ENVIRONMENTALLY AND SOCIALLY SUSTAINABLE DEVELOPMENT STUDIES AND MONOGRAPHS SERIES 20 Work in progress for public discussion Voices of the Poor Pouertyvand Soriýa/ Capital li inmzwiniaI overT AP ¾ 'XL> N KEV, Deepai Xarawyan ESSD Proceedings Series 1 Culture and Development in Africa: Proceedings of an International Conference (Also in French) 2 Valuing the Environment: Proceedings of the First Annual International Conference on Environmentally Sustainable Development 3 Overcoming Global Hunger: Proceedings of a Conference on Actions to Reduce Hunger Worldwide 4 Traditional Knowledge and Sustainable Development: Proceedings of a Conference 5 The Human Face of the Urban Environment: A Report to the Development Community 6 The Human Face of the Urban Environment: Proceedings of the Second Annual World Bank Conference on Environmentally Sustainable Development 7 The Business of Sustainable Cities: Public-Private Partnerships for Creative Technical and Institutional Solutions 8 Enabling Sustainable Community Development 9 Sustainable Financing Mechanisms for Coral Reef Conservation: Proceedings of a Workshop 10 Effective Financing of Environmentally Sustainable Development: Proceedings of the Third Annual World Bank Conference on Environmentally Sustainable Development 11 Servicing Innovative Financing of Environmentally Sustainable Development 12 Ethics and Spiritual Values: Promoting Environmentally Sustainable Development 13 The Self and the Other: Sustainability and Self-Empowerment 14 Meeting the Challenges of Population, Environment, and Resources: The Costs of Inaction 15 Rural Well-Being: From Vision to Action ESSD Studies and Monographs (formerly Occasional Paper) Series 1 The Contribution of People's Participation: Evidence from 121 Rural Water Supply Projects 2 Making Development Sustainable: From Concepts to Action 3 Sociology, Anthropology, and Development: An Annotated Bibliography of World Bank Publications 1975-1993 4 The World Bank's Strategy for Reducing Poverty and Hunger: A Report to the Development Community 5 Sustainability and the Wealth of Nations: First Steps in an Ongoing Journey 6 Social Organization and Development Anthropology: The 1995 Malinowski Award Lecture 7 Confronting Crisis: A Summary of Household Responses to Poverty and Vulnerability in Four Poor Urban Communities (Also in French and Spanish) 8 Confronting Crisis: A Comparative Study of Household Responses to Poverty and Vulnerability in Four Poor Urban Communities 9 Guidelines for Integrated Coastal Zone Management (continued on the inside back cover) ENVIRONMENTALLY AND SOCIALLY SUSTAINABLE DEVELOPMENT STUDIES AND MONOGRAPHS SERIES 20 Voices of the Poor Poverty and Social Capital in Tanzania Deepa Narayan The World Bank Washington, D.C. Copyright @ 1997 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing October 1997 This report has been prepared by the staff of the World Bank. The judgments expressed do not necessarily reflect the views of the Board of Executive Directors or of the governments they represent. Cover: This poverty map was drawn by a group of district officials in the Sumbawanga District in Tanzania as part of the Participatory Poverty Assessment. District workshops were held to get district officials' perspectives on poverty in their area and to give them the feedback provided by poor people. Deepa Narayan is senior social scientist in the Environment Department at the World Bank. Library of Congress Cataloging-in-Publication Data Narayan-Parker, Deepa Voices of the poor : poverty and social capital in Tanzania / by Deepa Narayan. p. crn. - (Environmentally and socially sustainable development studies and monographs series ; no. 20) Includes bibliographical references. ISBN 0-8213-4061-1 1. Poor-Tanzania. 2. Poverty-Tanzania. 3. Human capital- -Tanzania. 4. Tanzania-Economic policy. 5. Tanzania-Social policy. I. Title. II. Series. HC885.Z9P6157 1997 362.5'.09678-dc2l 97-34898 CIP The text and the cover are printed on recycled paper, with a flood aqueous coating on the cover. Contents Foreword vii Abstract viii Acknowledgments ix Acronyms and Abbreviations xi Chapter 1 Poverty, Social Capital, and Survey Methodology 1 Chapter 2 What Is Poverty? 7 Chapter 3 Inflation, Price Trends, and Coping Strategies 19 Chapter 4 Gender Perspective: Development for Whom? 30 Chapter 5 Constraints on Agricultural Productivity 40 Chapter 6 Social Capital 50 Chapter 7 Credit and Savings 66 Chapter 8 Conclusion 75 References 79 Boxes 2.1 Sliding into poverty: the story of Ndesheyo Esau, Maroroni Village, Arumeru District 15 2.2 Distant markets, poor roads, and poverty 17 3.1 Transport problems in Choma, Tabora 24 3.2 Declining health status 24 3.3 Coping strategies for food shortages 26 3.4 "It is disease that forces us into poverty." Kweshwpo, Tanga 26 4.1 Giving life to statistics: men's voices 32 4.2 "How can you face your children day after day, hungry?" Women's voices 32 4.3 "The water problem is very serious." Butandula, Tabora 33 4.4 "Once the husband is drunk, he beats the wife." Nsendakamoge, Tabora 34 iii iv Voices of the Poor: Poverty and Social Capital in Tanzania 4.5 "We are not living, we are just surviving." Women's group, Utende, Mtwara 35 4.6 "A woman cannot own anything valuable." Bumazi, Kagera 38 4.7 "Even the father hesitates in welcoming her after divorce." Tella, Kilimanjaro 38 5.1 "The good soil is lost." Butandula, Tabora 42 6.1 Trust and trustworthiness 51 6.2 Roman Catholic Church in Utende, Mtwara 56 6.3 Women's groups 57 6.4 The evolution of security groups in Butandula, Tabora 57 6.5 Profiles of youth groups 59 7.1 Women's savings institutions 70 7.2 Cows: high risk, high return 71 7.3 The ebb and flow of Upatu groups 71 7.4 Barriers to use of savings banks 71 7.5 Experiences with postal banks 72 7.6 Perceptions of CRDB 72 7.7 Experiences with NBC 72 Tables 2.1 Characteristics of poverty and wealth, Dodoma region 10 2.2 Characteristics of poverty and wealth, Kilimanjaro region 10 2.3 Importance of cash crops and crop yields by wealth category, Kagera and Rukwa regions 12 2.4 Importance of health and education services by wealth category, Tanga region 13 2.5 PPA national wealth ranking 13 2.6 Gender differences in findings between PPA and HRD studies 13 2.7 Well-being ratings by wealth compared over time 14 2.8 People's perceptions of changes in village prosperity and inequality 14 2.9 Causes of poverty as perceived by village groups 16 2.10 Vision for the future 17 3.1 Price trends in Tabora, 1985-95 21 3.2 Price trends in Shinyanga, 1985-95 21 3.3 Price trends in Kilimanjaro, 1985-95 21 3.4 Price trends in Kigoma, 1985-95 21 3.5 Measuring "access" to water underestimates the problem 28 4.1 National distribution of number-one problem by gender 31 4.2 National distribution of number-two problem by gender 31 4.3 Reasons for preferring to withdraw girls from school 34 4.4 Summary statistics on male- and female-headed households, PPA study 35 4.5 Ownership and distribution of assets upon separation/divorce, for Singida (Unyamikumbi and Nsunsu villages) 36 4.6 Ownership and distribution of assets upon separation/divorce, for Coast (Disunyaia village) 37 5.1 What villagers perceive as constraints in agriculture 41 5.2 Reasons for moving to current village 41 5.3 Use of inputs by wealth category, 1993-94 42 5.4 Reasons for not using fertilizers 43 5.5 Reported decline in soil fertility 43 5.6 Source of inputs by wealth category 43 Contents v 5.7 Source of credit by wealth category 44 5.8 Reason for not using agricultural credit 44 5.9 Amount of money borrowed 44 5.10 Reasons for borrowing 45 5.11 Most important crops grown 45 5.12 Where the most important crop was sold 45 5.13 Problems in marketing 46 5.14 Last message heard from extension worker 46 5.15 Outlook for agricultural operations 46 5.16 Gender differences in use of agricultural inputs 47 5.17 Gender differences in source of input 48 5.18 Gender differences in being informed about political change 48 5.19 Gender differences in sources of farming advice 48 5.20 Gender differences in last contact with extension worker 49 6.1 Are there more or fewer people you can trust now? 52 6.2 Reasons for declining trust 52 6.3 Change in trust toward different groups 52 6.4 Ratings in unity of the village 52 6.5 Membership in informal and formal groups and associations 53 6.6 Overview of associational activity 55 6.7 Groups important to people 55 6.8 Why people join groups 55 6.9 Thedisadvantages of group membership 55 6.10 Household expenditures per person and social capital, comparing village- and household-level and using HRD survey data for incomes 62 6.11 Household expenditures per person and social capital using SCPS data 63 7.1 The first thing you would do with 50,000 shillings 67 7.2 Where do you currently save? 68 7.3 Prevalence of savings institutions, national 68 7.4 Preferred attributes of savings institutions 69 7.5 Reasons for saving 69 7.6 Qualities new banks should have 70 7.6 Who would you first approach to borrow 2,000 shillings and 50,000 shillings? 73 7.7 Gender differences in degree of access to credit 73 Figures 2.1 Community map of Luana Village, Iringa 9 3.1 Seasonal analysis chart 20 3.2 National upward trends reported in villages 22 3.3 National downward trends reported in villages 23 3.4 Seasonal calendar, Tabora 27 6.1 Venn diagram of Bunazi village, Kagera 54 6.2 Venn diagram of Nsungu, Singida 58 嗡口〞 Foreword overty is a complex and multidimen- Second, the study demystifies the concept sional phenomenon. While it has long of social capital at the local level. Social capi- been recognized that poverty is more than tal and social organization have long been can be defined by income and expenditure data, recognized by sociologists and anthropolo- progress in going beyond purely economic indi- gists as critical building blocks of community cators has been stymied by a lack of measure- and society. This study provides quantifiable ment tools. This study is a major step forward in evidence that village-level social capital- the development of such tools in two ways. membership in groups with particular charac- First, it reveals the power and utility of par- teristics---significantly affects household ticipatory methods for policy research. It dem- welfare. In one telling statistic the study finds onstrates practices and processes that can tap that a one standard increase in village-level the knowledge the poor have about their own social capital increases household income reality. By comparing and contrasting findings per person by 20 to 30 percent. By compari- from participatory methods with findings from son, a one standard deviation in schooling- more conventional household consumption nearly three additional years of education and expenditure surveys, the study identifies per person-increases incomes by only 4.8 poverty indicators used by local people at the percent. Clearly, this finding is critical to village level. It also shows how using these formulating decentralization policies de- measures leads to different conclusions about signed to provide basic services to the poor the causes and nature of poverty and reveals and engage in management of local re- the policy and institutional choices that can sources. best address the problems. In addition the Voices of the Poor suggests that, in the global study focuses attention on the different "reali- challenge to eradicate poverty, the poor are the ties" of male- and female-headed households true "poverty experts." All of us, inside and and shows how development policies must outside the World Bank, would be wise to lis- take gender differences into account if they are ten to their voices and benefit from their expe- to be effective. rience and their wisdom. Ismail Serageldin Vice President Environmentally and Socially Sustainable Development vii Abstract In 1995 the Government of Tanzania, the in prices in rural areas, the widespread need University of Dar es Salaam, and the World for rural credit, and people's desire to save. Bank conducted a Participatory Poverty Further, the study found that while many face Assessment (PPA) involving more than 6,000 hardship as a result of cutbacks in government people in 87 villages across Tanzania. Specific subsidies due to liberalization policies, few ad- areas on which data were gathered included vocate returning to subsidies and other prac- such indicators as how the poor view poverty, tices of the past. Rather, they seek expanded perceptions of inflation and other trends over opportunities to improve their livelihoods time, the effets of liberalization on the rural (through access to transportation to move their poor, access to and use of formal and informal goods to market, for instance). credit and savings institutions, the role of so- Perhaps the most striking finding of the PPA cial capital in household welfare, and the rela- was the contribution of village-level social tionship between poverty and the natural capital to household incomes. A one standard environment. deviation increase in social capital at the vil- The PPA employed methods and tools de- lage-level was found to increase household ex- signed to facilitate "listening to the poor." penditures per person by at least 20 to 30 Three basic types of data collection methods percent. By comparison, a standard deviation were used: participatory tools (such as commu- in schooling of nearly three years per person nity mapping, group discussions, and Venn increased income by less than 5 percent. diagrams); key informant interviews; and The logistics of mounting a PPA are by no household survey instruments. Different sam- means simple, but one can be organized more pling techniques were employed for the vari- quickly and produce results faster than a tra- ous data collection methods used. ditional survey based on questionnaires. For The issue of gender was of central impor- this reason, the PPA technique may be very tance to the PPA, and significant differences useful for interim monitoring of poverty be- between the perspectives and actions of men tween major surveys. and women were documented. Particular at- In the case of Tanzania the aggregate results tention was paid to the priorities and coping of the two approaches were very similar. At strategies of female-headed and male-headed the same time, the PPA generated more subtle householos. and more detailed findings in a number of ar- Among the PPA's findings: the importance eas, and its policy conclusions are therefore of agricultural inputs, simple technology, and more discriminating, and in some cases differ- the inability of the poor to take advantage of ent, for example on questions of social capital, markets. The PPA also revealed the inflation gender, seasonality, and access to water. Viii Acknowledgm ents T he Participatory Poverty Assessmenttheir dedication there could have been no Par- (PPA) study was conducted by the Gov- ticipatory Poverty Assessment in Tanzania. ernment of Tanzania, the University of The team was composed of Tanzanian so- Dar es Salaam, and the World Bank, with fi- cial scientists from the University of Dar es nancial assistance from the British Overseas Salaam, government departments, and re- Development Agency and the World Bank. The search institutes. Field workers were drawn study was designed and led by Deepa from all walks of life and were largely re- Narayan. The study team was managed in cruited through a newspaper advertisement Tanzania by H. R. K. Amani, University of Dar and word of mouth. This volume draws pri- es Salaam. ODA's finanancial support was marly on the content analysis and field notes administered by Peter Llewellyn, British of the 29 field workers who generated more Council. than 100 notebooks of village-level data. A sec- Guidance was provided by several offi- ond source of information was a set of draft cials from the Government of Tanzania and regional reports written by the team supervi- the World Bank: Messrs. Kipokola, Kigoda, sors, as well as approximately 1,500 household and Zayuma, and staff of the Planning Com- questionnaires. mission; Rose Lugembe of the Department of The members of the regional PPA teams Statistics, the Ministry of Agriculture; and were: Luisa Ferriera, Alberto Agbonyitor, James Coats, and Emmanuel Malangalila from the Kagera, Kigoma, Rukza World Bank. Valuable comments on the draft J. Baghdellah were received from Ronald Brigish (Tanzania Andrew John Resident Mission), Albert Agbonyitor, James Restituta Marco Adams, and Caroline Robb, World Bank. Julianus Thomas Mutayoba Tanzania PPA Team Prudensiana We wish to thank the members of the Tanzania Dodoma, Singida, Tabora PPA team who conducted interviews, faaisi- Angelina Ballart tated discussions, and recorded data from Sabato Manyama more than 6,000 residents of 87 villages D. Meshala throughout Tanzania. The thoroughness and Christer Msuya seriousness with which they undertook the Geoffrey Mutashubirwa field work made this volume possible. Without Chambi Sasadnrka ix x Voices of the Poor: Poverty and Social Capital in Tanzania Arusha, Kilimanjaro, Tanga Along with the members of the PPA Team Venus Kimei operating in Tanzania, a number of people Raphael Kombo played invaluable roles during various phases Abdiel Mushi of the research project. The training of the team Rosemary Ruhinda members was conducted by Miranda Munro, Ebenezer Samuel Mungutti Kitui Kitua, and H. R. K. Amani. For- matting of data forms was guided by Lucy Mara, Mwanza, Shinyanga Goodhart. Data entry of survey results was Patrick Masanja carried out by Gemini in Nairobi. Sushenjit Victoria Chami Bandopadhyay was responsible for the quality Abdul Ibrahim of quantitative data and for statistical analysis; Selina M. econometric analysis was overseen by Lant P. Shoki Pritchett, Deon Filmer, and Jonathan Isham; Ferdinand Soka and Katie Youdel conducted the content analy- sis. Ted L. Howard was editor of the final Iringa, Mbeya, Ruvuma report, and Andrea Brunholzl was the copy- Mtatafikolo editor. Gaudencio Dizon desktopped the David Charles volume. Terraline Kaaya The most important acknowledgment, of Phina Lyimo course, is due to the thousands of people in the Raphael Mgowole villages of Tanzania who took the time to speak Theophilo Tweve with the PPA team, and who opened their lives to outsiders in the hope that their experience Lindi, Morogoro, Mtwara, Coast and information would reach the policymakers Charles Saanane who have the power to make decisions that Sarah George affect the well-being of the rural poor. Albert Kihunwa Gladys Kitilyo Christopher Mramba N.P. Mnkeni Acronyms and Abbreviations CCM Chama cha Mapinduzi (Party of NGO nongovernmental organization the Revolution) ODA Overseas Development Association CRDB Cooperative and Rural Develop- PPA Participatory Poverty Assessment ment Bank PRA Participatory Rapid Appraisal GDP gross domestic product SARAR self-esteem, associative strength, GNP gross national product resourcefulness, action GWP Global Water Partnership planning, and responsibility HRD Human Resources Development SCPS Social Capital and Poverty Survey Survey NBC National Bank of Commerce TFA Tanzania Fertilizer Authority xi % CHAPTER 1 Poverty, Social Capital, and Survey Methodology Poverty is a complex and multi-dimen- perienced by the poor. This report describes sional phenomenon. In order to inter- the approach, the substantive results, and the vene successfully in the cycle of poverty implications of the Participatory Poverty As- in which hundreds of millions of people sessment (PPA) conducted in Tanzania in 1995. around the world are trapped, sensible policies must be based on an understanding of its na- Background to the Tanzania Participatory ture and determinants. Poverty Assessment Traditionally, economists have used "pov- erty lines" derived from household consump- The East African nation of Tanzania is one of tion and expenditure surveys to measure the poorest countries. According to the most poverty. While providing a useful yardstick, recent World Development Report (1997), Tanza- poverty lines also involve a number of assump- nia has the third lowest GNP per capita in the tions and simplifications that may obscure the world-$120. This official figure, however, true nature of the problem. Other ways of may significantly understate national income measuring poverty, including measures de- and expenditure. It has been estimated that a vised by the poor themselves, lead to different minimum of 30 percent of total GDP takes concepts and, accordingly, to different policy place in the "second economy." This activity conclusions. includes production and sale of goods and Open-ended inquiry into poverty issues and services from the informal sector and barter the lives of the poor reveals many insights and transactions which are not captured by ac- findings not captured by traditional household counting and taxation systems, rent-seeking surveys. It is now also understood that in ad- activities, and illicit activities such as traffic in dition to human, physical and natural capital, exports and narcotics. Inclusion of this kind of social capital--the quantity and quality of as- income-earning activity would raise official es- sociational life and related norms--is an essen- timates by between 40 and 50 percent (Mal- tial factor in determining the prosperity or yamkono and Bagachwa 1990, World Bank poverty of a community. 1994). Accordingly, the World Bank in recent years About half of the population is estimated to has begun to employ participatory methods of be living at very low levels of welfare. As of research and data collection to elicit new infor- 1991 about half of the rural population was mation for understanding poverty as it is ex- living in poverty, with 42 percent of all rural 1 2 Voices of the Poor: Poverty and Social Capital in Tanzania Tanzanians living in households classified as analysis to explore the relationship between so- hard-core poor (Ferreira 1994). The average per cial capital and poverty. For these reasons, the person consumption expenditure in 1993-94 in study went beyond the lengths needed to sim- rural areas was 50 cents a day (Ferreira and ply inform the poverty debate in Tanzania. It Griffin 1995). raised and answered new questions altogether. Poverty in Tanzania is mainly a rural phe- To communicate these issues and the power of nomenon: rural households account for 92 per- participatory research methods, a film-"The cent of the poor. People whose main source of Poverty Experts"-was made of the PPA proc- income is their farm are five times more likely ess. This film is available from the World to be poor than those who receive a wage from Bank's Social Development Department. the public or private sector. For the past decade The Tanzanian PPA had four key building Tanzania has pursued a program of liberaliza- blocks: tion to reform the regime of state controls The view that the poor are the real poverty which, together with a series of external experts since they experience poverty shocks, had led to economic stagnation and a daily. When given the opportunity, the sharp deterioration of living conditions be- poor can provide both data and subjective tween the 1970s and early 1980s (World Bank insight into the causes, nature, and extent 1996). of poverty, and the range of solutions. In 1995, the Government of Tanzania, the e The use of open-ended participatory meth- University of Dar es Salaam, and the World ods to tap the knowledge and expertise Bank conducted a Participatory Poverty As- of the poor about their lives and percep- sessment (PPA), funded by the British Over- tions. seas Development Agency (ODA) and the e The involvement of a range of stakehold- World Bank. The study involved more than ers, especially policymakers, to build own- 6,000 people in 87 villages across Tanzania. The ership and commitment to using the study utilized three basic types of data collec- study's results. Stakeholders defined is- tion: participatory methods, key informant in- sues, discussed methodology, and were in- terviews, and household survey instruments. volved throughout the study process. The household survey had a dual focus-a So- The desire to better understand the role of cial Capital and Poverty Survey and a house- social capital. hold module devoted primarily to measuring household consumption and expenditures. Identiffing Issues The PPA was carried out as a primary input to the World Bank's Poverty Assessment for Tan- To narrow the focus of the study and to ensure zania (World Bank 1996), along with the Hu- that its focus had policy relevance, senior gov- man Resource Development Survey (Ferreira ement officials, NGOs, World Bank staff and Griffin 1995). working in Tanzania, and selected donors were Because the PPA methodology is relatively interviewed at the outset and asked "What do new and still gaining acceptance, the Tanzania you want to know that you don't already know PPA included a strong methodological element about poverty in Tanzania?" Over a period of designed to explore two issues. First, the study time, the responses to this basic question con- endeavored to determine the extent to which a verged around five broad themes which were PPA complements conventional expenditure subsequently used to define the Terms of Ref- and consumption surveys. In other words, to erence (TOR) for the PPA. These were: what extent does a PPA obtain information and provide insights that go beyond what can be Indicators ofpoverty. What are the indicators of discerned from national household expendi- poverty from the perspective of the poor? What ture surveys? Second, the study used statistical causes poverty? Based on local definitions of Poverty, Social Capital, and Survey Methodology 3 poverty, who are the poor? What do they do for Study Framework their livelihood? How do they survive? Are there any gender or regional differences asso- The PPA explored the role of social capital as ciated with the extent or degree of poverty? a determinant of household welfare alongside What are the poor's perceptions of changes in other determinants, such as human capital, poverty and inequality over the past ten years, physical capital, natural capital, and access to the past four years? What are their hopes for the markets. Social capital can be defined as the future? web of groups, associations, networks, and norms of trust at the community level that Effects of liberalization. What are the effects of form the social underpinnings of poverty and liberalization (especially of trade) on the agri- prosperity (Putnam 1993). cultural sector and on the devaluation of the The social capital question was tested us- shilling in rural areas? Is there price inflation in ing econometric methods which required col- rural areas? How do the poor perceive the lecting some quantifiable data. In-depth data changes resulting from liberalization policies? gathered through open-ended processes pro- How much do they know about the changes in vided the context for interpreting these find- the country? Are they better off than before ings. Chapter 6 discusses social capital in liberalization? Are there gender differences- Tanzania. both in perceptions about and in the impact of liberalization? Study Methodology Credit and saTings. What is the prevalence of Unlike econometric studies, there are at pre- informal and formal savings and credit institu- sent no hard and fast rules on how to conduct tions? How are they used and by whom? What a PPA beyond the maxim of "Do the study in are the preferred attributes of savings and credit a way that maximizes learning and ownership institutions? What is the need for such institu- at all levels, including policymakers." Within tions in the rural areas? Are there any gender this context, the Tanzania study was carried differences? out with two rules in mind: first, from the range of data collection tools available, choose Social capital. What is the role of culture in the most appropriate to answer the specific poverty? What are the groups, associations, questions raised by policymakers; and second, and networks functioning at the village level? choose sampling techniques that have credibil- What is the nature of these groups, what role ity in the eyes of the decisionmakers. do they play in people's lives? What is the Mounting a PPA is by no means simple, but role of these social organizations in coping it can be organized more quickly and produce with economic hardship? What does trust results sooner than a traditional survey based mean at the local level? Have levels of trust in on questionnaires. For this reason, the PPA groups and in formal and informal leadership technique often proves useful for the interim increased or decreased. Why? What are the monitoring of poverty, between major surveys. gender differences? It was reassuring that-at least in the case of Tanzania-the aggregate results of the two ap- Natural environment. What is the role of the proaches were very similar. But at the same natural environment in the coping strategies of time, the PPA generated more subtle and more the poor? How do the poor perceive environ- detailed findings in a number of areas, and its mental degradation, including decline in soil policy conclusions were therefore more dis- fertility and lack of available firewood? What criminating, and in some cases different (for actions do the poor take to protect the environ- example, on questions of gender, seasonality, ment? Are there any gender differences? and access to water). 