Zambia - Poverty assessment (Vol. 3 of 5) : Rural poverty assesment

Report No. 12985-ZA Zambia Poverty Assessment (In Five Volumes) Vo>lume IIl: Rural Poveny Assessment November 30, 1994 tluman Resources Division Southern Africa Department Africa Regional Office Document of the World Bank  Rural Poverty in Zambia: An Analysis of Causes and Policy Recommendations Prepared for Human Resources Division Southern Afrca Department The World Bank Washigton, DC Jdfry Aiwang Paul B. Siegel Departmedt of Agricltural Department of Agricultural & Applied Economics Economics & Rural Sociology Vria Tech University of Tennessee Blacksburg, VA, 240614401 Knoxville, TN 379014171 USA USA Januar 1994 Revised July 1994  ABBREVIATIONS AE Adult Equivalents RAP Research Acdon Program API Moderate Poverty Group RCDM Rural Community Development and AP2 Severe Poverty Group Motivadon Project ARPT Adaptve Research Panning Team SDA Social Dimensions of Adjustment ASIP Agricultuwal Sector Investment SEA Standard Enumeration Area Program SEG Socioeconomic Group CSA Consus SuevsoiyArea SEPA Socio-Eonomic Policy Analysis CSO Central Statistical Office SIDA Swedish International Development CUSA Credit Union and Savings Association Agency DCU District Cooperadve Union SIDO Small Industries Development DDF District Development Fwn Organization EAP Extension Action Progmm .T&V Training and Visit Extension System ESP Environmental Support Project UNICEF United Nations International Childrens PAO Food and Agricnlture Organizadon Emergency Fund FEWS Famine Early Waming System UNZA University of Zambia FINNADA Fi Internaional Development USAID United States Agency for International Agency Development GIS Geographical Ifformation Systems VIS Village Industrial Service GRZ Governmert of Republc of Zambia ZAMS Zanbia Agnbusiness and GTZ German Technical Assistance Management Support Project Ha Hectare ZAMSEED Zambia Seed Company HEIS Household Expenditure and Incomes ZAREP Zambia Agricultural Research and Survey Extension Program HEPS High Energy Phtein Supplement ZCF Zambia Cooperative Federation Progam ZCF-CS Zambia Cooperative Federation- IFAD Internatoal Fund for Agricultural Conswer Services Develolpment ZCF-FS Zambia Cooperative Federation- FPRI Imernte ionsl Food Policy Research Financial Services ZCRS Zambia Chriian Refugee Service DP Integrated Rural Development Program Km Klometer LNTCO Lint Company of Zambia LP Linear Programming MAFF Ministy of Agruure, Farming and Fisheries NAMBOARD National Marketing Board NATCO National Tobacco Company of Zambia NEAP National Enviromenal Action Plan NGO Non-governmental Orpnzatio NORAD Norwegian Agency for oeational Development PAM Program Against Malnutrition PCU Provincial Cooperative Union PPM Progran to Prevent Malnutition PRA Participatory Rural Appraisal PS Prioriy Survey PUSH Program for Urban Self Help PO Head-ount measure of poverty (incidence index) P1 Poverty gap measure (depth index) P2 Severity of poverty measure (severity index) i  EXECUTIVSUMMARY Many of the conditions associated with mral poverty in Zambia are examined in this report. The report bes by reviewing some recent stdies of mral and urban poverty in Zambia. Next, an analysis of the Priority Survey, carried out by the Central Statistical Office in 1991, is undertaken to create a povery profle. The profile focuses on the geogrplhical distribution of rural poverty and its association with various economic, social, and environmentl variables. This profile describes the conditions and helps identify some of the cntical constraints facing poor rural households. In the next section of the report, policies and institutions affecting the rural poor are reviewed. A review and critique of past policies and programs by GRZ, donors, and NMOs is included. The information gained from past studies, the poverty profile, and review of policies and institudons is then incorporated into a simple agriculture household model. The model helps quantify the relative importance of various conditions and constraints to income generatior. Using the model, several stylized policy scenarios are analyzed to better understand the response of smallholders to structural adjustment and market liberalization. Finally, the major findings of the report are summarized, and several specific recommendations are made on the types of policies and programs that may reduce rural povery. The major findings and specific recommendations are outlined below: MAJOR FINDINGS 1. Aniculture is P Agriculture is the main economic activity in rral Zambia and must be the focus of any stategy aimed at reducing mrual poverty. 2. Geouhv of Rural Povert Although there are some substanti geogphic variations in the incidence and depth, virtually all rural areas suffer from significant poverty. Rural poverty is most closely related with geographic isolation (i.e., remoteness). 3. lhe Strure of Agrdiur and Poe Most rural poor are smaUholders using low levels of technology, but, in most cases, linked to input and output markets. Agricultural policies and institutions have thwarted the development of efficient marketing, credit, etension, and research systems. In addition, seasonal labor shortages and the lack of traction for land prparation have been major constraits for smallholders. Smallholders with access to oxen traction and in less-isolated areas tend to use improved management practices and have higher incomes. ii  4. Distrution. Equitv. and Poverty Female-headed households and households with large numbers of dependents tend to be poorer than other households. At the same time, more poor people live in households headed by males. Problems related to poor health and nutrition, and natural resource degradation have strong detrimental impacts on income generation. Gender-specific allocation of agricultural tasks represents an important constraint to increased agricultural production. S. The Role of Education There is a positive return to education in rural Zambia. Members of households whose head has some formal education are more likely to be better off. 6. Institutions and Rural Poverty The policies and institutions of the past were not conducive to the reduction of rural poverty. In many ways, the opposite was true, and past policies and institutions led to the perpetuadon and worsening of rural poverty. SPECIC RECOMMEDATONS 1. Develop lndesadine of Markets and Hox to Take Advantaie of Market Onporunities Many rural poor lack undag of how to fmnction in a market economy. In isolated settings, information asmmetries and physical distance have created a sense of powerlessness and uncertainty about market forces. Market liberalization alone is not sufficient, and the poor need to articulate themselves in order to create rural markets. 2. btef ration of Remote Areas into Market Economv Developing an u snding of markets and how to take advantage of market opporunities (as described above) should help prepare rural poor as individuals and groups to function in the market economy. There needs to be a concerted effort to improve access to markets and lower transaction costs associated with agricultural marketng. This effort should be designed to promote competition in the provision of services to geographically isolated areas. 3. Provision of A,ro.nCate Technologies for Market-based Auriltgual Production and Post- harvest Activities Agriculural research and extension need to be reorented towards meeting the needs of flexible farming systems for smallholders that evolve according to market forces. In general, more ii  attention needs to be devoted to the problems of smallbolders. Also, links between extension and research need to be strengthened. 4. ,M1rove Capacty to Monitor National. ReAjonal. and Sectoral Policies and Programs and their Impact on Rural Poverty This report demonstrates the complex and diverse issues facing the poor in rural Zambia. It is important that an in-house capacity be created to mnonitor and evaluate policies and programs ard their inpact on the rural poor. iv  TABLE OF CONTENU ListofTables ............................ viii 1. Introduction ............................ 1 I. aracterizationofZambianPoverty . ........................... 3 A) Review ofZambia PovertyStudies ....... 3 i. Description of Studies . ..... 3 ii. General Rests ...... 4 iii. TheGegraphicalDistributionofPove.yS......5 iv. Socio-EconomicCharacteitcsan dPovety 7..... . 7 v. General Observations on Rural Povey ad Food hmacur ..ty ...12 vi. SumnaryofS tdies a. ..d. 14 M. Analysis of Priority Survey Da . . . . . .18 A) Oveew ...... 18 B) Data and Measurement 1.....8 C) Preliminary Ioformation and Conas ............................. . 19 1)) Poverty and Distance to Facilities .....26 E) Socia Factors ...... ..28 F) EcowmicFacors..... ...33 i. Enployment ....33 ii. Household Assets ....37 iii. Linkages to Markets ....41 0) RegressionResults ......................................... . . . .44 H) Sumay of Priority Survey Rets . ..... . 47 IV) Policies, Istitutions, and Donor Support .. . 49 A) Agricultural Policies .. 49 i. oview ................................... 49 ii. MaizeMarketing S1........51 (a) History S..... (b) Reform .....52 (i) Maize Prices and MIarrke 3.. . 53 (ii) MaizeConmption S7....57 (iii) Effect of LJ diead Iuk.fa .... . . .61 B) Istitutions ........ 63 i. Mnistry of Agricltue, FoodandFAFF . . ......... . 63 ii. ResearchInstitutions . . . . .64 iii. Extension I_ttutions ...... 66 iv. Infornation Institutions ....... 66 v. Agriculturd Markeingad InputSuply . ..... 67 vi. Land Tlenre .....68 vii. Environnx..e..... 68 viii. AgculturalCredit ..... 70 ix. Cooperatives... ........... ...... 71 v  x. Rural ndustry ............................... 73 xi. Local Govenents ..........................o..... 73 C) Donor Support ............................... 75 D) Sumnun, ............................ 81 V. A HouseholdModel ........................... 82 A) Description of Rurwa Households . .......................... 82 i. Types of Farm Households ............. .............. 82 ii. Access toLand .................................... 84 iii. Technologies for Smallholders ................................. 85 iv. Seasonal Patterns in Farming ................................. 86 B) Description of the Agricultural Household Model ......................... 87 i. Introduction ........................................... 87 ii. Objective Function of the Household ............................. 89 iii. Types of Households ..... ........ ......................... 89 iv. Composition of the Household ...... ........................... 90 v. Staple Foods .............. ............... 90 vi. Labor Availability and its Determinants ........................... 92 vii. Crop Activities and Prices of Outputs and Inputs ....... .............. 94 viii. Institutions and Infrastructure ................................. 95 xi. Land Availability ......................................... 96 X. Cash Availability ....................................... 96 C) Model Results ....................................... 97 i. Baseline Results ..97 a. Value of Objective Function and Retums per Household Member 97 b. Value of Staple Foods .98 c. Value of Ch nputs .101 d. Land Utilization and Crop Mix .102 e. Labor Utilization ad Constraints .102 ii. Agricultural Policies and Scenarios ..104 a. Available Cash .104 b. Market Liberalization .104 c. Remoteness .109 d. Input Supply .112 e. Health and Nutrition .114 f. Natural Resources .118 g. All Constraints .119 D) Suinmy of Model Resuits ................... 123 E) Comparison with conditions in Zambia Prevailing in April 1994 . . .128 1. Gender ..128 2. Labor and Animal Traction ..128 3. Remotemss ..129 4. Diversification ..129 VI) Summaly of Results and Implications for Policy ............................ 131 A) Agricutur is Paramount .................................... 131 B) Geogrphy of Rurl Poverty .................................. 131 C) The Structure of Agriculture and Poverty .............................. 132 vi  D) Dirbution,Equity,adPovety ........................ 134 B) The Environment and Poverty ........................ 136 F) TheRoleofEducaton .......................... . 136 0) nsitons and Rural Poverty ......................... 136 H) SpecificRnhmdalos ......................... 137 Referens ......................... 141 Acmowledgements ......................... 146 Amex I: Information on Survey Methodologies for Poverty Studies .................. AI-I Annex n: Projects Aied at Rura Poverty Reduction: An Overview and Plans for the Future AI-I Annx I: Information on Rural Households: Case Studies ...................... - Al- Annex IV: Description of Data and Assumptions Used for LP Model ................ AIV-1 Annex V ............................................... AV-1 vii  LIST OF TABLES Table II.' . Urban/Rural Comparisons of Poverty by Source ...................... S Table 11.2. Geographical Distribution of Rural Poverty ......................... 6 Table 11.3. Wihiin-Province Description of Poverty ............................ 8 Table 11.4. Prevalence of Moderate Poverty by Sex of Household Head .... .......... I I Table m. 1. Expenditure Shares in Rural Zambia by Poverty Group ..... ............ 20 Table 111.2. Mean Per-capita (Adult Equivalent) Expenditures in Rural Zambia, by Province (K/Month) ........... ....................... 22 Table 111.3. National Poverty Indices .22 Table II.4. Distribution of Rural Poor, by Province ............ .......... , 23 Table m.5s Rural Poverty Indices by Province ............ .................. 23 Table m.6. Rural Expendiure Shares by Poverty Group, by Province ..... .......... 25 Table 1.7. Average Distance in Km to Facilities by Poverty Group, by Province .... .... 27 Table 11.8. Average Distance in Km to Facilites Within Province by Low- and High-Poverty Districts ............................................ 29 Table m.9. Rural Poverty by Chaacteristics of Household Head ................... 29 Table 1I.10. Distribution of Rural Poverty by Education and Principal Occupation of Head ... 31 Table M. 1 1. Dependency Ratios by Rural Poverty Group, by Province ............... 31 Table 1.12. Average Rural Household Sizes by Poverty Group, by Province .... ....... 32 Table 1.13. School Attendance by School-aged Rural Children by Poverty .... ......... 33 Table 1.14. Rural Poverty Indices by Employment of Head ....................... 34 Table m1.15. Rura Poverty Indices by Employment of Spouse .34 Table m. 16. Employment of Head of Household and Spouse by Rural Poverty .... ....... 35 Table 11. 17. Percentage Distribution Employment Categories by Cbaractexistics of Rural Household Head .. ...............r........... 36 Table 111.18. Percent of Rural Poor and Non-poor Households with Children Employed ..... 37 Table M. 19. Percentage of Rural People Using Different Household Facilities, by Poverty ... 38 Table 111.20. Profile of Rural Asset Ownership by Different Poverty Groups .... ........ 39 Table 111.21. Rural Poverty Indices by Land Tilting Status of Household ............... 40 Table m.22. Percent Rural Households Producing Commodities by Province, by Poverty .... 42 Table M.23. Percentage of Commodity Produced That is Marketed, by Poverty, by Province . . 43 Table m.24. Regression Results for Household Expenditres ...................... 46 Table IV.1. Estimated Production and R,corded Maize Marketing 1992/93 .... ........ 55 Table IV.2. Real Producer Price of Maize .................. 55 Table IV.3. High and Low Reported At-Depot Price 90kg Maize, Various Dates (1993) .... 56 Table IV.4 Maize Meal Prices: Roller Meal (25kg) Average Urban Retail Price ....... .. 58 Table IV.5. Mealie Meal Retail Prices in Selected Areas (Kwacha per 25kg roller meal) .... 58 Table IV.6. Area Panted in Hectares 1992/93 and 1993/94 ........ .. ............. 60 Table Iv.7. Maize Area Plated in Hectares 1992/93 and 93/94 ....... ............ 60 Table IV.8. Funding for Proie.,ts in Zambia .. 77 Table IV.9. Production Oriented and Basic Needs .............. 80 viii  Table V.1. Cbaaeistic of Fann Units in Zambia .......................... 83 Table V.2. Aggegated Linear Programing Tableau .......................... 89 Table V.3. Types of Households Considered for the Model ...................... 91 Table V.4. Crop *roduction Activities ................................... 95 Table V.5. Results from Baseline Model .................................. 99 Table V.6. Results with Cash Constraint Imposed ........................... 105 Table V.7. Model Results Under Assumed IMacts of Market Liberalization .... ...... 107 Table V.8. Model Results Under Assmtion of Remote Household .... ........... 110 Table V.9. Model Results with Restrictd Supplies of nputs .113 Table V.10. Model Results with Declining Female Health .115 Table V.11. Effcts of Aids Dependency on Returns per Household Member .... ...... 118 Table V.12. Model Results Under Scenario Where Access to Natural Resources is Limited . . 120 Table V.13. Results With All Constraints Combined (Including Remoteness) .... ....... 122 Table V. 14. Conparison of Results: Market Liberalization With All Constraints Combined . 124 Table V. 15. Suamary of Results: Index of Returns per Household Member, Traditional Male-Headed Household 100 ............................ 124 Table V.16. Summary of Results: Comparison of Objective Value Baseline Objective Value = 100. ........... .......................... . 125 Table A.I.1. Correlation Between CSO Income Variable and Rural Household Expenditures by Province ............ ............................ A-4 Table A.I.2. Regression for Rental Imputations ............................. Al-5 Table A.IV.1. Maize Activities .. . AIV-7 Table A.IV.2. Summary of Assumptions About Maize Yield Responses ... AIV-8 Table A.IV.3. Other Crop Activities Included in Model .. . AIV-9 Table A.IV.4. Labor Reqduieent for LM ... AIV-10 Table A.WV.5. Labor Reqirements for HM9 . AIV-I Table A.J.6. Summary of Infonnaion on Crop iPoduction A . . AIV-12 Table A.W.7. Summary Ioforion on Crop Production Activities ... AIV-13 Table A.V. 1. Results: Labor Constraints Only (Beline w/o Food Security Constrint) ... AV-1 ix  1. INTRODUCTION Zambia is engaging in a macroeconomic stabilization program that might have a negative inpact on an already food insecre poor population. The last decade in Zambia was charcteized by a stark decline in economic conditions. Per capita income fell from $720 in 1981 to $420 by 1990 (World Bank, 1993b). This decline was precipitated by a fall in the price of copper, Zambia's main export, but it is widely believed to have been exacerbated by poor govenunent policy (World Bank, 1993b). In the last decade, the govenmment attempted to mainain an elaborate food subsidy program using parastatals and controls over food marketing and external trade. The result of the subsidy program was to reward urban consumers (40 percent of the population), tax agricultural producers, and create economy-wide distortions. The system distorted a number of incentives to farmers; of particular note was the heavy bias that promoted maize production. Agricultural research and extension efforts were largely aimed at maize producers, and internal maize prices were raised relative to alternative crops. In addition, because the agricultural pricing system was designed to provide low-cost food to a growing urban population, there existed a blatant policy bias in favor of commercial farmers, located along the line-of-rail and occupying the most productve lands. These farmers received the bulk of the goverment's resources destined for agriculture, including formal credit, subsidized inputs, and research and extension services. The subsidy system was maintained by borrowing money to substitute for falling copper earnings. Subsidization failed due to the inability of the government to sustain funding of its programs. The result of the failure is a desperate situation for Zambia's poor. Employment has fallen due to government retrnchment, inflation is high, and the poor, now without food subsidies, face serious food insecurity of a chronic and periodic nature. By all indications, poverty is widespread. Nutrition data indicate that long-term malnutrition (stuntig) is alarmingly high, placing Zambia among the countries with the highest levels of malnutrition anywhere in the world (Cogili and Zaza, p. 37). The goverrnment is committed to a privatization and stabilization effort to restore market incentives to the economy. In the short run, however, the situation of Zambia's poor is expected to worsen, and the critical need for programs to immediately improve their food security has been recognized (World Bank, 1993b, p. 74). The urban poor most likely will be more negatively impacted by the adjustment program in the short run, since rural/urban terms of trade should be improved, resulting in I ; er urban food prices. However, the adjustment program may have a negative short- run impact on the rural poor due to production and behavior patterns that have developed on the basis of distorted mnarket signals, and the inability of resource poor farmers to take advantage of improved terms of trade as input prices rise. In rural areas, where most of Zambia's poor lives, agricultural production has grown at a moderate pace from 1974-1991 despite declining terms of trade (World Bank, 1993b). Such growth has occurred despite an incentive regime that favored maize production at the expense of all other crops and widely-documnted underinvestment of public funds in agriculture. The sectoral growth leads many analysts to hope that agriculture night replace copper as the engine of economic growth. It is hoped also that agricutural exports can generate substantial foreign exchange earnings. The moderate growth in agriculture over the past decade has been uneven, mostly occurring in the favored areas along the line-of-rail where commercial farmers predominate. In many regions total food 1  production has not kept pace with population growth, and productivity has fallen. In fact, sector-wide most of the increase in production can be attributed to expanding acreage; yields have fallen for most crops (World Bank, 1993b). The agricultural sector is in a fragile and vulnerable condition. lnadequate public investment in research, extension, and infrastructure; continual over-emphasis on maize production; and crippled output marketing and input supply mechanisms characterize the sector. Some of this vulnerability was exposed during the 1991/92 drought when the maize crop, critically dependent on rainfall, failed in many areas of the country, leaving much of the rural population exposed to economic devastation. Over the long run, price incentives, in the form of improved terms of trade for rural food producers should benefit rural areas, but several caveats remain. First, two adjustments are occurring sinmltaneously, and these adjustments do not always work in the same direction. Liberalization of export markets and removal of price controls on agricultural products should increase prices to producers and thus benefit them. Small-scale producers may not feel many of these benefits since their marketed surpluses tend to be smal and they will face higher input prices following liberalization. Elimination of pan-territorial pricing and marketing subsidies, on the other hand, will benefit some producers while hurting others. The geography of relative price variations depends on infastuctre and distance to final markets, the density of market production, and availability of financing for agricultural marketing services. Second, there is continuing debate in the literature on adjustment whether price-based incentives are adequate to stimulate agricultural growth when infrastucture deficiencies and social conditons interact to create an environment that is not amenable to economic growth (Satm and Sarris). There is reason to believe that traditional Zambian agriculture suffers from such structual deficiencies and that broad-based public investment in the rural economy is necessary to stimulate growth. This report examines many of the conditions associated with rural poverty in Zambia. The report begins by reviewing some recent studies of poverty. It is hoped that by examining previous studies of poverty in Zambia, some clues about the causes of this poverty will be found. Next, an analysis of the Central Statistcal Office's (CSO) Priority Survey (PS) is undertaken to create a profile of rural poverty, and describe the geographical distributon of rural poverty and its association with various socila and economic variables. This profile helps identify some of the critcal constraints facing rural households as they struggle to survive. Institutions and policies affecting smallholders are then reviewed. Included in this review are policies and programs of the Government of Zambia (GRZ), and programs conducted by donor governments and non-governmental organizations (NGOs). The information generated is then incorporated into a simple agricultural household model in order to quantif the relative importance of these constraints to income generation. FinaUy, recommendations are made on the types of policies and programs that may reduce rural poverty. 2  H. CHARACTERIZATION OF ZAMBIAN POVERTY A) Review of Zabia Poverty Stuies Several studies have characterized poverty in Zambia and developed recommendations to increase the food security of poor Zambiams. Included among these studies are Caidwell; the Central Statistical Office's study (CSO, 1993a); Chipwende, et al.; Cogill and Zaza; Jensen and Luckett; Siandwazi; td World Bank (1993a, 1993b). This section reviews some of these studies with the intent of examning methodological approaches, common conclusions, and areas of disagreement in these studies. Results from the various studies will also be compared with our own analysis of the Social Dimensions of Adjustment (SDA) Priority Survey (PS) data that follows. i. Description of Studies In the reviewed reports, poverty is analyzed first by developing absolute or relative poverty lines or by describing the prevalence of undernutrition or malnutrition. Once the poor are identified, each study proceeds by examining the relationship between certain variables, at the household, district, or province levels, and the poverty outcome. The Jensen and Luckett report used the Household Expenditure and Incomes Survey (HEIS), conducted in June 1991 by the Prices and Incomes Commission of the GRZ'. The study used household expenditures to characterize poverty. Two relative poverty lines were used: The mean Zambian income, and one-half the mean income. Research has shown that in poor countries the mean income is a good approximation of the subsistence poverty level (Jensen and Luckett). The authors described the geography of poverty, and created a profile of poor households. The CSO report used the PS, conducted in October and November of 1991 by the CSO. The report used official absolute poverty lines and employed the income section of the PS to measure the prevalence and the depth and severity of poverty in rural areas. Poverty was broken down by urban and rural areas, by province, by socio-economic group (SEG), and by household demographics. The CSO report also used anthropometric data from the PS to describe the geographic and social pattern of malnmuition. Chipwende, et al. used both the HEIS and the PS to create a geographic and social profile of Zambian poverty. They used 2/3 of the mean expenditures from the HEIS as a household poverty cutoff. They then examied the prevalence, depth and severity of poverty by characteristics of the household and its head. They created two cutoffs with the income portion of the PS: 2/3 of mean household income for the moderately poor cutoff, and 1/3 of the mean for extremly poor households. The authors analyzed the relationship between poverty and distance and access to facilities, among other things. 