. World Bank, Washington, D.C. World Bank (2008a). WorldDevelopmentReport 2008: Agriculture and Development. Washington, D.C.: World Bank. World Bank (2008b).Doing Business Report 2008. Washington, D.C.: World Bank. World ResourceInstitute (2008).Data from http://earthtrends.wri.org/ 60 Appendix Growth accounting technique36 1. We assume that aggregate output can be expressed as a function of physical and human capital: Y = AF (K,H), where Y i s gross domestic product in constant 2000 purchasing power parity (PPP) prices; A i s an index o f total factor productivity; K i s gross domestic capital stock in constant 2000 PPP rices; H i s human-capital-adjusted labor input, defined as: H=LDPe l$S), where L i s population; D i s share of population age 15-64; P i s labor force participation rate; S is number of years of education per worker; q3 i s a parameter that measures the returns to education. 2. We consider two types of production functions. The first one i s a Cobb-Douglas production function with possibly non-constant returns to scale: F(K,H) = [PXH(I-~) I where a i s a parameter between 0 and 1that measures the relative importance o f capital, and y i s a parameter that measures the extent o f returns to scale. Reasonable values of a range from 0.3 to 0.5. If y=l (y > 1) (y e 1) there are constant (increasing) (decreasing) returns to scale. Reasonable values of y range from 0.8 to 1.2. The second one is a constant-returns-to-scale constant elasticity o f substitution production function: F (K,H) = [a+(1-a) H(I-p)] ('Ip) where p = (o-l)/a i s the elasticity of substitution between K and H. When o = 1 this reduces to the Cobb-Douglas case above with y= 1.Reasonable values of o range from 0.8 to 1.2. 3. To estimate the level and growth rate of A, we require data on Y, K, L,D, P, and S. These are drawn from the following standard sources. Real GDP and gross domestic investment in constant 2000 U.S. dollars adjusted for differences in PPP come from World Bank's World Development Indicators. Data on population, the share o f population aged 15-64, and the labor force participation rate are computed based on data from the World Bank's DDP. We assume that the labor force participation rate i s an average of the labor force participation rates for females and males. Data on the stock of years o f education are obtained from Barro and Lee (2000). Numbers for the other years were estimated assuming a constant annual growth rate in the human capital stock. The parameter q3, which measures the returns to education (i.e. the percentage increase in worker productivity due to an additional year o f education) i s assumed to be 10 percent. 4. Capital stocks are constructed using the perpetual inventory method in this worksheet. This requires information on the initial capital-output ratio in 1980, depreciation rates (d), and gross domestic investment inconstant U.S. dollars adjusted for 36BasedonPREMnote 42: Measuring growth in totalfactorproductivity by Swati R.Ghoshand Aart Kraay, August 2000. 61 differences in PPP (I). For most developing countries, reasonable values for the initial capital-output ratio range between 1and 2, and for the depreciation rate are between 0.04 and 0.08. We use initial capital-output ratio of 1 and 6 = 0.06 in all calculations. The capital stock i s calculated usingthe following formula: K (t) = (1- 6) K (t - 1) z (t). + wb22379 C:Documents and Settings\wb22379\MyDocumentsEambia Sharedgrowth analytics-July 9-08.doc 07/10/2008 2:46:00 PM 62