4 Voices of the Poor: Poverty and Social Capital in Tanzania Village Sampling their two-day stays in each of the 87 villages, certain redundancies were built into the data Many intensive participatory and anthropo- collection methods. Three basic types of data logical studies are dismissed because they are collection were used: participatory tools; key either conducted in just one village or in a informant interviews; and household survey handful of villages. To establish credibility and instruments. Some of these tools were created ensure that the study was broadly repre- specifically for Tanzania to address issues of sentative of rural areas as a whole, we selected concern to policymakers. Different sampling 100 villages spread throughout the country. techniques were employed for the various data These villages were part of the National Master collection methods used. Sample framework established earlier by the government's Planning Commission. All na- Participatory Data Collection Tools tional-level studies conducted in Tanzania are done in these villages to allow researchers to Two participatory methods were used in generalize their sample to the nation's rural the PPA. SARAR (Self-esteem, Associative areas as a whole, and also to ensure that find- Strength, Resourcefulness, Action Planning, ings from one survey can be compared and and Responsibility) primarily uses visuals as contrasted with data from other studies. stimuli to generate discussion and involve- Before the PPA team entered a village, ment. Participatory Rapid Appraisal (PRA) the community was informed in advance. This makes use of a family of methods relying on insured availability of village leaders and open-ended dialogue and community-level also enabled the team to meet with community analysis. The various participatory tools (be- members during its two-day stay. To ensure that low) were administered in discussion groups group discussions were not dominated by the of men and women usually segregated by gen- more powerful community members, the team der to conform to local culture and also to en- supervisor met with local elites while the re- sure that women felt the freedom to speak search team fanned out through the community. freely. More detailed descriptions of most of To maximize use of existing data bases, the these activities can be found in the World 15 households selected in each village for the Bank's Participatory Development Toolkit (1995) household survey were the same as those ran- and in Resource Kit for Participation and Social domly selected for the national agricultural Assessment (1997). survey. The final sample included 87 of the 100 selected villages because some of the villages Mapping. Groups were asked to draw a map were unreachable due to washed out bridges, of their community on the ground, marking the particularly in Mtwara, Lindi, and Morogoro. key features of the village and drawing in In addition, some villages in Kagera had been households which were labeled according to taken over by refugees and were dropped from wealth status later in the process. If it was rain- the study. Due to a communication problem, ing, the activity was conducted indoors using the household expenditure module of the markers on poster paper. household survey was administered only in 53 of the 87 villages. Poverty characteristics and wealth ranking. Once trust was established between the study team Data Collection Tools and the village group, people were asked to identify the characteristics of five different No single data collection tool can obtain the wealth groups, from the very rich to the very range of information needed in a study of this poor. Once consensus was reached on defining kind. To ensure the reliability and consistency characteristics, the groups used colored stickers of information gathered by researchers during classifying households into these categories. Poverty, Social Capital, and Survey Methodology 5 Seasonal analysis. The lives of poor people de- own farms. The discussion then moved on to pendent on agriculture change with the rhythm how to turn a poorly functioning farm into one of the seasons. People were asked to draw a that works well. A set of pictures was drawn to matrix on the ground, with the months on the represent the crops and ecology of different horizontal axis and activities or resources on the agro-ecological zones. vertical axis. The discussion focused on how people coped across years and during the peri- Gender analysis. A local artist drew three large ods of greatest stress in their lives. pictures of a man, a woman, and a couple. These were spread on the ground along a horizontal Trend and price analysis. Groups plotted axis. People were then handed more than 30 changes they had noticed in their lives (in such smaller pictures of different objects including terms as availability of services and food costs) land, a house, household items, babies, and during two time periods: 10 years ago (marking young children. The group was asked to ar- the beginning of the liberalization process) and range these pictures under the man, woman, or in 1991 (marking the liberalization of basic food couple to indicate who owned which posses- commodities). Prices of commodities consid- sions during marriage. Once the discussion was ered important were also collected. complete, people were asked what happened to property distribution during divorce or separa- Venn diagrams. People were asked to work tion, and finally, what happened to the division together in a group to draw a large circle on the of property if the husband died. ground representing the whole community. Circles of various sizes representing village- Key Informant Interviews level groups and institutions were then drawn. Size of the circle indicated importance and over- Much of the information obtained through lapping circles indicated overlapping mem- groups was also asked of key informants. bership. Lines between groups indicated These were generally village officials, includ- interaction between them. Discussion focused ing cell leaders, and others such as school on the membership, structure, functioning, and teachers who were residents in the community. relationships between village groups. In each village discussions were held with two or three key informants. Any information that Problem identification. To avoid introducing did not tally was further cross-checked. This sectoral bias, a set of 22 drawings was devel- provided a valuable means for cross-checking oped by local artists depicting a variety of prob- information and engaging officials while lems and issues. Blank paper was made group discussions were being held by other available so people could draw pictures of members of the study team. problems not depicted. Through a voting exer- cise using stones, people ranked the five most Household Surveys important problems affecting their lives. The issues were then discussed in depth. In some A two-part household questionnaire was de- villages, this activity was taken further. Groups veloped. The first part-the Social Capital and were given Polaroid cameras to produce photo- Poverty Survey-focused on social capital is- graphs depicting the most important problems sues, agriculture, savings and credit, environ- they faced. mental issues, and three measures of poverty (wealth ranking by cell leaders, self-ranking, Story with a gap. This activity consisted of two and an index of assets in the household). The pictures, one showing a poor, unproductive second part consisted of a household expendi- farm and the other a prosperous farm. Using the ture and consumption module of the 1993 Hu- pictures, people discussed the situation of their man Resources Development (HRD) survey 6 Voices of the Poor: Poverty and Social Capital in Tanzania (Ferreira and Griffin 1995). Some questions on whom held bachelor's degrees) and included demographic background and distance-to- both men and women. The work was super- markets were included. vised by a World Bank sociologist. District-level Workshops Data Analysis In Tanzania, district level officials implement All data gathered through participatory meth- many of the programs affecting the lives of the ods were subject to systematic content analysis. poor. Each team held a one-day workshop at Statistics were obtained at the end of content the district level. The workshop was used to analysis, where appropriate, through a step- brief district officials about the nature of the by-step aggregation of village data to reveal study, to explore their attitudes and knowl- the national pattern. This process took three edge about poverty, to learn the reach of their months. programs, and to feedback preliminary find- The quality of household survey data was ings from the field. After introductions were checked by the field teams soon after return made in small groups, officials drew maps of from the villages. The data then were entered the district, discussed poverty characteristics, into a computer by Genesis, a private firm in and then located areas of poverty on the map, Nairobi. Data were cleaned and checked for as well as where government services were accuracy. Statistical analysis also consisted of provided. The local agricultural picture was a step-by-step process, starting with frequen- also probed in depth using the Story with a cies, cross tabs, index construction, and, finally, Gap technique. model testing. PPA Process Report Writing The PPA field work was conducted by six teams Both data sets were combined and reports writ- whose members received intensive training for ten from field notes in Washington, D.C. As three weeks in a range of data collection tools. noted earlier, the PPA was initially designed The teams were drawn from the University of simply as a contribution to the Tanzania Pov- Dar es Salaam, other research institutes, and erty Assessment report produced in 1996. government ministries, as well as from respon- Hence, data were analyzed and information in- dents to a newspaper advertisement. Training tegrated into the main report with no plans for stressed behavior and the attitudes underlying a stand-alone PPA report. However, because successful use of participatory tools. Each team of the richness of the data collected, the present was headed by a senior social scientist, half of document was developed to inform the ongo- whom were men and half women. Teams con- ing conceptual and methodological debate sisted of five junior field workers (about half of about poverty and its determinants. CHAPTER 2 What Is Poverty? The one with nothing has everything taken from him, while the one with something has more given to him. - Oldadai, Arusha Poverty can be conceptualized and measured in different ways. The conventional economic ap- proach focuses on the quantifiable poverty line based solely on consumption and expenditure patterns (World Bank 1991a, 1992, 1993b). While the poverty line is an important measure of poverty in a country over time, poverty goes beyond income levels. It includes access to health care and education, respect, status, isolation within a community, and feelings of powerlessness and hopelessness. Poverty is multidimensional, and many of its dimensions are often hidden (Ravallion 1992, 1996, Moser 1996, Sen 1990). This chapter reports PPA findings showing how village people in Tanzania view poverty, what characteristics they use to categorize the poor, what major challenges the poor say they experience, and what they envision for their future. The 1993 Human Resources Development sur- their first birthday. While rural literacy has in- vey found that income distribution was quite creased (from 59 percent in 1983 to 73 percent uneven in Tanzania: the average adult equiva- in 1993), declining school enrollment rates im- lent expenditure of the richest quintile was ply that literacy gains may slacken or be lost more than six times greater than that of the (World Bank 1996, Ferreira 1994). poorest quintile. The inequality between rural To complement the poverty line data collected and urban areas was also considerable: while in the HRD survey, the PPA focused on gather- on average rural Tanzanians spent the equiva- ing information on poverty as defined by local lent of $193, their counterparts living in Dar es people. In a country where land is plentiful, yet Salaam spent on average $587. poverty is primarily rural, the other factors con- Access to safe drinking water has been on tributing to poverty must be understood. the decline since 1976. Rural households spent an average of 3.1 hours a day collecting water, Methodology with the brunt of the work borne by women. Infant mortality is high: for every thousand Three participatory data collection methods children born in Tanzania, 90 will die before were combined to assess the extent and nature 7 8 Voices of the Poor: Poverty and Social Capital in Tanzania of poverty in Tanzania's villages. In each vil- Average lage mixed groups of men and women were first asked to participate in a community map- The terms used to describe a person of average ping exercise. This activity initiated interaction wealth translated into someone who is ordi- and began to establish trust. As noted in the nary: ordinary folks, ordinary life, and ordi- previous chapter, community mapping in- nary farmers. volves people in drawing maps of the basic infrastructure, landmarks, and households in Poor the community. Once the mapping process was complete, people then engaged in a dis- Local terms generally revealed great sympathy cussion of poverty and its characteristics, clari- toward the poor: hardworking, toiling, those who fying for the PPA team local definitions for the sleep tired, one who has to sell his labor, and a five wealth groups: very rich, rich, average, cursed person-whatever he does, he loses. poor, and very poor. After agreeing on the key identifying features of each category, partici- Very Poor pants were asked to rank themselves and the 15 households that composed the household The most negative and ambivalent local terms interview sample. Although the initial ten- were used to characterize the very poor: those dency was to rank everyone as poor, people who are poor by choice, someone who cannot quickly became more discriminating in their even depend on himself, a really low person, categorization. Consensus was usually reached a stupid fool, people who rely on others, some- in the final ranking of households. Having one who is poor but invited it upon himself, completed the ranking exercise, discussion people who rely on luck and others, someone then turned to a discussion of the causes of who is completely disabled or physically poverty. When time permitted, two or three handicapped, and someone whom to employ villagers identified in the mapping process as is a threat. being poor were interviewed in-depth. Poverty-related proverbs generally revealed a strong work ethic, coupled with an attitude Local Terminology and Proverbs of shame and discouragement: Someone who discourages work, encourages Local terms used for different categories of poverty. (Bunazi, Kagera) wealth and well-known proverbs about poverty You have to work hard so in the end you will were revealing of popular beliefs and attitudes. become rich. (Bunazi, Kagera) If you want very good things, try to work very Very Rich hard. (Mititi, Rukwa) Poverty provides a very bad name to the family. The terms used to describe the very rich were (Kazflamihunda, Kigoma) surprisingly positive, with a few connoting a de- A poor person cannot take anything which is gree of envy. Literal translations of these terms attractive. If he takes it, he will be called a thief include: those who are very able, those who can Because he has nothing, nobody expects him to breathe freely, those who are lions, those who have good things. (Mititi, Rukwa) are stuffed, and those who set the prices. The one with nothing has everything takenfrom him, while the one with something has more Rich given to him. (01dadai, Arusha) The person who is poor when she wakes up in The terms for a "rich man" were less colorful: the morning has nothing to do, unless she goes a little bit rich, a rich person, a person who has to a rich person who can make her day start risen a bit. moving. (Bunazi, Kagera) What Is Poverty? 9 Poverty Characteristics ple's subjective psychological experiences (tables 2.1 and 2.2). The distinguishing characteristics of poverty were location-specific with clear regional Land as a Measure of Prosperity (agro-ecological) differences. Consistent with the economic definition of the poverty line, the In every region, land ownership was nearly characteristics reflected many consumption always the first characteristic mentioned in de- categories and people's capacity to spend. In fining categories of wealth. This despite the addition, they reflected social reality and peo- fact that land is still plentiful in Tanzania. The Figure 2.1 Community map of Luana Village, Iringa Se M o i aviCharles. 13N ZuA~~CiO Soure: appig poces failiatedby avidChales 10 Voices of the Poor: Poverty and Social Capital in Tanzania Table 2.1 Characteristics of poverty and wealth, Table 22 Characteristics of poverty and wealth, Dodoma region ilimanjaro region Wealth category Description Wealth category Description Very rich More than 40 cows Very rich Owns more than 10 acres More than 80 bags of food (sorghum) Uses fertilizer and improved maize seeds Has 3-4 wives Owns tractor House made of brick and galvanized iron Owns milling machine, large business Owns a donkey, bicycle cart Children go to best secondary schools Can own a shop Rich and responsible, children well behaved Rich Owns 5-10 cows Enough food for a whole year Rich Owns 4-8 acres, 2 dairy cattle, 10 goats, Can afford school fee in private school small shops Fairly good house Children attend private school Family is not happy, irresponsible Average No livestock Enough food, but not so good Average Owns 2-3 acres Cannot afford private hospital services Problem buying inputs Can cultivate land or has salaried Must sell labor employment Occasionally owns oxen plough House of mud walls, iron roof Children attend school up to standard VII but secondary school is unaffordable Poor Not enough food, skips meals Depends on piece work Poor Owns less than 1 acre of land Many children Uses hand hoe, first works for others, Uses second-hand clothing or waits until others have finished Brokn-dwn huseploughing so is late in ploughing and Broken-down houseplanting Very poor Mostly old, disabled Cannot afford inputs, poor harvest Lives by begging (3 tins from 3 acres) Women who have no children No cattle Mentally unfit, no home Children cannot complete primary school Eats once or twice a day if lucky PPA study found that landlessness among Very poor No land or very small plot of land the poor and very poor was increasing in Depends areas with high agricultural potential, particu- "The house has 7 doors" (broken walls larly where villages are close to urban centers, so they can exit from anywhere) have good road access to markets or where there are large farming estates (such as sisal in Tanga). nothing. I can never buy my land back because In Arusha and Kilimanjaro regions the poor the prices go up every year." said: "We have land but nothing to work it In villages of pastoralists or settled pastoral- with. All I have is a small hoe. You have to ists (Dodoma, Sindiga, Tabora, Shinyanga, bend your back to work with a small hoe and Kagera, and in areas of Arusha), cattle owner- after two hours you are tired." "The land is ship was deemed the most important sign exhausted from overcultivation. Ten years ago, of wealth. Yet even in these regions where I got eight bags per acre; today I get two bags." herd size was considered important, people "I cannot afford fertilizer or hybrid seeds, and compared the size of the herd with the size I do not qualify for credit. Without fertilizer, of the farms and success in farming. In the land does not produce even enough to feed Mwanza, access to motor boats and fishing nets the family." "I sold my land and now I have was a significant sign of prosperity. Growing What Is Poverty? 11 of cash crops was also used as a defining char- times per year because they can't afford them. acteristic of the very rich and rich in many They must sell whenever they can." The im- areas. portance of cash crops in Kagera and Rukwa regions is demonstrated by table 2.3. Agricultural Inputs Quantity and Quality of Food Consumed The 1991 liberalization of basic food commodi- ties eliminated many subsidized inputs sup- Throughout Tanzania, wealth was associated plied by cooperatives. Farmers now must pay with the ability to eat and drink as much as one full prices for inputs on open markets. This has desired. Poverty was associated with skipping had a strong impact on farming practices. The meals, reducing meals to one or two a day, in- ability to purchase needed inputs-such as hy- voluntarily changing diet, sending children to brid seeds, fertilizers, and insecticides-for neighbors' homes to eat, and poor performance food and cash crops is so important to pros- in school (hunger making poor children unable perity in rural Tanzania that it has become cen- to pay attention). Sugar, kerosene, and cooking tral to definitions of poverty and wealth across oil were considered luxury goods as "rare as the entire country. gold" for the poor. The very poor were de- While both the HRD and PPA studies revealed scribed as lucky to eat at all, totally dependent the importance of spending on food, hous- on the goodwill of others for food. ing, health, education, and clothing, the HRD survey-which focused only on household con- Access to Health and Education Services sumption-did not note spending on farming inputs as a distinguishing characteristic among Changes i government policy effected how the wealth categories. By contrast, the PPA found people associated access to health and educa- that lack of access to agricultural inputs was re- tion with levels of wealth (table 2.4). For the ported as an essential contributing factor to pov- poor, hospital care and a complete school edu- erty. The poor did not, by and large, recommend cation were perceived to be out of reach. a return to subsidized inputs as the solution to this problem. Rather, they focused on overcom- Power and Dependence ing constraints that put them at a disadvantage to the rich. Wealth was associated with respect, power, In areas with high agricultural potential like and leadership in all regions. In Arusha, for Kilimanjaro, the PPA found that people used example, people reported that "the very rich land, farming implements, and the ability to are very respected. Even their speeches are purchase fertilizer as key to defining wealth given the highest priority in the village. The categories (table 2.2). In the coffee-growing same with the rich-they can win any election area of Arusha, an average person was de- in the community." The rich were seen as pow- scribed "as someone who does not have the erful because they had the power to set the sprayer for coffee agrochemicals nor coffee prices for goods they purchased. This evoked drying wires, and who sells coffee in small both awe and resentment: "The rich buy goods amounts six times a year because he needs the from the poor at low prices after the harvest money immediately to solve other problems. and later sell them when prices are high." "The The very rich let the coffee accumulate and sell rich become rich by buying cattle during times once a year at good prices. The rich also get of hunger." The rich were also seen as power- better crops than we do, but they sell every six ful because they could hire labor to work for months. The poor on the other hand put agro- them and always had surpluses which they chemicals on coffee twice instead of seven could invest to make more money. 12 Voices of the Poor: Poverty and Social Capital in Tanzania Table 2.3 Importance of cash crops and crop yields by wealth category, Kagera and Rukwa regions Wealth category Kagera Rukwa Very rich Has shop, business More than 100 cows Farms 7-9 acres Has investments Cash crops, coffee, and plantains Permanent house, car Very healthy Rich Farms 4-7 acres 50 cows Cash crop, coffee 50 bags of maize per harvest Permanent house with metal roof, bicycle Good health Average Farms 2 acres 5 cows Mostly food crops 10-20 bags per harvest Thatched house Good clothing Has bicycle Poor Farms 0.5 acre 1 bag per harvest No hoe, doesn't farm Works as casual laborer Eats once a day Sells firewood Sells his labor to rich for food Has to buy food Often sick May eat up seeds Very poor Powerless Cannot cultivate Depends on others to survive Disabled Does not have hoe Begs, depends on others No land Steals, has nothing The poor were seen as dependent on others described someone with 80 cows as very rich, which brought shame to the family because they while men used the figure of 3,000 cows to could not take care of their needs. One poor describe the threshold of being very rich. person told the team: "I am not poor because I can still look after myself and my family." Female-headed Households Physical Disability Findings regarding female-headed households differed significantly between the PPA and the Physical disability was commonly associated earlier HRD survey. The PPA found that fe- with the very poor. The physically disabled male-headed households in rural areas were were considered a burden. Their survival de- substantially worse off at every level of wealth pended on the gifts and kindness of commu- compared with male-headed households (table nity members, both kin and non-kin. 2.5). The HRD survey found no differences be- tween them (table 2.6). In the HRD survey, Gender Dtfferences when the urban population was added, the findings actually reversed themselves, and fe- There was only one consistent gender differ- male-headed households emerged as better off ence in the various descriptions of poverty: than male-headed households at every level. women seemed to set their sights lower than The difference in findings was sharpest for men in defining categories of wealth. In one the very poor category. The PPA found that rather startling example, women in a village the number of female-headed households in What Is Poverty? 13 Table 2.4 Importance of health and education services Table 2.6 Gender differences In findings between PPA by wealth category, Tanga region and HRD studies Wealth category Description (percentages) PPA HRD Very rich Big person with position in community Male Female Male Female Hires laborers and supervisors Children attend private secondary schools Non-poor 52 29 49 45 Has better toilets Poor 32 40 33 37 Can afford private hospitals Very poor 14 29 18 18 Rich Healthy children, eat, sleep, and dress well Can afford school fees and uniform study and the PPA study seems to indicate that Children attend secondary school Chilrenattnd ecodar scoolfemale heads of households tend to engage in Average Cannot afford secondary school, sometimes petty hawking, trading, and "piece work" can afford technical schools No good food Cannot eat consistently ing in assets, these additional income-generat- May die of diarrhea, dysentery, and vomiting ing activities may account for their equality in Poor Food and clothing are a problem spending in the HRD study. Children do not complete primary school As already noted, the PPA study went be- Cannot afford health care fees yond consumption and expenditure to capture Verysocial concepts in defining poverty. During the Eats cassava boiled in water without salt poverty discussions, both men and women Etesckaabie nwtrwtotsl said that female-headed households tended to Often sick _____________________________ have less land and livestock and so depended greatly on their piece work and hawking ac- the very poorest category was more than dou- tivities. This lack of capital assets and lack of ble that of male-headed households. Why the children (who were left with the husband) pro- difference? Is one study more accurate than the duced social isolation which made them more other? vulnerable and powerless to change their situ- Explaining the discrepancy highlights the ation. For a fuller discussion of gender-related importance of interweaving findings from findings from the PPA see chapter 4. these two different survey approaches. The HRD survey revealed that rural female-headed Distribution of Poverty households had fewer permanent assets: male- headed households owned 6.08 acres of land Despite the local and regional differences in and had 4.7 years of education; female-headed poverty descriptions, the overall estimate for households possessed 3.74 acres and 2.45 years rural poverty in Tanzania was strikingly simi- of education. Evidence from both the HRD lar between the PPA data (50.3 percent) and a poverty line (based on the head count index Table 2.5 PPA national wealth ranking with the poverty line set at 114,187 shillings) (percentages) derived from the HRD survey (49.7 percent). Wealth Male Female Total Very ich 0 1Gini Coefficient and Perceptions of Poverty Very rich 1 0 1 Rich 10 4 8 ARage 43 278 A gini coefficient is a useful statistic (ranging Average 3 between 0 and 1, with higher values indicating Poor 32 40 34 Very poor 14 29 17 more inequality) that captures the distribution Tof a good or asset across a population. Accord- ing to the HRD survey, the gini coefficient for 14 Voices of the Poor: Poverty and Social Capital in Tanzania Table 2.7 Well-being ratings by wealth compared over time (percentages) Compared with 10 years ago Compared with 4 years ago Very poor Very ich Total Very poor Very rich Total Better off 50 37 47 35 63 35 Same 10 0 14 12 0 16 Worse off 40 63 39 53 37 49 Total 100 100 100 100 100 100 land for rural households was 0.46 and had when further liberalization of marketing of remained quite stable for the past decade. The commodities (wheat, rice, and maize) oc- gini coefficient for livestock was higher (0.85) curred. Since 1991, annual inflation had never a figure remaining fairly stable over the past been below 21 percent, with a high of 36 per- decade-indicating greater inequality in distri- cent in 1994. Important differences in percep- bution of livestock among households. The gini ton appeared by wealth level. For example, coefficient for income was 0.35, a decline from more than half of the very poor felt worse off 0.60 in 1991. However, the study concluded and more than half of the very rich felt better that "it is not clear whether this decline corre- off (table 2.7). sponds to the beginning of a new path, to a short-run change, or to a statistical and meas- Village-level Prosperity and Inequality urement error.... [It] requires more points of observation and secondary data" ( p.10). In the PPA people were asked to rate overall While the gini coefficient is useful, it also has levels of prosperity and inequality in their vil- its limitations. Livestock and land ownership, for lages (table 2.8). Overall 37 percent felt there example, are relatively easy to track and quan- was greater prosperity in the villages and 48 tify. Income levels in rural areas are more diffi- percent felt that this was accompanied by cult to measure. As a result, people's perceptions, greater inequality. These perceptions did not as revealed in the PPA, provide important in- differ significantly by wealth category. sight into the dynamics of local poverty. They may also provide a useful tool for monitoring Causes of Poverty poverty between large household surveys. In every village small groups of people dis- How People Perceive Prosperity cussed the causes of poverty (box 2.1). These As part of the PPA, a household survey waso conducted to collect data on people's subjective seine' 1 nn ofation v le prentyelod in equalit,y ihahgho 6pr judgments on poverty trends. People were (percenpes) asked to rate whether they were better or Prospet worse off than they were ten years ago and Less 45 four years ago. With the exception of the very Same 16 rich, the majority felt they were better off than More 37 10 years ago, "when the shops were empty" ato the beginning of the liberalization of the econ- s i nequality omy (table 2.8). O e 39 The percentages reversed, however, when More 48 people compared themselves to four years ago What Is Poverty? 15 Box 2.1 Sliding into poverty: the story of Ndesheyo Esau, Maroroni Village, Arumeru District "I came to this village in 1958 with my late husband an acre. Now I am caught in a circle. By 1990, the soil and two children. After coming here I had three more became tired, we had no money to buy fertilizer, only children. We came here to farm and raise cattle and a small hoe, no cattle, so to get food, the family life was good. In the early 1970s I could harvest up to started piece work. For example when my son works 20 bags of maize from one acre. The soil was very the whole day weeding someone's farm he gets one fertile and we did not have to apply any fertilizer. We tin of maize. We discovered that most of our time is had 40 cows and 400 goats. The government provided spent on other people's farms just for exchange of a dip free and so the animals were free of disease. But food for two to three days. We try to expand our own in the late 1970s the government had problems be- farm, but because we need food immediately we are cause of the war with Uganda, the free dips were pushed again to work for others. closed, the animals got diseased and died. None of us The only rich people are those who take their were educated about what to do in case of disease. By riches and invest them in a business. The traders are 1983, I was left with 5 cattle and 12 goats. doing well. They are not at the mercy of the rain. One In 1984 there was a drought which affected the man started by buying beans in bulk and selling it in cattle and there was hunger. To get food and because Dar es Salaam and bringing back other goods to the the pasture was dry, we sold the rest of the cattle and village. He has large farms now and hires labor to goats, and at that time I got eight bags of maize from farm for him." discussions underwent content analysis to de- we start with? We have to travel far to look for termine categories of response. Those catego- food, come back to farm for a few days, and ries of group responses were then collapsed then go again to look for food." From Kiliman- and frequencies were obtained (table 2.9). The jaro in the north to Mtwara and Lindi in the results share some similarities with the HRD south, the small handled hoe emerged as the study which singles out family size, education, symbol of being caught in the poverty trap: land availability, market integration through "All we can afford is the small handled hoe growing cash crops, and proximity to all- made in China. It costs 2,000 shillings. It breaks weather roads as key determinants. Results our back and it breaks within one season, both differ as well. the wood and the metal part." Farm Productivity Health A remarkable 47 percent of all responses about Poor health was seen as a quick way to descend the causes of poverty were related to being able into poverty. People said medicine was gener- to farm productively. Distant or poorly func- ally not available in the local clinics, and seek- tioning markets, price fluctuations, "exhausted ing health care in private clinics and hospitals land," and drought or floods all received men- was viewed as expensive. "If someone is sick, tion. Only 8 percent reported landlessness, and taking him to the hospital 25 km away is very among those responses four issues stood out: expensive. You can spend 10,000 to 20,000 shil- women not inheriting land (chapter 4), men in- lings. You have a choice either to take the per- heriting small plots of land, recent migrants not son to the hospital, spend all of your money having land, and the poor selling their land. and let your family members starve, or let the Reflecting the community-defined poverty sick person stay at home and just buy medi- characteristics, most frequently cited were the cines which you know about and hope will high cost or unavailability of fertillzer, insecti- work." As the elderly lose their physical cides, seeds, farming implements, and moder strength, they cannot farm. Thus, even when technology (box 2.2). "With no agrocheicals, healthy, the elderly often slide into poverty. we hardly get any yields. Next year what do The issue of malnourished and sick orphans 16 Voices of the Poor: Poverty and Social Capital in Tanzania Table 2.9 Causes of poverty as perceived by village groups their harvest was so poor that the food pro- Percentage Frequency duced was not sufficient to meet the family's immediate needs. Farming Lack of farm inputs 9 49Education Lack of farm implements 8 45 Lack of credit 2 13 Markets not working, distant 7 37 r D survey e st te econi r Droughtlflood 8 47 Infertile land 5 26 of schooling. But despite the fact that the poor Lack of land 8 43recognize the value of education and make Sub-total 47 260 much effort to send their children to school, Health the PPA found that lack of education was not Poor health services 3 15 seen by them as an important cause of poverty. Disease/disability 9 53 Elderly, malnourished orphans 3 17 Sub-total 15 85 Other Problems Social problems Laziness, selfishness 6 34 Among the other problems mentioned less fre- Alcohol abuse 6 31 quently were witchcraft, AIDS, poor extension Witchcraft 1 6 support, God's will, war, prostitution, poly- Sub-total 13 71 gamy, deforestation bride price, poor houses, Lack of education/education services 6 32 Large families 4 24 thef Bad government policies-cost sharing 4 21 Devaluation of shilling 2 13 Born into a poor family 3 17 Vision for the Future Livestock care problems 2 13 Water 1 7 When people give up hope and are trapped in Other 3 21 despair, social and economic development be- Total (groups) 100 564 come even more difficult. In the household in- terviews dealing with village-level institutions, people were asked about their vision for the came out particularly in Kagera which has future. Once again the difference between the been severely hit by AIDS and refugees. rich and poor was striking. The majority of the very rich and rich expected to do even better Social Problems in the future, the majority of the poor and very poor expected to be worse off (table 2.10). Two intertwined social problems most fre- Overall, 51 percent expected to be worse off in quently mentioned were drinking and laziness. the future. Given that inflation in prices of ba- Drunkenness was seen as causing poverty. sic commodities and farming inputs seems to During the height of the farming season, when have hurt the poor the most, it is perhaps re- fields have to be prepared, men were said to markable that approximately one-third faced spend time drinking, leaving women to do the the future with optimism. work. Because of the reduced manpower, few acres could be planted and harvested. The poor Policy Implications were said to drink to forget their problems. The ready availability of alcohol exacerbates the The poor do not advocate a return to gov- problem: poor women brew local beer at the ement subsidies. However, they cannot peak of the harvest to make money and to buy afford to buy farming inputs at current food, because either they had no harvest or prices without access to credit and other What Is Poverty? 