'More information on the survey methodologies and the different approaches to the measurement of poverty used in the studies discusses here is provided in Annex I. 3  Siandwazi used secondary data on the prevalence of malnutrition gathered from a number of sources, hospital and clinic admissions data, and several household consumption studies to paint a broad picture of regional and seasonal variations in poverty and malnutrition in Zambia. Caldwell used a number of data sources, including percent undernourished by geographic area, data on access to medical sources and other infrastructure, historical crop production and current food aid data to create a geographical profile of vulnerability in rurl Zambia. Using geographical information systems (GIS) technology, they overlayed and combined a number of these indicators to create scores for chronic and current vulnerability. Cogill and Zaza reported on results of the Nutrition Status Module, a part of the Zambia National Household Survey Capability Programme, implemented by CSO in December 1989 and January 1990. The Module represented a random sample of 2,133 children under 5 years old from 1,200 rural households from all provinces. The module, which collected anthropometry data and clinic attendance, was linked to the Crop Forecasting Survey, and the combined data were used to examine the geographic distribution of malnutiton, along with the relationship between socioeconomic characteristics and nutritional status. ii. General Results The results from the three income- or expenditure-based poverty studies are summarized in Table I. 1. Comparison across the studies is difficult ttcause the poverty lines employed are not always constant. There is a reasonable correspondence of results across the studies, even companng the HEIS-based studies with the PS-based ones. Estimates of the prevalence of moderate poverty range from a high of 77 percent to a low of 56 percent. The studies clearly show that nual areas suffer from a higher prevalence of poverty than do urban areas. Rural poverty is a widely-documented phenomenon in Zambia. There is a bias, however, in favor of overestimating rural relative to urban poverty for aU of these studies for two reasons. First, none of the studies cited using income or expenditures adjustd for cost of living differentials, and these differences are likely to be high, despite pan-territorial pricing of maize. If current (April 1994) maize prices are a proxy for intra- country variations in prices of other goods, then there were substantial variations in costs of living at the times of these studies.2 Second, income and expenditures are more accurately measured in urban areas; systematic undenasuremnt of these variables in rural areas is likely. The 1993 Priority Survey msures prices and, thus, can be used to eliminate the first source of bias. Without a concerted effort to measure all income and expendiures accurately, the degree of overestimation of rural poverty cannot be kmown. Despite these biases, the studies universally conclude that the prevalence, depth, and severity of poverty is greater in rural Zambia. Absolute comparisons between rural and urban areas are not possible because of the biases idendfied above, but it is reasonable to conclude that rural areas are relatively worse off in Zambia. WMore will be said about current maize price fluctuations in Section IV of this report. 4  Tamelc 11 1. Ufrban/Rural Comparisons of Poverty by Source Percent fallins into class Measure Rural Urban All Jensen and Luckett Moderate Poverty .79 .43 .69 Severe Poverty .42 .08 .32 CSO Moderate Poverty .85 .55 .71 Severe Poverty .78 .44 .61 Chipwende et al. Moderate Poverty' .88 .26 .68 Moderate Poverty2 .67 .29 .56 Moderate Poverty3 .86 .69 .77 Severe PoverW4 .74 .44 .57 ,Using data from "Situadonal Analysis of Children and Women in Zambia." See Chipwende, et al. for references. Using HEIS, one half mean expenditre as cutoff. 'Using PS, households less than % of average income. 4USing PS, households less than % of average income. iii. The Geographical Distribution of Povert= The geographical distribution of poverty emerging from the various studies is presented in Table 11.2. The ranking of each province in ascending order is also shown. Poverty in rural Zambia shows wide geographic variation. Rural poverty as measured by Jensen and Luckett varied from 89 percent in Eastern to 70 percent in Central province. The CSO (1993a) report showed slightly different orderings by province but the magnitudes of the estimates are similar. Moderate rural poverty in the CSO report ranged from a high of 93 percent in Northern to a low of 63 percent in Lusaka province. Luapula, Western, and Northwestern provinces are most subject to chronic food insecurity, while Lusaka, Copperbelt, and Central provinces are relatively most food secure according to the Caldwell measures (Table 11.2). A comparison of the cereal grain production per capita data with the other indicators shows that the former is a poor predictor of either poverty or malnutrition. Luapula, Northern, and Eastern provinces have the highest prevalences of malnutrition, according to the Cogill and Zaza study, while Northern is unambiguously the worst in the CSO study. 5  Table ].2. Geograpbical Distribution of Rural Poverty. Prevalewce of Poor _eople Nutrliton/AnthroDometR (4) (5) (6) (7) (8) (9) (1) (2) (3) Chronic Cereal % % Under- % % Under- Province HEiS Moderate Severe Score Prod/Cap Sunted nouished Stunted nourished Central 70 (2) 76 (7) 68 (7) 68.6 (9) 531 (9) 37 (4) 22 (5) 56 (2)22 (7) Copperbelt N/A 74 (8) 65 (8) 56.3 (5) 55 (2) 37 (4) 28 (3) 36 (6)21 (8) Eastern 89 (1) 85 (4) 81 (4) 60 (8) 416 (8) 41 (3) 28 (3) 48 (3)23 (4) Luapula 87 (3) 80 (6) 72 (5) 34 (1) 39 (1) 54 (1) 40 (1) 41 (4)29 (2) Lusaka N/A 63 (9) 56 (9) 57 (6) 125 (5) 24 (9) 22 (8) 34 (8)23 (4) Norhern 80 (4) 93 (1) 90 (1) 43.6 (4) 139(6) 44(2) 30 (2) 60 (1)35 (1) Northwester 88 (2) 92 (2) 86 (2) 40 (3) 78 (4) 28 (8) 23 (7) 32 (9)23 (4) Souhern 76 (6) 83 (5) 72 (5) 57.7 (7) 340 (7) 34 (6) 17 (9) 36 (6)21 (8) Western 72 (6) 89 (3) 86 (2) 36.3 (2) 70 (3) 33 (7) 27 (5) 41 (4)27 (3) Source: (1) Jensen and Lucket, using 1991 Household Expenditures and Inowme Study, and Ih of rural mean income as cutoff (1993a). (2 & 3) CSO (1993a) using 1991 PS with absolute cutoffs for povert (4 & 5) from Caldwell (6 & 7) Cogil and Zaza: Preschool chidren using established cutoffs. (8 & 9) CSO (1993a) usiDng PS anftropometry data Note: Rankings by province are in parentheses. 6  Comparison across studies allows some generalizations. Luapula suffers from the worst malnutrition and chronic food insecurity,3 yet according to the CSO report, poverty is significantly worse in other provinces. Northern province is always found to suffer from high prevalences of malntition and poverty. Eastern, Nordtwestern, and Western provinces also show relatively high rates of poverty, food insecurity, and malnutrition. Prevalences of poverty, malnutrition, and food insecurity in Central, Southern, Copperbelt, and Lusaka rural provinces are generally lower than those in the more isolaed provinces. Severe poverty also exists in these less isolated provinces, and the large populations in the centrally-located provinces mean that there are more poor. Some of the differences across studies may be attributable to the characteristics and actual variations of the different indicators; they may also be explained by systematic biases in measurement and survey coverage. The HEIS, for example, had a small rural sample size outside of Eastem province relative to the PS (see Annex I). Income measures employed in the PS-based studies have flaws whose impact varies from province to province. More will be said later in this report. The one firm generaization is that there are significant variations in poverty and malnutrition, and the more isolated provinces have higher prevalences of poverty. Significant rural poverty also exists in the more centrally-located provinces. Even though certain provmces have lower percentages of poor than others, they may have a higher number of poor due to large population size. This is the case in the Copperbelt where the absolute number of poor is exceeded only by that from Eastern province (World Bank 1993a). Copperbelt and Eastern provinces are heavily populated, and Copperbelt is highly urbanized. Poverty in Copperbelt is primarily an urban phenomenon. Within-province variation in some of the indicators is shown in Table 11.3. There is extreme variation in indicator values between districts within provinces. Exceptions occur in Luapula and Eastern provinces where measures are fairly consistent across provincial districts. iv. Socio-Economic Characteristics and Povertv To varyig degrees the different studies examined the relationship between socio-economic chacteristics and poverty. These results are summaized below. Household Size Poorer households are larger in size than wealthier ones in both rural and urban areas according to Jensen and Luckett. For rural areas, household size for the lowest expenditure decile is 6.42 members, while for the top decile it was 2.92 members. Size of household increases 3Chronic food insecurity is measured in the Caldwell study using per capita cereal producton, percent of population underweight, access to health facilities, and population within 12 km of a road (see Anmex 1). 7  Table 11.3. Within-Province Description of Poverty (Ranks are made in Parendeses). Average Per Capita Cmronic Cereals Percent Score (kgs/cap) Underweight Cutrnb 68.6 530.7 21.6 (9) (7) Sereje 52 i93.3 27 Kabwe Rural 68 335.4 28 Kabwe Urban 70 267.8 19 Mumbwa 73 438.3 16 Mkushi 80 1418.5 18 {QVg!jIt 56.3 54.9 19.4 (5) (9) Luanshya 44 15.3 19 Kitwe 46 3.7 16 Mufulira 46 8.0 16 Kalulshi 60 14.9 24 Chflilabombwe 64 11.4 19 Ndola Urban 67 5.9 14 Ndola Rural 67 225.6 28 4stetn 60 415.6 31.7 (8) (3) Chama 47 148.4 30 Petauke 54 364.6 36 Chipata 58 333.0 31 Ka4ete 63 462.8 32 Lundazi 66 559.3 31 Chadiza 72 625.4 30 Luw.uh 34 38.6 36 (1) (1) Kawambwe 29 35.6 42 Samfya 31 26.4 39 Mwense 33 19.4 39 Nehienge 33 18.3 36 Mansa 44 115.0 24 launka 57 125.4 22.5 (6) (6) Luangwa 55 117.2 23 Lusaka Rural 59 133.5 22 8  Table 11.3. A Geographical Description of Poverty (Continued). Average Per Capita Chronic Cereals Percent Score (kgs/cap) Underweight Nor_he_m 43.6 138.9 28.7 (4) (4) Kaputa 28 14.1 42 Chilubi 35 43.5 34 Luwingu 42 81.9 26 Mpika 42 144.2 30 Mporokoso 43 98.2 26 Chinsali 45 182.1 29 Kasame 49 109.7 21 Mbala 49 191.7 26 Isoka 59 384.3 24 Norlhwestern 40 77.5 34.2 (3) (2) Mwinilunga 34 16.2 42 Solwezi 39 58.2 35 Kasempa 39 160.0 35 Chizela 41 122.9 31 Zambezi 43 39.4 31 Kabompo 44 68.4 31 Souer 57.7 339.8 21.5 (7) (8) Gweinbe 36 86.2 29 Livigstone 49 96.5 21 Namwale 57 292.8 19 Alonze 61 318.3 18 Mazabuka 62 491.8 25 Choma 66 423.6 18 Kalomo 73 669.6 18 Western 36.3 70.08 28 (2) (5) Lukuu 31 44.1 32 Kalabo 33 30.2 28 Senenga 34 23.5 29 Sesheke 36 103.2 28 Kaoma 40 179.4 30 Mongu 44 40.1 21 Source: Caldwell. Numbers in parentheses are rankings by province from bad (1) to better (9) 9  significantly along with the depth of household poverty. The Jensen and Luckett study found that using the one half mean income poverty level, 4' percent of rural households are poor while for large rural families (those with more than 4 members) the prevalence of poverty increases to 50 percent. The CSO study4 was consistent with these observations; the likelihood of being poor increases as the number of people in the household grows. Interestingly, both the CSO and the Cogill and Zaza studies found a negative relationship between household size and the prevalence of either stunting or undernutrition of children. Thus, there is an inconsistency between the findings using income/expenditures and those using nutritional outcomes (anthropometry). There are a number of plausible explanations for this phenomenon, such as increased attention to children by older siblings (nutritional status should be closely linked to birth order in such a case), returns to scale in feeding and other nutrition inputs, etc. No study examined this discrepancy in a systematic fashion. More research on this issue is needed. Headship Households headed by females and by elderly of any sex were generally found by the studies to be more likely to be poor than households headed by younger men (Table II4). Jensen and Luckett found that households headed by mamed women are no more likely to be poor than the general popuation. However, other groups of households headed by women were significantly above the mean poverty level. Households headed by widowed and divorced women are very likely to be severely poor, and the contnrbution to total rural poverty of these households is substantial. The CSO study found that female-headed households are more likely to fall under both the moderate and severe poverty lines than are male-headed households (Table 11.4). The World Bank (1993c), also citing the PS, reported that 88 percent of female-headed rural households are moderately poor as compared to 81 percent for male-headed households. The figures are more stark for the emremely poor. Nationally, 70 percent of female-headed households are very poor (contributing to 23 percent of total poverty), while only 55 percent of male-headed households are extremely poor (76 percent of total poverty). Cogill and Zaza reported that children in households headed by females are more likely to be stunted than those from male-headed households, although rates of undernutrition are equal across headship (Table II.4). One of the authors' explanations for this finding was the smaller agricultural holdings of female-headed households. Once land holding size is controlled for, the relationship between headship and nutritional status disappears. The CSO (1993a) study also found a strong neaive relationship between female headship and nutritional status. Siandwazi (p. 25),reported that females who head households are likely to be much less educated than males, although no breakdown of poverty by household headship was provided in the study. 4The CSO report, however, did not break down the incidence of poverty by household size for rural areas alone. The study found an increasing prevalence of poverty as household size increases for all Zambia. 10  Table 1I.4. Prevalence of Moderate Poverty by Sex of Household Head. % Malnourished2 HEIS' PS2 Stuning Undernourished Male-headed .79 .67 51 24 Female-headed .82 .77 58 23 Elderly head .78 .83 NA NA Young head .79 .69 NA NA 'Source: Jensen and Luckett, elderly are defined as those older than 55 years. 2Source: CSO (1993a). Elderly are defined as those older than 50. Poverty is for all Zambia, not just mral areas. Education Jensen and Luckett found a relationship between education of the household head and household income. Secondary education significandy reduces poverty in both rural and urban areas. The effect is higher in urban areas and in rural areas prinary education is not associated with a significant decline in poverty. Using the one-half mean income poverty level in rural areas, 23 percent of those with secondary educations are poor and 6 percent of those with higher educations are poor. These mmbers are significantly below the average rural poverty of 42 percent. The CSO (1993a) report did not examine relationships between education of any member of the household and poverty, but did break down nutritional status by mother's education (presumably based on the theory that increases in mother's education translate to better care for the children, or better awareness of nutritous feeding practices). The study found that malnutrition (stunting and undermutrition, and to a lesser extent, wasftg) declines dramatically with education of the mother. For all Zanbian children whose mother has no formal education, the prevalence of stunting is 44 percent. This number declines to 41 and 30 percent as education of the mother increases to grades 1- 7 and 8-12, respectively.5 Undernutridon falls from 28 to 22 to 18 percent as the level of mother's education grows to more than secondary education. 5Children with mothers with 'A' Level and college education and higher have rates of stunting of 39 percent, but the cell size is extremely small, and the difference is not likely to be statistically significant. 11  Etployment Jensen and Luckett also analyzed the relationship between employment type and poverty. They found that the rural self employed (overwhelmingly farmers) have the highest poverty rates and contribute most to rural (and urban) poverty. Using the one-half mean income poverty level in rural areas, 46 percent of those self employed are poor. Such a level constitutes 85 percent of total rural poverty. Cogill and Zaza stated that off-farm income may be critical to some households, especially those headed by women. Unfortunately, the authors provided no hard data on this issue and only state that the income may be critical in creating household food security. Agricuture There was surprisingly little effort in any report to link land-holding size with poverty. The World Bank (1993b) found that poverty rates for small, medium, and large farms are, respectively, 77.9, 55.2, and 34.6 percent. The actual size of the farms was not reported in the World Bank study and land quality was not controlled for. Chipwende, et al., using the PS, reported average holdings for the extremely poor, the moderately poor, and the non-poor to be 2.62, 2.49, and 2.50, respectively. These data did not include rented land but the results still must be considered surprising. If the numbers are to be believed, there is no association between landholding size and poverty, and more analysis of this finding is clearly needed. The CSO report showed that rural small-scale farmers are the socio-economic group most likely to be poor and extremely poor. The prevalences of moderate and extreme poverty fall for the larger-scale farmer groups. Rural non-agricultural households have the second-highest rate of moderate and extreme poverty behind the small-scale class. Rural non-agricultual households had a rate of 68 percent falling into the moderate or worse poverty group, with 59 percent being extremely poor. Cogill and Zaza were unable to find a conclusive link between household landholding and malnutrition of children. Although height-for-age and weight-for-age increase significanly for girls whose families plant more than 10 hectares, there is no relationship between boys' nutritional status and land in production. For households with fewer than 10 hectares in production, there is no relationship between girls' nutritional status and planted land. v. General Observations on Rural Poverty and Food Insecurt The studies produced some general observations on the causes of rural poverty and food insecurity. Some of these observations are summrized below. The Role of Maize and Agncutue Zambians depend heavily on maize for their caloric requirements. Maize provides up to 70 percent of calories consumed (Siandwazi). Maize is particularly drought prone, and has historically 12  had very unstable levels of production in Zambia. Hybrid maize is less tolerant of water stress than local varieties. Farmers typically plant hybrid maize only after local maize is planted since there are well-documented labor constraints in rural Zambia. Planting local maize first is a strategy to miniimize risk of suffering a drought loss. However, late planting increases hybrid maize yield risk, and lowers yields. Several reports (World Bank 199Ma; World Bank 1993b; Chipwende et al.; Siandwazi) stressed that Zambian agriculture is such that access to vital inputs, particularly traction power, is the main constraint to production. The World Bank (1993b) stated that a typical family can hand-hoe only 2 hectares while the output from S hectares is needed for food security. It was also widely reported that most smallholders do not have access to adequate traction power, creating reduced yields and poor timing of farming operations. The inability to provide timely delivery of credit, seeds and fertilizer, transport, and payment for crops has been due, in large part, to policies that gave too much control to parastatal organizations that were given unreaistic mandates. Subsidies for credit, seeds and fertilizer, and pan-teritorial and pan-seasonal pricing for maize led to Wigh budgetary outlays that bankrupted the government and led to an uneconomic spatial distribution of maize production. Furthermore, the focus of maize production by all support services (credit, input and output markets, research and extension) led to a viral neglect of other crop and livestock activities. These issues are all analyzed in greater depth later in this report. Gender and Poverty Rural households with female heads are more likely to be poor and have inadequate access to land. These households are less likely to have land title due to divorce and inheritance customs, and are further discriminated against in credit markets (Siandwazi). They do not generally possess the physical strength necessary to hand-hoe an area sufficient for subsistence (World Bank, 1993c). Woman farmers typically achieve lower yields since they do not rotate crop land as fiequently as male farmers because cultivating new land by hand is labor intensive and arduous work. A frther problem for women and children in Zambia is their poor social status and the consequent low priority their mnrition receives (World Bank, 1993a). Siandwazi stressed this low social status in concert with the poor educational status of women as a factor in Zambia's high child malnutrition statistics. Poor fe4ing and weaning practices along with low stats malke malnutrition more prevalent for women and children. The conclusion that gender-based nutritional status differentials exist, however, is not universal. Both World Bank (1993a) and Cogill and Zaza found that female children have better numtritional status than male children. Impaas of Adjustent Only one of the studies examined explicidy discussed the impact of structural adjustment on rural poverty (World Bank, 1993b). Structur adjustment may improve the terms of trade for farmers in the long run. In the short run, hardship can be expected in rural and urban areas. The expected improvement in rural terms of trade will not affect all farmers equally as transport costs in isolated areas dampen price increases. Transport costs are expected to rise becuse of the demise of 13  publicly-supported maize marketing and because tradeable input prices (fuel, spare parts, etc.) will rise with adjustment (World Bank, 1993b, p. 74). All farmers will face increased input costs as a result of adjustment. Real interest rates can be expected to rise making it harder for marginal farmers to secure and service loans (World Bank, 1993b). Nevertheless, higher real rates of interest may increase access to formal sector credit by restoring viability to rural financial markets. The World Bank (1993b) noted that a coping strategy of the urban residents is to send away family metnbers and to reduce remittances to their home village. If large transfers are taking place then the rural poor will be hurt as urban incomes decline with adjustment. - The current market liberalization policies will lessen many of the policy-induced biases discussed above, but a question remains as to the responsiveness of indihidual farmers to new economic realities. For example, the impact of market liberalization on outlying regions is unknown, but it is obvious that transport costs for previously subsidized outlying regions will increase. vi. SuMmary of Studies and Recommendations The World Bank (1993b) asserted that Zambian poverty fits generic models of poverty quite well. They found low levels of human capital due to low levels of education and poor health care resulting in low incomes that perpetuate poverty. Poor access to real assets due to unfavorable land ownership laws, underdeveloped rural credit institutions, and price discrimination against agricultural products have also resulted in low returns on assets. The agricultural sector is further limited in its ability to respond to market signals by high transactions costs due to underdeveloped infrastructure and monopsonistic marketing channels. Cogill and Zaza noted that increased food production in rural Zambia is not a sufficient condition for reduction of malnutrition. They then stressed that the market reforms might not raise production or incomes, especially for the most food-insecure groups. Poor sanitation, poor water quality, infection, and poor health services are strucural problems that need to be addressed concurrently with pricing, marketing, and macroeconomic reforms. The reviewed studies suggested a wide range of reforms, policies, and programs to reduce poverty. This range includes direct and targeted interventions, and pro-poor policy reforms. Direct and Targeted Interventions Jensen and Luckett recommended direct programs to help Zanbia's poor. They concluded that targeted assistance to improve living conditions will have a greater impact in rural areas where poverty is most prevalent. They argued against direct food assistance since most of the rural poor produce their own food. They recommended programs that develop off-farm job opportmities in rural areas and that improve agncultural practices. They also recommended education above the primarily level as a priority. Siandwazi stressed the value of government services and programs to Zambia's poor. She showed that nutrition education (particularly for women), health services, sanitation and clean drinking 14  water supply decrease malnutrition; the relative effectiveness of each measure in reducing poverty was not, however, measured. Chipwende, et al. also emphasized programs to protect Zambia's poor. They stressed the need for basic social services, iacluding health care, education, and safe drinking water. They advocated public employment schemes to reduce unemployment. Siandwazi noted features of successfWl programs to target assistance to Zambia's poor. She found that nutrition education programs work because they have a personal approach that has a powerful effect on participants. She stated ta programs to enrich and fortify foods have been successful. Siandwazi recommended that the High Energy Protein Supplement program (HEPS, administered by the World Food Program) be extended to feed children in schools as well. Siandwazi found food for work prograns beneficial, particularly because women tend to be integrated into these programs. Like the World Bank (1993a), Siandwazi found that NGO- administered programs benefit from less bureaucracy and greater flexibility. She also recommended targeting of poor regions based on the seasonality of malnutrition. The World Bank (1993a) recognized the need for direct assistance to Zambia's poor, A number of programs that were supported by the donor community (particularly through the World Food Progran) but administered by local NGO's for drought relief were found to work very well. Such programs should be continued for targeted assistance to Zambia's poor and for infastructure improvement. Some examples of the programs are Program Against Malnutrition (PAM), Programn to Prevent Malnutrition (PPM) and Program for Urban Self Help (PUSH). While recognizing the need for targeted assistance to Zambia's poor, the study is pessimistic and not very specific about how aid should be targeted. It noted that households without able-bodied members and male heads should be targeted and that targeting snould be undertaken at the local level. However, it was pessimistic about this process due to bureaucratic probler s. Ag7iCua ral Poicy Chipwende, et al. recognized the need to develop a free market system but still see a role for subsidizing crops that the poor consume. They recommended that farmers be provided subsidies to cultivate these "self-targeted" foods. The authors stated that the rural poor should be provided title to land so they will have access to credit. Chipwende, et al. stressed that rural credit access should, in general, be improved and that such improved access might be accomplished by a targeted credit scheme. Cogil and Zaza recommended the following policy and program intervenions: Low cost technology for food production, storage, and processing (without being more specific); better availability of inputs; easier access to credit; minhiization (?) of marketing costs; and, improvemens in inu cture. None of these recommendations for interventions resulted directly from their analysis of malnutrition data rather they were developed through literature reviews and expert discssions. They also recommended a policy of expanding off-farm employment and income generation in rural areas. Once again, they were not specific about the form such a policy should take. finally, improvements in health service delivery and education were recommended. 15  The World Bank (1993a) strongly recommended that food security become a major focus of the government and stressed that the Policy Analysis and Coordination unit should be used to make government food security policy coherent. Many of the recommendations involve reducing the role of government and predicting the positive effects of decreased government interference in markets. The report (World Bank, 1993a) stressed agricultural policy as a vehicle for reducing rural poverty. They stated that hybrid maize production should still be encouraged due to its much greater productivity but with provisions that will make local processing and storage viable. The role of local hammermills is seen as critical to agricultural development. Without pan-seasonal pricing, local storage wiUl become profitable and millers are likely to become storage agents. The Bank also sees a role for the millers in supplying credit and inputs to farmers. This reconunendation represents a call for govement to stay out of markets and assigns to government the role of extending storage and milling technology. The World Bank (1993a) also stressed that current hybrid maize varieties are not satisfactory. Varieties that mature in a shorter period of time are needed to address labor shortages and drought- related yield variability. The authors stated that GRZ needs to take a role in diversifying agriculture away from maize production. The lifting of maize subsidizadon will decrease incentives to plart maize but the government needs to ensure that alternative crops such as sorghum, millet, cassava, sunflower, and soybeans receive research and extension priority so that local farmers will have alternatives to planting maize. This process can also be facilitated by encouraging the development of markets for alternative agricultural products. The report found the government programs that monitor production levels for advanced waning of droughts to be critical. It recommended that the government establish a 3 month supply of maize requirements and restore its macroeconomic health so that it can import food during times of extreme need. Reliance on food aid would be a mistake since current pressures in developed countries to limit agricultural surpluses will reduce future non-emergency food aid. In the agricultural section of the World Bank (1993b) report, the need to utilize Zambia's plentiful land and restructure the land ownership system was emp>hasized. This restructuring is perceived to be vital since secure ownership should make agricultural capital investments less risky. All of the studies mentioned the need for tenure security to assist the rural poor, but there has been little or no hard research supporting a link between tenure and poverty. Authors continuously mention that secure tenure is needed for access to credit, yet few seem to recognize that land tenure reform is a difficult means of guaranteeing collateral. The report (World Bank, 1993b) recognized that in the short run Zanwbia's poor in rural and urban areas will be adversely affected by the adjustment program. The urban poor will be hurt by rising food prices and fewer public sector job opportunities. The possible negative effects of adjustment on the rural poor have already been discussed. However, it did not provide detail on recommendations to alleviate these effects. The report stressed that donor assistance in the past was poorly coordinated and hence inefficient. This assistance should be allocated not by donor interest but by Zambia's development needs. The report recommended that health services should be provided closer to peoples' homes with teriy facilities receiving a higher proportion of resources. Efforts to make social services 16  more effective need to include decentralization of the Zamnbian bureaucracy. The report stressed that human capital needs to be improved through education. Swnmary of Recommendations There is general acceptance that subsidies to alleviate poverty need to be targeted. However, the specific means of targeting is never presented in detail. It is recognized that in rural Zambia, female-headed households and households without able-bodied nmnbers should be targeted but the mechanics of this targeting are not spelled out. The World Bank (1993b) was skeptical that targeting can be accomplished with much efficiency. A recurring theme of these reports is the need for land tenure reform so that poor farmers have collateral allowing them to gain access to credit. There is, however, no mention of alternative means of securing loans. The World Bank reports stressed the benefits to the agricultural sector from privatization though improved terms of trade. The reports all leave open the question of the net impacts of liberalization on the rural economy. Little is known about the likely evolution of rural input and output markets. Clearly their evolution will affect the poor. The geography of poverty is such that the poor are generally found in more isolated rural areas; these are the areas where the ultimate impact of liberalization is most uncertain. Deteriorating infrastructure, cutbacks in budgets for support services, and the general isolation of the poorest areas make it likely that tenns of trade will not increase dramatically in these areas, and they may actually fall. These forces place the poor in extremely vulnerable circumstances and may dictate govermnent interventions. 17  m. ANALYSIS OF PRIORITY SURVEY DATA A) Qekw The Priority Survey (PS)6 data analysis focuses on the attributes of poor and vety poor rural households and on the characteristics of their component members. Cross-tabulations and mean attributes of these families were computed for variables related to location, economic, social, and enviromnental factors. The analysis serves as a basis of classifying and categorizing the rural poor for purposes of policy design. Following the descriptive analysis, regressions were undertaken to control for the multiple influences affecting poverty. The analysis in this section does not address many of the issues surrounding the methodology of measuring the poor. It does not, for instance, compare results using relative and absolute poverty lines, or using income versus expenditures as a resource of well-being. These issues are explored in a number of theoretical and empirical studies of poverty (see, e.g., Alderman and Garcia; Ravallion; Ravallion, et al.). Instead, this study uses generally-favored indicators with the best available data to make poverty comparisons, such as where poverty is greatest, how the poor earn incomes, and other relevant attributes of the poor. The information from the analysis is used later to develop a simple model of rural household behavior. IB) E!l% anld Measreint In contrast to the CSO report (CSO, 1993a), market expenditures and imputed expenditures fiom consumption of home production were used instead of income to measure poverty. Household expenditures correspond most closely to welfare and are thus widely used in poverty analyses (Glewwe and van der Gaag; Grootaert and Marchart; Ravallion). Market expenditures were readily available from the PS instrument. Imputed expenditures on consumption of home production were obtained by subtracng the reported quantities of local maie, hybnd maize, and cassava sold from their respective quantites produced, and multiplying the result by market prices for each commodity.7 Rental expendiures were imputed using a simple linear regression. See Annex I for more information on the imputation procedure. Im-putations of home-consumed producton and the rental value of housing required valid observations for a number of variables. Absent observations and an analysis and deletion of outliers left 3709 valid household observations with 19875 individuals in rural Zambia. *The Central Statistical Office's Priority Survey was conducted in October-December 1991 (see Annex I for further information). The survey was designed to provide a short-term snapshot of the poor for the purposes of policy formulation (Grootaert and Marchant). Because of the dynamic nature of rural poverty (see, for example, Alderman and Garcia), much can be expected to have changed since the 1991 PS rounds. More information on these changes will be found using the 1993 PS, expected to become available in late summer 1994. 7Prices used were 1,200 Kwacha (K) per 90 kIlogram bag of maize, and K1,215 per 150 kg bag of raw cassava. District level prices will be available from the 1993 PS. 18  Once total monthly household expenditures were computed, they were converted into a per capita basis by dividing by the number of adult equivalents usually present in the household. The equivalence scales provided by the CSO were used to create adult equivalencies.8 Poverty Line The poverty lines used correspond to the CSO conventions of Kl,380.5/month/adult equivalent for poor and K963.5/month/adult equivalent for extremely poor households.9 Because of the systematic regional biases in expenditure measurement described in Annex I, poverty comparisons are generally made by province. The systematic biases result not only from problems related to home- consumed quantity measurement and cassava price variation described in Annex I, but also from the lack of provincial cost-of-living indices. The construction of such indices is necessary to ensure the accuracy of future poverty analyses, especially following the demise of pan-territorial pricing of key agricultural commodities. The poverty measures reported in this study include the head-count (P0), the poverty gap (P1), and the severity of poverty (P2) measures. These measures are widely used in poverty analyses (see Ravallion for a description of the indices). C) Prevmjn Information and Comnarions The expenditure shares for all rural Zambian households, broken down by poverty cutoffs, are shown in Table Im. 1. Engel effects are clearly present as food shares increase with poverty. Such effects may be contrasd with the CSO report which shows food shares increasing in urban areas at the same time that poverty measures show significantly lower poverty in these areas (see CSO, 1993a, p. 119, Table 10.2). The increase in food shares in urban relative to rural areas reported in the CSO report is mainly attibutable to not counting home-produced consumption during their measurement of income. In Table m1. 1, shares of clothing, transportation and remittances are higher for the non-poor households than they are for the poor. These results are expected since these items should have the highest expenditure or income elasticities of the expendiue groups considered. The maize shares of food shown in Table M.1 are higher than corresponding shares in the Household Income and Expditure Survey (HEIS), conducted in June 1991 (see Stampley, et al.). There are several explanations for the differences. Pirst, vegetables and frits accounted for 24 percent of the mral food budget in the HEIS. These items, which are mainly home-produced in rural Zambia, are largely uncounted in the priority survey since only home-produced consumption of maize and cassava was measured. Second, alcoholic beverages accounted for 8.5 percent of rural food expenditures in the HEIS; the PS did not measure alcohol expenditures. Finally, the HEIS asked for 'These age-dependent scales were: For a child under one year of age = 0 adult equivalents (AE); child 1-3 years old = .36 AE; child 4-6 years old = .62 AE; child 7-9 years old - .78 AE; child 10- 12 = .95 AE; adult female (13+ years) = .76 AE; adult mate = 1 AE. 'The higher cutoff (signifyig moderate poverty and worse) is denoted AP1 in this report, and the lower cutoff (severe poverty) is denoted AP2. 19  Table I1.1. Expenditure Shares in Rural Zambia by Poverty Group. API AP2 Item Non-Poor Poor Non-Poor Poor All Rural Food .6498 .6992 .6706 .6988 .6912 Housing .1063 .1488 .0998 .1575 .1419 Education .0096 .0173 .0093 .0185 .0160 Medical .0081 .0076 .0086 .0073 .0077 Clothing .1167 .0770 .1148 .0719 .0835 Trnportation .0635 .0361 .0600 .0333 .0406 Remittances .0460 .0140 .0368 .0126 .0192 Maize Share of Food .3754 .4458 .4059 .4449 .4324 Inputed Sha of Food .3633 .5199 .4211 .5215 .4900 Total Observations 708 3001 1126 2583 3709 Note: API is the cutoff for moderate poverty (monthly expenditure per adult equivalent of less than K 1380.5); AP2 is severe poverty cutoff (monthly expenditures per adult equivalent of less than K 963.5). the prior month's expenditures on food; in April and May (the months prior to the HEIS interviews) the maize share of food is likely to be lower than on average since at that time many rural areas are just enateing the maize harvest. The Engel effects shown here are, in contrast to the CSO results, consistent with the findings of Jensen and Luckett (p. 29). Other interesting results appear in Table M.1. First, imputed shares (home-produced cassava and maize consumed as a proportion of total expenditures) are higher for poor than non-poor families. Such a result is expected, and helps provide validation for using expenditures (both imputed and market) rather than incomes especially in measuring poverty gaps. The poor are less dependent on market-purchased food than the non-poor; further analysis of market linkages follows below. This result is consistent with results from the HEIS (Jensen and Luckett, p. 29). Second, food, housing, and education are necessities. The fact that poor households spend higher proportions of their total expenditures on education has important policy ramifications, and is dicussed in detail below. Finally, transportation shares are about twice as high for the non-poor as 20  they are for the poor. This result provides a hint that market access and use of marketing services may distinguish the poor from the non-poor in rural areas; further investigation follows below. Monthly adult-equivalent expenditures in rural areas by province are shown in Table 111.2. Households were divided according to poverty group (API and AP2) based on the computed expenditures and the CSO poverty line. The national distribution of poverty is found in Table 111.3. Sixty eight percent of all Zambians are from households that are classified as moderately poor, while 54 percent are from extremely poor households. The incidence of poverty is much higher in rural than in urban areas. In rural areas, 87 and 77 percent of all people are from moderately and extremely poor households, respectively, while in urban areas, 45 and 28 percent of the people are from moderately and extremely poor households, respectively. Notice that in urban areas, there is a higher proportion of poor between the API and AP2 cutoffs than in rural areas. For each of the indices (PO, P1, P2) the ratio of the API to the AP2 index is much smaller in urban areas than it is in rural areas. This result indicates that the indices are more sensitive to the location of the cutoff in urban areas and confirms that poverty is relatively more severe in rural Zambia. Such a disparity between prevalences of poverty in rural and urban areas is consistent with studies cited in the previous section of this report, yet there is a clear bias in the survey instrument that leads to an undercounting of expenditures on home-consumed production. This bias creates an overestimate of poverty in rural areas relative to urban areas, and makes it difficult to make comparisons across rural and urban areas. Similarly, costs of living are surely higher in urban than in rural areas; failure to account for these differences also leads to overcounting of rural relative to urban poverty. It is impossible, given the survey insumt, to estimate the degree of this bias, but it is important to realize that when the indices are sensitive to the cutoff, they will also be sensitive to the degree of measurement bias. The bias due to undercounting of consumption from production is most pronounced in the depth (P1) and severity (P2) indices. The simple headcount is less likely to be severely affected by the bias, and thus, subsequent breakdowns by poor and non-poor groups are valid despite the bias. The distribution of rural poverty by province is shown in Table M.4. The largest numbers of rural poor are found in the Northern and Eastem provinces, followed by Southern province. Fewest rural poor are found in Copperbelt and Lusaka provinces. The distribution of rural poor can be contrasted with the incidence (P0), depth (P1), and severity (P2) of rural poverty by province, shown in Table m.5. The highest incidences of nrral poverty are found in Westem, Northern, and Luapula provinces. Eastern and Centra provinces, although they have large numbers of poor (Table 111.4), have relatively low poverty incidences becase of their larger overall populations. About 94 percent of the rural residents of Western province are from poor households,'0 and about 88 percent are '0Notice that households are defined as poor and non-poor, but the headcount presented here is a headcount of the people residing in poor households. The headcount of people presented here best measures the overall incidence of poverty since people are poor, not households (Ravallion). 21  Table 111.2. Mean Per-capita (Adult Equivalent) Expenditures in Rural Zambia, by Province (K/Month). Province Number of Households Total Food Central 419 1143 752 Copperbelt 107 995 655 Eastern 733 993 636 Luapula 356 774 518 Lusaka 129 1409 897 Northem 612 817 596 Northwestern 266 671 421 Southern 632 812 499 Western 449 574 388 All Rual 3709 842 565 All Zambia 9873 1584 967 Table 111.3. National Poverty Indices. Total PO P1 P2 Aggregation Households API AP2 API AP2 API AP2 National 9873 .677 .544 .387 .290 .272 .196 Urban 6164 .451 .285 .200 .128 .127 .086 Rural 3709 .872 .768 .549 .430 .397 .291 Notes: API is the higher cutoff, i.e., for moderate poverty; AP2 is for severe poverty. PO is the headcount measure; P1 is the depth measure; P2 is the severity measure (see CSO, 1993a, or Ravallion). 22  Table III.4. Distribution of Rural Poor, by Province. Percentage of All Rural People in Poor Households Distributed by Province API AP2 Central 9.2 8.3 Copperbelt 1.5 1.3 Eastern 18.7 18.0 Luapula 12.1 13.1 Lusaka 3.0 2.7 Norhern 18.4 19.0 Northwestern 7.2 7.2 Southern 16.6 16.7 Western 12.5 13.5 Total 100 100 Notes: Table may be interpreted as meaning, for instance, that 9.2 percent of all people in rural Zambia from moderately poor households arc found in Central Province. API is the higher cutoff, i.e., for moderate poverty; AP2 is for severe poverty. Table m.5. Rural Povety Indices by Province. PO P1 P2 Province API AP2 API AP2 API AP2 Central .828 .662 .467 .343 .319 .221 Copperbelt .761 .601 .432 .321 .298 .210 Eastern .870 .742 .510 .382 .353 .246 Lupula .936 .858 .607 .479 .438 .318 Lusaka .722 .604 .400 .280 .250 .151 Northem .930 .857 .595 .460 .419 .294 Northwestem .920 .792 .589 .474 .438 .334 Southem .876 .775 .559 .438 .409 .307 Western .939 .878 .686 .588 .543 .438 Notes: AP1 is the higher cutoff, i.e., for moderate poverty; AP2 is for severe poverty. PO is the headcount measure; PI is the depth measure; P2 is the severity measure (see CSO, 1993a; Ravallion). 23  from very poor households. The depth and severity (P1 and P2) of rural poverty are greatest for Western, Luapula, and Northwestern provinces (Table 1.5). Northern province has a high incidence of poverty, but the depth and severity of poverty there are not as acute as in the other poor provinces. This result indicates that there is a large percentage of poor households in Northern province who are near the poverty cutoff, and helps validate the use of poverty indices in addition to the simple headcount. The poverty count in Northern Province is sensitive to the choice of cutoff. The geographic ranking of poverty is reasonably consistent with rakiangs from other studies (see Table 11.2), yet some differences emerge. Consistent with virtually all studies, Luapula, Northwestern and Northern provinces have high percentages of poor households. The CSO report showed Luapula with a relatively low proportion of poor, yet the HEIS and the nutrition studies are more in line with our findings of severe poverty in Luapula. The major areas of disagreement between the results in Table mA4 and those from previous studies are found in the results from Eastern and Western provinces. While there are large numbers of poor in Eastern province, we cannot conclude, for example, that poverty there is any worse than it is in Luapula or Northwestem provinces. The high dependence on maize production may, however, make residents of Eastern province more vulnerable to the macroeconomic adjustments taking place. The PS is likely to overestimate the prevalences of poverty in Western province since millet and sorghum are common subsistence foods there. A notable feature of Table 111.5 is the extreme depth and severity of rural poverty for all the provinces, except perhaps Lusaka. Depth and severity indices from other countries tend to be significantly lower than the headcount (PO) index, indicating relatively lower depth and severity in those countries. Results from, for example, India (Ravallion and Datt) and Indonesia (Huppi and Ravallion) show that the magnitude of the P1 measure is typically less than one-fifth of the headcount (PO) index, and the P2 measure is one-tenth to one-thirtieth the size of the PO index. Similar patterns were found in Ghana (Boateng, et al.) and Bangladesh (Ravallion). In Zambia, the P1 and P2 indices are persistenly high, indicating extreme depth and severity of poverty across the entire country. Such a result illustrates the severe inequality that prevails in rural Zambia; the mean per capita expenditure level is well above the mean consumption of the poorest households. Expenditure shares for rural households are presented by province in Table 111.6. The poor spend consistently higher shares of total expenditures on food and education than do the non-poor. A potentidal problem with the expenditure data becomes apparent when examining these shares. The food sbare in Western province, where the overall incidence of poverty is highest, is somewhat lower than in some provinces with lower poverty incidences (e.g., Eastern). This result is counter intuitive; we expect food shares to decline with increasing incomes and total expenditures. The poor in Westem province consume relatively higher quantities of sorghum and millet. Sorghum and millet home production is not available from the PS database, and thus food expenditures in Western province may be undercounted. Similar undercounting occurs in Luapula, Northern, and Northwestern provinces where millet and sorghum is also consumed. Education budget shares are, in both Tables m. 1 and m.6, consistently higher for the poor households than they are for the non-poor households. These results show that people consider education to be a necessity, and also that the increased move toward recovering costs in education (discussed in subsequent sections of the report) may be exposing poor households to increased financial stress. 24  Table 111.6. Rural Expenditure Shares by Poverty Group, by Province. Province 1 2 3 4 5 6 7 8 9 Food Poor .6886 .6290 .7524 .7214 .6950 .7400 .6418 .6208 .6647 Non-Poor .6578 .6488 .6743 .6315 .6278 .7019 .6209 .5730 .6416 Education Poor .0238 .0269 .0142 .0144 .0184 .0131 .0172 .0344 .0108 Non-poor .0044 .0097 .0059 .0076 .0182 .0031 .0075 .0299 .0058 Medical Poor .0050 .0053 .0105 .0073 .0102 .0055 .0015 .0059 .0122 Non-Poor .0060 .0196 .0080 .0041 .0068 .0044 .0040 .0196 .0081 Transportation Poor .0719 .0780 .0310 .0303 .0373 .0326 .0237 .0467 .0268 Non-Poor .0749 .0564 .0561 .0802 .0653 .0648 .0232 .0806 .0319 Matze Share of Food Poor .5000 .4059 .7556 .1489 .4602 .2320 .3499 .6249 .4745 Non-Poor .3533 .2057 .6438 .1932 .1529 .2619 .1962 .4811 .3355 imputed Share (Maize, Cassava, Rental) of Total Expenditures Poor .3463 .2786 .6032 .4472 .3625 .4715 .4610 .4873 .5022 Non-Poor .2256 .1164 .4720 .1600 .1431 .3370 .3379 .2852 .1743 Imputed Food, Share of Total Expenditures Poor .3211 .1896 .5252 .3838 .2608 .4354 .3297 .3832 .3292 Non-Poor .2155 .0911 .3916 .0976 .0682 .2971 .2367 .2808 .1240 Ipded Share (Maize, Cassava) of Food Ependitures Poor .4280 .2775 .6778 .4892 .3056 .5186 .5384 .4694 .4864 Non-Poor .3003 .1346 .6040 .1706 .0884 .4039 .4259 .3857 .1883 Note: 1=Central, 2=Copperbelt, 3=Eastern, 4Luapula, 5=Lusaka, 6=Northern, 7=Northwestern, 8= Southemn, 9=Westem. 