17 Box 2.2 Distant markets, poor roads, and poverty Kazilamihunda, Kigoma: "Poverty is caused by lack nent market and the influx of food for refugees has of cash crops which fetch a high price in the market. lowered the price of maize because the Rwandans do We started to cultivate cotton in 1979, but now some- not eat maize. So we are forced to sell to private times there are not even private traders to buy cotton. traders at 3,000 shillings for 100 kg of maize rather We have cotton ready by August and sometimes we than 10,000 before the food assistance surplus." cannot sell it until December. Sometimes it just sits Tsamasi, Arusha: "The road which leads out of and rots." the village is very bad and there is no transport to Marangu, Kilimanjaro: "During the rainy season Babati (district headquarters, 23 km away). So it is the roads are impassable. We can't get fertilizer, we very frustrating to try and get our produce to the have to travel far. There are some private traders who market. The rich provide tractor transportation, but come to sell fertilizer at the nearest town, but they they charge 500 shillings for a 100 kg bag of maize. I don't bring enough for the whole area. The rich farm- am forced to sell at a low price when I know if I could ers manage to buy it all and the poor remain helpless. just get it to town I would get a fair price. During the Improved seeds are also not available. We used to 1993-94 harvest one bag of maize sold at 3,600 shil- produce six bags of maize per hectare; now we get lings in the village and 6,000 in town. A tin of beans only one bag." went for 1,000 shillings in the village and 2,500 in Ahakishaka, Kagera: "The village has experienced town. We get many private traders here for the low a large influx of refugees from Rwanda and with prices and the poor have to sell to them. The worst them has come food assistance. There is no perma- is that sometimes these traders buy on credit." non-farm income-earning opportunities. of rural poverty by rural people them- The markets are working in rural areas, but selves, a social crisis is brewing. Once the imperfectly, and improvements are social fabric of a society is ruptured, eco- needed-particularly in roads and low- nomic development cannot proceed. cost transport-to help the poor take ad- 1 Female-headed households in rural areas vantage of them. are poorer than male-headed households v Improvements in and dissemination of ba- and yet may be more productive in all their sic farming technologies will benefit the income-generating endeavors. poor disproportionately. " Poor health and lack of access to effective Methodological Implications health care facilities exact a high cost on the poor and have a direct affect on the well- 1 The hRD and PPA studies arrive at similar being of individuals and families. statistics about overall levels of poverty. When social problems appear nationally as But the PPA captures some important gen- the third most frequently mentioned cause der differences and identifiers that are lo- cally specific and relevant. The PPA Table 2.10 Vision for the future approach, which yields data relatively (perwetages) quickly, has great potential as a method of ormonitoring poverty in conjunction with er et re me periodic expenditure and consumption Very rich 57 43 - surveys. Rich 56 33 While the income gini coefficient is Average 42 49 9 an important measure of poverty, data Poor 34 54 12 problems over time make some of its Very poor 31 implications questionable. More sub- Tot 39 51 10 jective measures of well-being create another level of data which can help to 18 Voices of the Poor: Poverty and Social Capital in Tanzania better predict the "social temperature" of a great implications for rural livelihoods. A nation. sequencing of methodologies would be * Household expenditure data captured in beneficial, with less extensive PPAs pre- the HRD do not include expenditures on ceding HRD-type surveys, to ensure all farming inputs. Yet these seem tobe clearly appropriate categories are incorporated in different among wealth categories with the survey instruments. CHAPTER 3 Inflation, Price Trends, and Coping Strategies We are not living. We are just surviving. - Women's group, Utende, Mtwara Life is getting tougher and tougher. Prices are going up every new moon. - A woman, Ulaga Kibaoni, Moragano Inflation is often assumed to affect rural people less than urban dwellers. This chapter examines the impact of price increases on the lives of rural communities. It also explores coping strategies of the poor in dealing with food and water shortages, and how such strategies difer for women and men. Tanzania's national consumer price index does across rows (figure 3.1). The discussion fo- not take into account prices of goods in rural cused on the ebb and flow of seasonal activities areas. To better understand the issue of rural and the stress periods in people's lives. In ad- prices and trends, three tools adapted from the dition, a household survey was conducted ask- PRA family of methods were used. First, ing a series of open-ended questions to groups of people were asked to recount the determine how much people knew about prices of goods important to them at three changes in government policies and about their points in time (1985, 1991, and 1995). Second, sources of information. they were asked about situational trends- ranging from levels of rainfall and firewood Price Analysis availability to the status of women and em- ployment opportunities-during the same Based on responses gathered from the three time periods. Third, in a PRA activity called exercises, tables plotting prices for basic com- Seasonal Analysis, a calendar was drawn on modities and services from 1985-95 were the ground on a horizontal axis, with a list of developed for each region (tables 3.1-3.4). The categories of activities or events on the vertical tables reveal high inflation rates, regional prices axis. The variation in months was plotted differences, and some gender and regional 19 Figure 3.1 Seasonal analysis chart Jan. Feb. Mar. Apr May June July Aug. Sep. Oct Nov. Dec. Weather Hot * ** a" Cold ** 4 0S Maize Prepare field see Ploughing *" *0 * 1 0 Weeding 0m* * 1 Harvest 0 Beans and groundnuts Prepare field Ploughing em e0 0e00*0 Planting 0 0 Harvest * 000 s 41 Kahana Insecticides 0 0 0 0 000 Livestock diseases 0011 0500 0e* Labor intensity em0 000 0106 0" 00 00 so 00 00 0m0 cm. 000 Transport 0e0 000 o00 0ee 0e0 0"0 000 0c. 000e 0"0 0"0 e000 Chicken pox e0 em Dysentery e0 e" e* em Market 00 00 00 em e0 em cm 00 00 09 0m 000 Water available 00 so em so e Diseases e41 em 0m0 00 00 Drunkenness em 0m 0m 0m ec 00 00 00 S0 00 Hunger 000 4066 000 0"0 00 00 0000 Extension agent c* em e0 em 00 se Goo 00 cm 0m 000 04 Inflation, Price Trends, and Coping Strategies 21 Table 3.1 Price trends in Tabora, 198"5 Table 3.3 Price trends In -9limanjaro, 1985-95 (in shillings) (in Shillings) Commodity 1985 1991 1995 1985 1991 1995 Paddy (20 kg) 400 800 1,200 Commodity Exercise book .50 10 50-60 Sugar (kg) 50 200 500 Cooking oil (20 kg) 1,000-2,000 5,000 20,000 Meat (kg) 300 400 800 Soap (one piece) 30 50 100-150 Kerosene 90 150 170 Sweet potatoes (20 kg) 150 300 400 Beans(kg) 60 100 350 Sugar (kg) 50-100 250 500-1,000 Maize flour 20 40 120 Kerosene (1tr) 18 100 250 Cooking oil(kg) 100 200 800 Bicycle 15,000 28,000 60,000 Farming items Secondary boarding school (per student) 1,500 8,000 15,000 Maize seeds (10 kg) 2,000 3,000 6,500 ____________________________________ Fertilizer-urea (50 kg) 2,000 3,000 6,200 Farming land (one acre) 4,000 5,000 7,000 Table 3.2 Price trends in Shinyanga, 1985-95 Building materals (in shillings) -Iron sheet 800 1,200 3,500 Commodity 1985 1991 1995 Cement 700 1,800 6,000 Phosphate (50 kg bag) 250 1,500 not seen Services 200 kg cow 15,000 50,000 100,000 Secondary school free 700 4,400 Water (20 Itr bucket) not sold 20 50 (per student) w/uniorm wfuniform Land rental (per acre) free 1,500-2,000 3,000 Malaria treatment - 2,000 Dagaa fish (in bowl) 50 100 150 Laundry soap (one box) 125 250 500 Piece of soap 50 70 100 Tabl ic d g 1 Soda 70 100 150 Sugar (kg) 300 450 800 Commodity 1985 1991 1995 Maize flour 300 1,000 1,000 Cooking oil (1tr) 150 200 800 Sugar (kg) 70 150 500 Hospital services free 900-1,500 2,800 Kerosene (1tr) 15 60 420 __________________________________ Maize (tin) 150 500 1,500 Cooking Oil (ftr) 150 240 900 differences in the commodities central in peo- Beans (tin) 250 500 2,500 ple's lives. For example, the escalation in the Ce () 50 10 30 price of fertilizer was mentioned most fre- Sorghum 5 30 250 quently in men's groups and in areas of high Cassava dry (bowi) 1 2 10 agricultural potential. In Mara the price of Bananas (for beer making) 51 50 1,500 chemicals for animal dips was mentioned most Meat (kg) 50 200 500 often. Khanga (Tanzanian) 250 1,200 2,800 When asked why prices had gone up, people Seand-han sr0 300 1,00 most frequently cited government policies: Second-hand srs 10 300 1,00 "Government economic policy changes are the Piece of cloth 1.25m 150 1,500 2,000 reason for the increase in prices. Now there is no bureau responsible for controlling prices, so the consumer is no longer protected like ten there are shortages at high prices, while we years ago." The transport situation-poor remain trapped in our villages." roads and expensive transportation-was also In Shinyanga a villager reported, "Prices often cited as causing inflation: "The people have increased because of devaluation of the who benefit are the rich traders who come here shilling, high prices at the source, and increases and buy our goods, and sell them later when in the price of consumer goods, while prices of 22 Voices of the Poor: Poverty and Social Capital in Tanzania Figure 3.2 National upward trends reported in villages National average trend National average trend for shelter for food 1 1 0.8 0.8 - 0.58 0.6 0.6 0.4 0.26 0.4 m 0.2 - 0.2 -0 0-0. -04-0.24 C C S-0.2 -0. -0.2 -4 -0.2 -0.4 -0.6 -0.6 -0.8 -0.8 -0.68 -1 I -1III 1985 1991 1995 1985 1991 1995 National average trend National average trend for education for clothing 1 1 0.8 0.8 0.6 0.6 0.4 0.35 0.4 0.2 0.2 0.11 0.1 0- 0- S-0.2 c -0.2 --0.4 --0.4 - -0.16 -0.0.5 -0.6 -0.8 -0.59 -0.8 -1 i I -1 I I I 1985 1991 1995 1985 1991 1995 cash crops have remained stagnant." Rises in While people mentioned the removal of prices were also linked to higher cattle prices. subsidies and dismantling of the coopera- In the same region, where ethnic groups are tives as central issues, their problem-solving known to place a high value on cattle, a com- strategies focused on improving rural infra- munity member said, "In 1985 a 200 kg cow structure to increase competition rather than cost 15,000 shillings, and today it costs 100,000, a return to direct subsidies of inputs. so naturally everything else has to cost more." People conceded that inflation also meant Welfare Trend Analysis higher sales prices for goods that they pro- duced, but as mentioned in chapter 2 the poor The trends people reported in their welfare felt that they could not take advantage of mar- were plotted on a scale ranging from minus kets to the extent the rich could due to poor 1 to plus 1 (figures 3.2 and 3.3). These trends transport and distance to markets. are based on a national sample covering Inflation, Price Trends, and Coping Strategies 23 Figure 3.3 National downward trends reported In villages National average trend National average trend for farming inputs transport fares 1* 1* 0.8 - 0.8 0.6 - 0.6 0.49 0.4 - 0.26 0.4 O 0.2 - c 0.2 c c :2 0 2 0 c -0.2 - -0.2 -0.4 -0.4 -0.6 -0.57 -0.6 -0.48 -0.8 -0.73 -0.8 -0.93 -1 1I I -1 1 1 1985 1991 1995 1985 1991 1995 National average trend National average trend for energy for health (kerosene, charcoal, firewood) 11 0.8 0.8 0.72 0.6 0.6 0.4 0.37 0.4 m 0.2 - 0.2 S 0 .~ 0 -0.2 i -0.2 -0.4 -0.38 -0.4 -0.6 --0.6 -0.8 -0.76 -0.8 -0.7 -0.84 -1 -1 I 1985 1991 1995 1985 1991 1995 16 regions. In general people had no prob- Take education. In several regions-Singida, lem recalling prices and speaking at great Dodoma, Mtwara, Lindi, and Tabora-people length about changes which affected their reported an increase in the number of pri- lives. Although the issues people focused mary schools and the number of children at- on varied, some were common to most re- tending them. In Mbonde, Mtwara, where a gions: new primary school had been completed re- cently, a person said, "We now have a new Positive Trends school with teachers who have courage and are qualified and highly motivated. They The trends for basic household needs, food, have many incentives, like donations from shelter, clothing, education, and roads gener- students' parents. So the students work hard. ally appeared to be improving or at least had Every year, one child is selected to join sec- begun an upward movement. ondary school." 24 Voices of the Poor: Poverty and Social Capital in Tanzania Boxmedicine when it was available; and wide- Bnom3. Tasortarblm spread disease (particularly diarrhea and ma- in Choma, Tabora laria) resulting from unclean and standing "In the old days, many roads surrounding our vil- water (box 3.2). lage were in good condition. Central and local gov- People's explanations for negative trends in ernment were very much paying attention to them. farming inputs and implements were similar Based on that, the spirit of self-help was very high to those reported in chapter 2-high costs or and people were contributing whatever they had in building or repairing roads. But nowadays the unavailability of fertilizer, insecticides, seeds, spirit has withered due to the central and local gov- farming implements, and modem technology, ernment failing to do their part. and distant or poorly functioning markets and "Two other problems have also emerged. Buses price fluctuations. are no longer available. Instead pickups have taken People in most regions reported more diffi- their place. And because the price of gas has been culty in meeting household energy needs with increased, the drivers now charge whatever fare they want." greater distances to travel for firewood and higher use of charcoal. In Choma, Tabora, one villager said, "Nowadays we have to travel far Although the overall trends for transporta- to get firewood. Land clearing, charcoal making, tion were negative (see below), improvements grazing, and browsing by animals has all re- in roads were mentioned in a few of the most sulted in the perishing of our trees and shrubs.' isolated regions, particularly in Kagera, Mtwara, Coast, and Lindi. The availability and use of second-hand garments (Mitumba), in- cluding school uniforms for the poor, had eased the burden of clothing. Mtiti, Rukwa: "Ten years ago medicine was avail- able and services were very good. We had one Negative Trends medical assistant and he had an aide. These services were provided free of charge. Pregnant women and People reported trends worse in several areas: children were treated here. Nowadays there is noth- transport fares and access to public buses, ing, there is no medicine. If somebody becomes sick, tranpor he has to go to Laela on foot, and if it is a serious health, farming inputs, farming implements, case we have to carry him on a stretcher because and energy requirements (figure 3.3). we have a transport problem. Also we have to pay Three reasons were most frequently cited for for the health service, so a person without money the worsening transportation situation: bus just stays at home and dies. Even the witch doctors fares had gone up dramatically, the "coopera- are after money." tive spirit" had died, and the government no Kasangezi, Kigoma: "Four years ago was the same as 10 years ago, except that medicines like longer maintained the roads (box 3.1). In tetracycline, metakelfin, and seprine were available. Mwanza people reported that a bus ticket cost- Today even where there are services, the medicines ing 300 shillings in 1985, cost 700 shillings in are rare and if you can get to the clinic on the right 1991 and 1,200 in 1995. In Mtwara bus fares day, you are lucky because the medical supplies last from Utende to Mtwara had increased from only for two days." 200Ulagu, Kibaoni, Morogoro: "Today there are 200 hiling in1991to 00 n 195.more diseases, especially diarrhea, dysentery, and Health was another area of decline. Both malaria during the rainy season and floods. Now men and women pointed to poor health status there is AIDS along the Morogoro Highway." of individuals and the lack of accessible and Mwamanongu, Shinyanga: "In 1992 fees in- affordable health care services. Three kinds creased and bribing started in the hospitals. Today of problems were outstanding: lack and high you have to pay 2,500 shillings first before a woman is allowed to enter to give birth. The situation is the cost of hospitals and dispensaries; lack of hardest for women." medicine in health clinics and the high cost of e Inflation, Price Trends, and Coping Strategies 25 Later chapters detail such other trends as de- season even for those who had access to tap clining status and landlessness among women, water during the rainy season. People coped particularly after divorce and separation; with this water shortage in a variety of ways: changes in crop production, including declin- by not bathing and washing; by getting the ing soil fertility and moving from cash to food entire village involved in water fetching, some- crops (particularly in Shinyanga and Coast due times from distances requiring a two to six to disease and low yields); difficulty in getting hour walk each way; and, for those who had land for farming; and, to a lesser extent, de- carts, bicycles, or donkeys, fetching and selling creasing employment opportunities in govern- water to others for drinking and cooking. In ment-run enterprises. Nyangili, Mwanza, people reported, "Water problems appear from late July to early Octo- Coping Strategies ber. At this time fetching water becomes a task for men because it involves getting up at 3 a.m. The large majority of households in rural Tan- and coming back after the noon sun. In this zania engage in farming, and almost all agri- period we use water only for drinking and culture is primarily rain fed. The fortunes of cooking." the rural poor, therefore, ebb and flow with the Similar problems were reported in getting weather and the rains, with each season bring- firewood. People cope with wood scarcity by ing its own stress and demanding distinctive cutting down on cooking, by using maize coping strategies. stems and cobs for fuel, and by traveling far- The most serious health problems occur ther to gather firewood. In Endaswold, Arusha, during the rainy season, when water borne a person said, "You have to rent a car or tractor diseases and malaria become common (box to collect wood because you cannot go on foot. 3.3). Everywhere the poor said, "If you are Those who have money rent the tractor and unlucky and fall sick during the clearing and those who do not make due with small pieces planting season, you will face hunger that of wood and maize stock. It is a constant strug- year." gle." And in Kasangezi, Kigoma, a villager said, "Firewood is a problem throughout the Rainy Season Stress year. Where we find firewood is about six miles from here. We start the journey at 7:00 Paradoxically, while farming depends almost in the morning and we will be back with a little entirely on the timing and amount of rainfall, firewood at 2:00 in the afternoon. This is our the rainy season is also a period of acute stress daily work." for the rural poor (figure 3.4). For example, the Another common seasonal issue was stress seasonal calendar in Tabora shows that the in cash availability just before harvest, from rainy season is from January to June, but this August to September: "The problem is not la- is also a period of low food availability, low bor availability but ability to pay laborers. Har- cash reserves, high disease outbreak, and high vest time is also the deadline for paying taxes. demand for labor on the family farm. If the So some people do not have the cash to pay rains are not good, people's livelihoods are the taxes. This results in poor people selling threatened; but when the rains do come, they their poultry to pay their taxes"(Oldadai, bring a host of related problems. Village re- Arusha). ports developed through the PPA revealed the determination of the poor to do whatever it Access to Water took to survive (box 3.4). The onset of the dry season brings a different The HRD survey recognized that the presence set of problems. In most areas, water availabil- of an available water source is a poor measure ity became an acute problem during the dry of whether people actually have access to safe 26 Voices of the Poor: Poverty and Social Capital in Tanzania Box 3.3 Coping strategies for food shortages In village after village the PPA found that the poor pie dig up wild cassava tubers (Mingiko), while in resort to six basic strategies to survive-strategies Morogoro, people switch to eating leafy vegetables which also ensure that the poor remain poor: only. Piece work: Women, more than men, take on Fewer meals: Reducing the number of meals is a "piece work"-casual work on the farms of the rich common coping strategy. In general, the family eats in exchange for food and sometimes cash. Searching its daily meal in the evening, so that they do not have for such work results in short-term migration out of to go to bed without "lighting a fire." the villages. Spending time away from home to earn Loans from traders: The poor also are forced to food to survive reduces the ability of the poor to take out "food loans" at exploitative rates from trad- develop their own farms, resulting in poor harvests. ers. A bag of maize or sorghum (100 kg) typically has Petty business: Those who have any cash or in- to be returned as three or four bags after the harvest. kind reserves turn to petty trading and beer brewing. Selling cattle: As stress becomes more acute, peo- Those who don't have such reserves turn to nature- pie must sell their livestock-unfortunately at a time collecting firewood, or when feasible, making char- when trade value is lowest. In Mwanza people indi- coal. cated that one cow fetches just one bag of maize. In Changes in diet: Food substitution becomes com- Shinyanga, people reported selling one big cow for mon. In Mtwara and Lindi regions in the south peo- two bags of maize. water. The PPA study provided dramatic evi- covered dug well, ponds, streams, and rivers. dence to support this hypothesis. Seasonality and water quality were the most The FIRD concluded that only about 22 per- frequently mentioned problems. As water ta- cent of the poor had access to safe water from bles fall in the dry season, people are forced to protected sources, indoor plumbing, stand- walk further and switch to unsafe water pipes, and covered wells with hand pumps. sources of questionable quality. The PPA collected information from the same villages, and people were asked about water Gender Differences access. In more than two-thirds of these "cov- ered" villages, they mentioned lack of water as As noted in chapter 2, female-headed house- a major problem. Table 3.5 compares data from holds were considerably more vulnerable than 13 villages across regions surveyed in both their male-headed counterparts. There were a studies. Even though the villages had "im- number of factors leading to this situation, proved" sources, people were forced to switch principle among them that women hold far to alternatives in the dry season, usually un- fewer assets (land, houses, and livestock) than Box 3.4 "It is disease that forces us into poverty." Kweshwpo, Tanga 01dadai, Arushw. "The main problem is dis- from relatives and friends to take the patient to the ease in both humans and livestock. For hu- hospital or buy the medicines. So now we use herbs. For chest pain Kitonge, to cure diarrhea Ngyelekana, for mans malaria is a problem in April to July eye problems Lungala, and other herbs for epilepsy. during the rainy season. The cattle usually get Mbonde, Mtwara: "During the dry season, the sick during the same period." drinking water problem is worse and we are forced Ilolo, Mbeya: "The main diseases which affect us to use the wells and water holes which are a source are malaria and diarrhea during the rainy season. of abdominal disturbances" (women). "Cholera and Measles is high during September and October." malaria erupt and disturb people in May, June, and Langino, Ruvuma "We cannot afford medical July. Tuberculosis normally occurs during the dry services. Sometimes we are able to get contributions season" (men). Figure 3.4 Seasonal calendar, Tabora Jan. Feb. Mar. Apr. May June July Aug. Sep. Oct. Nov. Dec. Food trading Local liquor brewing Rainfall availability Time to reach water points In minutes Food availability O10 0 0_0 : E E 0 0 OO 0+ 0 000017 000 000 0 1 0 0 00 0000 000 1111 11 Cash availability [00I 0 0 ]0 0 0000 0 0 0 xxxxxx xxxxxx xxxxxx xxxxxx xxxxxx xxxxxxX Diseases outbreak XXXXXX XXXXXX XXXXXX XXXXXX XXXXXX Labor demand Laborsupplyand 000 000 000 000 000 0000 000 000 000 000 000 0000 availability 000 000 000 000 000 0000 000 000 000 000 000 0000 Migration 28 Voices of the Poor: Poverty and Social Capital in Tanzania Table 3.5 Measuring "access" to water underestilmates the problem (percentages unless othenvse noted) HRD survey PPA Ssudy Mean Water source Water problem distance Inside Stand Well With Dry season Water Change Village, region (kilometer) house post hand pump problem problem in source Mubula, Mbeya 0.3 100 X X Dug wells Karaasi, Kilimanjaro 2.6 19 62 14 X Use river Silaloda, Arusha 2.0 4 88 X X Use lake Nuangili, Mwanza 1.0 33 56 X X Use hand-dug wells Disunyala, Coast 0.3 85 X Tap water Kasangezi, Kigoma 1.0 82 X X Use dug well Mwashigsti, Shinyanga 2.0 18 59 X X Use river-dug wells Kanga, Morogoro 1.0 X X Dig small holes Korinto, Tanga 1.5 7 64 X Taps dry do men (Fong and Bhushan 1996, Mbughuni Women also reported a great variety of in- 1994, Rwebangira 1995). formal work activities that they undertook to Caroline Moser (1996) has pointed out: "Al- make ends meet. This was less frequently the though vulnerability is often used as a synonym case with men. See chapter 4 for more on gen- for poverty, the terms are not the same. Because der differences and vulnerability. poverty measures are generally fixed in time, poverty is essentially a static concept. By con- Policy Implications trast, vulnerability is more dynamic: 'people move in and out of poverty' (Lipton and Max- e Inflation has a direct impact on rural ar- well 1992), and vulnerability better captures the eas. The consumer price index should be processes of change. In addition, the concept of based on prices in both rural and urban vulnerability introduces more complex aspects areas. of 'livelihood security' relating to survival, se- * Poor rural roads and lack of transportation curity, and self-respect, as well as people's sub- affect the poor disproportionately because jective perceptions of their poverty-what it they do not have the assets to overcome means to be poor (Chambers 1989)." transportation constraints. A few important gender differences in cop- e Lack of access to health care services in- ing strategies emerged across regions. Women creases the vulnerability of the poor. had more detailed information regarding fire- * Female-headed households in rural areas wood and water collection. While the calen- have fewer assets and are more vulnerable dars women created as part of the PPA than male-headed counterparts; they are indicated scarcity of fuel in the dry seasons, likely to be more averse to risk and thus men's diagrams often did not. In some regions less likely to take advantage of market op- women indicated they had greater cash avail- portunities without a safety net. ability very briefly right after the harvest, whereas men reported access to cash through- Methodological Implications out the year. As a woman in Magata, Kagera, said, "Even when we raise and sell the crops, Information on prices and inflation that our husbands take the money. How much they could inform policymaking can be obtained give back to us depends on their mood." easily in rural areas by talking to the poor. Inflation, Price Trends, and Coping Strategies 29 * Since poverty and vulnerability vary dur- beyond provision of services to issues of ing the year, seasonal analysis quickly re- how well the system functions and how the veals policy-relevant information. water is utilized. * Even an issue as well-studied as access to The complexity and dynamics of poverty water can be misinterpreted unless ques- become apparent through understanding tions probe seasonal differences and go the coping strategies of the poor. CHAPTER 4 Gender Perspective: Development for Whom? Men own everything because when they were born, they just found it like that. - Kanazi village, Kagera Signficant dtfferences between the perspectives of men and women are well documented. These perspectives are conditioned by cultural and social factors that eventually affect economic well- being. For this reason the differing perspectives of men and women need to be taken into account by policymakers, particularly in an era in which development policies are generally moving toward support of local demand, initiative, and willingness-to-pay. This chapter highlights three issues. First, it demonstrates the differing priorities poor men and women have as a reflection of their different life experiences. Second, it highlights the importance of participatory research and data collection methodologies in uncovering such gender differences. For example, the primacy of food and water in the lives of women emerges through the use of open-ended data gathering techniques in the PPA but is generally missed in structured economic surveys. Third, it documents the prevalence of tradition and customary law and the role they play in impoverishing women upon separation, divorce, or the death of a spouse. Gender Differences in Priority Problems To avoid introducing a sectoral bias, an ac- tivity called Problem Identification was devel- Expert planning is usually done from a