25  D) Poverty and Distance to Facilities Distances to public and private facilities are broken down by province and poverty classification in Table 111.7. In Western province, poverty appears to be closely associated with isolation. The poor in Western province are as much as twice as far from rnany basic facilities as are the non-poor." There are long distances from food markets and primary schools in this province, and these distances are larger for the poor. Notice, however, that the poor have equal or better access to hospitals in Western province dtan do the non-poor. The poor in Luapla are relatively more distant than the non-poor from primary and secondary schools, and transportation facilities, while the poor in Northem province are relatively far from food markets and post offices. In Copperbelt, poverty is associated with isolation from markets, post offices and water. Eastern province provides an exception to the finding that isolation is associated with poverty; isolation from any facility except transportation is not associated with poverty in Eastern province. Apparenty, there is a more even distribution of these facilities in Eastern province than in other provinces. In all other provinces, the poor are isolated in one way or another; distance from facilities and poverty are closely associated in rural Zambia. Most of the high-poverty provinces are also sparsely populated and have long distances to public facilities. Even within these provinces, however, there is a distinct relationship between distance and prevalence of poverty. Luapula, on the other hand, is a high-poverty province, with relatively good access to facilities. Even in Luapula, however, there is a stark contrast in access to facilities between the poor and the non-poor. Poor families in Luapula are significantly farther from virully all facilities as are non-poor, with distances to schools, hospitals, and transportation being about twice as great for the poor. While primary schools tend to be widely and evenly distributed in rural Zambia, the distribution of secondary schools is much less even. Northern province seems to be particularly short of secondary schools, although there is no association between distance to those schools and the incidence of poverty. An association between distance and prevalence of poverty dc>s not mean that increased access to facilities in poor areas will necessarily alleviate poverty. There may be a self-selection process occrrig, whereby facilities are constructed in areas where fewer poor live. Alternatively, better-off households might tend to locate near facilities. Or finally, access to facilities might imply more market transactions and, thus higher measured expenditures. None of these explanations could be examined using the PS data. The association betwcen distance and poverty remains strong. The result might be useful as a means of targeting poverty alleviation schmes according to distance to certain facilities. Alternatively, it can be argued that the self-selection process is relatively unimportant, and that proximity to facilities acually reduces poverty. Such a result would lead policyniakers to place facilities in areas with high poverty and poor access to the facilities. The information in Table m.7 "Some care must be made in making inferences based on these data. Since such high proportions of households are described as "moderately poor," the cell sizes for the APl poverty line are extremely unbalanced for the poorest provinces. 26  Table III.7. Average Distance in Km to Facilities by Poverty Group, by Province. Province 1 2 3 4 5 6 7 8 9 Food Market Non-Poor(API) 17.1 4.6 17.4 5.9 6.8 31.4 19.6 20.2 10.0 Poor(API) 16.7 8.4 17.0 8.6 7.1 41.1 20.3 22.9 30.4 Non-Poor(AP2) 15.8 5.2 18.6 7.1 7.2 33.0 17.7 21.4 13.9 Poor(AP2) 17.7 9.0 16.2 8.6 6.8 42.2 21.1 22.8 31.2 Post Office Non-Poor(API) 24.0 5.8 16.2 7.4 7.3 30.5 19.5 22.8 20.2 Poor(API) 21.8 18.5 18.1 10.2 12.3 37.4 20.0 26.8 38.8 Non-Poor(AP2) 23.2 9.8 17.1 8.1 7.7 29.9 18.1 24.2 26.3 Poor(AP2) 21.8 19.3 18.0 10.5 12.6 38.9 20.7 26.9 39.1 Primary School Non-Poor(API) 3.1 3.0 2.3 1.4 3.0 2.5 1.8 3.3 2.4 Poor(AP1) 4.0 1.9 2.6 2.7 2.9 4.4 2.5 3.5 7.2 Non-Poor(AP2) 3.2 2.2 2.5 1.6 3.2 5.6 2.0 3.1 4.1 Poor(AP2) 4.2 2.2 2.5 2.8 2.6 3.6 2.6 3.6 7.2 Secondary School Non-Poor(API) 19.9 13.8 15.3 10.5 9.7 48.5 22.3 21.2 27.1 Poor(AP1) 29.4 17.8 17.9 18.7 13.2 48.8 25.1 26.1 39.5 Non-Poor(AP2) 19.2 15.2 15.2 12.2 9.3 46.8 20.0 22.3 32.1 Poor(AP2) 33.1 17.9 18.6 19.5 14.0 49.5 26.2 26.5 39.5 Hospital Non-Poor(API) 13.4 3.8 7.2 3.1 5.7 16.1 5.0 10.0 20.8 Poor(API) 12.7 4.6 7.9 6.6 7.8 14.4 7.6 14.4 16.9 Non-Poor(AP2) 12.8 4.6 7.7 4.6 5.8 13.9 5.7 11.4 22.8 Poor(AP2) 13.0 4.3 7.7 6.7 8.0 14.8 7.8 14.6 16.3 Transpozaion Non-Poor(API) 6.6 2.2 6.9 3.6 1.2 14.4 14.0 9.9 5.3 Poor(API) 7.8 1.8 8.9 6.2 4.2 18.8 14.0 13.4 18.7 Non-Poor(AP2) 6.3 1.7 6.4 2.9 1.2 17.6 12.7 11.2 11.3 Poor(AP2) 8.4 2.1 9.6 6.9 4.6 18.4 14.6 13.5 18.6 Water Non-Poor(API) .08 .09 .54 .10 .00 .32 .19 .84 .23 Poor(API) .11 .22 .46 .11 .30 .32 .43 .66 .12 Non-Poor(AP2) .11 .06 .52 .07 .01 .32 .18 .92 .16 Poor(AP2) .08 .29 .44 .12 .34 .33 .47 .59 .13 Note: 1=Central, 2=Copperbelt, 3=Eastem, 4=Luapula, 5=Lusaka, 6=Northem, 7=Northwestern, 8= Soutrn, 9=Western. 27  can be used to identify these areas. Prior to targeting these facilities toward "high poverty" areas, however, further analysis is needed to document a causal relationship. Also, an analysis of financial viability of some of these facilities is needed as volume of use is likely to be significantly lower in isolated and poor areas. For the five high poverty provinces (Eastern, Luapula, Northern, Northwestern, and Western), distance to transportation is highly correlated with poverty. Liberalization of food and import markets will thus have a disproportionately strong negative impact on the poor in these provinces as transport costs will inflate the costs of inputs and of marketing production. Although the data are not representative at the district level, district-level comparisons were made to examine intra-province variations in distance from facilities. These comparisons are shown in Table 111.8, where districts are classified as being poor and non-poor depending on whether their prevalence of poverty is above or below the province mean prevalence. In Eastern province there is a stark heterogeneity between districts. Districts with high percentages of poor tend to be more distant from markets, secondary schools, and especially transportation than districts with low percentages of poor. Witiin districts in Eastern province there is no real difference in distance from facilities between the poor and the non-poor. This result suggests that placing some of these services or transportation infrastructure in the isolated poor districts may raise the level of living of a broad sector of the population. In Northern province, a similar picture emerges; access to post offices, primary schools, and secondary schools is district-specific. A different pattern emerges in Western province, where access to markets, transportation, and secondary schools within districts varies between poor and non-poor families. In Western province, for example, poor households in predominantly poor districts are 8 times as distant from transportation as the non-poor households in predominantly non-poor districts. This result reinforces the conclusion that poor households in Western province are characterized by inadequate access to key facilities. It is impossible to compare the results related to distance and poverty with those from other studies as none of the previously-cited studies examined the relationship between distance to facilities and poverty incidences. Caldwell's chronic poverty score (Table 11.2) contained a component measuring the amount of transportation infrastructure by province. His ranking of provincial poverty is very close to ours (compare Tables 111.5 and 11.2), so the transportation component helps isolate some of the effects we note here. Caldwell did not, however, make household-level comparisons, and thus the remoteness results can not be compared. Without controlling for all the influences that interact to deternine poverty, it is impossible to reach conclusions about the relationship between distances to facilities and poverty. Such an analysis is conducted below following a breakdown of poverty by social and economic conditions. E) Social Factors The distribution of poverty by the characteristics of the household head is shown in Table m.9. Female-headed households have a higher incidence, depth, and severity of poverty than male- headed households. At the same time, more poor and extremely poor people live in households headed by males. For example, 80.8 percent of the very poor people (AP2) in rural Zambia live in bouseholds headed by males; the remaining 19.2 percent live in female-headed households. There are important implications of this association between headship and poverty. Interventions that are 28  Table 11.8. Average Distance in Km to Facilities Within Province by Low- and High-Poverty Districts. Mean Disance To Food Post Primary Province Poverty Level Market Office School School Hospital Transporation Central High 20.5 22.5 2.7 36.5 15.1 11.1 Low 13.8 23.4 4.5 21.6 11.0 6.0 Copperbelt High 8.3 20.4 1.7 19.4 5.0 1.5 Low 4.9 6.0 5.4 7.5 4.7 4.7 Eastem High 39.8 25.2 1.6 44.4 6.3 39.8 Low 18.2 17.2 2.7 13.8 9.2 7.3 Lusaka High 2.1 8.6 2.6 17.6 14.2 .3 Low 8.4 11.9 3.9 13.2 6.9 4.6 Northern High 37.4 37.8 4.5 56.1 15.8 16.3 Low 35.1 27.6 2.6 32.3 15.6 17.9 Northwestern High 30.7 37.7 1.3 42.0 1.4 29.4 Low 18.8 17.5 4.5 20.0 9.1 11.8 Southern High 17.7 27.6 44.0 26.0 18.8 9.4 f LOW 28.9 40.0 4.6 39.6 18.3 20.0 Western High 29.6 36.9 6.8 40.8 16.1 22.2 Low 18.7 32.0 5.5 24.9 17.0 7.0 Notes: High= districts with above mean provincial poverty level (using API); low= below mean poverty. Table m.9. Rural Poverty by Characteristics of Household Head. Characteristics P0 of Head API AP2 API AP2 AP1 AP2 Sex Male .877(81.6) .766(80.8) .545 .423 .390 .283 Female .942(18.4) .873(19.2) .632 .514 .471 .356 Marital Status Married .927(86.3) .853(85.7) .614 .492 .454 .341 Single .879(13.7) .767(14.3) .548 .426 .393 .286 Education None .948(34.0) .883(35.9) .663 .552 .509 .396 < m6 years .906(30.4) .795(28.5) .556 .427 .393 .280 >6 years .818(35.6) .683(33.8) .470 .344. .319 .218 Notes: The numbers in parentheses represent the percentage of all poor (API) and very poor (AP2) people who live in households headed by a person of each characteristic class. The other numbers are headcount (P0), depth (PI), and severity (P2) by characteristic using each poverty cut-off. 29  targeted by headship have the advantage tha. the indicator (female headship) is very specific; virtually all female-headed rural households are poot. On the other hand, targeting of female-headed households will miss the large majority of rural poor who reside in households headed by males. Policies and programs designed to eliminate the specific constrints that female-headed households face in generating incomes will pay dividends by raising incomes for households that are universally afflicted with poverty. Later in the paper, different strategies are discussed for eliminating constraints to income generation for female-headed households. Members of households whose head has had no formal education are more likely to be poor or very poor than those living in households where the head has some education. Depth and severity of poverty for households whose head has had no education is significantly more pronounced than for bouseholds whose head has had even fewer than 6 years of education. At the same time, roughly equal proportions of poor families have heads with each category of education. Apparently, substantial rtrns to education exist in rural Zambia. To shed more light on the nature of these returns, education of head along with primary employment of the head are broken down in Table m. 10. Even for heads of households whose primay employment (self-employed or wage-employed) is agriculture, poverty rates decline as education increases. A household whose head is employed in agriculture and who has at least some formal education is as likely to be non-poor as a household whose head has no education but is employed outside of agriculture. Increased education causes a steeper decline in the rate of poverty for the non-agricultural households, but there is still a clear payoff to education for agricultural households. The reus to education found in this analysis are slightly different from the picture emerging from the Jensen and Luckett study that used HEIS data. Jensen and Luckett found substantial returns to secondary and higher education in rural areas, but no returns to primary education. Some further analysis of this point appears later in this paper. Here, we find substantial returns to all levels of education in rural areas, independent of the employment category of the household head. Mean dependency ratios for each group of poor and non-poor households by province are provided in Table M.I1. In each province, the proportion of dependent family members is positively associated with poverty, although the differences in some of the provinces such as Northern, Northwestern and Central are small. There is, in general, a positive association between household size, whether on a head-count or adult-equivalent basis, and poverty (Table HI. 12). This result is consistent with other studies (Jensen and Luckett, CSO, 1993a), yet stands in interng contrast to nutrition studies that generally find a positive association between household size and nutritional status of children (Cogill and Zaza; CSO, 1993a). A curious paradox results; larger families earn and spend less per person than smaller families, yet have children who are better nourished. More research is needed to examine why family size is positively related to nutritional status, but negatively related to average expenditures. 30  Table M.10. Distribution of Rural Poverty by Education and Principal Occupation of Head. PO P1 P2 Head's Head's Primary Education Employment API AP2 API AP2 API AP2 None Agrculture .951 .889 .670 .562 .517 .404 Other .892 .861 .602 .480 .447 .338 * = 6 years Agriculture .909 .798 .555 .426 .392 .281 Other .878 .690 .558 .450 .423 .333 > 6 years Agriculture .838 .715 .489 .360 .333 .229 Odier .609 .458 .308 .210 .195 .124 Table HI. 11. Dependency Ratios by Rural Poverty Group, by Province. Poverty Cutoff API AP2 Province Non-Poor Poor Non-Poor Poor All Central .3109 .4303 .3721 .4172 .3986 Copperbelt .2272 .4185 .2798 .5107 .4257 Eastern .3870 .4490 .3815 .4642 .4374 Luapula .3309 .4545 .3519 .4669 .4421 Lusaka .3003 .4289 .3639 .3972 .3842 Northern .3981 .4679 .4263 .4673 .4579 Northwestern .3304 .4178 .3482 .4233 .4067 Southern .3809 .4714 .4103 .4734 .4569 Western .4016 .4387 .4351 .4357 .4357 Note: Dependency ratios are calcuated as the ratio of non-working family members (aged 14 and under, and 60 and over) to total family members. 31  Table 11. 12. Average Rural Household Sizes by Poverty Group, by Province. Total Size Adult Equivalents Number of Adults Number of Children of Household in Household in Household in Household Province Non-Poor Poor Non-Poor Poor Non-Poor Poor Non-Poor Poor Central 4.23 6.08 3.35 4.79 2.36 2.77 1.87 3.31 Copperbelt 4.36 4.77 3.60 3.75 2.51 2.25 1.85 2.51 Eastern 3.66 4.87 2.93 3.92 2.10 2.27 1.56 2.61 Luapwula 3.43 4.51 2.53 3.76 1.78 1.97 1.65 2.54 Lusaka 4.36 5.63 3.64 4.73 2.49 2.69 1.87 2.94 Northern 3.12 4.88 2.44 3.89 1.60 2.12 1.53 2.76 Northwest 3.66 4.74 2.91 3.96 1.98 2.13 1.68 2.62 Southern 5.51 6.58 4.44 5.35 2.64 2.75 2.87 3.83 Western 3.18 3.97 2.69 3.37 1.68 1.90 1.50 2.07 Children = < 18 years old. Note: Poverty cutoff used to distinguish between poor and non-poor households is API (K1380.5/M/AE). A profile of the relationship between school atendance by school-aged children and the poverty indices is found in Table m. 13. There are higher prevalences of poverty among those households with children of school age who do not attend school than for those with school -aged children who do attend. Depth and severity of poverty is also higher for the families of non-attenders. A pattern of persistent poverty appears when combining the information in Table m. 13 with that in Table mI.9. Poor children, following patterns similar to those of the heads of the households where they reside, are less likely to attend school than children from non-poor households. If improved human capital is a necessary condition for escape from poverty (and education does seem to be, at least, associated with poverty), then the data indicate that poverty may persist from generation to generation in the poorest households. The earlier findings showing that poor households spend higher shares on education when combined with these results are informative, and suggest that the costs of schooling (both out-of-pocket costs and opporunity costs) may be forcing the poorest households to withdraw their children from school. 32  Table m. 13. School Attendance by School-aged Rural Children by Poverty. API AP2 PO Pi P2 PO P1 P2 Poverty Index by Attendance Attending .878 .539 .379 .840 .507 .359 Not Attending .935 .624 .469 .775 .412 .268 API AP2 Percentage of Poor Children Attending by Poverty Cutoff Poor .526 .521 Non-Poor .684 .628 F) Economic Factors i) E WI plment Information on the relationship between mrual poverty and the employment of the household head and spouse is found in Tables [1. 14 and 11. 15, respectively. There is a clear relationship between the employment of the head and all three poverty indices. People from households with an unemployed head are more likely to be poor and very poor than those from households with an employed head. This result must be 'lempered with the knowledge that the heads of virtually all the poor and non-poor households reportad working (Table m. 16). Thus, although the heads of poor households are more likely to be unemployed than are heads of non-poor households, the large majority of the rural poor are working, and most of them work in agriculture. In rural areas, agricultural employment is almost always available for people who lack other opportunities. There is very little evidence that the employment status of the spouse is related to household poverty; poverty indices are virtually identical for those households with an employed spouse compared to those with an unemployed spouse (Table 1.15). There is evidence that when the household head has a second job, the household is less likely to suffer from poverty, yet those households with a spouse who holds a second job are more likely to be poor (Table EII. 16). Rural poverty and prinary head employment in agriculture go hand in hand. Eighty-nine percent of the poor and 90.4 percent of the extremely poor households, respectively, have teir head employed in agriculture. There is no close correspondence between type of employment of the spouse and poverty. 33  Table 111.14. Rurl Poverty Indices by Employment of Head. API AP2 PO P1 P2 P0 P1 P2 Head Employed? Yes .869 .545 .392 .766 .425 .287 No .922 .646 .508 .832 .543 .410 Primary Employment Agiculue .898 .571 .414 .799 .449 .305 Other .689 .381 .255 .557 .274 .172 Head Second Job? Yes .854 .509 .354 .733 .383 .249 No .872 .552 .399 .772 .433 .294 Table M. 15. Rural Poverty Indices by Emnployment of Spouse. API AP2 PO P1 P2 P0 P1 P2 Spouse Employed? Yes .862 .530 .379 .745 .410 .275 No .849 .520 .367 .743 .397 .367 PFimary Employment Agriculture .870 .538 .385 .755 .416 .280 Other .718 .400 .277 .398 .297 .199 Spouse Second Job? Yes .860 .594 .456 .784 .497 .356 No .862 .526 .374 .742 .404 .270 34  Table 1I. 16. Employment of Head of Household and Spouse by Rural Poverty. API AP2 Percent Poverty Group With Non-Poor Poor Non-Poor Poor Head Employed 97.9 96.2 97.5 96.1 Head Employed in Agriculture 68.3 89.4 74.6 90.4 Head Holding Second Job 17.8 15.5 18.3 15.0 Spouse Employed 78.1 79.7 79.7 79.4 Spouse Employed in Agriculture 88.4 95.8 90.2 96.4 Spouse Holding Second Job 7.4 6.5 5.9 6.8 More information on the occupations of rural Zambians is provided in Tables m. 17 and 11. 18. Clearly, the wide majority of rural households are headed by someone whose primary employment is agriculture,12 including 94 percent of the female-headed households, In fact, of those households whose head is employed in agnculture, a higher proportion is headed by females than for the rural population as a whole (Table m. 17). As the education of the household head increases, there is a progressively lower dependence on agricultural employment. Professional and clerical services are the two employment categories in which most of the better-educated rural household heads work. Most female heads of households who are not employed in agriculture work in clerical services (Table M.19), and the majority of these are traders. Male household heads who work in clerical services are widely distributed among a variety of occupations (Table 1.20). The occupation characteristics of heads of households whose prinry occupation is not agriculture were examined within the following broad patterns emerging. Non-poor professional women are exclusively employed in education, and among the poor households headed by professional women, the majority are employed in education, although 41 percent are traders (mostly retail). Seventy two percent of the professional male heads not employed in agriculture work in education. Of those female heads of households who work in production, virtually all work in manufacturing of foods and beverages (100 percent of the non-poor females); mral beer-brewing accounts for alnost all the non-agricultural production activities of women heads of household. 12Note that the category "agriculture" includes fishing, forestry, and wage employment on other farms. 35  Table 111.17. Percentage Distribution Employment Categories by Characteristics of Rural Household Head. Employment Category Male-Headed Female-Headed Education of Head of Household Head All Rural Households Households 0 <6 6+ years Agriculture 86.7 84.8 93.5 92.1 91.5 78.8 Professional 3.0 3.4 1.6 1.1 1.0 5.9 Clerical Services 5.1 5.8 2.8 3.6 2.3 8.2 Produedon 3.7 4.2 1.7 2.4 3.9 4.6 Other 1.5 1.9 .4 .7 1.3 2.4 % Male Female Of those in Headed Headed 0 <6 6+ Agriculture 77.1 24.0 37.1 26.9 36.1 Professional 88.4 11.6 12.4 8.8 78.8 Clerical Services 87.9 12.1 24.8 11.6 63.6 Production 89.9 10.1 23.4 26.9 49.7 Other 93.6 6.4 16.1 21.7 62.2 AU Households 77.8 22.2 34.9 25.5 39.6 Second Job Characteristics Employment Category AU Male Female of Head (2nd Job) Rural Headed Headed Agriculture 46.5 55.8 7.4 Professional .7 .9 0 Clerical Services 15.1 13.7 20.9 Production 31.3 24.1 61.6 Other 6.4 5.4 10.5 N= 261 211 S0 Agricultral development is, thus, a necessary condition for the sustained reduction of poverty. Viraly all rural poor households are headed by a person whose primary source of employment is agriculture. A large majority of the households whose head's primary employment is agriculture is poor (Table M.14). Women who head households are more frequently employed in agriculture than are males; therefore development of small-scale agriculture on a gender-neutral basis world disproportionately benefit this vulnerable group. In practice, as will be discussed in the following section, agricultural programs have been heavily biased toward male-headed households. The wide variety of activities practiced by heads whose primary job is not agriculture makes it difficult to focus poverty reduction efforts on a specific economic sector outside of agriculture. More than 90 percent of the poor households have a head whose primary employment is agriculture. Targeting based on employment status of the spouse would be much less precise.  Table m. 18. Percent of Rural Poor and Non-poor Households with Children Employed. API AP2 Age of Childrenl Type of Employment Non-Poor Poor Non-Poor Poor Age 14-18 Any Employment 33.0 42.5 38.8 42.0 Paid Employment 4.9 7.9 6.6 7.8 Family Enmployment 28.1 36.4 32.2 34.3 Age 6-14 Any Employment 8.7 17.0 11.2 17.5 Paid Employment 1.9 1.3 1.6 1.3 Family Employment 6.8 15.7 9.5 16.2 The percentage-. of poor and non-poor families whose children under 18 who are employed are shown in Table M..18 For each age group, poor children are much more likely to work than are the non-poor. Most of the difference in employment patterns between children in and non-poor families can be traced to the use of children and young children as unpaid family laborers. The reported rate of employment of very young children in unpaid family jobs by poor families is nearly double the rate for non-poor families. This information is consistent with the school attendance rates reported in Table 1.13; poor families, perhaps due to farm labor constraints, but almost certinly due to economic conditions, tend to remove their children from schools to work in unpaid family jobs. When added to the higher dependency ratios found among poor families, these results provide further evidence of a poverty trap, whereby poor families have high fertility rates and produce children whose opportunities to remove themselves from poverty are limited by economic circumstances. ii) Household Assets The rates of ownership and use of different types of household facilities (drinking water, lighting and cooking sources, etc.) for poor and non-poor families are shown in Table m. 19. The poor are less likely to receive their drinkng water from a protected well and ate more likely than the non-poor to use a surface water source. Poor households are also less likely to use kerosine for ligbting. Both mrual poor and non-poor families are highly dependent on collected wood as a source of cooldng fuel; use of charcoal increases only slightly as poverty declines. Reliance on income growth as a means of reducing pressures on the forest (wood for cooking) does not show much promise. Observations and interviews from the field suggest that fuelwood collection for use in rural households is not a major source of stress on Zambian forests. Charcoal production for urban consumers is, on the other hand, associated with increased deforestation. More is said about fuelwood and deforestation in subsequent sections of this report. Whatever stress 37  Table m. 19. Percentage of Rural People Using Different Household Facilides, by Poverty. Poverty Measure API AP2 Facility Non-Poor Poor Non-Poor Poor Water River, Lake 25.0 39.7 27.8 40.9 Protected Well 24.1 14.9 22.9 14.0 Unprotected Well 35.4 36.8 35.3 36.4 Public Tap 5.9 3.4 4.5 3.5 Own Tap 2.9 0.3 1.7 0.3 Other 6.7 4.8 5.9 4.8 Lighting Kerosine 86.9 78.0 87.4 76.6 Electricity 8.2 2.1 6.2 1.8 Candle 1.2 0.2 0.9 0.2 Other 3.6 19.7 5.4 21.4 Cooking Fuel Collected Firewood 82.1 92.3 87.7 91.9 Purchased Firewood 4.1 1.4 2.7 1.5 charcoal 9.9 5.4 7.4 5.5 Kerosine -.