sectoral oped for use in the PPA. In this exercise, a set perspective: educators conduct surveys to find of 21 simple drawings was given to a group of the condition of services and willingness to pay men; an identical set was given to a group of for education; health care officials commission women. Once the pictures were correctly iden- research on health, water officials on water. The tified, people were asked to discuss them and perspective of the poor, however, is not sectoral then through voting select the five most severe in nature. Just as central planners are forced to problems faced by the village. make choices when resources are limited, those The frequency distribution of the number-one who live in a world of limited resources are problem by gender is presented in table 4.1. The forced to make tradeoffs every day to survive. different priorities of men and women are strik- Like planners, they also know that every prob- ing, reflecting their different world views and lem canoot be solved simultaneously. realities in rural Tanzania (Creighton and Oman 30 Gender Perspective: Development for Whom? 31 Table 4.1 National distribution of number-one problem ier said than done. Even in carrying out the by gender PPA, the gender bias against including women (percentages) was evident at every step. In recruiting re- Problem Men Problem Women search supervisors to act as regional team lead- Transportation 28 Food shortage 21 ers, we were provided with a long list of men. Farming 22 Water 20 Only when the study was delayed until quali- Drunkenness Health 16 fled women were found did the search inten- (crime, laziness) 15 sify; eventually a gender balance was reached. Health 10 Drunkenness However, despite serious effort, the team in- (crime, theft) 16 cluded many more male field workers than fe- Water 8 Farming 9 Deforestation/ Transportation 4 fuelwood 7 interviews with groups of men (87) than Food shortage 6 Deforestation/ women (56), despite instructions to meet with fuelwood 4 both equally. As a consequence, the study's Drought/flood 2 Drought/flood 3 aggregated results from group discussions Education 1 Labor (women, children, elderly) 3 Poor housing 1 Education 2 the results were adjusted to compensate for the Livestock 0 Livestock 2 sample's gender imbalance. Labor (women, Unlike the men-who cited the process of children, elderly) 0 Poor housing 0 farming as their number-one problem (box Total 100 Total 100 4.1)-women focused on the consequences of poor farmidng-low yields, food shortage, high prices, lack of cash, migration, and hunger (box 1995, Rwebangira 1994, Ferreira 1994, United 4.2). Republic of Tanzania and UNICEF 1990). For men the top-ranked problems were transpor- Table 4.2 National distribution of number-two problem tation (28 percent), farming 22 (percent), by gender drunkenness (15 percent), health (10 percent), (percentages) and water (8 percent). The most important Problem Men Problem Women problems mentioned by women were food shortages (21 percent), water (20 percent), Health 20 Health 25 health (16 percent), drunkenness (16 percent), Drunkenness Drunkenness 14 and farming (9 percent). From the methodo- (crime, laziness) 16 logical viewpoint, the most striking lesson Transportation 15 Food shortage 11 learned is that differences between men's and Farming 15 Water 10 women's views are most pronounced at the Food shortage 7 Transportation 9 highest level of problems. Differences are not Water 7 Deforestation! nearly so pronounced on what are considered D e oLo to be secondary problems, as indicated by table fuelwood children, eldedy) 8 4.2. When the top five problems are aggregated Livestock 7 Education 6 to give an overall ranking, gender differences Education 3 Farming 5 disappear completely. Poor housing 3 Poor housing 3 These findings highlight the importance of Drought/flood 0 Drought/flood 0 gender disaggregated data and the importance Labor (women, Livestock 0 of talking directly to women to ensure that children, eldery) 0 their perspectives and needs are reflected in Total 100 Total 100 development planning. This, of course, is eas- 32 Voices of the Poor: Poverty and Social Capital in Tanzania Box 4.1 Giving life to statistics: men's voices lack of water as a major problem. Even though the villages had improved sources, people In the PPA men were more likely than women to were forced to switch to alternatives in the dry cite problems of farming, farm inputs, and trans- season, usually uncovered dug wells, ponds, portation. Almost everywhere, they also spoke of streams, and rivers. Seasonality and water declining soil fertility. Among the responses by quality were the most frequently mentioned farmers to questions concerning problems they problems. When tables fell in the dry season, faced were: Kigombe, Tanga: "The land is tired. In 1985 an people waled e farther a sw itd acre produced 20 sacks of paddy, but now the same acre produces about 10 bags. In 1985 one could Studies demonstrate that when distance to cultivate maize and be sure of harvesting all of it, the water source is greater than 200 meters, the but now only 1.5 acres are left after monkeys and quantity of water used sharply declines. This wild pigs come." issue is seasonal. In most areas, even where Igombe, Mwanza: "In the past-4 to 10 years ago-we used to harvest 10 sacks of maize. Lately, we get only two sacks." Kapewa, Rukwa: "Ten years ago we used no Box 4.2 "How can you face your children fertilizer because the soil was fertile and our yields day after day, hungry?" Women's voices high. Four years ago we did not use fertilizer even though there was a decline in the soil fertility. Now Kazilamihunda, Kigoma: "The biggest problem in the fertility and yields are down even more and we the village is that most of us cannot feed our chil- still do not use fertilizer because it is not available. dren enough, both quantity and quality. This is be- Also, the extension agents who gave us good advice cause of our very low income, as well as laziness have no inputs left." and the men spending what little we get on beer -------- drinking rather than on food. There is also a short- age of meat and fish in this area. Because of this problem children are attacked by disease during the Collecting Water Is a Woman's job infant stage and many die." Singida, Dodoma: "The problem is hunger Water collection is a woman's job, assisted by caused by drought, laziness, and drunkenness. It is children, particularly girls. Men help in collect- a very serious problem. How can you face your ing water only when the circumstances are par- childre day after day, hungry? To cope, we go to ticularly wealthier villages and do piece work, use drought- ticuarl harh (uch s wen te wter resistant varieties, plant fast-maturing crops, sell collection trip takes 6 to 12 hours) or when firewood, and do charcoal making." water can be purchased from a vendor and Mititi, Rukwa: "Hunger is caused by infertility collected by cart or donkey. According to the of land, non-use of fertilizers, unpredictable poor HRD study, women spend six times the rains, hence poor harvests. The children are the amount of time that men do on collecting ones who are most affected. The consequences are amountpoor growth in children, stealing crops from people water. who have food, and selling our labor just to get food Water is arguably the most basic need for to eat." survival. According to the HRD study, only Sanga, Mwanza: "The prices of goods keep ris- about 22 percent of the rural population had ing day after day. Worst of all the prices of irpor- "access" to safe water from protected sources, tant commodities like cooking oil and sugar keep indoor plumping standpipes, and covered going up. Sugar now sells for 500 shillings per kilo. covere We try all kinds of things to get money: we sell local wells with hand pumps. Yet even this figure brew, collect firewood, and even grow tomatoes to appears to overestimate access because the raise our income." HRD did not measure and take into account Kashambya, Kagera: "AIDS has been a killer reliability, quality, or seasonality of the source. leaving many orphans. Most of the households in In more than two-thirds of these supposedly the village are headed by older women. There have been many deaths of parents in the last four years. covered" villages (with 22 percent of the rural The family size has grown with orphans." population, as noted above), people mentioned Gender Perspective: Development for Whom? 33 there were pipe water systems, water sources wife-beating was closely related to drunken- dried up in the dry season making water col- ness, and both said that sometimes women de- lection an all-consuming challenge (box 4.3). In served to be beaten to discipline them for such Mtwara, Coast, Tanga, Kilimanjaro, and Arusha behavior as rudeness, answering back, and not lack of water was a severe problem. The last listening to the husband (box 4.4). three areas lie in the north with high agricul- tural potential. During the dry season, water Where Is Education? vending emergee as an industry, with men managing the water vending when a mode of Some findings are important by their absence. transport was involved, and women taking the Investment in primary education and human lead with head loading. capital formation is an essential part of any long-term development strategy. Although it Gender Dfferences in Attitudes is well established that investment in women's toward Social Problems education yields high returns-the HRD study concluded that increasing each household A number of social problems were mentioned member's education by one year has a greater by both men and women, among the most impact than increasing land holding by one common: drunkenness and alcohol abuse. acre per adult-in the PPA study investment Other problems mentioned to a much lesser in education by the poor, while mentioned, extent were laziness, crime, and drug abuse. was not given the highest priority anywhere. Crime was mentioned both as internal break- The rural poor in Tanzania have more im- down of discipline and cross-border thefts, mediate survival needs such as food, water, particularly in Mara (from Kenya) and Mbeya and health. They also are dissatisfied with (from Zambia). the rising costs of primary education, the qual- Women mentioned wife-beating to a greater ity of education, the perceived low returns on extent than men. Both men and women said investment in education, and the impossibility Box 4.3 "The water problem is very serious." Butandula, Tabora 01dadai, Arusha: "Women are most affected because (men) use an ox-cart to carry water and some carry they have to lose most of their time searching for water on their heads from 7 to 10 hours away." water. During the dry season, from August to Janu- Msolwa, Coast: "We dig wells and small water ary when the springs become dry, we have to walk holes to get water and at other times we use a small 2 km to Kishimbo to get water. When we reach there, water pond 5 km away." we find that there are so many people lined up Mbonde, Mtwara: "Drinking water comes from for water, it takes six hours to get one bucket of local wells and water holes made by the villagers but water." the water is not safe. We dream of clean tap water. Mtii, Kilimanjarw. "Water is scarce. There is only During the dry season the water problem is worse one stream which becomes dry during the dry sea- and forces us to drink unclean water. This causes son, June to September. The stream is about 1.5 km abdominal disturbances. In the rainy season we use from here. Searching for water affects women and rain water and the problem is temporarily lifted." children, especially school children. When they come Mwamonongu, Shinyanga: "There is a river pass- back from school they immediately have to leave to ing through the village, but it is salty and it dries up look for water rather than do their studies. You can- from November to June. In the dry season, women not build proper mud-brick houses because you need have to buy water, which is very expensive: one water. People have to buy water at 400 shillings per drum of 20 liters goes for 50 shillings. But they have tank (200 liters)." to buy, otherwise there is no place to get water." Butandula, Tabora: "The water problem is very Ilolo, Mbeya: "Water is available during the rainy serious since there are no wells, no taps, no nearby season, but during the dry season our taps dry up. river, and there is drought much of the year. We We have to go to the river to fetch water." 34 Voices of the Poor: Poverty and Social Capital in Tanzania Box 4.4 "Once the husband is drunk, he and leave the home (25 percent), might become beats the wife.' Nsendakamoge, Tabora pregnant (24 percent) or were viewed as a waste of money since they did not bring in an Nsendakamoge, Tabora: "Men beating women is income (7 percent). Only 16 percent said they an extraordinarily common problem here. Once the had no preference in whether a girl or boy husband is drunk, he beats the wife. Women get should be withdrawn from school. beaten when the husbands are drunk, and this hap- pens all the time." Mbula, Mbeya: "The clubs (bars) are open from Female-headed Households: Doing More the morning; men stay there all day boozing, aban- with Less doning their farm work." Nsunsu, Singida: "Alcohol is a problem for eve- The PPA study found that female-headed ryone. People waste their time drinking rather than households were less well off than their male farming. A lot of food is used in the process. The counterparts at all levels as measured by sub- village government has finally taken action by re- stricting the time people are allowed to drink. Those jective wealth ranking methods. But this find- who break the rules have to pay fines." ing was inconsistent with the HRD survey Utende, Mtwara: "Wife beating is a family prob- based on consumption and expenditures. Why? lem not to be discussed publicly. Sometimes the The answer is in the data gathered for the cause is that women are rude and arrogant with HRD study which established that female- their husbands who beat them to discipline them. But some men are just oppressive and like to mis- headed households had less land, less live- treat their wives." stock, more dependants, and fewer major assets (watches, bicycles, radios). The PPA data confirmed the pattern of these findings-fe- of sending their children on to secondary male-headed households appeared to do more school. with fewer assets through a range of coping strategies (box 4.5 and table 4.4), particularly Why Girls Are Withdrawn from School petty trading. Hence, while their consumption patterns may be on a par with male-headed According to the HRD survey, the enrollment households, female-headed households re- of girls drops as children move into higher mained more vulnerable because they pos- grades at school. In the PPA study, when par- sessed fewer capital assets and experienced ents were asked if their children had dropped social isolation. out of school, the answer was almost always "no." Further exploration revealed a pattern of Divorce and Women's Property Rights periodic drop out of school to cope with eco- nomic stress rather than a permanent with- The PPA findings also provided some insights drawal. For example, parents pulled children into the main stress points in women's lives- out of school seasonally when there was insuf- ficient food or cash to pay for school-related Tab4e 4.3 Reasons for preferring to withdraw girls expenses. This pattern obviously undermines from school learning and the retention of knowledge. In this process of pulling children out of school either temporarily or more perma- Girls will get married 25 nently, girls were likely to be kept home for a Risk of pregnancy for girls 24 variety of reasons (table 4.3). Overall, 20 per- No difference 16 cent preferred educating boys as they brought Boys bring income 12 in income and were the guardians of the home. Boys are home guardians 8 Educating girls was not considered as good an Girls are a waste of money/no income 7 investment because they would get married ter Gender Perspective: Development for Whom? 35 Box 4.5 "We are not living, we are just Exploring Women's Property Rights surviving." Women's group, Utende, Mtwara survvin." ome's roup Utnde Mtara How property is divided after divorce or sepa- Msolwa, Coast: "When there is hunger and food ration is a culturally sensitive issue in Tanza- shortages, we do piece work. We go very far--one nia, so it was explored through a visual gender and one-half hours each way to fetch water; then analysis activity. A series of pictures was we try and sell it." drawn and then adapted to different regions Utende, Mtwara: "If the rain is poor or too by two local artists. Three large pictures were heavy, some of us do piece work; others go into the bush searching for Mingioko, root tubers like cas- sava. Five pieces sell for 10 shillings, so in one day couple. A group of men or women was then we can make 100 shillings and have some left over provided with 25 smaller cards depicting dif- for food. When we weed farms, we receive 20 shil- ferent assets, including children. The task they lings for a plot 10 meters by 20 meters after com- were given was to place each asset under the pletion. When we cut firewood, we can sell one ture of the bundle of 30 pieces for 120 shillings. But now be- cause the school is also doing this business it is ple is married. The process generated much difficult for us to find buyers. We also try and make discussion and immediate involvement. Once coconut oil to sell. If the coconut oil is well pre- this sorting activity was completed, the group pared, it will sell for 350 shillings. Some of us collect was asked "How are the assets distributed if traditional materials like Makuti and sell it to some- there is a divorce?" one who is building or repairing a house. One piece sells for 5 shillings. Still, sometimes the children protest and refuse to go to school because they have Ownership during Marriage not had a meal that day." During marriage women have rights to use al- most everything except those possessions a dissolution of marriage and widowhood. In man identifies as his, either because he bought principle women in Tanzania have equality be- them or uses them exclusively. For example, in fore the law. But in practice division of prop- Butundula, Tabora, men said that the radio, erty upon divorce or separation is primarily bicycle, cattle, and house are owned by the dictated by customary law and prevailing cul- man and cannot be used by others without his tural norms which vary across regions. Often, permission. The men also claimed spears, ar- these customs leave women destitute follow- rows, bows, beehives, and children because ing divorce. they had paid the bride price. However, they saw kitchen pots, ornaments, and vegetables Table 4.4 Summary statistics on male- and female-headed as belonging to women because women either households, PPA study made them or were the primary users. A very _____________________________ similar pattern emerged in Mtwara where men Male-headed Female-headed are considered exclusive owners of fishing nets Indicators household household and boats (tables 4.5 and 4.6). Expenditure per adult (tsh) 230,362 206,208 During marriage a woman owns her kitchen Expenditure per capita (tsh) 157,896 152,042 utensils, gourds, vegetables, necklaces, and Physical assets score 41.4 25.5 bracelets as these are things she uses in her Family size 6.8 5 daily activities. In some areas this was sped- Age of head of household 44 55 fled further. So long as food crops and vegeta- Years of schooling 5.5 3.5 bles were for home consumption the woman Percent literate 74 33 had "ownership rights," but when there was a Expenditure on food (tsh) 89,779 84,388 surplus, the men took possession. In other Expenditure on health (tsh) 2,677 1,824 words women did not have the right to decide Expenditure on water (tsh) 3 0n ro h t ___________________________ _ b to lcl ayhnr to kheep cafrmpitresaes. 36 Voices of the Poor: Poverty and Social Capital in Tanzania Table 4.5 Ownership and distribution of assets upon separation/divorce, for Singida (Unyamikumbi and Nsunsu villages) Who owns what duing marriage Who owns what after divorce Assets Man Woman Joint Man Woman Joint Land X X Children Xa Xb X Xa xb House X X Crops X X Donkey/cart X X Goat/sheep X X Cow X Xc X Chickens X X Beehive/honey X X Radio X X Bicycle X X Furniture X Xd Hand hoe X X Kitchen utensils X X Panga (knife) X X Personal clothing X X X X Baskets X X Decorations X X Necklaces X X Bracelets X X Sugarcane X X Fruit X X Vegetables X X a. Older than 7 years. b. Younger than 7 years. c. Owns cows milk. d. Woman gets a few chairs she brought with her. Distribution of Property at Divorce times got a portion of that year's crop but usu- ally were asked to leave just with their personal When it comes to ownership of property fol- belongings (box 4.6). lowing divorce, men were generally the win- In Mtwara and Lindi both men and women ners and women the losers, with some reported some joint division of assets. In a few regional differences. The impact on a woman villages in Dodoma women and men reported and her degree of destitution was said to vary that sometimes land was divided equally be- depending on her age, whether she had chil- tween them. In Nsunsu, Sindiga people said a dren, whether a bride price had been paid, cow's milk belongs to a woman because she whether the woman was seen to be at fault, milks the cow, but if the cow dies, the meat her ethnicity, and the temperament of the kin belongs to the husband. network. In most areas a woman lost everything in Where Divorced Women Go divorce and was said to be "lucky to leave with her clothes." The main reason: the man had Leaving their marital home with nothing, paid a bride price and so owned her, the prod- few women have the luxury of setting up uct of her labor, and any children they had independent lives and starting over. Many produced (tables 4.5 and 4.6). Women some- first try to return to their parental home if Gender Perspective: Development for Whom? 37 Table 4.6 Ownership and distribution of assets upon separation/divorce, for Coast (Disunyala village) Who owns what during marage Who owns what after divorce Assets Man Woman Joint Man Woman Joint Land X X Children X X Huts/houses X X Arrows/spears X X Money X X Cows X X Donkeys X X Chickens X X Fishing nets X X Bicycles X X Radio X X Furniture X X Personal clothing X X X X Gourds (for water shortages) X X Cooking utensils/pots X X Baskets X X Necklaces X X Bracelets X X Coffee X X Cotton X X Maize X X Maize cobs X X Sugarcane X X Cassava X X Fruits X X Vegetables X X their parents are willing to have them (box said, "It is tiresome for the legal process to 4.7). Parents give shelter to a daughter only reach conclusion; and there is a possibility that if they perceive that she was clearly not at the woman can fail to get her rights. This is fault in the divorce. If the parental home is because the man can give a lot of money to all closed to her, a woman may turn to petty the people dealing with legal rights to make trading or prostitution and find temporary sure that the woman fails." shelter wherever she can. In some villages a The younger and more educated women in woman had been given a piece of land by the some villages had chosen to fight their baffles village administration in which she was born, with the support of women's organizations. In "so she can settle down and look after her a handful of cases these women received some children." of the marital property. Some women appeared to know that they had legal recourse to ensure fair division of Widowhood Also Leads to Destitution property on divorce. The majority chose to avoid legal action, however, for a variety of If a woman is widowed, her life prospects reasons-most did not believe that the legal immediately change for the worse. Every- system would benefit them. As one woman where people said that when a woman is 38 Voices of the Poor: Poverty and Social Capital in Tanzania Box 4.6 "A woman cannot own anything valuable." Bunazi, Kagera Bunazi, Kagera: "A woman can't own anything Nsenda Kanoge, Tabora: "All the cows are left valuable. On divorce or separation a woman can with the husband unless they have an older daugh- take a young child with her until he reaches the age ter who has gotten married and has brought in bride of seven. Then she must return him. The children price. If the daughter was paid four cows, then the belong to the father. If she has no children, she gets mother on divorce will get two cows. The children nothing except what she brought when she got mar- belong to the father, and if they are young, they go ried." with the mother; the father will come for them at Kapewa, Rukwa: "The man owns everything. He age seven. If the husband is not cruel, he may give must get three-fourths of the crops because the chil- her some of the food kept in storage." dren stay with him and he must feed them." Kigombe, Tanga: "If the fight has not been so bad, Disunyala, Pwani: "A woman is allowed to move a woman may get a few more things, like a radio and out of the house only with baskets, cooking utensils, a hoe, especially if the family is well off." bracelets, and her clothes. In rare cases the clan may Kasangezi, Kigoma: "In this village men have the decide she is worthy of assistance and give her half bad habit of chasing the women away after the har- the crop of that year's harvest." vest, so they can have a good sale for that year, and then try later to get them back." widowed, she can stay on as a custodian if witch. Relatives ensure that she leaves with her children are young. This was the best case nothingbutherclothes."If awidowedwoman scenario for her. Even then, all property be- becomes involved with another man while longed to the children and she had no entitle- staying at her former husband's home, she is ment. Since the children were understood to immediately asked to leave the family home. have belonged to the father, the father's kin If the children are old at the time of widow- assumed his rights to "guard" the property hood, they have the power to state that their and often the widow was thrown out or "en- motherwill continue living with them. couraged to leave." Occasionally if the hus- band's kin were particularly kind-hearted, Policy Implications they let the widow stay on and cultivate a small plot of land. The priority problems of men and women, Women reported that "If the woman has even in the same village, are often different no children at the time of widowhood she because of their distinct roles and respon- is asked to leave immediately, sometimes sibilities. As development policies shift to blamed for the death, and even labeled a support demand orientation, taking these Box 4.7 "Even the father hesitates in welcoming her after divorce." Tella, Kilimanjaro Tella, Kilimanjaro: "It is tragic for a woman, because prostitution. Many lacking education do not know when she comes back with nothing, even her father their legal rights and end up moving with drivers of hesitates in welcoming her because she cannot inherit long-haul trucks along the Dar-Malawi or Rwanda anything from the family. A divorced or separated roads. They come back when they get pregnant." woman will be buried at the church compound, not Kanga, Morogoro: "The parents will accept her on her father's farm. In some areas they bury her at only if she is thrown out and cannot be blamed for the boundary of the farm, as she has no place in the bad behavior and has not made a legal case and made farm. The farm is for the sons." more trouble and shame. If this happens the divorced Msolwa, Pwani: "For a woman it is a problem to start woman will have a hard life. There are very few such life afresh, because the parents are older and cant control women who can cope with life and maintain moral their own lives. Sometimes women engage in businesses behavior. Many begin to prostitute themselves or beg like selling food in the open markets, do piece work, or from their neighbors." Gender Perspective: Development for Whom? 39 gender differences into account becomes Methodological Implications even more critical for success. * For poor families primary education is not It is critical to gather gender segregated a top priority in the short run. Hence, poli- data to identify priorities for any demand- cies that require financial contributions, oriented interventions. even if associated with improvements in A combinationofquantitativeand in-depth quality of education, are likely to prevent qualitative data is important in under- poor children from receiving even a pri- standing the dynamics of family decision- mary education. Targeted scholarship pro- making which determine how families grams may be needed. cope with primary education costs and * Despite the legal framework, women often who stays in school and who drops out. experience isolation and destitution after How poverty is conceptualized and meas- divorce or widowhood because of the ured helps determine who is identified as dominance of culture and traditional law poor and what policy remedies are ap- This is unlikely to change unless women plied. It is particularly important to cap- are supported through local women's ture gender differences in poverty, groups to claim a share of the household's including those based on head of house- capital assets. hold in rural areas. CHAPTER 5 Constraints on Agricultural Productivity Ten years ago we harvested ten sacks of cassava and eight sacks of maize per acre. Today, because of decline in soil fertility and rain and because we do not use fertilizer or improved seed, some of us get three or four sacks of maize while others harvest nothing. - Farmer, Hingawali, Lindi Why are so many Tanzanian farmers poor when land is so plentiful? The PPA study found that while the poor's access to land was an issue, the primary constraint was their inability to purchase agricultural inputs and implements to use the land productively. This constraint separates the country's rich farmers from the poor, making it one of the defining characteristics of poverty in Tanzania. This critical factor was not captured in the HRD survey. The PPA study also established that the worlds of rich and poor farmers are strikingly different- in particular, the opportunities and challenges they face, even when both groups live in the same village and where the farming potential is conditioned by the same agro-ecological features. Similarly the farming worlds of female- and male-headed households are quite different. To obtain information about the agricultural ices, and availability of land. Destruction of sector, the study used a variety of data collec- crops by wild animals emerged as an issue pri- tion methods and samples. Data concerning manly in Tanga and the Coast region. Theft differences associated with wealth, gender of emerged in Shinyanga, Kagera, and Mara. heads of households, and agro-ecological Table 5.1 reports constraints on agriculture zones were drawn primarily from the house- derived from systematic content analysis of hold survey. While the overall picture that group responses. Since the impact of policy var- emerged from these diverse processes points ies by poverty level and gender, the findings are in the same direction, the details vary. discussed below within these two categories. More than 80 percent of villages identified the Differences are highlighted where appropriate. following as the most important constraints: credit, crop destruction due to pests and disease, The World of Poor Farmers lack of availability of implements, and price and availability of inputs. Other frequently men- Although there are overlaps between the lives moned constraints were markets, extension serv- of rich and poor farmers, the poor face many 40 Constraints on Agricultural Productivity 41 Table 5.1 What villagers perceive as constraints Table 5.2 Reasons for moving to current village to agriculture toarclueReason Percent Number of villages Percentage of that highlighted village that high- To purchase or rent land 34 Constraint constraints Ighted constraints To use clan land 14 Credit 65 89 For emoyment 11 Cres ddit es 48 To start business other than farming 11 Pests and disease Moedb4gvenen8 Availability of implements 63 86 To b n rves 6 Price of inputs 62 85 Refugees 3 Availability of inputs 60 82 Other 12 Drought 55 75 Marketing 50 68 Extension services 46 63 Availability of land 46 63 Price of implements 42 58ment related. Those in the middle class (the Delays in payments 24 33 "average") moved for both land and employ- Wild animals 19 26 ment-related reasons. Theft 10 14 Weeds 9 12 Land Trading There was no significant difference in reported constraints not faced by the rich: lack of access ease of selling land between the rich (41 per- to land and credit, the inability to afford the cent) and the poor (40 percent). Not surpris- high cost of agricultural inputs, and