- 0.1 -.- 0.1 Gas -.- 0.1 -.- 0.2 Electricity 3.6 0.1 2.0 0.1 Crop Residues 0.2 0.3 0.1 0.4 Other 0.2 0.3 0.2 0.3 Toilet Flush 10.6 5.9 8.6 5.9 Pit Latrine 57.4 49.7 56.8 48.8 Bucket -.- 0.2 -.- 0.2 Aqua Privy 0.2 0.3 0.3 0.3 Other 31.8 44.0 34.3 44.9 House Owned 79.8 93.3 84.0 93.8 Rented 12.9 1.6 8.2 1.5 Free of Charge 7.2 4.6 7.2 4.2 Other 0.1 0.6 0.5 0.5 38  Table m.20. Profile of Rural Asset OwnersWhip by Different Poverty Groups. Percent Percent Owning Overall Poor Owning Who Are poorb Percee Asset Owning API AP2 AP1 AP2 Bicycle 26.9 24.1 21.5 81.4 63.5 Car 1.7 1.0 1.1 53.3 51.3 Boat 2.8 2.8 3.0 90.2 85.9 Motorcycle 0.5 0.3 0.1 51.1 35.6 Tractor 0.7 0.5 0.4 72.6 48.7 Sprayer 11.6 9.7 8.2 76.3 56.7 Hammermill 1.4 1.1 0.8 73.6 48.4 Handmill 3.9 3.5 3.3 77.7 64.6 TV 1.1 0.4 0.4 49.7 49.5 Radio 27.7 23.7 21.0 76.3 60.4 Refrigerator 1.4 0.9 0.7 53.4 35.8 Plough 25.2 23.8 22.4 83.7 70.0 Colunmns are the percentage of poor households (classified according to poverty cutoff API and AP2) who own each asset. bcolhns are the percentage of all rural households ownig specific assets who are poor (according to each cutoff, API or AP2). demand for fuelwood by rural households does place on forests will not be alleviated by incos e growth. The high reliance of the poor on 'other' sources of toilet facilities provides some evidence of sanitation problems, but more information on the nature of the 'other' category is needed. Asset ownership by poor and non-poor households is presented in Table 1.20. Rural households are, in general, not very likely to own these assets. As expected, poverty rates for families who own assets are lower than those for the population as a whole. Households owning assets that can be used in agricultural production (ploughs, tractors, etc.) are much less likely to be poor than households without those assets. The critical productive asset in agriculture is land. The data on landholdings or on land devoted to production by household from the PS were unusable. Instead, poverty indices are broken down by ownerip of any land (a binary variable) in Table M.21. Households with title to land are less likely to be poor than those lacking title. Despite these results, we cannot infer from these data 39  Table M.21. Rural Poverty Indices by Land Titling Status of Household. All Households Head Agriculturalist % Households % Households Province PO P1 P2 Holding Title P0 Pi P2 Holding Title Centr Titled .668 .120 .054 3.4 .668 .120 .054 3.9 Non-titled .834 .482 .328 .852 .489 .332 Cort Titled .438 .311 .223 11.8 .961 .541 .305 2.5 Non-titled .778 .433 .395 .830 .460 .315 Eastern Titled .635 .428 .308 2.1 .635 .428 .308 2.4 Non-tided .870 .516 .359 .879 .528 .372 Luapula Titled .946 .519 .295 1.9 .946 .519 .479 2.6 Non-dtled .939 .613 .444 .982 .656 .221 [usaka Titled .356 .275 .222 7.8 1.00 .799 .680 3.8 Non-titled .718 .381 .235 .955 .507 .319 Nortiem Titled .944 .653 .474 2.5 .942 .658 .481 3.0 Non-dded .934 .607 .432 .956 .622 .441 Northwestern Titled .306 .151 .074 2.3 .306 .151 .084 3.0 Non-titled .939 .599 .445 .963 .661 .548 Soudern Tided .791 .406 .234 4.7 .790 .407 .235 5.4 Non-tidtled .870 .562 .444 .868 .554 .401 Westen Tided .938 .469 .404 1.5 1.00 .700 .613 1.8 Non-titled .936 .685 .541 .978 .730 .584 All Rural Titled .743 .401 .260 3.3 .776 .422 .273 3.3 Non-titled .889 .566 .409 .914 .587 .425 Note: P0 is the head count index, Pi is the depth index, and P2 is the severity index. 40  that land titling would be an effective means of reducing poverty for two reasons. First, there is a low frequency of yes responses to the question of whether title to the land is held. The unbalanced cells make it difficult to make inferences, and call into question the validity of the survey question. Second, ownership of land is more a sign of wealth than a means of raising agricultural productivity. There is a much smaller difference between poverty rates by land titling status for agricultural compared to non-agricultural households. People own land because they are wealthy, they are not wealthy because they own land. Our study, like all of its predecessors, was unable, because of survey design, to examine the link between landholding anid poverty. The link found here between land title and poverty is not conclusive enough to warrant recommendation of an ambitious land titling program. Later in the report some of the links between access to land and income in agriculture are examined using a household model. iii) Linkages to Markets Information about linkages the poor have to markets is contained in Tables 111.22 and M.23. The percentage of households producing each of the main agricultural commodities is presented in Table M.22. Higher percentages of poor households produce local maize and cassava, and the poor are less likely to produce hybrid maize. Despite this result, large percentages of poor households do produce hybrid maize. Hybrid maize has been widely adopted by farmers in many provinces; in Central, Copperbelt, Luapula, Lusaka, Northern, and Southern provinces, hybrid maize is produced by a higher proportion of all households than is local maize. In all of the provinces where maize production is widespread, except Eastern, poor households are more likely to produce any maize than are the non-poor. Poor households market much lower percentages of their crops than do the non-poor (Table 111.23). Hybrid maize is obviously more heavily marketed than is local maize, but hybrid is also widely consumed by the poor and the non-poor alike. Cassava is largely a subsistence crop, and in the 3 provinces where production of cassava is high (Luapula, Northem, and Northwestem), very low percentages of cassava production are marketed. A pattern of semi-subsistence agriculture" emerges for poor rural households (Table 111.23). Less than 50 percent of agricultural output is marketed, and as poverty grows the proportion marketed declines. This information, combined with the earlier analysis showing that most poor are primarily agricultural, provides some lessons about the likely impact of adjustment on the rural poor. 13Semi-subsistence in the sense that hybrid seed and fertilizer purchases are common, but large proportions of the output are consumed in the household. 41  Table 11.22. Percent Rural Households Producing Commodities by Province, by Poverty. Loco Mauie Hbri Maine Cassava Non-Poor Poor All Non-Poor Poor All Non-Poor Poor All Central 22.79 37.89 32.83 80.43 51.75 61.37 1.55 3.37 2.76 Copperbelt 25.96 37.60 33.06 42.13 38.15 39.70 9.36 5.45 6.98 Eastern 92.19 84.62 86.49 45.14 22.79 28.31 0 0 0 Luapula 16.21 9.58 10.61 38.53 17.47 20.91 41.90 69.15 64.90 Lusaka 21.48 31.41 27.66 27.46 48.08 40.29 0 0 0 Northern 9.32 21.79 19.77 64.41 35.59 40.25 53.81 60.44 59.37 Northwestrn 31.34 52.43 48.98 35.48 10.36 14.47 59.45 35.77 39.64 Southern 32.70 39.08 37.34 79.43 54.88 61.55 0 .04 0 Wester 33.47 51.43 49.16 31.45 21.08 22.39 28.63 23.82 24.43 All Rural 37.86 43.70 42.32 59.87 34.68 40.62 12.81 21.94 19.79 Firt, increases m input pnces could have nmajor negatve impauts in some of the poorest provinces. In Northern, Luapula, and, to a lesser extent, Northwestern and Western provinces, hybrid maize production is widespread even among the poor. Even though in all of these provinces except Western, the poor market higher percentages of hybrid maize than do the non-poor, the proportion of production retained by the poor for home-consumption still averages between 40 and 50 percent. input price increases could impose hardship on these farmers, especially if chanmels for marketing their output are not formed. Disruptions in input marketing channels may also exacerbate poverty; the large percentage of hybrid producton even among poor farmers leaves them vulnerable. Second, the demise of pan-territorial pricing and govermment-supported marketing channels will lead to major reallocations of agricultural resources. The relative isc ;adon of the poor in Western, Northern, and Northwestern provinces will translate into costly marketing charges and falling prices for aU marlketed commodities. Producers in such areas will pay higher input prices and receive lower prices, especially for maize."4 Such a shock may drive them away from market- '4Maize prices are expected to fall relative to other commodities in distant provinces, since the maize final market is farther away than the final market for other commodities. Historically, government-supported marketing supported maize prices relative to the prices of other commodities. 42  Table M123. Percentage of Commodity Produced That is Marketed, by Poverty, by Province. Local Maize H id Maine Va-sava All three Commodities Non-Poor Poor Non-Poor Poor Non-Poor Poor Non-Poor Poor Central 7.81 17.50 47.28 46.00 100.00 32.66 43.78 36.59 Copperbelt 8.78 5.79 37.18 37.77 19.93 46.46 27.45 23.07 Eastem 1.53 3.21 76.32 73.31 . . 29.S4 10.24 Luapula! 52.13 16.56 17.52 60.30 9.08 3.51 20.00 11.67 Lusaka 4.31 2.57 50.53 37.35 . . 15.16 27.93 Northern' 2.97 14.69 41.92 52.52 1.63 1.88 27.54 18.32 Northwestern' 14.64 13.90 42.59 53.60 12.32 10.24 24.26 19.56 Southern 12.07 2.71 39.92 28.09 . 47.50 33.24 18.88 Western 0.19 5.48 35.33 16.56 29.59 8.74 24.07 8.28 All Rural 5.99 7.13 48.68 43.21 9.20 4.81 30.85 16.76 Provne where large quanities of cassava are produced. Note: AP1 is the cutoff used to distinguish between poor and non-poor households. orented production and toward subsistence. The effects of price changes on smallholder producer behavior are examined in detail using a household model later in this report. The bulk of the evidence from the PS says that even in relatively isolated areas of rural Zambia, many poor farmers have a heavy market orientation. Historical promotion of hybrid maize (discussed in the following section of the report) has led to deep penetration of production even in areas where such production is economically inefficient (especially, for example, Western and Northwestern Provinces). As market signals change with liberalization, production patterns will change, and institutions are needed to disseminae information and knowledge about different crops. 43  A critical issue that cannot be examined with the PS is the net food sales/purchase position of the rural poor. The poor are selling substantial amounts of maize and cassava, but it is not known if they sell at one time of the year and return to purchase substantial quantities at a later date. Seasonal sales and purchases were not recorded in the priority survey, and sales of other crops are unknown. The net sales position is obviously critical because product price changes can have different effects depending on sales position. G) Ression Results In order to control for the separate influence of the different factors discussed above, a multiple regression analysis of the determinants of household expenditures per adult equivalent was undertaken. Such a regression might be motivated by considering a household production process that takes human capital endowments and constraints faced by a household and uses them to produce well-being, proxied by expenditures. The model estimated was: InEXP = f(lnAGEH, InAGEH2, SEXH, DEPRATIO, hnADEQ, ED1, ED2, HAGRIC, DFOMA, DTRAN, DPSC, DHOSP, OWNL, WHRMIL, WTRACT, WRADIO, PROVINCE) Where EXP are adult equivalent household expenditures, AGEH is the age of the household head, SEXH is the household head's sex, DEPRATIO is the dependency ratio, ADEQ is the nunber of adult equivalents, EDI is a dummy variable for some education for the head (less than 6 years), ED2 is a dummy variable for education of the head for more than 6 years, HAGRIC is a dunimy that takes a value 1 if the head's principal employment is agricuture, and OWNL is a dummy for title to land. DFOMA, DTRAN, DPSC, DHOSP are the distance to the nearest food market, transportation, primary school, and hospital in Km, respectively. WHRMIL, WIRACT, WRADIO, are dummies for ownership (by anyone in the respondent's ward) of a hammermill, tractor, and radio, respectively. The distance variables are included to measure the impact of distance to critical facilities on welfare. The presence of an asset in the ward is included to measure the effects of spillovers of these assets beyond the owning household. The influence of the province could have been modeled in a number of ways. Based on the information presented earlier in the report, it was decided to begin with a general specification of provincial effects. Initially, full slope and intercept differences by province were allowed. 15 Following estimation of this "full" model, restricted versions were run, where provincial intercepts shifters were permitted, and the slope shifters were tested for inclusion block by block. The full model was subjected to a series of misspecification tests (see McGuirk, et al., for details). It satisfied all of the underlying statistical assumptions. The results from this estimation by province available from the authors upon request. To test more restricted versions of the model, first the slope shifters for the human capital variables (EDI, ED2, OWNL, and HAGRIC interacted with the province dummies) were tested. We failed to reject the null hypothesis that these human capital effects were equal "5The variables representing assets available in the ward of residence (WHRMIL, WTRACT, WRADIO) could not be allowed to have a variabla slope by province, since in some of the provinces there were no such facilites in any ward, and linear dependencies would have resulted. 44  across province (p=.3241). Thus we deleted all the human capital interactions with the province dununy variables. Next, we tested the block of variables related to distance to hospital and prinary schools (the interactions between DHOSP and DFOMA and provincial dummy variables). Once again, the null hypothesis that the effect of distance from these facilities on household expenditures was equal across provinces could not be rejected (p-.7401). Thus, this block of interactions was deleted. Finally, we examined the effect of distance to transportation and distance to primary school by province. We reject the null hypothesis of equal slope by province (p= .041). We were thus left with a model with provincial dununy variables and no slope shifters for the primary school variable. The regression results are presented in Table 111.24. The impacts of the human capital variables all follow expectations. Age of head is not a significant determinant of expenditures. Households headed by females have per adult equivalent expenditures that are, at the sample mean, 81.8 percent of expenditures of male-headed households. Larger households are worse off than smaller ones. Even controlling for dependency effects, a one percent increase in adult equivalents leads to a .25 percent decrease in per capita expenditures. The coefficient on InADEQ (it can be directly interpreted as an elasticity) confirms that larger households are more likely than small households to be poor. By controlling for dependency, however, we see that much of the relationship between household size and poverty found in the bivariate analyses earlier in the paper and in other studies may be attnbutable to dependency. The small elasticity associated with household size could not lead to the degree of differences in household size by poverty class found in the previous analysis. Having a head whose primary employment is in agriculture leads to increased poverty. The coefficient on HAGRIC, although not statistica1ly significant, indicates about a 3.4 percent difference in per capita expenditures depending on employment of head. Households headed by agriculturalists are likely to have lower expenditures. Land ownership (OWNL), consistent with the bivariate analysis above, shows that land ownership is associated, even controlling for other factors, with wealth. The education of the household head continues to have a strong influence on household expenditures, even controlling for many of the other factors associated with income generation. The coefficients for EDI and ED2 indicate, respectively, that at the sample mean having a household head with some elementary education and with more than 6 years of primary education is associated with a 42 and a 113 percent increase in per capita family expenditures. Returns to education in rural Zambia are, thus, substantial. Increased distance from various facilities persists in having a negative impact on per capita household expenditures and thus a positive impact on the likelihood of a household being poor. Increased distance from hospitals was the only of these variables whose impact was non-significant. 45  Table 111.24. Regression Results for Household Expenditures. Dependent Variable: ln(EXP) Variable Parameter Esdimate Standard Error Intercept 6.943 (1.755) In AGEH -0.190 (0.200) In AGEH2 0.034 (0.269) SEXH -0.201 (0.046) D)EPRATIO -0.379 (0.069) In ADEQ -0.255 (0.032) EDI 0.354 (0.048) ED2 0.757 (0.050) HAGRIC -0.034 (0.043) OWNL 0.497 (0.122) DFOMA' -0.248 (0.087) DTRAN* -0.249 (0.115) DPSC -5.300 (1.660) DHOSP -0.243 (0.141) W1HRMIL -0.039 (0.043) WTRACT 0.363 (0.048) WRADIO 0.038 (0.232) COPPERBELT 0.212 (0.141) EASTERN 0.105 (0.087) LUAPULA -0.177 (0.101) LUSAKA 0.024 (0.129) NORTHERN -0.023 (0.089) NORTHWESTERN -0.236 (0.107) SOUTHERN -0.177 (0.082) WESTERN -0.147 (0.097) COPPERBELT * PSC 0.056 (0.025) EASTERN * PSC 0.047 (0.022) CENTRAL * PSC 0.043 (0.020) LUSAKA * PSC 0.062 (0.019) NORTHERN * PSC 0.044 (0.019) NORTH1WESTN * PSC 0.038 (0.022) SOUTHERN * PSC 0.045 (0.018) WESTSrERN * PSC 0.031 (0.017) RI .203 N 3464 * Parameter *100. 46  At variable means, the coefficients can be interpreted to indicate that a 1 percent increase in distance from a food market, transportation, and primary schools leads to a .002, a .002, and a .054 percent decrease in per capita expenditures in rural Zambia.'6 Distance from facilities is clearly associated with increased poverty; controlling for other influences, distance has the same impact across all provinces. The presence of a tractor in the ward of residence has a positive impact on household expenditures. This variable could represent spillover effects from having commercial farmers in the region, or it might be interpreted as representing the existence of a rental market for plowing services. Alternatively, the positive sign on this variable might be interpreted to suggest that for political or other reasons, tractors were placed in wards with more wealthy residents. In any case, wards with tractors have, at the sample mean, households with 44 percent higher expenditures per capita than wards without tractors. Much more investigation is needed to evaluate the impact of this variable; the data necessary to conduct such an analysis were not available in the PS. The information could, however, be used for targeting a program; there is unlikely to be significant poverty in wards with tractors. H) Summary of Priority Surev Results The information from the analysis of the PS data, combined with the other studies of Zambian poverty and other literature on Zambia, allow us to make some generalizations about poverty in rural Zambia. These generalizations will be used in the subsequent section to develop a stylized model of a poor rural Zambian household. Generalizations can be made about geography, social factors, and economic factors associated with poverty. First, although rual Zambia is overwhelmingly poor, higher prevalences (number of poor divided by total population) of poverty and severe poverty are found in the most isolated areas. Western, Northwestern, Luapula, Northern, and Eastern provinces have the most severe problems. The simple analysis showed, and the regression results confirmed, that distance from key facilities such as food markets, transportation, and primary schools is positively associated with poverty. The ranngs of provinces by prevalences of poverty found here are largely consistent with similar rankings from the other studies. Second, female-headed households are most likely to be poor. This result is so widely confirmed that it bears repeating only to stress its urgency. Poor households headed by women are almost completely dependent on agriculture, for their primary source of income, and those that are not dependent are small-scale processors of agricultural products, traders, and teachers. Third, more dependents in a household makes it more likely that the household is poor. Larger households are also more likely to be poor, although much of the household size/poverty linkage is caused by dependency effects. There is a remaining conundrum: Poverty studies show a positive relationship between poverty and household size, while nutrition studies show that larger '6Average national distance in rual areas to food markets, transportation, and primary schools were, respectively 21.7, 11.69, and 3.58 kilometers. Notice that the effect of distance from a primary school varies substantially between districts. 47  households are less likely to have malnourished children. This study is consistent with the other poverty studies in finding the positive link between household size and poverty. Fourth, education is strongly associated with reductions in poverty. Results using the HEIS, the PS's nutrition module, and the PS expenditure data are all consistent. Households headed by persons with some education are better off than those with none. More education reduces poverty even more. The return to education in Zambia occurs in all occupations. While there are more returns to education when the head of the household is employed in a non-agricultural activity, there are significant returns to education to those employed in agriculture. At the same time, children of poor families are less likely to attend school and are more likely to be employed by their family in an agricultural job without pay. The link between poverty and education is blurred by the results showing higher education shares of expenditures for poor households. The poor do see returns to education and are trying to educate their children accordingly. Short-run considerations such as the need to withdraw children during periods of farm labor shortages may outweigh long-run benefits from education. Fifth, despite their isolation, the poor are linked to input and product nmarkets, and should be relatively sensitive to changes in these markets. The poor market up to 50 percent of their total output of cassava, and hybrid and local maize. The widespread production of hybrid maize, even among the poor is indicative of the market integration of this group. This integration into the market is most likely a result of historical policies that were biased in favor of maize production. Sudden shocks to the market resulting from adjustment, or a decline in (newly-privatized) marketing institutions reaching remote areas can have a strong negative impact on these households. Sixth, agriculture is the main industry in rural areas, and virtualy all the rural poor list their primary occupation as agriculture. Female-headed households, as mentioned above, are predominaly agricultural. Outside of agriculture, there is a wide variety of occupations in rural areas. This variety makes it difficult to categorize these occupations, or target them for possible interventions. The picture of severe rural poverty that emerges is one of isolated agriculture, with other economic activities being undertaken as the agricultural calendar permits. The information generated from the PS data analysis is employed later in this study (Section V) where a household model is constructed. The model helps clarify some of the relationships found during the data analysis. In order to ensure that the model accurately reflects the institutional realities of rural Zambia, a description of the institutions is found in the following section of the report. This description of the institutions also defines the issues and constraints that the rural poor face, and is thus, helpful in motivating some of the scenarios that are simulated with the household model in Section V. 48  IV) POUICIES, INSTITUTIONS, and DONOR SUPPORT Rural poverty is widespread in Zambia, and many authors point to past policies and institutions as main contributors to this poverty (e.g., Jansen; World Bank 1992a; 1993b). Agricultural policy, for example, has been biased in favor of large-scale commercial farmers and toward the production of hybrid maize. The rigidity of the government-controlled marketing system prohibited farmers from adjusting to changing economic conditions and created inefficient patterns of production. Agricultural institutions such as extension and credit were also biased in favor of maize. As reform occurs, changing policies and institutions will help determine the rural poor are able to teact to economic forces. The household model presented in Section V represents an attempt at modeling these reactions. In order to develop the model and determine the issues that it should be used to address, the policy and institutional conditions faced by the poor must be described. We begin by outlining the current policy environment, then discuss institutions, and finally describe the various types of donor-supported projects found in Zambia. More information on the scope and effectiveness of donor-supported projects is found in Annex I. A) AficutuS rles i. Overview The Ministry of Agriculture, Food, and Fisheries (MAFF) prepared a strategy paper A Frameworkfor Agricultral Policies up to Year 2000 and Beyond in December 1992. Specific agricultural policies and programs will be set out the Agriculture Sector Investnent Program (ASIP) which is being drawn up by a Task Force on Agriculture. The World Bank is actively involved in the ASIP that is currently being finalized. The ASIP has the following objectives:17 1. To ensure national, regional, and household food security through dependable annual production of adequate supplies of food staples. 2. To ensure that the agricultural resource base is maintained and improved upon. 3. To generate income and employment to maximum feasible levels in all regions through full utilization of local resources and realization of domestic and export market potential. 4. To contibute to sustainable industrial development. 5. To significantly expand the sector's contibution to the national balance of payment. "7This sub-section on agricutural policies draws heavily on a Discussion Paper on the Agricultural Development Situation in Zambia (Potentials and Constraints), which was prepared in mid-1993 by C. Chabala, Deputy Secretary General, Zambia Cooperative Federation, and member of the Task Force on Agriculture. 