difficulties ingly, there were differences between wealth in transporting their crops to market. groups in the ease with which they could pur- chase land. Overall, 41 percent reported that it Land Availability was easy to buy land. This was more true for the rich and very rich (50 percent) than it was The land issue was explored in different ways. for the poor and very poor (41 percent). People were asked the reasons why they had The strategies the rich and the poor used to migrated into their current village. Overall, 25 acquire land also differed. Among the very percent of the population had moved to their poor, 27 percent had no idea how they would present village in the past 20 years. The most acquire land, while the PPA study team en- frequent reason cited for migration was land: countered not a single case of a rich person 34 percent moved to purchase or rent land for who was unclear about this. The land acquisi- farming, and another 14 percent moved to use tion strategies among the poor were slightly clan land (table 5.2). more diverse than they were among the very Responses differed by agro-ecological zone. rich. These included inheritance, renting, being The primary reason for movement into the given land by the village government, and a Coast region was to find employment, while variety of local agricultural processes such as in the southern and western highlands, 50 per- crop-sharing, informal planting rights on land cent of migration was associated with attempts owned but not used by the rich, and opening to find new land. In the northern highlands of the bush. where people reported the inability to get land The very rich reported using two primary as the most important constraint to food pro- strategies for gaining land: clearing the bush duction, 23 percent of migration was for a land- for new land (68 percent) and purchasing land related reason. (32 percent). The emphasis the very rich placed Of those who moved to rent or purchase on clearing the bush was apparently due to land, 50 percent were poor or very poor. The their having mechanized equipment. Equip- 42 Voices of the Poor: Poverty and Social Capital in Tanzania Box 5.1 "The good soil is lost" Butandula, Tabora Butandula, Tabora: "Many people become poor be- some of us get three or four sacks of maize while cause land exhaustion leads to poor production of others harvest nothing." both cash and food crops. Ten years ago, the land Silaloda, Arusha: "Land fertility has dropped a lot was fertile because people used only the hand hoe because of soil erosion. Ten years ago we used to which does not cut deep into the soil and expose the harvest 17 bags of maize per acre. But now we get nutrients. Now many people use the plough and cut four bags per acre due to poor soil and unreliable deep into the soil, and the good soil is lost. There is rainfall." also a shortage of livestock and so we do not have Muleba, Kagera: "Some of our crops fail to grow much manure to increase the fertility of the soil. Ten or grow very badly because of lack of fertilizer, years ago, a bag of fertilizer was available and cost monoculture, drought, and tired soil. This has af- 550 shillings. Four years ago it cost 2,500 shillings. fected every farmer in the area. Most people don't Today it costs 9,000 shillings." get satisfactory food. This forces people to depend Ulaya Kibaoni, Morogoro: "Maize production has only on one type of food, bananas. So we are forced decreased compared to 10 years ago. This is because of to buy food from other places." heavy rains which cause floods every time they come. Disunyala, Coast "Ten years ago we grew a lot of Farmers do not have money for buying fertilizer." cassava and cashew nuts, enough for food consump- Hingawali, Lindi: "Ten years ago we harvested 10 tion and to sell. Today, there is not enough cassava sacks of cassava and 8 sacks of maize per acre. Today to sell and the cashew nut trees are tired. We have because of decline in soil fertility and rain and be- no money to buy chemicals for the trees which are cause we do not use fertilizer or improved seeds, infested by pests." ment makes clearing much easier than it is for the cooperatives are being dismantled, how- the poor, who, at best, must use a small han- ever, they are no longer able to supply inputs dled hoe or rent implements from others, at the same level. The private sector has yet to fill this demand so affordable inputs are not Use and Sources of Agricultural Inputs widely available. The PPA found important differences in input constraints among wealth As previously noted, the liberalization of agri- categories. In the case of fertilizer, for example, cultural markets began in 1985 and was the poor and very poor lack the financial speeded up in 1991 when the marketing of ba- wherewithal to purchase the input, but for the sic food commodities was liberalized. Prior to rich, lack of access was the key problem. Lack this, agricultural cooperatives and other gov- of availability was perhaps the orndy constraint emtent-managed or -supported corporations felt by everyone. A fairly large number (15 per- were the main sources of support in the agri- cent) either did not know how to use fertilizer cultural sector. Almost everywhere, people re- or believed that fertilizers were of no use (table ported declining land fertility (box 5.1). 5.4). Use of inputs. Nationally, approximately 20 percent or ren(by percentages) es, agrochemicals, and improved seeds. Use of these inputs was lowest among the very poor Fertlizer Improved seeds and poor (11 percent) and highest among the Veoor 11 16 very rich (33 percent) (table 5.3). Poor 16 13 A number of barriers prevented farmers Average 27 22 from using agricultural inputs, but the major Rich 11 28 one was lack of easy availability. The tradi- Very ich 33 tional source of agricultural inputs for Tanza- Tot 20 18 nian farmers was the cooperatives. Now that Constraints on Agricultural Productivity 43 Table 5.4 Reasons for not using fertilizers (percentages) Not easily Prices Don't know No No Input of Wealth category available too high how to use transport finance no use Very poor 34 12 5 1 31 10 Poor 32 15 4 - 32 9 Average 39 14 4 - 19 14 Rich 45 18 2 4 8 6 Total 36 15 4 1 24 11 Fifty-five percent of all farmers obtained agrochemicals, and improved seeds for all in- needed inputs within a radius of 3 km of their come groups, primarily the rich received im- homes; 20 percent had to travel up to 10 km; proved seeds from cooperatives. Private trade 25 percent traveled more than 10 km to obtain stores-whih were used more by the poor than inputs. Given the condition of roads in Tanza- the rich-supplied more seeds than fertilizer or nia and the lack of affordable public trans- other agro-chenicals. In general, these stores port (except along major highways), these provided inputs to the poor at high rates of distances posed a major constraint on the abil- interest. The Tanzania Fertilizer Authority ity of the poor to raise farm productivity (TFA) provided more fertilizer to the rich than through inputs. the poor. In some parts of the country, donor-fl- Over the long run differential access to fer- nanced projects such as Global 2000 and other tilizer is reflected in soil fertility. Both the rich donor/NGO-financed activities aimed at small and the poor linked the decline in soil fertility farmers were reaching the poor. to declining use of fertilizer. When asked about changes perceived in the quality of farm soil, Table 5.6 Source of inputs by wealth category more than twice as many of the very poor (47 (percentages) percent) reported large declines in fertility than Source of input did the rich (21 percent) (table 5.5). This inter- Coope Develop- pretation is further strengthened when exam- Weaith category rative TFA Trader ment Other ining the distribution of reasons people gave for improvement in soil fertility. Among the Fertil2zer rich, 21 percent linked it to the use of fertilizer, Poor 1 22 18 5 3 compared with 13 percent of the poor. Average 41 15 18 26 Rich 19 60 9 12 Source of inputs. Farmers used a variety of Total 33 20 17 30 1 organizations to obtain different types of agri- Agrochemicals cultural inputs (table 5.6). Although the coop- Very poor 56 17 13 14 eratives appeared as sources of fertilizer, Poor 36 21 16 26 Average 53 15 16 16 Table 5.5 Reported decline In soil fertility Toth 46 21 16 (percentages) Wealth category Declined a lot in 10 years Improved seeds Very poor 42 20 18 20 Very poor 47 Poor 27 23 23 27 Poor 34 Average 33 24 20 23 Average 27 Rich 71 16 4 9 Rich 21 Total 28 21 35 30 44 Voices of the Poor: Poverty and Social Capital in Tanzania Table 5.7 Source of credit by wealth category (percentages) Family, Savings and Trader/ Private Wealth category friends NBC Cooperative credit societies store indtiiduals NGO Other Very poor 61 - 4 - 15 20 - - Poor 30 - 41 1 6 7 - 15 Average 22 2 56 1 7 9 2 1 Rich 33 - 67 - - - - - Total 27 1 49 1 7 9 1 5 Credit for Agriculture primary reason for not borrowing money was due to lack of availability (45 percent). Other Almost 90 percent of all villages surveyed in reasons included not needing credit (20 per- the PPA mentioned lack of credit as a con- cent), not wanting to go into debt (12 percent), straint in agriculture. Nationally, 7 percent of lack of know-how about credit (13 percent), all farmers had borrowed money for agricul- and inability to qualify for a loan (5 percent). ture in 1991-92 and 8 percent in 1993-94. The Again, there were differences by wealth. The main source of agricultural credit for the 1993- poor were constrained primarily by lack of 94 agricultural season continued to be the co- knowledge and lack of collateral to qualify for operatives (49 percent) while the remainder credit; the rich either did not need credit or felt was from private sources, family, friends, rich it was not available. Overall, 78 percent of all people, and traders. loans reported were under 500 shillings, or less There were some important differences in than US$1.00 (table 5.9). sources of credit among the different wealth There were sharp differences among the rea- categories (table 5.7). The very poor do not par- sons for borrowing between the rich and the ticipate in any formal lending systems, includ- poor (table 5.10). The very poor borrowed only ing cooperatives. They either received credit to purchase seasonal inputs, fertilizer, agro- from family and friends or at high rates from chemicals, and seeds (100 percent). By contrast traders, trade stores, and private money lend- the rich borrowed primarily to hire labor to ers. The cooperative was the most important work on their land (69 percent). source of credit for better-off farmers. Since the borrowing rate was so low nation- Market Issues ally, people were asked why they did not use credit for agricultural purposes (table 5.8). The In probing the issue of opening markets-the central feature of the agricultural reform Table 5.8 Reason for not using agricultural credit process-the PPA study found that neither the Reason Percentage Table 5.9 Amount of money borrowed Not available 45 Tanzanian shillings Percentage Not needed 20 Did not want to go into debt 12 0-500 78 Do not know how to get credit 11 1 Could not qualify 5 10015000 8 Was not aware of credit 2 5001.10000 4 Standard rates too high 1 10001-50000 8 Other 4 50001-up 1 Constraints on Agricultural Productivity 45 Table 5.10 Reason for borrowing being more important (21 percent). Root crops, (percentages) cassava and sorghum were more important Reason Very poor Rich among the poor than the rich. Purchase seasonal input 100 31 Markets. It appears that the rich are better able Purchase tools, equipment - - Pay for labor - 69 to take advantage of market conditions, for ex- ample in cashew. The rich seemed to have quickly moved out of cotton; people com- poor nor the rich yearn for the state socialism plained that marketing cotton at present was of the past. However, the poor felt that they extremely difficult. The different abilities of the were at a decided disadvantage in the era of rich and the poor to take advantage of present open markets due to their low bargaining day markets also was reflected in where people power, lack of transport facilities, and inability sold their crops (table 5.12). The rich-who deal to obtain credit. As a result, they were unable in larger quantities and had greater access to to exploit the advantages of open markets as transport facilities-sold primarily to private the rich could. During the transition between traders (51 percent). The poor sold primarily the two economies, the situation of the poor through local markets (39 percent), to other may in fact be worsening. For example, al- farmers (12 percent), to the cooperative (27 per- though there has been a slight overall increase cent), and to private traders (15 percent). Group in the use of agricultural inputs from 1990-91 discussions revealed that whereas the rich and 1993-94, the use of inputs has declined farmers negotiated and set the prices when sell- among the poor and increased among the rich. ing to traders in large quantities, the poor deal- ing in small amounts were forced to sell at lower Crops and marketing channels. Detailed informa- prices. In one region, the rich were described as tion was collected about the most important crops those who "set the prices" and the poor as grown and the marketing channels for each. Na- "those who are forced to accept the prices set by tionally, maize was most frequently named as the others." most important crop (34 percent) followed by cof- All households reported a variety of market- fee (18 percent), and cotton (11 percent). There are ing problems. Among the very poor and poor, important differences by wealth (table5.11). While given their subsistence farming, the most fre- 22 percent of the poor feel maize is their most quent problem mentioned was that the quality important crop, the rich rank beans and pulses as of the crop produced (maize) was too poor to sell. All groups reported the disincentive of Table 5.11 Most important crops grown low prices (table 5.13). (percentages) The constraints felt by the rich were different Crop Very poor Rich Total from other groups, since they were most able Maize 22 17 34 Coffee 8 6 18 Table 5.12 Where the most important crop was sold Cotton 13 11 11 (percentages) Rice 6 8 7 Where sold Very poor Rich Beans and pulses 12 21 7 Sorghum/millet 12 7 4 Sold at local market 39 26 Cassava 9 6 4 Another farmer 12 6 Cashew 5 13 2 Cooperative 27 15 Tobacco 5 5 2 Small-scale processor 1 - Other root crops 7 4 1 Private trader 15 51 Other 1 2 10 Other 6 2 46 Voices of the Poor: Poverty and Social Capital in Tanzania Table 5.13 Problems in marketing Table 5.14 Last message heard from extension worker (percentages) (percen"es) Very Within Problem poor Poor Average Rich past 3 3 months 1-2 2 Never Income months to 1 year years years heard Production insufficient to sell 42 34 32 26 Price too low 29 36 30 26 Verypoor 11 7 6 16 60 Transport problem/cost 8 9 15 22 Poor 14 20 12 12 42 Delay in payments - 2 - Average 20 21 16 13 30 Could not find buyer at right Rich 14 18 25 8 35 time - 3 3 8 Total 17 19 14 12 38 Price too variable/risky to sell 5 5 5 13 Inefficient marketing - 1 2 - Poor crop quality 5 2 3 - Other 11 8 9 4 The very poor reported the least contact with the extension service. Sixty percent had never received an extension message. The PPA found to take advantage of markets. The constraints that the extension service seemed to most often they reported were low crop production, low reach the average fanmer and, to a lesser extent, prices, transport problems, highly fluctuating the poor and the rich (table 5.14). prices, and inability to connect with buyers at the right time. Outlook for Future Agricultural Operations Extension Services Farmers reported that their outlook for the future was optimistic despite the present pe- Tanzania's agricultural extension service gen- niod of rapid change and uncertainty. This erally is respected and the technical knowledge was true for all wealth groups including the of its extension workers valued. Several dis- very poor. Forty-five percent expected to tricts have adopted the Ministry of Agricul- either expand or remain working at the same ture-sponsored training and visit (T&V) level; 37 percent anticipated reducing or system. stopping some operations; and 18 percent People did report, however, a general de- were uncertain about their future level of ac- cline in the quality of the service in terms of tivity (table 5.15). Planned expansion areas Ucorruption," being charged for advice, and a were largely in food crops; planned reduc- decline in the presence of extension workers. tions were in cash crops, particularly coffee. Despite this, most said that if extension work- This was true for all groups except the very ers were available, they would turn to them for rich. advice. In their absence, farmers generally turned to other farmers in the locality, and the Table 5.15 Outlook for agricultural operations poor turned to the village leaders. (rcentages) People were also asked when they last re- ceived a message from an extension worker. MylokVr Only 17 percent nationally had received an ex- OuoIokpr PorAeaeRc Ttl tension message in the previous three months. Remain same 8 12 15 15 14 Another 19 percent had received a message Start new enterprise 3 2 2 3 3 during the past year. Sixty-four percent of the Expad 36 24 30 31 28 population had either never received an exten- Reduce 12 10 9 9 '10 sion message or had not done so for more than Stop 17 31 20 26 27 Uncertain 24 20 24 15 18 a year. Constraints on Agricultural Productivity 47 The World of Female-headed Households Table 5.16 Gender dferences In use of agricultural Inputs (percentages) Overall, 13 percent of rural households were Female-headed Male-headed headed by women. The size of female-headed households households households tended to be smaller (5.1 people per Use of fertilizer (-91) 10 24 household) compared with male-headed Use of fertilizer (93-94) 3 22 households (6.8). Female heads of households Use of agrcultural chemicals 14 25 had significantly less education (2.8 years) than (90-91) their male counterparts (4.5 years). Among fe- Use of agrcultural chemicals 6 24 male-headed households, the number of adult (9394) Use of improved seed (90-91) 4 21 years of education in the household (3.8 years) Use of improved seed (93-94) 4 19 was also lower than in male-headed house- holds (4.5 years). Sixty-eight percent of female-headed house- Source of Inputs holds reported agriculture as their primary ac- tivity, as opposed to 76 percent of male-headed While cooperatives generally remained an im- households. More women declared themselves portant source of inputs in the 1993-94 agricul- as farm wage-laborers or self-employed in a tural season, female-headed households variety of trading and mass production activi- reported much less use of cooperatives and the ties. This was reflected in women's reasons for Tanzania Farmers Association than did their moving out of their own villages. The primary male counterparts. Instead, they were reached reason for migration among male-headed by development projects of NGOs and church households was land-related (45 percent), groups (table 5.17). while among female-headed households the primary reason was to seek employment (40 Distance to Inputs percent) or land (25 percent). This demographic background should be The general lack of availability of agricultural kept in mind when assessing agricultural data. inputs affected male and female farmers alike. The data reported here draw upon interviews Women in rural areas, however, were gener- with women who were heads of households ally much more restricted in their physical mo- rather than women farmers in all households. bility. While 27 percent of male-headed The differentiated access to agricultural inputs, households traveled over 10 km to purchase markets, and sources of information between inputs, only 5 percent of female-headed house- female- and male-headed households is strik- holds did so. ing (table 5.16). Reasons for Not Using In puts Use ofAgricultural Inputs The primary reason for lack of use of inputs The use of all inputs was lower among female- among female-headed households was lack of headed households. While levels of use re- financial resources (46 percent), compared mained relatively stable among male-headed with 30 percent of male-headed households. households, as prices increased use dropped This confirms earlier findings that female- dramatically among female-headed house- headed households had fewer capital assets holds. This was especially true for fertilizers than male-headed households and were and agrochenicals; use of improved seeds re- judged by themselves and by community mained at its low level throughout the 1990-91 groups as poorer than male-headed house- and 1993-94 periods, holds. 48 Voices of the Poor: Poverty and Social Capital in Tanzania Table 5.17 Gender differences in source of input (percentages) Tanzania Farmer Category Cooperative Assoc. Trader/store Development project Other Fertilizer Female 21 20 22 37 - Male 29 24 19 28 - AgricuRural chemicals Female 18 33 - 48 - Male 41 26 18 24 1 Improved seeds Female - 35 19 46 - Male 31 20 26 23 - Sources of Information Sources of Farming Advice Both men and women were asked about the Men and women turned to different people many changes sweeping the country (table with their farming problems (table 5.19). 5.18). A surprising number of people had Women were more likely to seek out family heard about changes in the political system. members than were men. Men were much However, they were most knowledgeable more likely to first turn to extension workers about changes that directly affected their lives. for guidance. This included cost-sharing or user fee policies While female-headed households had con- in education and health care. tact with extension workers, it was generally Both men and women rated radios as their less so than men. Sixty-one percent of women most-used source of information (33 percent). had either never had contact with an extension Other sources mentioned were neighbors (18 worker or had no contact for more than two percent); newspapers (13 percent men, 8 per- years, while among men the percentage was 46 cent women); village leaders (12 percent); (table 5.20). groups (9 percent); and trips into town (7 per- Women also had greater difficulty in buying cent men, 3 percent women). land or getting credit, although both women and men agreed that the difficulties were be- Table 5.18 Gender differences in being informed about cause of social and cultural practices rather political change than law. Both men and women were con- (percentages) cerned about environmental degradation and Issue Female Male said that "bush clearing" was the single most MultipaWtism 17 23 Table 5.19 Gender differences in sources of farming aMice Trade liberalization 5 (percentages) Cast-sharing for education 22 19 Cost-sharing for health 28 24 Other Farmer morelta li to 8family in Village Extension Increaeibuuare 10 8 Gender Spouse member village leader Trader wonker Parliament 1 2 Women 2 38 17 13 - 30 Other 3 2 Men 5 14 20 8 1 52 Constraints on Agricultural Productivity 49 Table 5.20 Gender differences in last contact * Poor farmers need access to durable low- with extension worker cost implements and advice on sound oMnages) farming practices. Last contact Female Male o Despite much dissatisfaction with their Within past 3 months 13 17 performance, the cooperatives still remain 3 months to 1 year 15 22the most iportant source of iputs, particu- 1 to 2 years 11 15 larly for the better off. It is important to en- More than 2 years 12 11 courage competitive private sector provision. Never had contact 49 35 The farming world of female-headed households is quite distinct from male- headed households. Gender sensitive destructive practice. More women than men strategies, such as those used by some reported overgrazing and monocropping as NGOs and church groups, are important in practices having detrimental effects on the reaching households headed by women. environment. Methodological Implications Policy Implications *Once again, the findings reveal the critical * Improved rural infrastructure, particu- importance of marrying quantitative ap- larly rural roads and low-cost public proaches with the kinds of open-ended in- transport, is essential to the effective func- quiry methods used by anthropologists tioning of markets. and participatory practitioners. CHAPTER 6 Social Capital The prosperity in our village has increased. Before, the two tribes in our village used to be divided. Now they belong to the same group which brings them together. No two people have the same intelligence or resources, so when people come together they can solve many problems. - Village resident, Mararoni, Arusha Does social capital--the prevalence of norms of "civic-ness" and the vibrancy of horizontal ties in associational life--make a difference in household welfare? After two decades of research in Italy, Harvard University political scientist Robert Putnam concluded that the most important determinant of government performance and regional prosperity was social capital. Does this hold true in developing countries? Does the contribution remain significant even after controlling for the effects of physical, human, and natural capital? Is the relationship causal, that is, does higher social capital build prosperity, rather than the reverse? And, if social capital is indeed a key determinant in household welfare in developing countries, what are the policy implications? This chapter looks at social capital in Tanzania, and the magnitude of social capital's impact on incomes. The chapter also shows some of the ways social capital affects incomes: better public services, greater use of modern agricultural inputs, more community activity, and greater use of credit in agriculture. Measuring Social Capital governing economic interactions in the market- place. Social capital differs from human capital Social capital is the rules, norms, obligations, in that it is relational, embedded in social struc- reciprocity, and trust embedded in social rela- ture, and has public good characteristics. "As tions, social structures, and society's institu- an attribute of the social structure in which a tional arrangements which enable its members person is embedded, social capital is not the to achieve their individual and community private property of any of the persons who objectives. benefit from it" (Coleman 1990). This broad definition of social capital sub- The greatest potential contribution of the sumes both social capital at the micro-institu- concept of social capital is that it adds a social tional level and the rules and regulations dimension to the development equation of 50 Social Capital 51 capital that has been mostly ignored in eco- been impacted, and how that may, in turn, nomic explorations of determinants of poverty have had an impact on household welfare. and household welfare. Trust, Unity, and the Spirit of Participation Social Capital at the Community Level The psychic and transactional costs of interact- Communities are not simply geographic enti- ing with people we trust is lower than the costs ties but social entities. The behavior of individu- of interacting with someone we distrust. Trust als within a cohesive group is governed by and trustworthy behavior are, however, cultur- implicit and explicit sets of norms and expec- ally determined. When levels of trust are low tations which build up with repeated inter- among a group, associational activity and col- actions. Over time these norms and expectations lective action are inhibited. become part of the pattern and stability repre- Trust and trustworthiness that increase the sented by informal and formal organizations. chances of exchange among people without In the past few decades Tanzania has expe- written contractual obligations. Instead people rienced social upheaval (through the nation- rely on expectations of mutual obligation, hon- wide village Ujama program in the 1970s), esty, reciprocity, mutual respect, and helpful- economic decline, and a drastic structural ad- ness. In this environment, if there is a perceived justment program. Against this backdrop of in- need, cooperative action is more likely to occur stability, the PPA study sought to define and than when trust is low among people living in measure how rural communities assess them- the same village (box 6.1). selves in terms of trust, unity, and spirit of participation, and how they relate to and assess Trust over Time the function of formal and informal village- level institutions. In other words, the PPA at- People were asked a series of questions to tempted to understand how social capital had gauge their levels of trust in various groups Box 6.1 Trust and trustworthiness The PPA study teams asked people to describe what focused most often on measures of relations. Descrip- trust meant to them. People's responses focused pri- tions included the following: a custodian of other marly on adjectives describing qualities in an indi- people's money; someone who is not a thief and will vidual's openness and kindness: keeping someone's deliver things safely to someone without interfer- secret; truthfulness without tricks; talks well with ence; someone who loves others as much as they love others; does not misuse public utilities; a person themselves; the one who respects others and is re- whose behavior does not lead him to violate agree- spected by the village; someone who does not use ments; when two people believe in each other; trust abusive language, who speaks with others in a good means soul; to respect people; state of being confident manner and advises them in good directions; some- with someone; no cheating; an act of helping some- one who helps others with their problems; someone one who normally helps you; to be rewarded with who keeps his or her word; ready to ten the truth respect by another; people who live in peace and help even if it endangers someone else's interest; a person one another; ability to take care of public property; who pays back what has been given by others; one returning other people's property when you have who believes in transparency; one whose lifestyle is borrowed it and not taking it without permission; known; a hard worker; a person with good discipline; having no favorites; faith in others and in oneself; someone who you give something and who pays it when someone is given something to deliver and back; someone who meets obligations; someone who they deliver it safely; someone you can turn to for fulfils promises, has a settled mind; one in whom help. people can be confident that he will not cheat coop- People were also asked to describe a trustworthy erative properties; someone who trusts in god and person and his or her characteristics. The descriptions has love for everyone. 