49  The agricultural sector is perceived to be central to overall (agricultural and industrial) developm of the economy. To achieve these policy objectives, the following strategies are envisaged: 1. Liberalization of Markets - This entails allowing market forces to function, with prices reflecting the interaction of supply and demand. Liberalization will mean that direct involvement by GRZ in agricultural marketing and input supply will be reduced and eliminated. Constraints to international trade in agricultural commodities will be eliminated. 2. Diversification of Crop Production - This recognizes that not all areas in the country are suited for maize production. Farmers will be encouraged to diversify their crop patterns. 3. Development of the Livestock Sector - Increased productivity of Zambia's livestock will play a major role in improving the welfare of the rural people both through expanded food supplies and increased incomes. Livestock development could also help make more effective use of land that is not suitable for continuous cropping. 4. Emphasis on Services to Smallholders - Medium- and large-scale commercial farmers will quickly adapt to the new policy environment of liberalization. However, smaliholders will require special attention if they are to benefit fully from opportunities available to them. Areas in which they require assistance include esearch, extension services, technology, credit, land tenure, marketing, and processing. 5. Expanded Opportunities for Outlying Regions - Better developed infrastructure and closer proximity to urban centers place line-of-rail provinces in a position to take advantage of liberalized markets. Special measures are needed to help farmers in less developed provinces become active participants in the economy. Such measures include improvemens in inftastructure and programs to improve access to inputs and markets. 6. Improvement in the Economic Status of Women - Women's economic status in agriculture must be improved through restructg of agricultural research, extension, credit, and land policies. 7. Improved Use of Available Water Resources - Improved management of available water resources can make a major contribution to improving the economic performance of the agricultural sector and ensuring adequate food supplies, even in years of severe drought. 8. Full Utilization of Land Suitable for Agriculture - Zambia is utilizing only about two million hectares out of 18 million hectares of potentially arable land. Arable land under cultivation also suffers from low productivity per unit area. The full utilization of these lands requires public investment in basic infrastructure in areas where these do not exist. Support services are also needed. 50  9. Helping Farmers Deal With Natural Disasters - There is a need for programs (including public grain reserves) that help people affected by disasters to survive, recover, and continue with their agricultural activities. Farmers must be helped with technologies (and possibly crop insurance) that will help guard against such disasters. 10. Emphasizing Sustainable Agriculture - Some of the above strategies can make major contributions to the establishnent of sustainable agricultural and food systems. However, efforts will be needed to ensure adequate forest cover in certain areas, prevent soil erosion, and minimize any adverse effects of changing fanning technology on the environment. ii. Maize Marketine Structural adjustment and market liberalization are expected to have some positive effects on the agricultural sector and on rural households. Liberalizing agricultural marketing implies that farmers can negotiate directly with millers or any other maize traders for the best price. However, for many smallholders, the lack of a guaranteed market and lower farmgate prices are likely to discourage them from producing maize. Relative price changes will create incentives for growing alternative crops. For most Zambians, maize is the major food staple. It is, therefore, critical to ensure that maize marketing becomes as efficient as possible. Maize pricing and marketing is a poverty matter. Poor households allocate a substantial part of their budgets to maize, so their real incomes are significantly affected by maize prices. In addition, maize is the most common crop produced by Zambian farmers, providing a source of cash income for most poor rural households. Although poor urban households are net maize consumers, poor rural households can be either net producers or purchasers. (a) History Since Independence, Government policy promoted maize production throughout the country using pricing policies that, in particular, benefitted farners located off the line-of-rail (Jansen). These outlying areas were the poorer regions in Zambia (and remain so despite the policy). Pan-territorial maize producer prices were fixed for each harvest year, and only sanctioned government agents were allowed to participate in maize marketing. In addition to uniform producer prices, subsidies were provided for productive inputs and tnort. As a result, some maize production shifted into high- cost areas. Government also subsidized maize to the consumer by regulating milling and sales. Large- scale milling operations were located in urban centers and processed maize was sold throughout the country. During this period, large-scale millers controlled most of the market for mealie meal. Small-scale mills (hammermills) were not allowed to participate in maize marketing and functioned only as service mills for home consumption. Prices were set in such a manner that it was often cheaper for rural consumers to sell their harvested maize and exchange it for cheap milled maize. 51  Numerous state and parastatal organizations were established to administer maize policies. While these historical policies did have some positive aspects from a poverty perspective, they were achieved at a considerable cost. The costs involved both direct budgetary costs and (possibly more important) the hidden costs of encouraging an inefficient pattern of maize production and consumption. Although the price system provided benefits for maize farmers in remote areas, it trapped them into a pattern of production and consumption that has proved vulnerable to climatic uncertainties and that has very limited potential for long-term income growth. Pricing policy in conjunction with agricultural credit, research, and extension institutions promoted crop specialization. This specialization has had a deleterious effect on household food security. Alternative crops and cultural practices that mnight have positive nutritional, environmental, and economic outcomes have been neglected. Problems associated with the lack of crop and diet diversity affect the poor most in times of external shocks resulting from unpredictable weather, and economic conditions. For example, droughts in maize-growing areas have led to extreme stress for both the urban and rural poor. Maize pricing and marketing policies have meant that large sums of money (the lion's share of resources allocated to agriculture) went to subsidies rather than to investments which could have improved the productivity of snallholders (e.g., fuiding for transportation and communication improvements). A good part of the subsidy did not benefit the consumer, but instead served only to prop up an inefficient producton and distribution system. In practice, outlying areas suffered from untmely and poor quality service under pan-territorial pricing. Thus, rural poor in outlying areas did not benefit as much as may have been expected from the policies. Past maize pricing and marketing policies distorted urban and rual consumption and production pates, and wasted scarce resources. Maize policies left a legacy of dependency by consumers and producers on centrally-controlled or directed organizations. The past pricing and marketig policies were clearly unsustainable from all perspectives and a major impediment to economic growth in Zambia. (b) Reform Since assuming power in 1991, the MMD government has committed itself to reform of maize pricing and marketing, including removal of international trade restrictions. However, reversing the long history of distortions is complex and challenging. Replacing centrally-controlled structures with decentralized ones is a dauing task. While progress has been far from smooth, some significant changes have occurred. Most important are the official dissolution of pan-territorial and pan-seasonal pricing, the elimination of seed and fertilizer subsidies, and the decentralization of maize marketing and processing. Proliferation of hamemills and the development of unrecorded maize marketing are evidence of movement toward a more flexible and competitive system. A concerted effort is still, however, needed to create macroeconomic and institutonal conditions for a decentralized marketing system. Many vestiges of the old system remain, and many consumers and producers continue to act as if the old system wfll survive. The refonns should help enable poverty reduction by promoting growth in smallholder agriculture, and (through improvements in the efficiency of the marketing system) lower prices to conumers in the medium-term. However, transition to a new system potentially increases the short- 52  term vulnerability of two important groups of poor: Low income maize consumers, and high cost (in terms of location and transport costs) small-scale producers.' The impact of reforms on maize production and marketing during the 1992193 and 1993/94 seasons will be discussed briefly, with attention focused on these vulnerable groups. (i) Maize Prices and Markets Reforms scheduled for 1991/92 were postponed due to a severe drought. Record levels of emergency aid were required to avert a national disaster. Significant quantities of food aid were still available when the 1992/93 maize harvest commnenced in May 1993. Farmers and potential traders awaited a market signal to detenmine a price for the newly harvested maize. However, private agents (farmers and traders) did not respond to public proclamations of an open market system. A price signal from Government was anxiously awaited. In June 1993, Government responded by announcing a pan-territorial "floor"" price of K5,000/bag. At the same time it was indicated that the into-mill price should not exceed K6,500/bag if the millers provided transport (K7,000 if delivered to the mill). In addition, several "principal buying agents" of maize were designated.20 A KIS billion credit at commercial interest rates was released by Government for maize purchase by principal agents. It was expected that the K15 billion would cover the purchase of 3 million bags for the strategic reserve. The arrangements with the agents were, however, unclear as the agents proceeded to purchase quantities far above the requirements for the strategic reserves. The rationale behind engaging the lending institutions, which had hitherto not participated in maize marketing, was to help them recover the input credit given and guaranteed by Government (a total of approximately K18 billion, including K6 billion in interest payments). It was envisaged that the finance for the remaing maize marketing would come from private sources and banks. The failure, however, to control inflation in the beginning of the year and the subsequent efforts to suck up excess liquidity with treasury bills earmng high interest rates meant that most private traders, and some of the principal buying agents, preferred to invest in treasury bills rather than purchase maize.2' "8See results of the household model in Secdon V. 191n the public debate the term "floor" price was used. It was, however, not a real floor price as Government was not obliged to buy all maize at that price. It was intended to be the Government purchasing price, i.e., the price that Government was willing to pay for the maize that it wished to procure for the strategic reserve. The public perception was, however, that it was the pan-territorial floor price. 20The major principal maize buying agents were government-supported agicultural lending institutions-ZCF/IFS, CUSA, and LIMA Bank. 211t has been estimated tbat the anlized compounded interest rates of 328 day treasury bills went as high as 700 percent in July 1993 falling towards the end of the year. 53  The floor price and arrangements with principal buying agents caused confusion, as farmers, conditioned by many years of Government price setting and control over marketing, regarded this as continuation of the old policies. Thus, maize price formation in the 1992/93 season was greatly influenced by the announced official floor price. About three-fourths of recorded maize purchases were made by the principal buying agents (see Table IV. 1). The floor price and government arrangernents with principal buying agents also effectively limited the possibilities of private traders, who found it hard to cover costs and still earn a profit. Traders feared continued government intervention in price setting to protect urban consumers. Likewise it meant that the principal agents found it very hard to cover transport and handling costs as they faced the same problems. Trading agents were engaged to cover designated areas and were provided with authority to make purchases and bring the maize to the depots. A commission of K950/bag was offered to the trading agents to induce them to service high-transport cost areas. Registered traders ("other agents") purchased little maize from farmers in the outlying provinces of Luapula, Northern, Northwestern, and Western (Table IV.1). By the end of October 1993, the principal buying agents had made payments of K15 billion, but fanners were still owed about K22 billion. Government introduced promissory notes to cover the debts owed to famers. The notes earned 125 percent annual interest rate from October 15, 1993 to February 15, 1994 and were, in theory, negotiable. In practice, the promissory notes were not very negotiable. Most farmers did receive payment shortly after February 15, although as of mid-April 1994, not all the notes had been discounted. At the K5,000 floor price, which in many cases was taken to be a pan-territorial price set by govermnent, maize producer prices in September and October 1993 represented a fall in real prices from prior years (see Table IV.2). The K5,000 price brought the real producer price roughly back to its 1989 level. At the same time, the real price of fertilizer increased significanly since 1992. Therefore, based on prices in September and October 1993, the rural/urban terms of trade did not Improve over the last year. As can be seen from Table IV.3, producer prices diverged from the floor price of X5,000 during 1993. Several factors influenced the disparity of producer prices through the 1993 marketing period. Regional differences based on transport costs impacted outlying maize-surplus areas. For example, parts of Eastem province reported sales at prices lower than K5,000/bag. However, the failure of efFective market reforms were a factor causing some frmers to receive low prices. Some farmers were willing to sell at loss-maiEng prices in order to obtain cash. Lack of payment from the buying agents left many farmers constrained by cash needs for inputs and essential goods. Only members of the principal trading organizations (e.g., customers of or borrowers from the banks) were allowed to sell to the principal traders. The smaller, poorer farms sell their maize to members who, in turn, charge a commission for this service. Under such circumstances, the real price received by smallholders has been lower than the official "floor" price. 54  Table IV.1. Estimated Production and Recorded Maize Marketing 1992/93 (thousands of 90 kg bags). Province of Purchased by Purchased by Total Market Purchases as Forecast Total Prod. Principal Agents Other Agents Marketed Forecast % Forecast Central 4406 1702 517 2219 3406 65% Copperbelt 882 248 339 587 640 92% Eastern 3771 1423 380 1803 1215 148% Luapula 419 184 22 206 300 69% Lusaka 952 646 388 1034 638 162% Northern 1336 643 38 681 905 75% Northwestern 326 158 0 158 181 87% Southern 5140 1616 74 1690 2813 60% Western 520 188 37 224 232 97% Total 17752 6808 1795 8603 10330 83% Souce: Preliminary crop forecast, CSO, Feb. 1994. Table IV.2. Real Producer Price of Maize. Maize Producer Year Price (90 kg) CPI Real Price 1985 28.32 100 28.32 1986 55.00 154.8 35.53 1987 78.00 227.6 34.27 1988 80.00 320.6 24.95 1989 125.00 800.4 15.62 1990 284.20 1,677.2 16.94 1991 800.00 3,243.0 24.67 1992 3,000.00 9,448.5 31.75 1993 (September) 5,000.00 33,449.3 14.95 1994 (February) 9,500.00 38,157.5 24.90 Source: Consumer Index, CSO, (Feb. 1994) p. 21. 55  Table IV.3. High and Low Reported At-Depot Price 90kg Maize, Various Dates (1993). May 31 June 30 July 26 August 30 September 27 October 18 October 25 High 5000 6733 7500 7000 7000 7000 7500 Low 4500 3750 3500 3300 3500 3600 3600 Source: Agricultural Market Information, MAFF. in addition to spatial differences in maize producer prices, there have been seasonal differences. Maize producer prices averaged about K5-6,000/bag from September to November 1993. They increased by about 25 percent in December as most maize had already been collected from farmers, and principal agents started selling maize to millers. The producer price of maize stayed at about K9-10,000/bag from Decemnber through April 1994. The nominal increases in maize producer prices since November do not, in general, cover the storage costs and interest charges owed by principal buying agents. One reason that maize producer prices have not increased to reflect these costs are imports of maize from Zimbabwe.' The principal buying agents are still (in April 1994) negotiating the price at which they will transfer ownership of some 3 million bags of maize earmarked for the strategic reserve. These agents are suffering a liquidity problem and potential financial crisis as they attempt to cover all the purchase, trasport and hauling, storage (and most importantly) interest charges on the 1992/93 maize harvest (while also dealing with low recovery rates of input loans). Of the total estimated production of 17.7 million bags for 1992/93 it was estimated that 7.4 million would be retained for home consumption while 10.3 million would be marketed. Actual marketed maize is only about 83 percent of forecasted quantities (see Table IV. 1). While purchases by the principal agents are recorded, trade going through other channels is unknown. It could, however, be a substantl part of the 1993 harvest as the recorded marketed quantities in some provinces are significantly below the expected level, most notably in Central, Luapula and Southern provinces. It has been suggested that the reasons for the low marketed output is increased retention and apparent unwillingness of farmers to sell to buying agents because they are wary of being cheated and because of the promissory notes. Output from hammermills also increased significantly. Some export took place, particularly from Eastern and Northern provinces, although the export prices are not favorable with the relatively high value of the Kwacha. Exr. The forecasted 1992/93 harvest exceeded domestic consumption needs, and some exports were expected. An export quota of 2 million bags was introduced and licenses were issued on first come first served basis. However, fewer than 500,000 bags of official exports were recorded. There is, however, unrecorded cross border trade, especially to Zaire and Malawi. Two reasons why official exports have been sluggish are: 22Some argue that Zimbabwe is subsidizing these exports to Zambia. 56  (a) the appreciation of the Kwacha made Zambian maize uncompetitive. This unexpected development has become a problem for traders who paid high prices for maize and were unable to dispose of it and at the same time faced high real interest rates with which to fnance their inventory. (b) the lifting of restrictions did not permeate the official system, so the export licensing de facto was not freely granted. (ii) Maize Consumption Demand. The major markets for maize and mealie meal are in the provincial towns and the towns in the line-of-rail provinces, mainly Lusaka and Copperbelt. They account for over 63 percent of marketed maize consumption. Thus, a considerable portion of the maize harvest must be moved fom rural to urban areas. The real consumner price of mealie meal has more than doubled since December 1991, the st of liberalization (Table IV.4). With the recent price increases for mealie meal and the continuing falling purchasing power of the population, the demand for mealie meal is declining. The real price in Lusaka has exhibited some month-to-month variability, however, it has stayed more or less constant since September 1992. These mealie meal prices are averages for 25kg bags sold in Lusaka, and do not necessarily reflect prices paid"by poor consumers. There is an increasing trend in the poorer compounds to buy less than the standard 25kg bag. Traders sell portions as small as 100 to 200 grams raising the price considerably.23 In addition to seasonal variability, there have been noticeable differences in mealie maize prices between provinces. These differences are highlighted in Table IV.3. As can be observed, the lowest mealie meal prices were in Chipata, located in the maize-surplus Eastern proviIce. Highest mealie meal prices were in Mongu and Solwezi, the outlying maize-deficit Western and Northwestern provinces, respectively. Thus, the market is working, but poor residents of Mongu and Solwezi are negatively impacted by the higher mealie meal prices. Fcessing. Hammermill operators are experiencing an increase in throughput as traders and consumers buy maize directly from famers or at market and use hammermills for processing. An estimated 75-85 percent of rural consumption of mealie meal goes through the hammermills and a substantial part of the urban consumption does too. The exact market share between large-seale mills and the hammermills is not known, but at present, it is estimated that the former are running well below capacity (approximately 30 percen as opposed to the normal 60-70 percent). Although there may have been a decline in total consumption of mealie meal the hammermills are increasing their market share. Despite becoming more competitive and introducing more flexible packaging, some large-scale mills have a hard time competing with the hammermills. Large-scale mills are at a cost disadvantage because ftey have outdated capital-intensive plants, whereas the hammermills minimize %In Kamanga compound in Lusaka, the )-ice of a 25kg bag, if split up in smaller portions, can fetch up to double its original purchase price. Thus, some poor consumers must pay twice the reported average price of mealie meal. 57  Table IVA. Maize Meal Prices: Roller Meal (25kg) Average Urban Retail Price. Price CPI (1985 = 100) Real Price March 1988 14.85 205.3 7.23 March 1989 41.00 440.8 9.30 September 1989 82.30 859.2 9.58 March 1990 82.30 1,145.4 7.19 June 1991 158.00 3,056.2 5.17 December 1991 207.00 4,941.6 4.19 March 1992 464.28 7,429.4 6.25 June 1992 760.00 8,968.3 8.47 September 1992 1,371.00 10,923.7 12.55 December 1992 1,629.00 14,388.2 11.32 March 1993 2,023.00 19,487.9 10.38 June 1993 (Lusaka) 3,350.00 27,937.1 11.99 Sqepmber 1993 3,500.00 33,449.3 10.46 February 1994 4,700.00 38,157.5 12.32 Source: Consmner Price Statistics, June 1990 (CSO), Consumer Price News First Quarter 1993 (CSO) Index Numbers, World Bank, October 1993. Table IV.5. Mealie Meal Retail Prices in Selected Areas (Kwacha per 25kg roller meal). Town Province January 24, 1994 February 23, 1994 Chipata Easten 3000 3350 Choma Southern 4200 4500 Kabwe Central 3850 4500 Kasama Northern 3325 4500 Kitwe Copperbet 4600 4400 Lusaka Lusaka 4200 4700 Mongu Western 5100 5225 Solwezi Northwesten 3150 5000 Source: Food Security Bulledn January-February 1994, published by the Zambia Famine Early Warning System. 58  overheads and are more labor intensive. The large-scale mills are part of the formal economy, and provide higher wages and benefits to their workers than do hanunermills. This is an example of how lower-paying labor-intensive techniques are replacing capital-intensive higher paying jobs in the economy. Clearly there are winners and losers from such changes. Because of their cost advantages, hanunermills can reduce the price that consumers pay. This is a clear benefit for poor urban and rural consumers. For example, in October 1993 the price of 25kg of hammermilled meal in Lusaka was between two-thirds to three-fourths the price of roller meal from large-scale mills.2Y Use of the hammermill option, however, depends on the consumers' access to grain markets and to hamermills. If substantial transport costs and time are needed for the consumer to take advantage of hammermills, the savings are lower. At present, there are 4,700 hannermills in the country and the sector has witnessed a subsantial growth over the last few years. As part of the National Hannnermill Progranme, the sector has been supported by Small Industries Development Organization, various NGOs (e.g., Village Industries Services), and lending institutions (such as ZCF-FS and LIMA Bank).- There is great variation in the operational efficiency and the profitability of the business, especially in rural areas where capacity utilization can be low. The structure of the business also affects efficiency as most mills are so called service mills. They simply accommodate the needs of residents wanting to have their grain miiled. There are normally no storage facilities on site. The study by Skonsberg (1994) in Kefa Village confirms the increased use of hammermills. The purchase of maize for hammermilling has also been identified as a coping strategy for the urban poor. 1993/94 Season. The lingering effects of the not-so-liberalized 1992/93 maize marketing season are still being felt. Some observations can be made. It seems that less hybrid maize has been planted by some small farmers without a cash surplus as the promissory notes were not recognized by itput sellers as valid. Limited availability of hybrid seed and other inputs have contributed to movement away from hybrid mame. The results from Kefa village in Eastern province are astounding; no one is growing hybrid maize there anymore as no one could buy inputs with the promissory notes (Skonsberg, 1994). Crop forecast data indicate that nationally some crop diversification has taken place (see Table IV.6.) Plantings of maize decreased, while plantings of altenative food crops such as groundnuts, sunflower, soybean, sorghum, millet, and mixed beans increased. Most maize plantings took place in Southern province (Table IV.7.) It is known that many commercial farmers in Southern province switched from maize to other crops. '"Although it is possible that the quality of some hammermilled mealie meal is lower than that milled in large-scale mills. 