52 Voices of the Poor: Poverty and Social Capital in Tanzania Table 6.1 Are there more or fewer people you can Table 6.3 Change In trust toward different groups trust now? (percentages) Level of trust Percentage Four years ago Now Fewer 76 Group Low High Low High Same 9 Family 3 66 3 59 More 15 Fellow trbesman 1 27 3 20 Cell (ward) leader - 30 3 25 Village chairman 3 29 5 21 starting with the family and moving through Extension worker 9 22 15 17 tribesmen, village chief, central government, Health clinic nurse 3 29 6 23 and strangers. Traditional leader 19 15 28 10 Overall 76 percent said that there were fewer District officials 6 17 9 12 people whom they could trust today than ear- Cooprative officials lier (table 6.1). Forty-three percent attributed Trae 8 11 1 3 7 this decline to greater economic difficulties, Strangers 22 11 32 5 while 35 percent spoke about a decline in values-increasing corruption, greater selfish- ness, and lack of wisdom. Another 11 percent trusted now than four years ago, which is focused on the increasing lack of youth disci- largely explained by growing monopolistic pline (table 6.2). practices by traders in remote areas where During the past several years there had been competition is constrained by poor roads and a consistent trend toward decline in trust of lack of transport. different groups (table 6.3). As expected, fam- ily members were the most trusted and strang- Village-level Unity ers the least trusted. People valued extension workers as important sources of information People were asked to rate the level of unity of but gave them a low trust rating because, in their village now compared with four years some areas, government extension workers ago (table 6.4). Overall, 46 percent rated their had "privatized " their services and provided current level of village unity as high, com- advice only to those who could pay a fee. pared with 71 percent four years ago. The Of all service providers traditional healers number of villages reporting increased tension were the least trusted (28 percent), followed and quarrels had more than tripled, from 4 to closely by cooperative officials (23 percent). It 14 percent. is interesting to note that traders are less The increased tension was also evident from people's perceptions of declining secu- Tabl 6. Reson fordecinig tustrity and increased crime in the villages. Half Table 6.2 Reasons for declining trustof the respondents (51 percent) reported Reason Percentage high levels of insecurity and crime in their vil- Life more difficult 25 Greater economic problems 18 Table 6.4 Ratings in unity of the village Decline in values, corruption 13 (percentages) People more selfish 11 Unity rating Four years ago Now Youth indiscipline 11 Loss of belief in God 8 Extremely united 23 13 Hunger 3 United 48 33 Lack of wisdom 3 Average 25 40 Many newcomers 2 Some quarrels and conflicts 3 12 Other 6 Great tension/conflicts 1 2 Social Capital 53 lages. Once again, the primary reasons given tions, the size and centrality of the circle were increased economic insecurity and pov- indicating the relative importance of the or- erty (48 percent), laziness (18 percent), and in- ganization, group or institution in people's creased influx of strangers (7 percent). In many lives. Overlapping circles indicated interaction areas, however, associational activity involv- with other associational activities. Information ing security guards had increased to combat about village-level institutions was cross- crime in the absence of government police checked with key informants, primarily the vil- forces. lage leaders. Detailed information about The decrease in unity had yet to kill the membership and characteristics of groups was "spirit of participation" in the villages. Overall, also obtained through a household survey 78 percent rated their villages as average or questionnaire. above average in the spirit of participation on a five-point scale. People were also asked Overview ofAssociations if the village undertook any local develop- ment activities collectively. An overwhelm- On average, people belonged to 1.5 groups. ing number (89 percent) reported they did. Overall, 71 percent of the people interviewed The two most frequent activities were build- for the household survey belonged to at least ing schools (34 percent) and repairing roads one group; the maximum number of group (32 percent). memberships for an individual was seven. Membership was the most common in the Associational Activity church (21 percent), the CCM (17 percent), bur- ial societies (15 percent), Moslem groups (9 Every village had formal and informal organi- percent), and women's groups (9 percent). zations, groups, and networks. In Tanzania for- Schools and health facilities were conspicuous mal organizations include the village office and by their absence (table 6.5). chief, the cell or ward leader, the CCM (the Fifty-six percent reported that membership political party at the time of the study), UWT had increased in groups. Only 10 percent (the CCM-affiliated women's organization), reported the functioning of groups as poor schools, health clinics, hospitals, cooperatives, or very poor, while 64 percent rated group primary societies, shops, churches, mosques, functioning above average. A small percent- the police, and the marketplace. age perceived groups as receiving assistance Villages also have a range of informal village groups and networks, including: rotating grous an nework, icludng:rotaing Table 6.5 Membership in informal and formal groups credit societies (Upatu), women's groups, and associations men's groups, youth groups, sports clubs, Percentage reporting elder groups, security guards, burial societies, Group type membership and cultural and dance groups. Church 21 Institutional Mapping CCM (political party) 17 Burial society 15 Muslim group 9 A complex picture emerged in the PPA about Women's group 9 village-level associational activities and the Farmers'group 8 role they play in people's lives. Data were de- Youth group 7 rived from three sources. The first involved Primary society 4 discussions with groups of men and women Cooperative 2 using Venn diagrams (figure 6.1). In this tech- Rotating credit societies 2 nique, participants drew circles representing Darycafte 1 wthter asoitoa5ciiis nomto the village's key formal and informal instituc- 54 Voices of the Poor: Poverty and Social Capital in Tanzania Figure 6.1 Venn diagram of Bunazi village, Kagera Bank O Primary O 0 court MACOP School Health center Market HESAWA Red Cross Betania Church ' UNICEF Ito, Women's Muslim group women's \j group pWawata Note: Wawata, Roman Catholic Women's Group; Betania, Lutheran Church Women's Group; MACOP, Intemational NGO for Orphans; HESAWA, Danish- assisted Water Program. from outside the village. The most frequent Why Do People Join Groups? outside supporters of community activi- ties were NGOs (45 percent), church organiza- Membership in a group, participation in its ac- tions (19 percent), and the government (19 tivities, contributing resources in-kind, and percent). payment of fees all impose costs on members. Most people (72 percent) reported that there In times of increased economic hardship, it was at least some overlap in membership be- would be expected that membership would de- tween groups, indicating increased interaction dine. Yet in Tanzania, membership in at least among people engaged in different activities some groups is increasing, both among the' (table 6.6). poor and the wealthy. Why? Social Capital 55 Table 6.6 Overview of associational activity Table 6.8 Why people join groups (percentages) Reason Percentage Questions 1 2 3 4 5 Economic support 25 Has membership declined (1), Brings people together, belonging 19 same (2), increased (3)? 17 27 56 - - Strengthen religious beliefs 13 How are groups functioning, Provides consolation 11 1 = very poor; 5 = excellent 2 8 24 48 18 Provides encouragement, emotional support 8 Does group receive outside To share informaton, ideas, news 8 support? NGOs = 1; govem- 6 ment = 2; church = 3; foreign donors = 4; federation = 5 45 19 19 5 1 Provides credit 2 Degree of overlap in membership, Provides marketing opportunities 2 1= no overlap, 2 = little overlap; Other 6 3= much overlap 28 39 33 - - groups for religious support. Another 8 per- The importance of groups in people's lives cent mentioned sharing news and gaining ac- was probed in different ways. Six types cess to information (table 6.8). of groups were most frequently mentioned (ta- People reported disadvantages to group ble 6.7). If people could join only one group, membership as well (table 6.9). The most fre- the frequency distribution was as follows: quent disadvantages mentioned were waste of church, farmers' groups, burial societies, time (18 percent), poor group skills and man- mosques, women's groups, and the political agement (12 percent), costs and fees (10 per- organization CCM. Among female-headed cent), delay in profit-sharing (8 percent), and households, the church received the most men- poor information flow (7 percent). tion (35 percent), followed by women's organi- zations (24 percent). Among male-headed Types of Groups households, the church also was most fre- quently mentioned (24 percent), followed by If you engage yourself in many groups, how are farming groups (14 percent) and burial socie- you going to work? One needs to survve... ties (15 percent). But if you don't join a group how can you cope Many people (44 percent) joined groups for with a difficult life? emotional support, encouragement, a sense of - Villager, Tabora belonging, for consolation during misfortunes, and to feel important. Direct economic support The Venn diagram technique yielded rich in- or advantage (including credit and marketing sights into history, structure, functioning, and opportunities) was mentioned by 29 percent of variation among groups over regions. In gen- the respondents. Thirteen percent joined eral, the number of groups reported in a village Table 6.7 Groups Important to people Table 6.9 The disadvantages of group membership (Rrcentages) Most important If you could join only Disadvantage Percentage group in your life one group, which Group at present? one wouldit be? Waste of time 18 Poor group management and skills 12 Church 29 24 Costs, fees 10 Burial society 19 14 Misuse of funds by leaders 10 Mosque 11 8 Delay in profit-shaing/payments 8 Farmer's group 8 16 Poor information flow; members not kept 7 Women's group 5 8 informed Political organization 10 3 Other 31 56 Voices of the Poor: Poverty and Social Capital in Tanzania was larger than the average number of groups support to members, most churches did offer to which a household belonged. On average, assistance to the poor and needy. In Kagera, a villages reported having seven groups within region afflicted with AIDS, the church gave or- their community. The region of Kagera, bor- phans basic food and second-hand clothes. dering Rwanda, Burundi, and Uganda, had both the village with the least number of Women's groups. Tanzania has a variety of groups (two, the church and burial society) and women's groups. The most prevalent, UWT, the village with the largest number of groups was aligned with the official political party (20, including several international agencies prior to the advent of the multiparty system sponsoring work with refugees and AIDS). (box 6.3). Another type of women's group fo- Throughout the country, membership was cuses on economic activities, including garden- increasing in church groups, mosque groups, ing, salt making, small animal raising, tailoring, youth groups, and burial societies. Groups and other microenterprise activities such as whose membership was declining were agri- beer brewing, rotating-credit societies, and tree cultural cooperatives, informal credit societies, planting. A third type of group focuses on relig- and groups for the elderly. In some areas, ious, cultural, and social activities, and pro- membership in the CCM was reported to be vides emotional support. These include a declining. variety of church and mosque-related women's Women's groups were plentiful, but had a groups, burial societies, traditional women's mixed record of growth and decline. Those groups, and cultural, dance, and entertainment functioning well or associated with the church groups. Some of these, including the so- were reported to have increasing membership cieties, incorporate economic functions. while others were declining because of poor leadership and misuse of funds. Burial societies. Everyvillage shares one group in common: the burial society. Burial societies Church groups. By definition, churches are help people cope emotionally and financially open only to Christians. In Tanzania's villages with the"calamity of death." Sometimes church church membership can be as small as 30 people or mosque groups perform the function of bur- or as large as 300 (box 6.2). The main purpose of ial societies. While there are no fees, all mem- the church was reported as: to provide spiritual bers provide support to perform the burial rites direction and instruction; to strengthen one's when a member dies. One villager said that "for belief in God; to have faith; to pray. While some a relative you pay 100 shillings, and for commu- church groups provided no specific material nity members you pay 50 shillings or contribute what you can-goats, cows, firewood, sugar, oil, wrapping cloth." In many places, people Box 6.2 Roman Catholic Church reported increasing membership to "cope with in Utende, Mtwara the expenses of death." Burial societies dearly divide labor based on The Roman Catholic group has 35 parishioners. gender. Men dig the graves, collect the cash Church members work 2 two acres of maize and and sometimes the firewood, and send mes- millet to support church activities. Income from the crops is deposited in the National Bank of Com- merce. raising contributions if necessary to buy the All church members can participate in the choir coffin and other items needed for burial. group, which now has 15 members. The choir tends Women contribute and cook the food (ugali, 1.5 acres of millet and maize which earns the group maize flour, cassava, and millet) and draw the 6,000 shillings a year. Loans are sometimes made to water for the ceremony. Consoling the family choir members who are experiencing family or takes place ritualistically for three to five days. health problems.gops Teost pre at UWT ______________________cusesI Kiono economa acivepl inclutdn tgatrden- Social Capital 57 Box 6.3 Women's groups We help any member of the group when they fall into ing their rights vis-a-vis their husbands and the gov- problems; we provide consolation to the one who has a ement. The group has a farm which generates problem. 8,00010,000 shillings a year. This income is used to - UWT women's group, Butandula, Tabora expand the farm and to help members in need. The group has 35 members and imposes an entry fee of Women's groups provide a variety of services, in- 100 shillings. cluding emotional, social, cultural, and economic Nyakasanda, Kigoma: This group describes itself support activities. Following are some illustrations: as "a tree planting group." There is no fee, but eve- Kisangiro, Arusha: All married women automat- ryone has to contribute seeds and animal manure. ically belong to this group when they wed. There is no The group started in 1994 sponsored by KIDEP which fee for membership. The leader of the group is selected gave us seeds; at that time there were 20 members by the members based on her good behavior since child- but now there are only 10. The group said it was hood. The main activity is to educate girls not to become trying to solve the firewood problem. Currently it pregnant before marriage. The group also prepares and gets firewood from the neighboring country, Bu- sends food to the "youth warriors" who have the re- rundi, 6 km away. sponsibility of preventing raids on the village's cattle. Mindda, Dodoma: This salt-making group re- The group is also responsible for building a traditional stricts its membership to 20 people to keep its logis- church and mobilizing special prayers. tics simple. Every member pays a small entry fee, and Ulaya Kibaoni, Morogoro: This group brings each has specific duties to help manage daily activi- women together to cooperate on securing and keep- ties around the salt ponds. "stay with the dead person for five days if the were considered training for good citizenship dead person is a man and four days if the dead (box 6.5). person is a woman." Other forial institutions. Links with external Sungusungu or traditional security guards. One organizations were noted, particularly in asso- of the unexpected discoveries from the PPA was ciation with mission hospitals. In Kagera and the presence and importance of Sungusungu Kigoma UNICEF-supported activities received groups in many areas, particularly Tabora, Sin- diga, Dodoma, and Mara (box 6.4). These groups are often linked to youth groups, and, in Box 6.4 The evolution of security groups the absence of effective police forces, they have in Butandula, Tabora become an important village institution "toi the e e ge e combat crime, theft, and increasing distrust." Alrennut d wo menin the ely groupv ie age are tresponsible for security and at night patrol the vil- crops, livestock, and household belongings. Oc- lage to make sure that there people are not loitering casionally the activities extend to stopping around. They are also on guard against witches, witchcraft by chasing away people identified as whom they chase out of the village if found. witch doctors (figure 6.2). The women take turns preparing food for the guards every evening before the patrols begin. They also assist in tracing and giving reports to the Youth groups. Youth groups were common guards on who is a thief and who is a witch. throughout the country and primarily geared There are many advantages to the Sungusungu toward young men. The groups sometimes besides security. According to villagers, the security formed part of a larger organization such as a group provides discipline for local youth. As they political party, church or mosque groups, tradi- become involved in security work, they no longer tturn to theft, and the death rate goes down. While groupsr ra no one is paid village leaders sometimes give tional groups, such as football clubs. While guards in-kind contributions. they covered a range of activities, youth groups 58 Voices of the Poor: Poverty and Social Capital in Tanzania Figure 6.2 Venn diagram of Nsungu, Singida Mosque Church School '3un90sUngu Traditional d Burial SocietY frequent mention; in Shinyanga, it was Cargil measured by the average number of years of and Oxfam. In the Arusha area small farmers' schooling for males and females age 20 in the groups were linked to Global 2000 and FAO- household. Physical capital included house- financed projects. Other village-level institu- hold assets and access to markets, which was tions mentioned were schools, health dispensa- measured by presence of roads and proximity ries, village courts, markets, hammer mills, to markets. Agro-ecological zones were the wells, and dipping tank groups. only measure of natural capital. Based on the literature, the study assumed Econometric Analysis that levels of generalized trust and reciprocity preceded the emergence of voluntary groups, The broad working hypothesis underlying the although increased participation in groups and PPA study was that household welfare is a func- associational activity obviously feeds back into tion of human capital, physical capital, natural higher levels of trust. Through the household capital, access to markets, and social capital. survey, people rated their degree of trust in Household welfare was measured by con- others, ranging from family to strangers, on a sumption and expenditure per adult equiva- five-point scale at two moments in time. lent in the household (Ravallion 1992, 1996, The rich descriptions obtained by the PPA Ferreirh and Griffin 1995). Human capital was study reveal how trust and associational activ- Social Capital 59 Box 6.5 Profiles of Youth Groups Mararoni, Arusha: "A football club began in 1984 the interaction among the young people in the vil- with 18 members; now it has 38. Six men started the lage. It mobilizes funds from cultural shows, contri- club when they asked the village officials for assis- butions by members, and farming activities." tance with buying a football. The officials agreed Utende, Mtwara: "The group started six months and supervised the club, but it did not do well be- ago with six members-one woman and five men. A cause it lacked resources. In 1994, the club decided membership card costs 200 shillings, and annual fees to become independent so as to mobilize resources are 100 shillings. The group started with the purpose from members to buy uniforms and balls. The mem- of sharing ideas and doing business together. So far bers now make all the decisions and the club is rap- we have been engaged in poultry raising and we now idly expanding." have 13 hens. Some of the members are also members Kamsekwa B, Tabora: "The youth group is very of the village mosque group. We are saving every important because it brings all the youth together and shilling because we have plans to obtain a milling informs them of all kinds of activities. It improves machine." ity play an important role in people's social, The HRD study was particularly rich in data emotional, and economic lives. They reveal about village-level social services and house- that trust, the number of people considered hold preferences for such services. The PPA's trustworthy, and the nature and importance of SCPS was conducted using the same dusters particular groups change over time. In other used in the HRD study. Data were entered words, like other forms of capital, social capital coded in such a way that exact matching of as measured by associational activity can be dusters (villages) was possible from the two created, but it is not permanent. Social capital data sets. gains or loses value over time depending on Within each duster, 15 households were se- how people use and invest in it. lected to match the households from an earlier Since social capital is embedded in the social national agriculture survey using random sam- structure, it was measured primarily at the vil- pling conducted in the same clusters. In addi- lage level, with much less attention paid to so- tion, up to five non-agricultural households cial capital ratings of individual households. A were randomly selected and added to broaden Social Capital Index was developed to capture the sample. both the extent and nature of associational ac- tivity in a village. The Social Capital Index The Data Sets The study developed a Social Capital Index inspired in part by Robert Putnam's work in Does social capital have economic payoffs? To Italy. The Index is an arithmetic average of answer that question, the PPA used economet- both the number and characteristics of groups ic analysis drawing on two data sets: to which a person belongs. The result is more S The Social Capital and Poverty Survey a measure of the nature of groups than purely (SCPS) segment gathered data on associa- the extent of associational activity. tional activity collected from 87 villages The survey queried household respondents comprising 1,376 households. about three dimensions of social capital: first, STShe second data set was derived from the their membership in groups; second, the char- household consumption-expenditure data acteristics of those groups; and third, individ- borrowed from the 1993 HRD survey of a ual values and attitudes, particularly in light representative sample of 5,000 households of respondents' level of trust in various groups in Tanzania. and their view of social cohesion. 60 Voices of the Poor: Poverty and Social Capital in Tanzania For each group in which individuals re- results of the econometric analysis are dis- ported membership questions were asked cussed in detail in Cents and Sociability from about that group's characteristics in several di- which the rest of this section draws directly mensions relevant to its contribution to social (Narayan and Pritchett 1997). capital, including: * Inclusion of non-kin or non-clan members Results * Inclusion of members from different eco- nomic groups Analysis of the data using regression analysis * Leadership deriving livelihood from the leads to the following conclusions: same activity as members e Social capital is indeed both capital (in that * Overall functioning of groups it raises incomes) and social (in that house- * Extent of membership fees and sanctions hold outcomes depend on village and not * Degree of participation in decisionmaking just household social capital). * Membership by birth or other criteria. 9 The effect of social capital on incomes is Using the data on the frequency of member- impressive: a one standard deviation in- ship and the characteristics of groups, an index crease in village social capital increases of the village associational life (a proxy for so- household expenditures per person (a cial capital) was created. The survey also ex- proxy for income) by at least 20 to 30 per- plored the role of social and civic norms and cent; by comparison, a one standard devia- the degree of trust individuals felt toward so- tion in schooling-almost an additional cial groups such as family, village or tribe, and three years per person-increases incomes government authorities at local, district, and by only 4.8 percent. national levels. * The impact of social capital is due to vil- lage-level effects and not to wealthier Other Variables households having higher social capital. . Some of the proximate channels through The key dependent variable was household ex- which social capital affects income: better penditure per adult equivalent, a proxy for publicly provided services, greater use of household income. The analysis included modem agricultural inputs, more commu- among household characteristics the levels of nity activity, and greater use of credit in schooling of all adult males and females in the agriculture. household. Other variables were the total The coefficient of the village-level Social number of household members adjusted for Capital Index is empirically large and strongly full residency and household servants, gender statistically significant. In other words, house- of the head of the household, self-employment holds in villages with higher levels of social in agriculture, and the distance of the village capital have significantly higher expenditures to the nearest road and market (used as a proxy than households in villages with low levels. for market integration). But while the association is large and strong, A household assets index was created by as- is the relationship causal? That is to say, does signing weights to a set of household items higher social capital lead to higher incomes, or (sewing machines 20; bicycle 16; radio 8; table vice versa? 6; clock, watch, or bed 4; chair 3; lamps 2). The To establish the direction of the causality of items received half the score if they did not the relationship, the analysis had to demon- work. Originally a weight of 400 was assigned strate three things: to motor vehicles, but because very few were The effect is social. Higher incomes are observed, they tended to dominate the vari- attributable to village-level social capital ation; eventually, it was decided to arbitrarily and not to the social capital of particular reassign them the same score as a bicycle. The households. Social Capital 61 * Higher social capital causes higher in- The estimated impact is almost exactly the comes and not the reverse. same using HRD incomes as using SCPS data * The relationship is not due to the incidental for both incomes and social capital. That is, the association of both social capital and village level social capital derived from aggre- higher incomes with some other variable. gating the social capital of households inter- viewed in the SCPS has exactly the same Is Village- or Household-level Social impact on the incomes of other households in Capital Critical? a village (surveyed two years previously) as on the incomes of the original set of households. Two compelling strands of evidence establish This is like finding that one household's land that the effect is social, and not simply due to the or asset ownership is important not only for its fact that richer households have more leisure own sake but for its neighbor's as wel. It is time to participate in associational activity. hard to overstate the importance of this finding Table 6.10, column 1 shows the result of re- as it imples that at least some significant frac- gressing household expenditures (a proxy for tion of association life creates capital that is income) on the social capital of the village locally social. (excluding the household's contribution to vil- lage-level social capital) and on the house- Determining Causality hold's social capital. Again, the quotient for the village-level social capital is large and statisti- To tease out the direction of causality, a proce- cally significant; for the household's own social dure known as instrumental variable analysis capital, however, the quotient is insignificant. was used. To estimate the impact of shifts in In other words, almost all the effect is due to social capital, instrumental variables use the the village-level social capital and almost none correlation between social capital and another to the household's own social capital. This instrument which is not determined by-and finding is especially compelling because most does not directly determine-income. This of the variation in the Social Capital Index oc- eliminates the difficulty of simultaneous deter- curs across households in the same village and nination of income and social capital and iden- relatively little is due to village variation in tifies the effect of social capital on income. The social capital. drawback to this approach is that one must The second, even more compelling way to have valid instruments, and even more impor- demonstrate the social nature of the impact is tant, that the validity of an instrument depends by testing if the same relationship holds true entirely on theoretical arguments about the for different sets of households from the same structure of the model, since the "identifica- villages. If the effect is because of the village- tion" assumptions cannot be directly tested. level social capital, then other households in In the PPA study trust was posited as the the same village should similarly benefit from instrument that was not directly affected by the village-level social capital. household income and which did not affect Because of the sampling framework, it was income directly, except through social capital. possible to repeat the same equations using the In other words, greater levels of household HRD survey expenditure data on 15 different trust did lead to greater village-level social households and other demographic and village capital or associational activity of an inclu- characteristics with the village-level social sional nature. As mentioned before, trust was capital SCPS data. Column 2 of table 6.10 measured both qualitatively and quantitatively shows the results of regressing household in- through ratings on a five-point scale of the de- comes from the HRD survey on social capital gree of trust that an individual felt toward dif- from the SCPS using instrumental variables ferent groups in society. The instrument set estimation. included: trust in tribesman, trust in cell leader, 62 Voices of the Poor: Poverty and Social Capital in Tanzania Table 6.10 Household expenditures per person and social capital, comparing village- and household-level and using HRD survey data for Incomes Column 1 2 3 4 Source of data SCPS Human Resource Development Survey Level of data Household Cluster Household Type of estimationa OLS IV IVe (A) IV (B) Cluster-level social capital 0.084 0.208 0.193 0.227 1.10 2.56 2.31 1.71 Household-level social capitalc -0.020 0.526 Household size -0.077 0.019 -0.080 -0.079 7.61 1.04 10.5 10.3 Average adult schoolingb 0.019 -0.057 0.021 0.021 1.43 1.42 2.87 2.79 Female head of household (1=yes) -0.041 0.345 -0.009 -0.010 (0.448) 1.19 0.150 0.173 Asset ownership (In)b 0.253 0.245 0.143 0.143 4.40 3.88 5.26 5.20 Self-employed in agriculture (1=yes) -0.193 -0.325 -0.068 -0.069 2.36 1.19 1.69 1.68 Distance to nearest market (cluster)b -0.0036 -0.004 -0.0087 -0.0087 0.243 1.05 2.21 2.21 Agro-climate zone dummiesd Regression statistics Number of observations 846 84 1,505 1,505 Adjusted R-squared 0.215 - - - First stage incremental R2 - 0.092 0.116 0.061 Instrument test (p-level) - 0.618 0.783 0.786 a. The t-statistics are based on Huber corrected standard errors that are heteroskedasticity consistent and account for stratified sampling. b. If any of these variables were missing, then a value was imputed for that household and a missing dummy variable is set equal to one. c. Cluster level social capital index excludes household's own response. d. Included in the regressions, but not reported, are dummy variables for each of six agro-climatic zones and the three missing value dummy variables. trust in village chairman, trust in district offi- supporting the view that social capital is an cials, and trust in central government. exogenous determinant of income. Table 6.11 shows the independent variable The second argument against causality mov- estimates of the social capital impact using ing from income to social capital is the similar- cluster-level data based on the following as- ity of results whether one uses income from sumption: trust in strangers and trust in gov- the SCPS data set or the HRD survey. If indi- ernment officials are not affected directly by viduals with higher incomes have greater so- household incomes nor do they affect income cial capital because they have the luxury to directly; rather, greater levels of trust lead to participate in such activity, then one would higher village-level social capital. Column 2 of expect that the results would be stronger when the chart shows that the estimated effect of so- matching household income to household so- cial capital is strongly statistically significant, cial capital than when linking the social capital Social Capital 63 Table 6.11 Household expenditures per person and social capital using SCPS data Column 1 2 3 4 Source of data Social Capital and Poverty Survey Level of data Cluster averages Household Estimation methoda: OLS IV (A) IVc (A) IV (B) Social capital 0.119 0.496 0.559 0.345 1.80 2.75 5.37d .1.29 Household size -0.011 -0.033 -0.073 -0.075 0.210 0.424 8.66 6.21 Average adult schoolingb -0.030 -0.105 0.0087 0.013 0.789 1.66 0.512 0.761 Female head of household (1=yes) -0.439 -0.458 -0.090 -0.070 0.714 0.566 0.810 0.676 Asset ownership (In)b 0.102 -0.038 0.176 0.207 1.24 0.297 2.78 3.58 Self-employed in agriculture (1=yes) -0.99 -0.975 -0.207 -0.203 2.76 1.92 2.12 2.32 Distance to nearest market (cluster)b -0.023 -0.015 0.0062 0.0018 1.57 0.801 0.286 0.090 Agro-climatic zonesd Regression statistics Number of observations 53 53 846 846 Adjusted R-squared 0.272 - - - First Stage incremental R2 - 0.119 0.099 0.072 Instrument test (p-level) - 0.274 0.004 0.345 a. The t-statistics are Huber corrected standard errors that are heteroskedasticity consistent and account for stratified random sampling. b. If observations for any of these variables for any households were missing a value, a missing value dummy variable is set equal to one (not reported). c. The instrument sets are a: trust in strangers, tribesman, cell leaders, village chairman (govemment), district officials, central govemment; instrument set b excludes strangers. d. Included in the regressions, but not reported, are dummy variables for each of six agro-climatic zones. of one set of households to the incomes of an- The impact of village-level social capital stays other set of households. But this interpretation strong even after controlling for village-level is not supported by the results in table 6.11. variables including level of education in the village, fraction of self-employed in agricul- Village-level Effects ture, agro-ecological zones, and the village's access to markets. Other indicators, including Yet another possible explanation for the strong people's perception of the level of unity of the effect of village-level social capital is that it is village and their vision for the future, did not a proxy for some other village-level charac- explain household incomes nor change the es- teristic. Since social capital is a village-level timated impact of the Social Capital Index. measure, it was not possible to test for village- level effects. However, it was possible to test Why Higher Social Capital Leads to see if all good things go together at the vil- to Higher Incomes lage level by observing what happens to the social capital measure after controlling for a What are the proximate determinants through whole range of village-level characteristics, which village social capital impacts house- 64 Voices of the Poor: Poverty and Social Capital in Tanzania hold income? Why is it that households living portant role in overcoming the "tragedy of the in villages with higher social capital have commons"- the tendency of individuals to higher incomes? The literature suggests four maximize their own individual good at the cost mechanisms. of the collective body (Ostrom 1990). In the SCPS, households were asked if they partici- More Effective Government Services pated in communal activities aimed at main- taining certain community assets, such as Putnam suggests that higher levels of associa- building roads. Villages with higher social tional activity lead to more effective govern- capital were more likely to have engaged in ment services. The HRD survey included much community road building. This suggests an- information on quality of government-pro- other link between social capital and higher vided schools and health services. The correla- income levels. tion between school quality and village-level social capital was modestly significant. The re- Market Transactions lationship between village-level social capital and parental participation in school-related ac- It has long been recognized that market trans- tivities was significant. actions based on trustworthy partnerships fa- No link was found between quality of health cilitate economic performance. Villages with facilities and village-level social capital. This is higher social capital had significantly greater not surprising given that the key factors affect- use of agricultural credit. Since only 8 percent ing health centers are national in scope-such of households reported using credit, the one as lack of qualified doctors and drugs-and standard deviation effect is to increase credit beyond the control of village communities. use by about two-thirds (67 percent). Information Dtfhusion Conclusion People coming together can lead to sharing of The nature of village-level organizations influ- information. Households in villages with ences local development outcomes. However, higher social capital were much more likely it has proved difficult both conceptually and to use fertilizer, agrochemnical inputs, or in measurement, to include such variables in improved seeds. Analysis shows that a one the calculus of determinants of household wel- standard deviation increase in social capital in- fare or poverty. As a result, household welfare creases the probability of agrochemical use by has been measured by physical capital, human 42 percent (6.7 percentage points above the capital, and natural capital. mean of 16), fertilizer use by 38 percent (5 per- The PPA study in Tanzania shows that the centage points), and improved seed use by 17 concept of social capital is rich in potential sig- percent (2 percentage points). As with income nificance for socio-econoic development. By effects, these results are surprisingly strong. using econometric analysis-and even after These positive associations of increased use controlling for household education level and of agricultural inputs with social capital hold physical assets, and such characteristics as true even after including an individual's con- agro-ecological zone, distance to markets, tact with an extension worker and his or her population density, and ratings of unity of the own reported land quality. village-the study found that higher village- level social capital is associated with higher Management of Common Property levels of individual incomes. The quantitative effect of social capital is sur- The ability of local groups to cooperate and prisingly large: a one standard deviation in- come to acceptable solutions can play an im- crease in village-level social capital predicts a Social Capital 65 20 to 30 percent increase in expenditure per decentralization strategies and is a pre- person for each household in the village. This condition to sustainable change at the local impact is as large as tripling either the level of level. However, the technical orientation of education or stock of non-farming physical as- sectoral ministries, coupled with the pub- sets. While much remains unknown about so- lic goods nature of social capital and the cial capital in different economic and cultural lack of demand orientation in most public contexts, social capital needs to be included in sector agencies, results in "mining" exist- the household welfare equation in Tanzania. ing social capital. This often erodes or de- stroys what exists because of a lack of Policy Implications understanding and an unwillingness to in- vest in social capital as a resource with * The mere creation of groups by outsiders productive returns. Public sector agencies builds neither trust nor effective social must be educated to recognize that invest- capital. Rather, as both the qualitative and ing in social capital yields positive re- quantitative data show, groups that are turns-both to themselves and to their embedded in the social context, which clients. evolve based on community-determined There are important gender differences in needs, that include people from different the kinds of groups men and women form kin and clan groups, that involve people and join. Some groups have mixed mem- engaged in different livelihoods, and that bership and some are gender specific. function effectively, can help individuals These differences need to be taken into meet objectives that would otherwise be account to ensure that development efforts beyond their reach. reach women. * Local-level institutions have remained largely invisible to larger bureaucracies Methodological Implications such as national policymaking bodies, sectoral ministries, and international aid Many social and psychological dimen- organizations. In formulating an action sions often considered too "soft" to be agenda associated with social capital, measurable can be measured to increase the first principle should be to map what understanding of social, political, and in- local institutions already exist and their stitutions factors linked to poverty. characteristics. Qualitative data can reveal important pat- * The investment of time and resources in terns through systematic content analysis building trust and self-organizational ca- and aggregation; quantitative econometric pacity of groups and municipalities is techniques are useful teasing out magni- linked to successful implementation of tudes of impact and causality. CHAPTER 7 Credit and Savings We need funds for fertilizer and implements. We were advised to create joint ventures to get credit, and we did, but we are still waiting in vain. Borrowing from banks or the government is so bureaucratic and almost impossible. We have been filling out forms since 1992, and we are still ignored. - Farmer, Kasangezi, Kigoma Mobilizing financial resources is critical to stimulating development among the poor. In Tanzania, formal financial institutions are virtually nonexistent in rural areas. As a result, many people are forced to save in their homes or by investing in cows. Little is known about other types of informal financial institutions, nor about rural people's preferences and attitudes about savings and credit institutions nation-wide. The PPA used open-ended discussions and various participatory proc- esses to discover how the poor view savings, the network offormal and informal credit and savings institutions, and the different perspectives and practices of men and women. In 1991 the government of Tanzania issued a foster a more effective system of financial in- policy statement on financial sector reform de- termediation (World Bank 1996, Ferreira 1994, signed to stimulate competition and lessen gov- Bagachwa 1994, United Republic of Tanzania emnment interference. In an important reversal 1994). Against this background of reform of of policy, the Banking and Financial Institutions financial institutions the PPA study attempted Act became law allowing private banking. In- to explore issues surrounding savings and terest rates were liberalized, and banks could credit from the perspective of the poor. set lending rates below an announced maxi- The issue of credit was raised through open- mum and set deposit rates freely. Attempts ended discussions focusing on priority prob- were made to restructure the National Bank of lems (Priority Problem Identification). Credit Commerce (NBC), but with little success. The questions related to agriculture scenarios were Cooperative and Rural Development Bank raised using the technique Story With A Gap (CRDB) remains financially weak even after its and mapping of local-level institutions using privatization. Currently, more reforms are be- Venn diagrams. ing pursued to protect depositors of NBC, to Several questions related to savings and stem the losses experienced in the past, and to credit were included in the household ques- 66 Credit and Savings 67 tionnaire. Some were open-ended: "What do Table 7.1 The first thing you would do with 50000 shillings you understand by savings?" "What do you do category Percentage when you have surplus resources?" "Some people feel that saving for bad times or for Buy farm inputs, extend farming 22 other needs is important while others feel it is Use it for nonagricultural business/trade 13 not important. What do you think?" In addi- Build 13 tion a series of structured questions was asked Buy livestock 9 about saving habits and use and sources of Save it 6 credit. Farm tools/equipment 5 Buy/rent land 4 Savings Institutions Buy other consumer item 4 SavngsPay school fees 3 Pay medical expenses 1 In the villages of Tanzania savings meant put- Other 9 ting something aside for emergencies, or for major or special expenses. Although a common assumption is that poor people cannot manage greater vulnerability of female-headed house- money responsibly, the survey found other- holds. Women also favored investing in chil- wise. The idea of savings seemed to be wide- dren's education. spread and well understood to "help overcome the bad times in a difficult life." Overall, 96 Savings Practices percent said that savings to meet a variety of needs was very important or important in their Rural villages largely lack effective financial lives. Sixty percent of the respondents said that services. Evidence of this emerged in a var- they currently saved; 40 percent said they did ety of findings including those already re- not because they had no surplus. ported in chapter 5. When asked "Where do The subsistence level at which most rural you currently save your surplus resources?" people live in Tanzania and their investment 82 percent reported saving money at home. orientation was also evident in people's re- Investing their money in surplus grain which sponses to the question "If you were to sud- could be sold when the need arose or after denly get 50,000 shillings, what would you do prices had risen was the most common with it? Give three uses in order of importance" method. The next most frequent method was (table 7.1). by buying cattle and other livestock includ- Overall, 70 percent would invest the money ing goats. Cattle were considered "walking in productive activities or increase their per- banks and goats and sheep liquid cash." manent major assets, buy or renovate their Twenty-two percent said they saved money house, or place it in savings. Forty percent around the house or yard-under the mat- would invest the money in farm-related activi- tress, in a box, under the roof, or buried in ties including buying livestock, 13 percent in the yard (table 7.2). trading or other non-farm businesses. Thirteen Although Upatu-informal credit societies- percent said they would buy food because they are quite common, they were mentioned by frequently experienced hunger. only 1 percent of respondents. This figure ob- There were some important differences tamed from the household sample survey is based the on the gender of the head of house- surprisingly low and does not fit with infor- hold. More female-headed households (18 per- mation obtained through other methods. It is cent versus 12 percent for male) reported they possible that many people think of Upatu more would spend the additional money on food as a credit than a savings group. In fact, in and invest in non-agricultural business and some regions the literal translation is "giving trading activities. This once again reflects the credit in turn." Some respondents said that use 68 Voices of the Poor: Poverty and Social Capital in Tanzania Table 7.2 Where do you currently save? existed about rural people's preferences and Category Percentage savings habits, and gender differences were likely, a visual activity that involved people Grain storage 35 voting with stones was developed and used in Cattle 25 gender-segregated groups. In the house 22 The activity explored people's knowledge NBC 11 about savings and credit facilities and which Cooperative Rural Development Bank 2 institutional attributes they considered impor- Savings and credit 1 Post office 1 Upatu 1of five pictures which were laid out on a hor- Children's education2 zontal axis in front of a group. They repre- sented a Postal Bank, CRDB, Upatu, NBC, and cows. Pictures of attributes-such as distance of the system was declining because of decreas- from village, interest rate, safety, convenience, ing trust. and access to service-were placed on a verti- The most likely reason for the low figure is cal axis to form a matrix. Each individual in that 90 percent of the heads of households in- the group was given 20 stones to rate the insti- terviewed were male. Upatu is primarily an tutions and the attributes. investment vehicle for women (see below). The female-headed households interviewed were Village-level Data on Prevalence less able to save than their male counterparts and reported much lower use of all savings Data from group discussions surrounding institutions, both formal and informal. this activity in 77 villages were collated (table 7.3). The results were quite different from Factors Affecting Current Use, Explored those obtained from the household survey. through Voting Activity Saving in cows was most frequently reported (92 percent). Upatu was the next most preva- Use of particular facilities or institutions is af- lent (26 percent). Among the formal institu- fected by a variety of factors, the most important tions the prevalence in declining order was being their presence. But beyond presence, use Postal Bank, CRDB, and NBC. Given the lack is affected by an individual's knowledge, and to of formal institutions in most villages, a sun- some degree understanding, of the working of prisingly large number had heard about differ- an institution and by individual preference. ent types of banks. The most common source Unless much is already known about a par- of information was the radio. Overall, 40 to 50 ticular topic, household survey questionnaires percent of the villages reported no knowledge may not be the best way of getting reliable of the three formal and one informal institution information. Since little reliable information (Upatu). Table 7.3 Prevalence of sahinga Institutions, national institutions Postal Bank CRBD Upatu NBC Cows Knowledge and use n % n % n % n % d i Currentguse 14 18 7 9 20 26 10 3 71 92 Heard about but don't use 30 39 33 43 16 21 37 48 2 3 No knowledge of institution/practice 33 43 37 48 41 53 30 39 4 5 Totalnumberofzvillages 77 100 77 100 77 100 77 100 77 100 Credit and Savings 69 Table 7.4 Preferred attributes of savings Institutions Institutions Postal Bank CRBD Upatu NBC Cows Total Attributes n % n % n % n % n % n % Distance 123 19 99 21 100 22 158 31 126 8 606 17 Interest 172 27 121 26 46 10 106 20 538 37 983 28 Safety 185 29 112 24 81 18 138 27 206 14 722 20 Convenience 84 13 86 19 114 25 62 12 247 17 593 17 Access to service 80 12 45 10 110 25 54 10 345 24 634 18 Total 644 100 463 100 451 100 518 100 1462 100 3538 100 Preferred Attributes when needed, convenience, and distance. The ratings on safety were identical to the data The most helpful information to planners in- from village groups (20 percent). Another di- terested in encouraging domestic savings is mension of security which appeared to be im- knowing which attributes people most want in portant was the psychological feeling of safety, their savings institutions. Table 7.4 reports the captured in the notion of I can see it." Only combined attributes rating given by groups of 11 percent mentioned receiving interest as an men and women, collated per village. Obvi- important reason for saving. ously these ratings could only be made by groups if they had some personal knowledge Attributes Wanted in New Banks about the working of an institution. There were no dramatic gender differences, hence these are At the end of the section on financial services not reported separately. in the household survey, people were asked People most valued easy access to service (18 "What qualities should new banks have so that percent), convenience (17 percent), and lack of you are likely to use them?" In addition to the distance (17 percent). There is much overlap in attributes already discussed, people focused on these categories, which together account for 52 the need to remove the entry barriers prevent- percent of attribute preference. Other impor- ing their access to banks. Three specific recoi- tant criteria were interest rates (28 percent) and mendations: reduce the minimum balance to safety (20 percent). 500 shillings (41 percent); decrease the amount Cows were rated the highest in providing of land needed as collateral (18 percent); and the best rate of return on investment (37 per- greater transparency in management, rules, cent), but they were rated the lowest on safety (14 percent). Upatu rated the highest when dis- Table 7.5 Resons for saving tance, convenience, and access were combined (72 percent), but lowest in terms of interest. By and large, the formal banks were rated the Access (I can get it out whenever I want) 21 safest at around 25 percent. Convenient 16 Safe and secure 18 Household Data Gains interest, multiplies 11 Ican see it 14 Questions about attributes or reasons for say- Don't have t travel dit ne 2 ing in particular ways were also asked in the Don't know any other way 4 household survey (table 7.5). Similar to the rat- Other 9 ings from village-level data, 42 percent of the 100 attributes related to ease of access to savingsi 70 Voices of the Poor: Poverty and Social Capital in Tanzania Table 7.6 Qualities new banks should have specific knowledge about the formal savings categoryinstitutions, especially CRDB and NBC. Category Percentage Women often said that they did not personally Reduce minimum balance to 500 shillings 41 handle money, but that their husbands took Community participation in banks 20 care of it. Women primarily saved money in Decrease amount of land requested as collateral 18 small livestock, grain, and around the house Other 21 (box 7.1). Again, with the exception of a few areas, Upatu was primarily a source for women. and regulations through greater community participation (20 percent) (table 7.6). Different Forms of Savings Gender Difierences Investment in Cows The PPA study uncovered two levels of gender Cows are clearly the most common method of differences-one at the level of gender of head savings, mentioned in 92 percent of the vil- of household level (through data gathered lages. The main advantages, mentioned again from the household survey) and the other as- and again, were that they reproduce, give milk sociated with differences in practice (which to drink and sell, are convenient and accessi- emerged from group discussions with men and ble, and they provide manure and skins. The women in separate groups). Using people's main disadvantage was the insecurity of poverty rankings, the household survey re- the investment due to illness, death, or theft vealed that female-headed households are (box 7.2). poorer and more vulnerable than male-headed households. Accordingly, they are less able to Upatu save (50 percent versus 60 percent of male- headed households), and they have less access Nationally, Upatu was the most prevalent to both formal and informal savings groups. method of savings and credit in 26 percent of Group discussions with women all over the the villages, with usage most common in the country revealed that while knowledgeable Mtwara, Coast, Tanga, Mara, Mwanza, and about attributes one should consider in choos- Kagera regions. While it was most often used ing a savings institution, women had much less by women, in a few places men participated as Box 7.1 Women's savings institutions In addition to Upatu, Tanzania's informal revolving the harvest and sell when the prices have risen, you credit societies women keep any savings they have can make a good profit." close to home. Nyakasanba, Kigoma: "We put our savings into Igombe, Mwanza: "We have no information about goats, sorghum, millet, and groundnuts. Women can the bank (CRDB and NBC). We just keep our money keep up to three goats and some cereals as well. in holes and crannies around the house and under These investments are good because they are safe, our mattress. Most of us have no money. If we do we accessible to anyone who wants to do it, and there is just walk with it in our pocket. We die with it." no inconvenience. Magata, Kageral "In most cases, cows belong to Igaga, Shinyanga: " keep my surplus in my the men and none of us women have an account in house. I buy cattle for savings purposes. We have any bank. We get savings by keeping some of the heard of banks, but there are none in the village. we coffee because we participate in the farming. Then, save in cows, because cows reproduce and we can with the consent of the husband, we can sell it and get milk. We can sell the milk and that helps us in use the money for family uses. If you wait until after buying small items like salt." Credit and Savings 71 Box 7.2 Cows: high risk, high return trust and trustworthiness. If you cannot trust one another, Upatu cannot work. We men be- Ikigijo, Shinyanga: "Most people save in cattle and gin'to quarrel easily." very few save in NBC or the post office because the banks are 20 km away. Security nowadays is nil in Formal Credit Societies cows because of theft, disease, and death. But for the sake of the return on their investment, people buy cows. You may buy one cow for 30,000 shillings and after some time the price has doubled to 60,000. ported the use of formal credit and savings Cows increase through multiplication, they give societies, which appeared to charge very high milk, they provide manure for our farms, they are rates of interest. In Shinyanga, a case was re- used as tractors, ox ploughs, they are used in trans- ported in which a credit society charged 200 port, they provide food, beef and cooking oil, they shillings monthly for every 1,000 shillings provide skins which can be sold or used as a mat- tress, and they can be used to marry, to pay bride borrowed. price." Kigombe, Tanga: "When an emergency arises, Savings Banks you can sell cows quickly to get money." Mititi, Rukwa: "The disadvantage of cows is that In general people had heard about banks; men there are no good pastures and places to wash had more specific information about them them." ____________________________ than did women. Most people felt that to use a bank for savings, "You have to be rich, well. Convenience, access, and familiarity were have a lot of money and be influential, and be the positive attributes of Upatu most fre- very determined to overcome the bureaucracy quently cited (box 7.3). and delays. In addition most banks are too far" Upatu usually involved 10 to 20 members (box 7.4). who already knew one another. Beyond that number, people felt it was difficult to trust eve- Box 7.4 Barriers to use of savings banks ryone. Ironically, the trust basis of Upatu was the main reason given by men for not using IGgombe, Tanga: "If I have a surplus of 1,000 shil- these informal institutions: "Upatu is based on lings and I want to keep 700 shillings in the bank and the bus fare to get there is 600 shillings, I don't see the use of going. So I decided to keep my money Box 7.3 The ebb and flow of Upatu groups at home. Some people say the minimum amount you can keep in NBC is 3,000 shillings. We don't In some areas Upatu groups are viable because no have that, so we keep our money at home." other savings and credit institutions are available Kamsekwa B, Tabora: "We don't use banks be- to women. In others, Upatu groups are on the cause they are in town, so one has to travel for the decline., service. The banks are not aimed at delivering serv- Kweshindo, Tanga: "We once used Upatu as a ices to rural people as most of us cannot afford to means of savings, but now we have stopped be- maintain the minimum balance. We are also not cause of difficult economic conditions. It is not easy knowledgeable about how they operate so we fear to trust someone with your money." that our money will be lost." Maroroni, Arusha: "Upatu is convenient but not Mararoni, Arusha: "There are no banks in the very safe now. Due to economic hardships someone village. Most people who keep their money in NBC may easily get tempted and run away with the in another village are rich-mainly the ones who money." have coffee plantations, shops, or those who work Nyangili, Mwanza: "In Upatu you get profit. In for the government." the bank we are not told how much our savings Bukwimba, Mwanza: "Transport costs to the will grow. In Upatu I put in some money, and each bank are 3,000 shillings, that is the problem. If the time somebody gets it. When it is my turn I get a bank came to our village and gave us credit we lot of money, and I can see it each time." would save our money in it." 72 Voices of the Poor: Poverty and Social Capital in Tanzania Box 7.5 Experiences with Postal Banks Box 7.6 Perceptions of CRDB Marangu, Kilimanjaro: "You can get your money lgombe, Mwanza: "The CRDB is the best because any time you need it and at any Postal Bank, not the amount of collateral needed is very little com- only at the one where you initially deposited the pared to NBC. Because NBC delays the provisions money. But they are not very convenient because of loans, it takes very long to get credit. Obtaining you have to go a long way to reach them." credit at NBC is more bureaucratic and is much Ilulu, Iringa: "It is a lot of inconvenience withdraw- easier for those with a lot of money. Also interest ing money from the Postal Bank." rates on loans are higher at NBC compared to Nyangili, Mwanza: "We have no idea, but we CRDB." heard that when you put your money in, it gets lost. Marangu, Kilimanjaro: "CRDB is good because We know nothing about it." it is safe and has high interest. It is not convenient Magata, Kagera: I know a man who kept his because you have to go a long way to reach the money in the Postal Bank in Muleba. One day he banks. We have heard that the bank gives out credit had a problem, a very urgent problem, and when but we have not been able to get it because we don't he went for his money, they simply told him there even know the process of obtaining credit." is no money and he would have to come back after three days, and that was that. He needed only 2,000 shillings out of 5,800 shillings he kept there." among women was the interest rate; among men it was safety (24 percent). In several places CRDB was rated more favorably than NBC Postal Banks (box 7.6). Approximately 13 percent of the villages re- Use of Postal Banks was reported in 18 percent ported use of NBC for savings and 10 percent of villages (however, only two villages sam- for credit. NBC was valued for safety of money pled in Sindiga, Tabora, and Dodoma used and its proximity to at least some of the vil- them). For men, the most important charac- lages. NBC ranked low in convenience and teristic of Postal Banks was their safety and ease of access to money (box 7.7). security; for women, the interest rates. The pic- ture that emerged about the value of Postal Credit and Credit Institutions Banks was mixed (box 7.5). While people expressed the need for credit for Cooperative Rural Development Bank a variety of purposes, formal institutions were remarkably absent in rural areas. Approxi- CRDB was the least used formal institution (9 mately 7 percent used credit to support agri- percent). The most valued attribute of CRDB cultural activities in the 1993-1994 agricultural Box 7.7 Experiences with NBC Kasangezi, Kigoma: "NBC officials once came to our Kanazi, Kagera: "Our women's group keeps village to advertise their business. One old farmer money at NBC 15 km away so it is not very conven- agreed to keep his money with them in 1993. He ient and the interest rate is low. We do it as a group deposited 25,000 shillings with them because they because the minimum balance is 3,000 shillings and said that was the minimum amount to qualify for a no one woman has that much money. All the money loan. Then he asked for credit without success. He is held by men. The procedure is complicated; you wrote four letters but to this day he has not received are always told to come back tomorrow or the next a reply. Others from our village also joined, and no week, and there is a very complicated procedure to one got credit. So now we do not trust any of the follow, forms to fill out, and letters to write. NBC is banks. We still feel the banks are safe but they are the only bank around here. We believe the bank gives not for the poor farmers, and the closest one is very credit, but we have no idea; we wish somebody far away, 50 km." would tell us." Credit and Savings 73 season; and 80 percent of these loans were un- Table 7.7 Gender ditlerences In degree of access to credit der 500 shillings. Approximately 50 percent Category Percentage borrowed from family, friends, and relatives, 30 percent from cooperatives. Banks were men- Easier for women 12 tioned by only 0.5 percent of respondents. ao me 2 In addition to interviewing people about Dot tea 3 sources of credit for agricultural activities, peo- ple were asked whom they would approach first if they suddenly needed to borrow 2,000 less dramatic for 50,000 shillings, reflecting the shillings and 50,000 shillings (table 7.6). Re- general lack of cash surpluses in the rural econ- sponses varied depending upon the amount omy and lack of credit institutions. In group requested. For 2,000 shillings, most people discussions women said, "Loans and credit are would turn to family and friends, particular for the rich and for men. Most of us do not individuals, and finally to traders. Formal in- meet collateral needs. We have no permanent stitutions were mentioned by 3 percent of the houses and no right of occupancy certificates group. Nine percent said they would not be for the land we own" (Nyangili, Mwanza). able to access 2,000 shillings from anyone. This When probed, most