2'More details on the funding for the National Hammermill Program are provided later in this section. 59  Table IV.6. Area Planted in Hectares 1992/93 and 1993/94 (thousands of ha). Crop 1992/93 1993/94 % Change Maize 633 555 -12% Groundnuts 71 104 +46% Sunflower 39 40 +3% Soybean 20 32 +60% Sorghum 47 51 +9% Millet 53 62 +17% Mixed Beans 38 42 +11% Source: Preliminary Crop Forecast, CSO, February 1994. Table IV.7. Maize Area Planted in Nectares 1992/93 and 1993/94 (thousands of ha). Province 1992/93 1993/94 % Change central 113 101 -11% Copperbelt 30 28 -7% Eastem 129 130 +.I% Luapula 15 17 +13% Lusaka 29 26 -10% Northem 50 78 +56% Northwestem 18 21 +17% Southem 203 119 41% Wester 46 35 -24% Total Zambia 633 555 -12% Source: Preliminary Crop Forecast, CSO, February 1994. 60  1994 Maize Harvest. For the 1993/94 marketing season it is the stated policy of MAFF that the prices for maize will be left to be determined by the market and that international trade will also be liberalized. Entering the 1994 maize harvest, Government is considering the following options: a. no announced floor price; b. a strategic maize reserve (with maize purchased primarily from outlying areas); c. leasing of government-owned storage facilities; d. improved market information services; e. no restrictiors on maize imports or exports; f. establishment of a revolving fund to provide liquidity for maize traders; and, g. establishment of a Food Security Agency to oversee maize pricing and marketing to prevent supply shortfalls in domestic markets. The proposed Food Security Agency, charged with preventing supply shortfalls, will be able to control maize pricing and marketing if it is not carefully structured and monitored. There is a forecasted maize shortfall for the 1993/94 harvest, due to late planting and drought. By controlling imports and food stocks, the Food Secrity Agency will be able to influence domestic maize prices. (iii) Effects of Liberalized Marketing Conclusions. There is reason to believe that in the longer run the principles behind the maize pricing and market reforms are sound and that they will probably help reduce poverty. However, the short-run effects of each component of the changed maize marketing policy are negative on many of the poor. The execution of maize pricing and market reforms is still a major hurdle that the government must overcome. There is a persistent lack of confidence in the market. The costs of mnarket liberalization seem to be outweighing the benefits for nmy poor rural and urban households. To fully realize these benefits, government must be willing to have more faith in the market and not intervene at the first sign of distress. Most important, however, is the need to really get the macroeconomic prices right and bring down real interest rates and inflation. The short-term measures taken in 1993 to redress the problems described above did not do much to help the situation for the poorer farmers, but did provide a subsidy of some magnitude to the better-off farmers. The floor price was supported by only the three principal buying agents, and they only bought crops from their members. In turn, these (relatively better-off) members then extracted rent from the other (poorer) farmers in the communities by charging a fee for selling the maize of the non-members. Even the better-off farmers that did sell their crops were given promissory notes that were not discountable de facto, and as a result, they were not able to purchase desired quantities of inputs for the 1993/94 crop. In short, the interventions did litde to help ease the pain of the refonns, and (like the on-off adjustment policies of the 1980s) the fact that Government wavered on its commitment to market-oriented policies has created uncertainties about the sustainability of the reforms. 61  There have been some positive effects of liberalization. These include: (a) the removal of the subsidy to and the monopolistic protection of the large mills have led to increased use of hammermills and have helped lower the real consumer price of maize meal during the last year. (b) some private traders did enter into maize marketing. Uncertainty about goverment pricing policies, high interest rates, and export restrictions, however, constrained such market activities. (c) the official removal of the implicit subsidy inherent in the pan-territorial pricing and the transport subsidy has led to some diversification away from hybrid maize; and, (d) increased on-farm storage by smallholders as more maize is being mnilled in hammermills. Increased storage should increase household food security and allow farmers to sell some of their crops later in the season at higher prices. The negative effects include: (a) lower cash incomes for the poorest segments of society, the rural poor in remote provinces. The general lowering of real farm-gate maize prices due to the market-determination implied by the removal of pan-territorial prices and the transport subsidy has been compounded in the short run by the credit squeeze caused by stabilization policies, the lack of competition in maize trading in rural areas. The fact that the 1993 harvest was a bumper crop also helped depress maize prices as exports were not liberalized quickly enough and were hit by the temporary appreciation of the Kwacha in 1993. (b) increased risk of national maize shortages as diversification out of maize takes place. This is mainly a political issue of "self-reliance." The economist's answer would be to create a special fund in foreign exchange to import maize if a deficit situation should arise. (c) even though diversification should help in decreasing risks for rural areas, the fact that prices will now fluctuate with market forces across regions and over the year will increase vulnerability as information costs increase and farmers have to begin bargaing under new circumstances. The most important negative impact of the last marketing year is the widespread perception that reform of maize pricing and marketing was a failure and that government should once again take an active role. In many ways the 1993 maize harvest was handled like in the past, with a pan- territorial "floor price" and marketing agents collecting maize, and restrictions on maize exports. 62  Market signals are still quite ambiguous, and the combination of price and non-price constraints limit the responsiveness of smallholders.26 Many smallholders are continuing to grow maize with fewer inputs and poorer management techniques, which may actually increase their food insecurity (and vulnerability). In contrast, farmers with access to technology, credit, extension, and marketing services have been quicker to shift from maize to alternative crops, where profitable alternatives exist. B) Institions i. Ministry of Agnculture. Food and Fisheries (MAFF) MAFF has policy responsibility for agriculture and fisheries. There is currently a separate Ministry of Food Security but its status in the post-drought situation is unclear. A substantial reorganization of MAFF will be required to meet its reduced, but essential, responsibilities in the post-liberalization era. Its mission needs to be redefined to serve a market-based economy. This restructuring of MAFF is awaiting finalizadon of the sectoral program by the Task Force on Agriculture. MAFF has managed a large portfolio of donor-assisted projects in recent years. Such fragmentation has severely burdened its limited technical and management capacities, resulting in inefficient use of scarce managerial and technical resources, and inadequate prioritization of development activities in the sector. A consensus is emerging within GRZ and among donors that fragmented and uncoordinated donor assistance has contnibuted to less than satisfactory results and poor sector performance. The ASIP currently being prepared should provide the basis for individual donors to support specific components of a coordinated sector strategy. A major restructuring effort at MAFF is already underway. In late 1988 a Socio-Economic Policy Analysis (SEPA) Section was established in the Planning Division. The SEPA group was established to promote a multi-disciplinary target group perspective in agricultural planning to balance the economic and agronomic perspectives in the Planning Division. This new perspective is recogntion of the need to address smallholder agricultural production in the context of household-level food security. SEPA also brought the issue of gender into the policy arena. The most comprehensive restructring effort taking place at MAFF is occurrng under the auspices of the Zambia Agricutulral Research and Extension Project (ZAREP). ZAREP is a major project that aims at improving the overall effectiveness and efficiency of research and extension, and to specifically re-orient them to !he needs of smallholders (World Bank, 1992b). ZAREP is supported by the World Bank, along with NORAD and others. ZAREP led to the establishment, in 1991, of a National Research Action Plan (RAP) and an Extension Action Plan (EAP). A substantial reorganization of MAFF will be required to meet its reduced but essential responsibilities in the post-liberalization era. Restructuring has already begun to orient the ministry toward its new role as facilitator for agricultural production and marketing in a liberalized economy. 'See Section V on rural household model. 27As witnessed in the Kefa results and as prediced by the household model in Section V. 63  ii. Research Isitutions Most agricultural research in Zambia is conducted by the Research Branch of the Department of Agriculture, MAPP. Zambia has had an extensive network of agricultural research stations since the 1960s. There is at least one provincial station and one or more sub-station in each province. As of 1990, there were 171 scientists in the research service, but only 108 established posts; the balance has largely been met by donor-funded projects that recruit expatriates. Donor support has accounted for 30-50 percent of total recurrent and capital expenditure in the Research Branch. The agricultural research system has produced numerous technologies suitable for adoption by medium- and large-scale commercial farmers. The crop research system has made considerable progress in varietal and hybrid development. Higher-yielding varieties of maize, sorghum, soybeans, and other crops suitable for agro-ecological conditions in Zambia have been developed. In many cases these technologies were forwarded to smaliholders without addressing their particular constraints and needs. The result has been suboptimal utilization of such techmologies by smallholders (World Bank, 1992a). Furthermore, the focus on high-yielding, high input technologies, and monocropping has led to problems of environmental degradation, especially soil acidification. The livestock research system has focused on exotic breeds and cattle diseases, instead of native breeds. Irrigation research has developed technologies mainly suitable for adoption by medium- and large-scale farmers. Appropriate adaptive technologies for smallholders remain a strategic issue that the research program nust address. Adaptive Research Planning Teams (ARPTs) are based at research stations in all provinces (except for Southem province, which is planned to be started in the near future). The ARPTs specifically focus attention on smallholders. The ARPT system, which is based on a fanring systems approach, was initiated as part of the reorganization of the Research Branch in 1980. Participatory rural appraisals are used to identify smallholders' objectives, resources, and constraints based on agronomic, economic, social, and mntritional factors. The ARPT system is currently examining household food security and mtrition. Major themes for ARPT include crop diversification, decentalized seed banks in outlying areas, and low-input crop rotations that maintain soil fertility. These themes are in reaction to the past emphasis by all agricultural institutions on hybrid maize production. Other priority areas are labor-saving techniques for staple food production and processing. For example, groundnuts are a staple food crop that can also be a high-value cash crop if mechanical lifters and improved oil presses can be introduced. Decentralized seed banks, low input crops, and labor-saving technologies for staple food production and processing can help households achieve improved food security and nutrition. The Animal Drought Power Program has developed tillage, harvesting, and transport technologies that are being tested in conjunction with ARPT. Animal traction shows promise as a labor-saving device in areas and among farming systems where it is appropriate. Applied research into animal traction technologies for a variety of smallholders is required. In areas of the country where animal traction is widespread, the issue of animal health is extremely important and requires more research and investment. Research is needed on small livestock that can be integrated into existing cropping systems. Poultry and pigs are complementary to crop production, and can make better use of household labor (especially in the off-season), provide additional income to finance crop production, make use of 64  Sorghum and Milet: Bilding on Tradition Sorghum and millet are traditional food staple crops in Zambia. Since they are more drought tolerant than maize, they have been promoted in crop diversification strategies. However, a sorghum and millet development strategy need not be based solely on this perception of drought tolerance. These crops are versatile and can be used for food, feed, and beer. Research is needed into using these crops as substitutes for other grains such as maize and wheat. Most sorghum and millet is either retained by households or marketed in unofficial markets (often used as a barter good). Formal markets and improved technologies are needed to provide a steady supply of sorghum and millet to industrial producers. Technologies exist for smallholders and larger commercial farmers. However, only smallholders that are currently food and income secure, and located closer to urban and industrial centers, are in a position to adopt the new technologies. Some of the largest commercial farmers around Lusaka already grow improved sorghm varieties. These farmers tend to have livestock operations and food processing capacities that allow them to vertically integrate and reap maximwm benefits from this "traditional" crop. In some cases larger commercial farmers implement out-grower schemes with smallholders to produce sorghum. In these schemes, land preparation, seed, fertilizer, credit, and guaranteed markets are provided to smallholders. This type of a contractual arrangement is an example of how private economic agents can provide services traditionally provided (inefficiently) by the public sector. (Drawn from Rohrbach and Mwila.) surplus crop production, and improve household nutrition. Farming systems that integrate food security and nutrition, soil conservation and fertility, livestock production, and fuelwood and water issues are needed. Market liberalization will probably lead to increased on-farm storage of food staples in rural areas. The Food Storage and Conservation Program has developed low cost structures for safe on- farm storage. More research is needed in this area, especially in conjunction with food processing. 65  iii. Extension Institutions Extension activities are the responsibility of the Extension Branch of the Department of Agriculture, MAFF. In 1978, Zambia initiated an extension approach similar to the Training and Visit (T&V) system. The T&V system has only been partially introduced because of funding problems. Budgetary cut-backs have left the service with few resources to allow the relatively large staff to operate effectively. It is estimated that extension reaches only 25 percent of smallholders. Most disadvantaged farmers are not in touch with extension workers. Of the 2,400 extension staff only about 5 percent are women. Zambia has had a wide range of agricultural development programs and projects involving extension services as a major component. In recent years, these included four integrated rural development programs (IRDPs), nine area-based agricultural development projects and ten general projects. These projects have accounted for about 40-60 percent of the total budget allocation to extension services. MAFF is now seeking to develop a national, coordinated extension service in line with its EAP. The Lima program was introduced to facilitate the adoption of proven technological packages by smallholders. The program encouraged the use of inputs for the major crops (notably hybrid maize), on a standard unit of land using standardized measures of seed, fertilizer, and insecticide. The Lina program did not adequately address the labor constraints faced by smallholders to plant, weed, and harvest hybrid maize, but was very successful in spreading the adoption of hybrid maize. In cases where there may be appropriate messages to smanlholders, the ARPT system has had a major problem communicating results of its research to end users. Part of the problem is due to the lack of coordination with the extension service. ARPT has limited personnel and needs the cooperation of the extension service to spread its messages. ARPT messages, however, are not always compatible with those of the extension service. The extension service has been more closely linked to the mainstream research system that focuses its efforts on high-input agriculture. Coordination between ARPTs and the extension service needs to be strengthened to successfully target smallholders. iv. Information Institutions Within MAPF, the Agricultural Marketing and Logistics Information Center is responsible for collecting information on crop production, including regional crop and input prices, and producing a weekly report. This information is, in turn, made public. It is expected that the restructured MAFF will devote considerable resources to providing market and logistics information to both farmers and traders. Thus, the clientele of the ministry will be expanded to include non-farmers. The restructured MAFF may also take a broader view of agricultural development in the context of community development to help farmers learn how to obtain and use information, and deal with market forces in a liberalized economy. More attention is being given in MAFF to the interrelationship between agricultural production, food security, and nutrition. The Household Food Security and Health Monitoring System is a manifestation of this attention. The latter collects data on food availability and prices, livestock health, human health and nutritional status, and water quality and availability. 66  The National Agricultural Information Service within MAFF operates at provincial and district levels. In an era of market liberalization, new types of messages about prices and production, processing, and marketing possibilities is needed. More widespread information can only be achieved through a dynamic information collection and dissemination effort that is flexible enough to reach all types of producers. v. Agicultural Marketing and Input Suoply As discussed above, the present Government has taken some measures to liberalize grain marketing. The proposed Food Security Act will abolish Government's general authority to control the production and pricing of agricultural connodities and will establish a Food Security Agency. This agency will be prohibited from trading itself, except in relation to the national food reserve and donated commodities. It will operate a mnarket information system and register traders. This Agency wIll control all Goverwment-owned storage facilities and will be empowered to lease or sell them to persons involved in the production, marketing, or processing of agricultural commodities. There is concern that the proposed Agency will become a de-facto parastatal; care needs to be taken to ensure that the Agency is conpatible with the stated goal of market liberalization. Prior to 1989, the National Agricultural Marketing Board (NAMBOARD) marketed most commodities, especially staple cereals (maize, wheat, rice, sorghum and millet). Marketing of fertilizers and seeds was dominated by parastatals (NAMBOARD and ZAMSEED), and other agricultural inputs such as farm chemicals and machinery were marketed by cooperatives. Specialized parastatals (National Tobacco Company - NATCO, and Lint Company of Zambia - LINTCO) controlled the marketing of tobacco, cotton, soybeans, and sunflower. Other parastatals handled beef, milk and pork, although poultry and fish marketing was entirely private. For controlled commodities, both procurement and sale prices were regulated, and losses to the parastatals were covered by govermment subsidies. Transport rates were also regulated and subsidized. Subsidies for inputs and transport costs were a major drain on the GRZ budget. Also, because of the subsidies and losses absorbed by various parastatal and private businesses, the inputs and processed food such as maize meal and cooking oil did not consistently reach outlying areas. In June 1989, NAMBOARD was dissolved and the cooperative system was given the exclusive right of procuring and selling maize. The cooperatives were to act as buyers of last resort, at officially guaranteed floor prices, for some other commodities for which marketing had been liberalized. Nitrogen Chemicals of Zambia retained the monopoly of fertilizer production (25 percent of the country's requirements), import (commercial and donations) and distribution down to provincial level. Intra-provincial distribution of fertilizers was left to the cooperative system. Rural transportation in Zambia is made difficult by low population densities and weak institutional and financial capacity for road planning and maintenance. Rural transport is deficient because of poor roads and shortages of vehicles and spares. Poor rural roads and the lack of appropriate vehicles lead to high transport costs in remote areas. Market liberalization will cause transport costs to increase in the short-run, placing remote areas at a disadvantage. Increasing transport costs will affect the crops grown by farmers and the technologies they use. In addition, there will be greater incentives for local processing of food staple and oil-bearing crops, and searching for higher-value crops that can cover higher transport costs. 67  vi. Land Tenure Land ownership in Z:ambia is vested in the State. Farmers acquire usufruct rights either under the leasehold system or under custoomary land tenure. State land in agricultural use is leased mainly to large-scale commercial farmers for 99 years. Under customary tenure, access to Trust and Reserve land is granted by the traditional authorities; with the consent of the chief, Govermment can lease this land to individuals for an initial period of 14 years, which can be extended for up to 100 years. Customary tenure provides broad access to land. This type of tenure meets the needs of traditional farming on 1-2 ha, but does not necessarily meet the needs of commercial farmers for larger fields and the security of tenure and credit collateral. It is necessary to institute a land tenure system which provides incentives for long term investment in on-farm improvements (e.g. irrigation). Women's access to land varies from area to area; married women normally receive a separate field for their own cultivation, while single women theoretically have the right to acquire land for cultivation also. Even if theoretically available, land accessible to women may be less fertile, more distant or need clearing. Tradidonal authorities and District Councils often require a married women to obtain her husband's consent to acquire title to land. There is a great deal of debate about the appropriate land tenure policy for Zambia. There are fears that the rural poor might suffer from privatization of tribal land, because wealthier Zambians, and foreigners could buy up the land. These fears are fueled by the infusion of ex-civil servants to rural areas and immigrants from Zimbabwe and South Africa who are purchasing State Lands from private landowners or the government (in the case of abandoned state-farms), and even tribal lands from chiefs. The relationship between landholding and rural poverty is, as stated in the previous sections of this report, uncertain. Many of the arguments in favor of tenure reform are based on a perceived need for more investment to continue the growth of commercial agriculture. vii. Environment Changes in agricultural practices in Zambia have led to more permanent methods of agricultural production, and the subsequent abandoning of shifting cultivation and land fallowing. Fallowing allowed regrowth of vegetation and was thus a system of soil conservation. Fallow periods have been reduced or are non-existent, and soil fertility has declined. In addition, use of fertilizer to maintain fertility has led to acidity problems. The removal of fertilizer subsidies and decontrol of maize prices should help reduce fertilizer use in areas of high soil acidity and where its application has marginal impacts on yields. Clearng land for agricultural production is an important cause of deforestation, since expansion in land cuttivation has been the major source of agricultural growth. In addition, in many areas slash-and-bum (chitemene) is practiced. Cutting of forests for fuelwood and charcoal is a major problem. Since about 90 percent of all Zambians use fuelwood or charcoal, the problem of deforestation is critical, especialy in rural areas around Lusaka and Copperbelt towns. In Eastern and Southern provinces overgrazing is also a problem. Despite the abundance of land, continued deforestation and forest degradation cannot be sustained forever. The problems with deforestation are most critical in areas with relatively dense population. Thus, there is actually a situation of "induced" land scarcity in some areas. 68  The Supply System for Oxen Equipment Oxen equipment is generally available in the urban centers of Southern, Eastern, and Central provinces. in most of the towns there are shops that sell oxen equipment and spares. Some cooperatives are active in the trade of oxen equipment and other farm supplies. Whether these shops and institutions are able to meet peak demands after a good agricultural year is, however, questionable as stocks are limited. In other provinces the availability of equipment and spares is much more limited. In provincial capitals there is usually a shop with a stock of oxen equipment, and a cooperative or project may provide assistance in the procurement of oxen equipment. A more detailed look into the types of oxen equipment available at retail outlets reveals qualitative shortcomings in the supply systeir.. First, the range of equipment is often linmted to plows and cultivators, whereas farmers may require additional types of equipment like ridgers, harrows, and planters. Second, the equipment is generally of one brand, and the brands available in the shops vary from year to year. This leaves farmers with no opportnity to select the most suitable design, and difference sin designs of brands are substantial. The change of brands available over the years leads to problems with the availability of spares. Within the various types of oxen equipment a distinction must be made between equipment like plows, cultivators, harrows, and ridgers on the one hand, and ox-carts on the other. The first group of equipment is almost exclusively produced by large manufacturers and sold through traders and insiutions in the formal sector. Ox-carts are produced by larger anuactrers as well as small-scale manufacturers encounter serious difficulties in obtaining suitable wheels, hubs, and axles. Therefore, it would be very beneficial if the formal oxen equipment supply system would include the supply of these components. This would enable local nanufacturers and craftsmen to meet the widespread demand for ox-carts by farmers. (Drawn from Moll and Kahokola.) A National Environmental Action Plan (NEAP) and an Environmental Support Project (ESP) to implement the NEAP are currenly under consideration. It is expected that ASIP will be closely coordinated with NEAP/ESP. Under ASIP the expansion of cultivated area will mostly take place through the rehabilitation of existing under-utilized or abandoned farmland rather than through an expansion of agriculture on virgin land. Program activities would emphasize agricultural conservation practices, and restrict the use of program funds for environmentally sensitive areas. viii. Agricultmal Credit 69  The formal agricultural credtt systemn in Zambia consists of two separate components. One is operated by the commercial banks (mosdy private) that lend money primarily to medium- and large- scale commercial farmers and service operators, such as crop processors and truckers. The other is the publicly-supported smallholder credit system, in which, for the most part agricultural inputs rather than money are provided to farners. Commercial banks tend to have stricter collateral requirements, notably the requirement that the borrower possess title to the land or a formal lease. In contrast, the publicly-supported credit system is much more flexible in setting collateral requirements. Different collateral requirements also affect the types of loans available to borrowers. Only seasonal loans for relatively small amounts are granted with flexible collateral requirements, whereas larger sized medium- and long-term loans are only available with tangible collateral. Thus the existence or absence of titled land as collateral is more of a constraint with respect to the type of credit available to smallholders, rather than the avaUlability of credit. The fact that seasonal loans are provided as inputs rather than cash has two effects: It provides smallholders a set package of production inputs, and it constrains their ability to adjust to market conditions. Credit is provided mostly for one crop-maize-and credit programs have helped create and sustain the maize culture. Two commercial banks have small lending programs to smallholders, but they cover fewer than 2,500 farmers. In contrast, the three insdtutions oriented towards smallholders serve over 120,000 farmers. Other sources of funds are also important for certain crops, like LINTCO for cotton (and soybeans) and NATCO in the case of tobacco (and sunflower). The three major sources of formal credit for small-scale farmers are: a) the parastatal LIMA Bank, b) the Zambia Cooperative Federation - Financial Services (ZCF-FS), and c) the Credit Union and Savings Association (CUSA). In recent years, the ZCF-FS program has reached the largest number of farmers (through the cooperatives), but in very small amounts and for just over one hectare on average. CUSA also lends a small amount per hectare, but for somewhat larger areas - about 2.4 hectare on average. LIMA Bank tends to lend more per hectare and for larger areas, compared to ZCF-FS. Negative interest rates and low repayment rates make loans attractive for individual farmers, but make financial institutions unsustainable. These rates also result in credit rationing. Rationed credit has traditionally gone to more prosperous and politically well-connected farmers. Thus, a major issue in Zambia has been not the cost, but the availability of credit. The movement toward positive real interest rates will increase the cost of borrowing and probably lead to more attempts at self-financing and to changing crop mixes and technologies by some smallholders. A number of legislative changes are proposed to assist the liberalization measures and to provide a well-defined set of rles for market transactions. Credit legislation is also being promoted that would allow the lending institutions to register charges against the assets of borrowers. A loan guaraee program is being considered to encourage banks and other businesses lend to farmers and traders. Also, credit for medium- and longer-term loans need to be available to smallholders, especially to purchase oxen and equipment. Such a credit program will require smallholders to provide some type of collateral to banks. Alternative methods for providing collateral by smallholders including land titling and group lending need to be explored. ix. C2gggmav-es 70  The cooperative movement has a four-tier structure, starting with primary societies at village or grassroots level. Each lower level in the cooperative structure swns the higher level by owning shares. Prinay societies are affiliated, through membership. to District Cooperative Unions (DCU) which, in tum are affiliated, to Provincial Cooperative Unions (PCU). Fmnally, all PCU are affiliated through nembership to the Zambia Cooperative Federation Limited (ZCF), which is the apex organization for the movement in Zambia. The Ministry of Cooperatives and Marketing has historically been used to transform GRZ policies and directives into practice. This Ministry needs to become more a partner with ZCF and a facilitator-of improved decentralized cooperative organizations and market systems. Primay Cooperative Sodeties. There are more than 1,300 registered primary societies with a membership in excess of 400,000. The main economic activities at this level include: (a) Grain marketing and storage; (b) Marketing of agricultural inputs; and, (c) Provision of services to members such as consumer shops, agro-processing facilities and agricultural credit. Distict Cooperatives (DC). There are about 30 registered DCUs. The DCUs were formed more recently in response to government policies aimed at decentralization of agricultural marketing. The main activities at this level include: (a) Crop marketing; (b) Transport operations; (c) Sale of equipment and farm requisites; and, (d) Promotion of cooperative development. Provincil Cooperative Unions (PCU). Each of the nine provinces has a PCU. Their major economic activity was to carry out crop marketing but most have diversified into activities such as haulage and passenger transport operations, motels, farms, etc. The majority are now concentrating on developing agro-processing facilities. Mhe Zambia Cooperative Federation limited (ZCI9. ZCF, the apex body, is essentially a service organization in relation to its member cooperatives. It has received considerable external support from donors and it has developed as a provider of specialized services, through its subsidiary companies. There has been unbalanced development within the cooperative system. The Primary Cooperative Societies have the weakest resource bases and management capabilities. The PCUs have been the focal point for most economic activities. The PCUs are better managed and receive most of the Government and donor development effort and support. Like other agricultural institutons, the role of ZCF and other tiers of the cooperatve structure will undergo a transformation to integrate into a market economy. Primary Cooperative Unions need to shoulder much more responsibility as the economy evolves from centralized to decentralized market fimctions. Also, to efficiently deal with market liberalization, more small cooperatives (or groul -) must be formed to pool resources to reduce transactions costs. By reducing transaction costs, support services can be provided to more households at lower private and public costs. ZCF, especially 71  Agricultmal Credit: L1DA Bank In 1992/93 there were 43,000 customers, mostly for seasonal loans (K5 billion was disbursed). About 90 percent of the customers were small-scale (1-2 ha) and 10 percent medium scale (2-5 ha). Loan applications are receved by June 30 and processed from June to November. Problems facng the LIMA Bank include: 1. Not enough account officers to process loans. One loan officer appraises about 3,000 loans. 2. Long distances traveled to lead and monitor loans. 3. Poor recovery rates. Historically, loan recovery rates have averaged about 40 percent. Once a farmer defaults, he/she is not allowed to apply for a new loan from LIMA Bank. 4. LIMA Bank receives funds from Government which, in turn, dictates the rate of interest to be charged on loans. In 1992/93 the interest rate was set at 48 percent/year, despite inflation rates of over 100 percent. S. Deficient management information systems (MIS). LIMA Bank really has no idea about the performae of its loan portfolio because the MIS is a mess. Audited accounts are four years behind. Because of high transaction costs, poor recovery rates, and negative interest rates, LIMA Bank depends on Government contributions to replenish its coffers. However, because of the poor MIS, the extent of the problems are not really known. Although LIMA Bank is not a sustainable insttuton, as a provider of cheap credit it is very popular among farmers. In 1992/93 there were 72,000 loan applications and 43,000 loans. For the 1993/94 season there were 130,000 applications. Government is trying to direct donor-provided seed and fertlizer to LIMA Bank to lend in- lind to farmers. The bank is already accepting payment in-kind for past loans. This is compounding LIMA Bank's poor financial position, as it must absorb the cost of handling (i.e., ransportng and storing) maize and other crops. Plans for the Future: The new anag director has previous experience in commercial banks. His vision for LIMA Bank is to resuure it so that it can operate on a commercial basis. The lending program must be operated on a commercial basis, with cost recovery for adminirative costs, sound loan appra techniques and aggressive recovery of loans, flexible interest rates that guarantee borrowers pay real interest on loans. Obviously a sound management information system is rqied. This ciaizaton would change the target of LIMA Bank's lending. Small loans are cosly, and hard to monitor. Thus, there would be a bias against small-scale farmers trying to obtain individual loans. There would be a bias towards groups of farmers that receive and guarantee loans as a group, and towards medium-scale farmers. A major effort by agricultural institutions and donors to help smallholders (especially resource-poor smallholders in remote areas) organize into groups is needed. through DCUs, needs to help small groups organize into cooperatives. 72  x. Rural Industn There is increased awareness in Zambia that the lack of non-agricultural economic activities has exacerbated the rural poverty problem. Problems include the high cost in time and human energy that women devote to hand processing of staple foods, the food insecurity that results from selling staple foods to centralized buyers and then purchasing processed meal, the lack of activities linked to agricultural production that can provide income generation in rural areas, the lack of consumer goods that can provide incentives for households to produce larger marketable surpluses using improved inputs and technologies and infrastructure, and the lack of production, repair, and maintenance facilities for farm implements. Much of the non-agricultural activities that exist in rural areas are related to food processing, especially the milling of staple foods. There has been considerable effort to stimulate ecor.nmic activity in rural areas based on food processing, notably hammermills. The lack of rural industry is due to several factors. Most important are the historical centralization of industrial, financial, and administrative activities in urban areas, and the low-level of technology and small amounts of marketable surpluses generated by smallholders. Another constraint to rural industry is the lack of credit. There is no tradition of commercial loans for non-agricultural activities. Many new enterprises have the potential to be financially solvent, but access to short- and medium-term credit is a mnajor constraint. Other constraints to rural industrialization in Zambia are the lack of electricity and infrastructure. The Small Industry Development Organization (SIDO), a Government-funded agency, promotes small-scale industry throughout Zambia. SIDO's projects provide direct and indirect benefits for the rural poor. They have increased availability and lower prices of consumption goods, and new income-generating activities. Village Industry Services (VIS) is an indigenous Zambian NGO dtat promotes rural development by focusing on labor-intensive cottage- and village-based industries. VIS, which was established in 1976, is mostly involved in agro-processing enterprises. Since 1990, USAID-funded Zambia Agribusiness Management Services (ZAMS) has worked together with SIDO, VIS, and ZCF to provide commercial loans (20 percent down and repayment over two years), and technical assistance in maintenance and repair (to both owners and mechanics). In addition, hammermill owners receive training in business principles and are encouraged to expand activities to include storage, sale of inputs, and the sale of consumer goods. Thus, the hammermill program has been successful not only in terms of providing milling services, but also because hammermills have served as a catalyst for other economic actiities. Some hammermill operators also provide credit facilities. In addition, these centrally-locat6' facilities are used as centers for the distribudon of information on health. xi. Local Governments The Local Administration Act of 1980 was a first step toward improving the capacity to plan and make decisions at local levels. The Act was, however, inadequate because there was no effective strategy for implementing its objectives. It also did not explicitly recognize the importance of district govermnents. The lack of effective local governments is a major constraint to market liberalization. 73  National Hammenmlfl Prrgram - A Plan for Rural Revtalization Hammermills are important for household-level food security in rural areas for several reasons. First, hammermills can save a considerable amount of time and energy for women. Hand-pounding of maize is time- and energy-consuming. This time and energy could be used for food production and caring for children. Second, with hammermills nearby, rural households have more incentive to store food at home and thereby control their food supply, rather than sell staple foods and buy processed meal and oil. Govemment support for hammermills in rural areas began in the raid-1980s, with SIDO, VIS, and ZCF. At that time Goverment interfered in the selection of individuals or groups that would receive hammermills and the loan terms (which were highly subsidized). In the end of 1989, Government officially introduced a National Hammermill Program. The program was resuscitated with assistance from the Zambia Agri-business Management Services (ZAMS) Project, funded by USAID. ZAMS provides technical assistance and acts as a facilitator, and does not lend money. In 1990, there were a few hundred hammermills in rural areas in Zambia. By 1993, there were about 5,000, of which about two-thirds were in working order. Prior to ZAMS participation, decisions on site selection for hammermills were largely political, and loans were highly subsidized. Many of the hammermills needed maintenance and repair. Since 1990, ZAMS has worked together with SIDO, VIS and ZCF to provide commercial loans (20 percent down and repayment over two years), and technica assistance in maintenance and repair (to both owners and mechanics). In addition, hammermill owners receive training in business principles and are encouraged to expand activities to include storage, sale of inputs, and sale of consumer goods. The hammermill program has been successful not only in providing milling services, but also because the hammermills have served as a catalyst for other economic activities. Some hammemnill operators provide credit facilities. Also, these centrally-located facilities are used as ceters for the distribution of information on health and nutrition. Commercialization and expansion of private enterpnse are critical for rwual development and poverty alleviation under the new regime of market liberization. ZAMS and VIS are devoting considerable resources to training workshops that educate individuals and groups in business and en neuial skills, in addition to technical skills. Teaching rural Zambians business principles is a major challenge. Training-teaching people how to tplay the free market" has, thus, been identified as a major objective. Many of the new hammermdl owners are retired or active civil servants, people with assets, human capital, and contacs in urban areas. These individuals act as agents of commercialization. More attention is now being devoted to establishing links between urban business people and rur ameas to increase the level of commercialization (that is, to decentralize economic activity from urban centers), and to increase rral-urban linkages. To provide suitable conditions for a market economy and rural development, local govenments (preferably at the district level) need to be able to raise revenues from an expanding 74  agricultural base. This revenue generation will require technical assistance for local governments to learn how to function as autonc nous financial units. Despite symbolic support, central government institutions and donors are more reluctant to experiment with political decentralization than with economic decentralization. The desire to control, both by central ministries and donors, is a major impcdiment to adopting rational decentralization policies enabling district councils to lead development efforts in rual areas (Pudsey, et al.). C. Donor Support More than twenty official donor agencies and a large number of NGOs have been involved in providing either financial or technical support to Zambian agriculture in recent years (see Table IV.8). Many donors are now awaiting clarification of policies and institutional responsibilities before committing to major new programs. The World Bank supports ZAREP, a coffee project in Northern province, and the new Marketing and Processing Infrastructure Project (detailed in Amnex II). The World Bank is also supporting the ASIP. UNDP/FAO supports the Early Warning, Extension and Storage projects; the agricultural extension and storage projects will be extended from Southern province into Eastern and Western provinces. The World Food Program is supporting an umbrella project to provide a safety net for vulnerable groups under structural reform (see descrption of Program Against Mal-nutrition in Annex II). The principal components of the safety net are food-for-work projects, health and nutrition programs, and micro-projects. The European Community has provided substantial support in the livestock sector, and also supports area-based projects in Central and Copperbelt provinces. Major bilateral donors, such as Sweden, Norway, The Netherlands, and Germany, are continuing to support area-based projects.21 USAID is concentraing its support on policy reform and capacity building. There have been many projects designed to reduce poverty in rural Zambia. A detailed summary of these projects is presented in Annex II. Some have been successful and some have not. A broad overview of past and ongoing projects in Zambia is presented in Table IV.9. The typology and descriptions found in these tables convey a perspective on the changing nature of projects. Four types of projects have been identified: 1. Production-Oriented Proiects, aimed at increasing agricultural production in rural areas by introducing improved inputs and technology. Most projects are related to introducing hybrid maize producing using fertilizer, with or without oxen; 2. Basic Needs Projects, aimed at providing basic social needs such as education, health, water, and sanitation; 3. Infrastructure Projects, aimed at providing infrastructure such as roads, bridges, and 28Sweden has projects in outlying areas of Luapula and Eastem provinces, Norway in the Luangwa Valley, The Netherlands in Western province, and Germany in Gwembe Valley. 75  storage sheds are often carried out in conjunction with production-oriented projects; and 4. Emergency Aid Projects, have been provided to rural and urban areas during times of serious food shortages brought about by drought or other natural disasters, or brought about by political upheaval. More than twenty official donor agencies have been involved in providing either financial or technical support to Zambian agriculture in recent years. Project implementation thus far has been extrenely disappointing for most externally-funded projects. This disappointment can be attributed to two major factors: Excessive centralization and fragmentation of donor support. While there have been new attempts at devolution of responsibility, most government spending continues to be managed by central ministries, and state, and parastatal agencies, with virtually no involvement of local populations. Ex-ante control over expenditures and procurement remains largely centralized. Each donor continues to negotiate special procurement and disbursemnent procedures, undennining any effort at improving the national fiscal control mechanisms. Donors have provided generous project support to Zambia's agricultural development. However, this assistance has not always been used efficiently. Among donors there is a consensus tht providing project assistance to agriculture has failed to achieve improved agricultural growth. Donor support to agriculture has been characterized by fragmentation, where each donor has provided support on a project-by-project basis in isolation from, and sometimes, in competition with other donors. Donor support has also been poorly coordinated, both among donors and with respect to Government strategies and priorties. For example, there are over 100 independent donor-funded projects in agriculture. A major objective of the ASIP will be to coordinate donor-funded projects. The problem, however, is much broader than just a lack of coordinated activities. The objectives and standards of the projects often exceed local management and financial capacity. Projects are therefore staffed with long-term expatrates that contribute little to the development of local capacity and project sustainability. Lack of capacity at the district level and historical centralization of service provision has led many donors to conclude that the only effective delivery mechanisms for their programs are NGOs. Many NGOs support projects in rural Zambia. These NGOs are heterogeneous, with some specializing in certain activities and in certain areas. Some of the larger organizations like World Vision and OXFAM are active in a number of areas in Zambia while smaller NGOs tend to concentrate their efforts in localized areas. Most, if not all, receive funds and resources either from large bilateral donors, such as the Nordic countries or multilateral organizations such as UNICEF and other UN agencies. NGOs were successfiul at delivering drought relief during 1991/92, and this success has propelled them into the forefront of discussion about delivery of programs for the poor. Many NGOs have real presence in the villages and districts, and are thought of as an alternative to Government institutions in reaching the poo:. Stories of successful micro-project implementation, capacity building, and local empowerment have helped raise their stature. 76  Table IV.8. Funfdt for Projects in Zambia. Ministry of Agulre Food and FiLeries A Schedule of Foreign Asanc and GRZ Contribution for the Year 1993 DONOR GRZ TOTAL DONOR PROJECT TLE DONOR (K) (K) (K) % tegrated Rura Development Pogranmn SIDA 90.400.000 4,320,000 94.720.000 95 Integrated Rural Develpmet Progamme, NW Provue GZ 40,000.000 8,432,640 48,432,640 83 North Western Province Area Development Project IAD 234,938,000 66,019,200 300,958,200 78 Gwambe Intgted Disict DI-' elopment Project 07Z 40,96,000 44,2f4,720 85,246,720 48 Nationl Ea ly Warming and Census of Agcultre FAO/DUTCH 118,337,000 34,560,000 152,897,000 77 Support to Agri:ult Planning SIDA 44,000,000 34.560.000 78,560,000 56 Gwembe South Develpment Project GOSSNER 37,863,000 48,917,760 86,780.760 44 ZAREP ADF/IDA/NORWAY 1.000,000.000 67,392.000 1,067,392,000 94 coffee (11) Project IDA 73,500,000 6,350.400 79.850.400 92 Palabana Daiy Trahing Instiu BEC 16,109.000 3,067.20D 19.176,200 84 Regional Tset aMTndT omiai Control Pect EEC 444.498.000 103,680,000 548,178,000 81 Animal DLase Control Project W/provlnce DUTCH 41,150,000 21.729,600 62,879,600 65 SADCC Regi. Tste & Tiparwsomiasis Tining Cnr. EEC 785,598,000 25,920.000 811,518,000 97 Livestock Coopeadve Sociey EEC 477,00,000 0 477.000.000 0oo Soybean Research and Deveklment Project CIDA 32,700,000 10,972,800 43,672.800 75 Adapdve Resa Plg Team SIDA 63,830,000 8,984.640 72,814,640 88 Cropping ResearchPrramme ITA 811,000 9,293,760 9,904,760 6 Cereals Researh Progamne CIDA/SIDA 147,364,000 17.280,00 164,644,000 90 Oil Seeds Reseh Prgamme CIDA 9,693,000 22,025,280 31.718,280 31 Pln Goen. Resomuce Progmme SADCC 1,082,000 9,789.120 10.871,120 10 Root & Tubers Resach Programe SIDA 31,500,000 8,640,000 40,140,000 78 Vegetable Research Programme SIDA 16,625,000 23,328,000 39,953,000 42 Livestock and Pasure Research Pogrmme SIDA 21,525,000 17,280,000 38,805,000 55 Soils ResarchPmme NORAD/IBIAMMEA 48,771.000 25.920,a00 74,691,000 65 Golden Valley Reseach Staon Devel. Pogamme SJDA 196,350,000 17,280,000 213,630,000 92 Vetriary Research UNDP/LAEA 63,816,000 11.476,800 75.292.800 85 Econ. of Tics & Tickbome Disease Control Project FAOIDANNIDA 10D,000,000 0 100,000,000 100 NFDC SIDA 15,500,000 46,656.000 62,156,000 25 Fab Culture Devdopmen Prject DUTCH 19,149,000 4,320,000 23,469.000 82 Fisb Culture Adaptve Reach NORAD 7,099,000 7,688,640 14,787,640 48 77  ,~~ |# r o o ! . oW W 8u S 8 & X X^ t 00 -J~ ~ o