men and women stated group rose to 35 percent when the amount that it was possible for women to have their needed was 50,000 shillings. The number of own bank accounts, but in reality it was more people who could turn to family members for difficult for women than men to borrow money such an amount dropped dramatically and (table 7.7). rose a bit for formal institutions. Women's groups generally felt that credit fa- cilities were not open to rural women. "Money More Difficult for Women to Get Credit is an important element in our farming activi- ties because you can fall sick and need to hire There were significant gender differences in labor. If the season comes and you are sick and the responses, pointing once again to the you have five hectares, you will still remain greater vulnerability of female-headed house- poor." In many places, women's groups were holds. Twenty-four percent of female-headed bitter about promises made but not kept: "We households would not be able to access 2,000 really want to change our agricultural prac- shillings anywhere, whereas only 6 percent of tices, but we need funds for fertilizer and im- the male-headed households indicated that to plements. We were advised to create joint be the case. This difference remained but was ventures to get credit, and we did, but we are still waiting in vain. Borrowing from banks or Table 7.6 Who would you first approach to borrow 2,000 the government is so bureaucratic and it's al- shillings and 50,000 shillings? most impossible. We have been filling out Percetageforms since 1992, and we are still ignored." Percentage Category 2,000 shillings 50,000 shillings Why People Need Credit Family, relatives, friends 82 32 Private individuals 3 7 People sometimes got upset about the issue of Trader 3 10 credit and said, "Do not ask us about credit. NBC 1 6 How can you ask us about something which Savings and credit society 1 1 you know is impossible to get?" Where women CRBD .5were willing to talk about credit, they said, Cooperative .5 1 "We don't know anything about credit because Mission, NGO 4 Mossioure NO 3 there are no credit facilities for women. If the NA bk hes to 23 less dramt for 00illgs, efwilen te 74 Voices of the Poor: Poverty and Social Capital in Tanzania happy, but the minimum balance should be The attributes most needed in rural finan- 500 shillings. We will use credit for so many cial institutions are safety, low entry re- things to raise our income, cattle rearing, ho- quirements in terms of collateral or tels, restaurants. But we may also need some savings, ease of opening accounts and business training so we do not lose our withdrawing funds, attractive rates of in- money." In the absence of credit people indi- terest, and proximity to users. cated that they turned to greater exploitation Women have less knowledge about formal of the natural environment because it was the savings institutions, hence information only resource that was free. They mentioned campaigns should target them. Customary charcoal making, firewood, wildfruit, and for- laws which limit womens control over as- est product collection, and in coastal areas, sets thereby also limit their access to collat- coral collecting. eral and credit. If it is not possible to lessen Many people saw lack of credit as the main the force of customary law in the short reason why they could not lift themselves out term, steps must be taken to improve of poverty. In Mindola, Dodoma people said, women's access to credit by other means. "The government does not care about the small Efforts should be made to learn from the farmer. A small loan could boost farm produc- experience of savings and credit schemes tion. A large number of us remain poor because in Tanzania and elsewhere, and to encour- there is no alternative way of getting farm im- age experiments or pilot schemes. Experi- plements or inputs. To cope we turn to charcoal ence in many countries suggests two rules making." In Mwanza people said, "There are of thumb: it is best to start with mobilizing few ploughs in the village. Those without savings before moving on to the provision ploughs have to provide labor in kind before of credit, and there should be no subsi- they can receive ploughing services from the dized interest rates-any subsidy should rich, so they are always late in planting. If be confined to the cost of building local there were credit facilities and ploughs were organizational capacity. reasonably priced, we could buy and share a few." Methodological Implications Policy Implications fOpen-ended exploratory methods are par- ticularly inportant to uncover all the dimen- tOther than the Postal Banks, formal rural sions of a problem and its solutions. Once financial institutions are not much in evi- the issues are better understood, household dence. Because transactions are small and surveys can be used to obtain data. costs high, formal institutions tend to limit When dealing with sensitive issues, estab- their rural activities to market centers. lishing trust and rapport are particularly Nevertheless, rural savings are available to important. This is easier in free-flowing be mobilized and there is a high demand conversations than when filling out ques- for small-scale credit. tionnaires. CHAPTER 8 Conclusion In assessing the usefulness of Participatory significant impact that strong and vibrant Poverty Assessments, two questions must associational relationships can produce on be asked. First, does the participatory, income. The lesson for policymakers should community-level approach to research about be clear: when social capital is hidden from poverty uncover data missed by more conven- view, not only is its effect on the development tional survey techniques-data valuable to process unacknowledged, but poverty allevia- conceptualization and policymaking? Second, tion programs can actually be damaging when does the concept of "social capital" yield useful they unknowingly reduce social cohesion or insights into the nature and determinants of when resources are captured by particular poverty? This volume argues that the answer groups. Understanding local-level social capi- to both questions is clearly "yes." tal is particularly important in designing The results of the PPA in Tanzania, as well decentralization processes to deliver basic as in other countries in which PPAs have been services. conducted, demonstrate that listening to the In sum, a Participatory Poverty Assessment "poverty experts"-the poor themselves-can can ensure that the voice of the poor is taken lead to a different understanding of the dy- into account in developing policy and pro- namics of poverty over time. In Tanzania, the grams. When their voice is heard, poverty al- PPA revealed, for instance, the significant gen- leviation efforts will have a much greater der dimensions of poverty and the very differ- likelihood of producing the tangible results to ent perspectives and priorities of poor men and which they aspire. women. Similarly, the importance of agricul- tural inputs, simple technology, and the inabil- Policy Recommendations ity of the poor to take advantage of markets was revealed. Even in an area about which Five broad areas of recommendation to policy- much is understood-the water challenge makers emerged from the PPA study: faced by the poor-the PPA showed that the e Create rural savings and credit institutions- magnitude of the challenge, in particular its for farm inputs, implements, and house- seasonality, was completely underestimated. hold emergencies The PPA also revealed the inflation in prices 9 Improve rural access-for input supplies, in rural areas, the widespread need for rural marketing, credit, extension, health emer- credit, and people's desire to save. gencies The importance of social capital to the qual- * Target education and health services for the ity of life in communities was confirmed by the poor-who often do without these services Social Capital and Poverty Survey, as was the at present 75 76 Voices of the Poor: Poverty and Social Capital in Tanzania * Differentiate by gender-women do more be made to employ (or contract with) local with less and are invariably worse off than communities to maintain roads, and to remove men any regulatory disincentives which inhibit * Rely on and invest in local social capital- healthy competition among transport provid- which must be viewed as a productive as- ers. set, but at present is deteriorating. Target Education and Health Services Create Rural Savings and Credit Institutions for the Poor Poor farmers in Tanzania are no different from There is extensive evidence to show the value farmers everywhere else in the world-they to society of primary education, particularly need credit to buy agricultural inputs and for girls. Yet primary education is not a priority equipment, and to avoid sales of their crops at for the poor, especially education for girl chil- distress prices. They also need secure and con- dren; the benefits are simply too remote to venient ways of saving their small surpluses. count in a time of severe stress. Consequently, Experience elsewhere suggests it is likely that even though raising school fees may generate conventional banks or other formal credit in- funds for quality improvements while enhanc- stitutions operating on commercial principles ing local accountability, it is also likely to re- will find the costs of these small transactions duce the number of poor children, and in dispersed locations simply too high to be of especially poor girls, attending school. The so- interest. On the other hand, there are many lution is not to abandon school fees, but to examples of savings and credit schemes which target fee relief for the poorest families. The function well in rural areas. Typically these are PPA found that it is not difficult for a commu- run by community-based organizations, nity to agree on who is poor or very poor. NGOs, or specialized banks, and focus initially Clearly there would be some temptation for on savings rather than credit. When they do manipulation of who would benefit from relief, lend, they charge market rates of interest. but this could probably be minimized by trans- To be sustainable, schemes will have to be parency and publicity; in any case, the costs to financially self-supporting, but it will be essen- society of large numbers of uneducated girls tial at the outset to provide seed money for greatly exceeds any adverse effects of such capacity building of both village people and manipulation. sponsoring organizations. In contrast to education, demand for health services is high among the poor. But access is Improve Rural Access severely limited by distance and cost, with ad- verse consequences for family well-being. An People can only make good use of rural savings arrangement limiting fees for the same families and credit institutions if they can obtain inputs who benefit from education fee relief might be and tools, and take their produce to market. In possible. This would not, however, solve the many areas, rural roads have deteriorated, in problems of transportation or inadequate part because of reduced maintenance efforts by medical supplies at health posts-these have both the state and local residents. Bus and bush to be addressed separately and urgently. taxi fares have risen sharply and now seem beyond many poor peoples' reach. There is no Diferentiate by Gender easy solution to this problem. In the long run, rehabilitation and proper maintenance of roads The PPA uncovered many examples of what will lower transport costs directly and at the can happen when gender differences are not same time encourage competition in transport taken into account: what is "demanded" turns services. In the short term, renewed efforts can out to be what men want, which is very often Conclusion 77 significantly different from what women want. other villages and a variety of confidence- It is also clear that women characteristically do building measures are also important elements more with less than their male counterparts, of capacity building. and that women's contributions to social capi- Outside agencies do not usually see any ad- tal are especially valuable. This is not to argue vantage in such steps, since in the short run it for special preferential treatment for women is easier to "mine" a community's existing so- (although that may be warranted in some cial capital or ignore it altogether. As part of cases), simply to urge that proper attention decentralization strategies, sector agencies be given to men's and women's different (and others involved in rural development) needs. Such differentiation will become even must be educated to recognize the positive re- more important as government agencies adopt turns from investments in social capital, and to "demand-oriented" approaches, and increas- ensure that the necessary provisions are in- ingly gear their programs to "what the clients cluded in service delivery costs. These added want." costs are unlikely to be large; in any case they The study also draws attention to the will usually result in eventual net savings as a continuing strong grip of customary law which result of more effective project implementation is often seriously biased against women. Be- and sustainability. cause customary law embodies power relation- ships which will not be easily relinquished, it Methodological Implications cannot be expected to fade away quickly. As a of the PPA Study step toward moving beyond the confines dic- tated by such laws, government might con- There are several key methodological lessons sider partnerships with elements of civil to be drawn from the Tanzania PPA: society to take up the challenge of women's * PPAs arefaster than traditional surveys, but rights, in particular the right to a just share of give similar results at the aggregate level, household assets upon the dissolution of a while producing more nuanced detailed marriage. findings. 9Subjective measures are valid and useful for Rely on and Invest in Local Social Capital taking the "social temperature" and also can be used for hard analysis. The study provides powerful support for the e PPAs can improve surveys by revealing idea that social capital is a productive asset, flaws in survey methodology and identify- with significant impact on the well-being of ing areas needing special focus. households. One implication is that outsiders, whether government, private sector, or NGOs, PPAs Are Faster must pay careful attention to existing local in- stitutions to understand their strengths and The logistics of mounting a PPA are by no weaknesses, avoid destroying what is valuable, means simple, but one can be organized more or reduce their value through their interven- quickly and produce results sooner than a tra- tions. The aim should be to strengthen and ditional survey based on questionnaires. For build on them, recognizing that this will some- this reason, the PPA technique may be useful times take more initial time and effort than for interim monitoring of poverty between ma- simply calling for the establishment of a new jor surveys. It was reassuring that-at least in committee. Another implication is that inter- the case of Tanzania-the aggregate results of ventions must include provisions for capacity the two approaches were very similar. At the building including training in group manage- same time, the PPA generated more subtle and ment, including skills such as record-keeping more detailed findings in a number of areas, and how to run meetings. Exchange visits to and its policy conusions are therefore more 78 Voices of the Poor: Poverty and Social Capital in Tanzania discriminating, and in some cases different (for tion, such techniques initiate the process of example on questions of gender, seasonality, change by raising awareness of local problems and access to water). and possible solutions. There is no reason why hard analysis Subjective Measures Are Valid (econometric techniques) cannot be applied to soft data (answers to open-ended questions). People in rural areas, whether rich or poor, are On the contrary, a great deal can be learned in well-informed and have good judgement this way not only about what happens but also about the issues that concern them. Their rec- about why it happens. ollections may be a sounder basis for assessing change over time than statistical measures de- PPAs Can Improve Surveys pendent on unreliable data. Participatory tech- niques engage their active involvement in an PPAs are not a substitute for traditional data inquiry, and by building trust help ensure that collection methods, but they can complement responses are thoughtful and accurate. The such surveys in numerous ways. One use of process of informal discussions using draw- PPA-generated data is as a precursor to a larger ings, voting with stones, and participatory survey. The PPA can help identify areas requir- community mapping generates information of ing special attention and prevent flaws in sur- relevance to national policyiakers. In addi- vey design. References Bagachwa, Mboya. 1994. Financial Integration and Fong, Monica S., and Anjana Bhushan. 1996. Development in Sub-Saharan Africa: A Study of "Toolkit on Gender in Agriculture." Gender Formal Financial Institutions in Tanzania. Report Toolkit Series 1. Gender Analysis and Policy, of a study sponsored by the World Bank Re- Poverty and Social Policy Department, World search Committee and administered by the Bank, Washington, D.C. Overseas Development Institute. University of Isham, Jonathan, Deepa Narayan, and Lant Dar es Salaam. Pritchett. 1995. "Does Participation Improve Booth, David, and others. 1993. Social, Cultural and Performance: Establishing Causality with Sub- Economic Change in Contemporary Tanzania. Report jective Data." World Bank Economic Review 9(2): to Swedish International Development Authority 175-200. (SIDA), commissioned through Development Lipton, Michael, and Simon Maxwell. 1992. The Studies Unit, Department of Social Anthropol- New Poverty Agenda: An Overview. IDS Discus- ogy, Stockholm University. Stockholm: SIDA. sion Paper 306. Sussex, United Kingdom. Carvalho, Soniya, and Howard White. 1996. Com- Maliyamkono, T. L., and M. S. D. Bagachwa. 1990. bining the Quantitative and Qualitative Approaches The Economy in Tanzania. James Currey: London, to Poverty Measurement and Analysis. Washing- United Kingdom. ton, D.C.: World Bank. Mbughuni, P. M. 1994. "Gender and Poverty Chambers, Robert. 1989. "Editorial Introduction: Alleviation in Tanzania: Issues from and for .Vulnerability, Coping and Policy." IDS Bulletin Research." University of Dar es Salaam. Dar es 20(2): 1-7. Salaam, Tanzania. Coleman, James Samuel. 1990. Foundations of Social Moser, Caroline 0. N. 1996. Confronting Crisis: A Theory. Cambridge, Mass.: Harvard University Comparative Study of Household Responses to Pov- Press. erty and Vulnerability in Four Urban Communities. Creighton, Colin, and C. K. Omari. 1995. Gender, Environmentally Sustainable Development Family and Household in Tanzania. Aldershot, Studies and Monograph Series 8. Washington, U. K.: Avebury Ashgate Publishing Ltd. D.C.: World Bank. Ferreira, Luisa. 1994. "Poverty and Inequality dur- Narayan, Deepa. 1995. The Contribution of People's ing Structural Adjustment in Rural Tanzania." Participation in 121 Rural Water Projects. ESD Research Paper Series, Paper 8. World Bank, Occasional Paper Series 1. Washington, D.C.: Washington, D.C. World Bank. Ferreira, M. Luisa, and Charles C. Griffin. 1995. Narayan, Deepa, and Lant Pritchett. 1997. "Cents "Tanzania Human Resource Development Sur- and Sociability: Household Income and Social vey: Final Report." Population and Human Re- Capital in Rural Tanzania." Policy Research sources, Eastern Africa Department, World Working Paper 1796. Social Development and Bank, Washington, D.C. Development Research Group, Poverty and 79 80 Voices of the Poor: Poverty and Social Capital in Tanzania Human Resources, World Bank, Washington, Tanzania. Ministry of Agriculture, Cooperative D.C. Department. Dar es Salaam. Ostrom, Elinor. 1990. Governing the Commons: The . 1995. Country Report to the Fourth World Evolution ofInstitutions for Collective Action. New Conference on Women: Beijing, September 1995. York: Cambridge University Press. Part One: Women in Tanzania Mainland. Dares Putnam, Robert. 1993. Making Democracy Work: Salaam. Civic Traditions in Modern Italy. Princeton, N.J.: United Republic of Tanzania and UNICEF. 1990. Princeton University Press. Women and Children in Tanzania: A Situation . 1995. "Bowling Alone: America's Declin- Analysis. Dar es Salaan. ing Social Capital." Journal of Democracy 6(1). World Bank. 1991a. Assistance Strategies to Reduce Ravallion, Martin. 1992. Poverty Comparison: A Poverty. A World Bank Policy Paper. Washing- Guide to Concepts and Methods. Living Standards ton, D.C. Measurement Study Working Paper 88. Wash- . 1991b. "Women in Development The Case ington, D.C.: World Bank. of Tanzania." Washington, D.C. .1996. "Issues in Measuring and Modeling . 1993a. "Tanzania: A Poverty Profile." Poverty." Policy Research Working Paper 1615. Washington, D.C. Policy Research Department, World Bank, . 1993b. Implementing the World Bank's Strat- Washington, D.C. egy to Reduce Poverty: Progress and Challenges. Rwebangira, M. K. 1995. "The Status of Women Washington, D.C. and Poverty in Tanzania." Background Report . 1993c. Poverty Reduction Handbook. Wash- prepared for the Workshop on Socio-Economic ington, D.C. Growth and Poverty Alleviation. Arusha, -- . 1994. "Tanzania Agriculture Sector Tanzania. Memorandum" Vol H: Main Report. Agriculture Sen, Amartya. 1990. "Gender and Cooperative and Environment Operations Division, East Africa Conflicts." In Irene Tinker, ed., Persistent In- Department, Africa Region Washington, D.C. equalities. Oxford: Oxford University Press. . 1995. World Development Report 1995: Shome, Parthasarathi, Carlos J. Moreno, and Workers in an Integrating World. New York: Kavita Rao. 1996. "Quantitative and Qualitative Oxford University Press. Methods to Social Science Inquiry." Economic . 1996. "Tanzania The Challenge of Re- and Political Weekly (July 27). forms: Growth, Incomes and Welfare." Vol 1: United Republic of Tanzania. 1994. An Approach to Main Report. Country Operations Division, East- the Development of Cooperative Banking System em Africa Department, Africa Region Washing through Savings and Credit Cooperative Societies in ton, D.C. Distributors of COLOMBIA GERMANY ISRAEL NEPAL PORTUGAL SWEDEN A D istributor ofclwia. RMOVft YoRAEtLLiterature Ltd. Eeres Mae Intermational Services (R) Ud. iari Prugal W ENfren-Williams AS W orld Bank Camera 6 No.51-21 Poppedlcorelee i5 P0. Box 5SSS GPO Box 5443 Apartado 2681, Ru Do Carmo 70-74 PO. Box 1305 Apartado Aereo 34270 53115 Bonn 3 Yohtanan Hasardar Street Katonan* 1200 Lisbon S-171 25 Soha Publications Saof6 de Bogot, D.C. Tel: (49 228) 949020 Tel Aviv 61560 Tel: (977 1) 472 152 Tel: (1) 347-4982 Tel: (46 8) 705-97-50 Fax:(97 1)224 31 ax:(1) 47-264Fox: (46 8)327--71 Prices and credit terms vary fro Tel: (57 1) 285-2798 Fax: (49 228) 217492 Tel: (972 3) 5285-397 Fan: (9771)224431 Fax: (1) 347-0284 county to country Consult your Fax: (571) 285-2798 URL: htpJ/www.uno-verlag.do Fax: (972 3)5285397 NETHERLANDS ROMANIAEma wwi. local distributor biefore placing an E-OTEl uDIeraYOREco order. COTE D'IVOIRE R.O.Y Intemational Do LiboomInOr-Publikaies Compani Do Librail Bucuresti S.A. SWITZERLAND Center cEdition ot do Diffusion Africares GREECE PO Box 13056 P.O. Box 202,7480 AE Haaksbergen Str Lipscani no. 26, sector 3 Librairie Payot Service Institutionnel ARGENTINA (CEDA) Papasolirou SA. Tel Avi 61130 Tel: (3153) 574-004 ucharest C6tes-cde-Montenon 30 fARGE IN i04 B.R 541 35. Stoumara SIr Tel: (972 3) 5461423 Fax: (31 53)572-9296 Tel: (40 1) 613 9645 1002 Lausanne Oicia del Libo intemacional tn Fa: (9723) 541442 lincdebootwrdolione.n Fax: (40 1) 312 4000 Tel: (4121) 341-3229 An. Cordobra 1877 Ebidjin 104 (4121)341-3220(723) 4644 112 Coooos aloa Te (225)246510246511 Tol: (30 1) 34-1826 E-mail: royil0netvision.not.l URL httpyww .wodrdontne.nt-indUoRo Faa: (41 21) 3413235 11ot m (54) r15-35 Fan: (225(25 0567 Fax: (28 1) 384-8254 RUSSIAN FEDERATION Tel: (54 1) 815-854 Palestinian Authorily/Midle East NEW ZEALAND Isdatelsivo ADECO Van Diermen EditionsTechniques Fax: (54 1) 8154156 CYPRUS HAI Index Information Services EBSCO NZ Ltd. 9o, Lolpachnly Poroulok Ch. do Lacuez 41 AUSTRALIA, FIJI, PAPUA NEW GUINEA Center for Applied Research Calore Dission ROB. 19502 Jomslm Nrvae Mal Sag 99914 Moscow 10 931 CH(7 Blonay SOOOISADS AUT N Cyprus cologo 5, Ruo Caloie Tl (0722) 6271319 Now Marot Tel: (7 095) 917 87 49 Tot: (4121) 942873 SOOMONISL S, VANUATU, AND 6, Diogones Street, EngDmi C.R 257 Fax: (972 2) 6271634 Auckland Fax: (7 095) 917 92 59 Fax: (41 21) 943 3605 DAESTERNnla SMviOs0802 Port-au-Prince Tol: (84 9) 534-9119 84A.nfsoionrvis ila xl: (509) 2 9260 ITALY Fax: (649) 524-807 SINGAPORE, TAIWAN, THAILAND 648 Whitehorse Road T: (357 2) 44-1730 Fax: (509) 23 4858 Licoosa Commissionaria Sansoni SPA MYANMAR, BRUNEI Central Books Distribution Mitclam Fax: (357 3)46-2051 Via Duce Di Calabria, 1/1 NIGERIA Asahgate Publishing Asia Pacific Ple. Ltd. 306 Silom Road Victoria 131 r36r HONG KONG, MACAG Casella Postale 552 Uoivoerity Prn Limited 41 Kallang Podding Road #04-03 Bangkok 10500 Tel: (61) 3 9210 777 CZECH REPUBLIC Asia 2000 Ltd. 50125 Firenze Three Crowns Building Jericho Golden Wheel Building Tel: (66 2) 235-5400 Fax: (61) 3 9210 7788 National Information Center Sales & Circulation Department Tel: (55) 645-415 Private Mat Bag 5095 Singapore 349316 Fax: (66 2) 237-8321 URml: sepirvweftdadirect.con. as pmodopra, Konvintska 5 Soabind Houso, anit 1101-02 Fax: (55)8641-257 lbadan 1Tl (85) 741-5168 i hu CS - 11357 Prago 1 22-29 Wyndham Streot, Contral E-mail: licosa0Itbcc.f 191: (234 22) 41-1356 Fao: (5) 742-935 TRINIDAD & TOBAGO, AUSTRIA Tel: (42 2) 2422-9433 Hong Kong URL: httpJAvww fbccilicosa Fax: (234 22) 41-2056 E-mail: ashgato@ansconect.com AND THE CARRIBBEAN Geind Co. Fa: (423) 2422-1484 Tl: (952) 2530-1409 Systemaics Studies Unit Woithburfigasse 20 URL: hoypl/www.nis.cz/ Fan: (852) 2526-1107 JAMAICA NORWAY SLOVENIA 9 wags streow E-mail: sales@asia200.com.hk Ian Rondo Publishers Ltd. NIC Info NS Gospodarski Veslnik Publishing Group Curope A-1011 Wien DENMARK URL: htlpilww .asia2000.com.hk 206 Old Hope Road, Kingston 6 Book Department, Postboks 6512 Efterstad Dnajska costa 5 Trinkiad, West Indies Tel: (43 1) 512-47-31 0 SamundsLitteratur Tel: 876-927-2085 N-06M Oslo 1000 Ljubljana Tel: (809) 662-5654 Faa: (431)512-47-31-29 Rosenoems AII6 11 HUNGARY Fax: 876-977-0243 Tel: (47 22) 974500 Tel: (386 61) 133 83 47; 132 12 30 Fax: (809) 662-5654 URL: htlpJiww .gerold.colatonline DK-1970 Frederiksberg C Euro Info Service E-mail: irpl@colis.com Fax: (47 22) 97-4545 Fax: (386 61) 133 80 30 E-mail: lobe@trinidadnet BANGLADESH Tel: (45 31) 351942 Margtszgeti Europa Haz E-mail: repanselj@gvestnk.si Micro Indosties Development Fax: (45 31)30782 H-113110doot JAPAN PAKISTAN SOUG D AFRICA, BOTSWANA st d. A9ostaco oclly MIDS) URL: httpJWw.sl.cbs.dk Tel: (3681) 111081 Eastomn Book Service Mirza Book Agency SW FIA OSAACsr d Assistace Society (MIDAS) Fax: (36 1) 302 5035 3-13 Hongo 3-chome, Bunkyo-ku 65, Shahrah-e-Quald-o-Azam Forsingle titles: PO Box 9997, Madoni Building House 5, Road 16 ECUADOR E-mail: eourooi@mail.matay.hu Tokyo 113 Lahore 54000 Oxford University Press Southem Africa Plot 16/4 Jinja Rd. Dhonrd R/Area Libir Mundi Tel: (81 3(3818-0861 Tel: (92 42) 735 3601 Vasco Boulevard, Goodwood Kampala Dhaka 1209 Libreria Intemacional INDIA Fax: (81 3) 38184864 Fax: (92 42) 576 3714 P0. Box 12119, NI City 7463 Tel: (256 41) 251 467 Tel: (880 2) 326427 PO. Box 17-01-3029 Allied Publishers Ltd. E-mail: orders@syt-ebs.co.0 Cape Town Fax: (256 41) 251 468 Fan (880 2) 811188 Juan Leon More 851 751 Mount Road URL: htpilww .bekkoame.or.ipl-svt-ebs Oxford University Press Tl (27 21) 595 4400 E-mail: gus@swiligands.com BELGIUM Quito Madras - 609 002 5 Bangalore Town Fax: (27 21) 595 4430 JEan eL Tel: (593 2) 521-606; (593 2) 544-185 Tel: (9144) 852-3938 KENYA Share Faisal E-mai: oxiord@oup.co.za UNITED KINGDOM Lannoy Fax: (593 2) 504-209 Fax: (91 44) 8524049 Africa Book Service (EA.) Ltd. PO Box 13033 Microino Ltd. Av. du Rol 22 E-mail: ibrimul obrimund.com.ec Quaran House, Miangano Street Karachl-75350 For subscrition orders: PO. Box 3. Alon, Hampshire GU34 2PG 100 Brussels E-mal: Ibritu2librimuni.com.ec INDONESIA PO. Box 45245 Tel: (92 21) 446307 Intemational Subscription Service England Tel: (32 2) 53-5169 Pt. Indra Liniled Nairobi Fax: (92 21) 4547640 RO. Box 41OD5 Tel:(441420)86848 Rue (32 2) 538-84 EGYPT, ARAB REPUBLIC OF Jalan Borobudur 20 Tel: (254 2) 223 641 E4-al: ouppak@l eOficenet Craighal Far (44 1420) 89889 BRAZIL Al Ahram Distribution Agency RO. Box 181 Fax: (254 2) 330 272 Johannesburg 2024 E-mal: wbank@uko.rddemon.coak PRAi T AI Galaa Street Jakarta 10320 Pak Book Corporation Tel: (27 11) 880-1448 URL- htlpi/www microkdoco.uk Ron Poio o T in a Cairo Tel: (62 21) 390-4290 KOREA, REPUBLIC OF Adz Chambers 21, Queens Road Fax (27 11) 8800248 R0an Potoom, 29 Tel: (25 2) 578-083 Fax: (62 21) 390-4289 Daejon Trading Co. Ltd. Lahore E-mal: issOls.coza VENEZUELA 01409 Sa Paulo, S Fax: (20 2) 578-6833 RO. Box 34, Youida, 706 Seou Bldg Tel: (92 42) 636 3222; 636 0885 Tocni-Ciencla Libmo, S.A. Tel: (5511) 25964 IRAN 44-6 Youldo-Dong, Yeongchengpo-Ku Fax: (92 42) 636 2328 SPAIN Cntio Cuilad Comercial aarao Fa i ( 11) 258-6990 The Middle East Observer Ketab Sara Co. Publishers Seout Email: pbclbrain.net.pk Mundi-Prensa Llbros, SAL Nivel C2. Caracas E-mail: postmaster@pti.ol.br 41 Sherif Street Khaled EsIambrli Ave., Oth Street Tel: (82 2) 785-1631/4 Castello 37 ld: (58 2) 959 5547; i035; D016 URL: tpjAww .ol.br Cairo Delafrooz Alley No. 8 Fax: (82 2) 784-0315 PERU 28091 Madrid Faxu (58 2) 959 5636 CANADA Tel: (20 2) 393-9732 PO. Box 15745-733 Editorial Desanollo SA Tel: (34 1) 4313399 RcoND PAlishing Co. Ltd. Fax: (20 2) 393-9732 Tehran 15117 MALAYSIA Apartado 3824, Lima I Fax: (34 1) 575-3998 ZAMBIA Renou9 Publishing Roatd . Tel: (9B 21) 8717819; 8716104 University of Malaya Cooperative Tel: (51 14) 285380 E-mail: Ibrelia0 m rensa.es Uoiverosy Bookshop, University of Zambia 5369 Canotek Road FINLAND Fax: (98 21) 8712479 Bookshop, Limited Fax: (51 14) 286628 UL hap1rww mundilmonsa.esi Great East Road Campos 391: O6ta3)45 K6J 5 J3 Akaeeminen Kirjakauppa E-mail: ketab-sara@neda.nel.ir .O. Boo 1127 PL I. Boo 32379 Te:(63 75065R. Boa 128 Jan Partal Barn PHILIPPINES Mun&diPfensa Barcelona Luakas Fax: (613) 745-7660 FIN-00101 Helsinki Kowkab Publishers 59700 Kuala Lumpur Intemational Booksource Center Inc. Consell do Cot, 381 Tel: (260 1) 252 576 E-mail; order.dlept@renoulooks.com Tel: (358 0) 1214418 RO. Box 19575-511 Tel: (60 3)756-5000 1127-A Antipolo St. Barangay, Venezuela 0809 Barcelona Far (260 1) 253 952 URL htlp// www.renoufboks.com Fax: (358 0) 121-4435 Teihran Fax: (60 3) 755-4424 Makati City Tel: (34 3) 488-3492 E-mil: akalaous@stockmann.1i Tel: (98 21) 25-3723 Tel: (63 2) 896 6501; 6505; 6507 Fax: (34 3) 487-7659 ZIMBABWE CHINA URL htfliAww akaleomInen.com1 Fax: (98 21) 258-3723 MEXICO Fa (63 2) 896 1741 Emal: barcelonaemundiprensa.e8 Longman Zrnbabwe (Pvt.)Ltd. Chia Financial & Economic INFOTEC Tourfe Road, Ardbennie Publishing House FRANCE IRELAND Av. San Femando No. 37 POLAND SRI LANKA, THE MALDIVES PO. Box ST125 8, Do Fo Si Dong Jie World Bank Publications Govenment Supplies Agency Col. Todello Guena Intemtional Publishing Service Lake House Bookslop Southerton Beilig 66, avenue di6na OilIg an tSoldthair 14050 Mexico, D.F UI. Piekna 31137 100, Sir Chilar palerm Gardiner Mawalhe Harare Tet (86 10) 6333-8257 75116 Pads 4-5 Harcourt Road Tel: (52 5) 624-2800 00-077 Warzawa Colombo 2 Tel: (263 4) 6216617 Fax: (86 10) 6401-7365 Tel: (33 1) 4069-30-5657 Dublin 2 Fax: (52 ) 624-2822 Tel: (48 2) 628-6089 Tel: (94 1) 32105 Fax: (263 4) 621670 Fax (331) 40-9-30-61 Tel: (353 1) 661-3111 E-mal: infotec@rtn.net.mx Fax: (48 2) 621-72i5 Fax: (94 1) 432104 Fax: (353 1) 475-2670 URL: htlpl/rtn.netma Email: books%ps0klhp.atm.cm.pl E-mal: LHL@sd.fannot , UAL: httpilww .ipsg.wawpSII. lieAport re 10 Enabling the Safe Use of Biotechnology: Principles and Practice 11 Biodiversity and Agricultural Intensification: Partners for Development and Conservation 12 Rural Development: From Vision to Action. A Sector Strategy 13 Integrated Pest Management: Strategies and Policies for Effective Implementation 14 Rural Finance: Issues, Design, and Best Practices 15 The Economics of Involuntary Resettlement (forthcoming) 16 Social Assessments for Better Development: Case Studies in Russia and Central Asia 17 Expanding the Measure of Wealth: Indicators of Environmentally Sustainable Development 18 Five Years after Rio: Innovations in Environmental Policy 19 Advancing Sustainable Development: The World Bank and Agenda 21 20 Voices of the Poor: Poverty and Social Capital in Tanzania 21 The Risks and Reconstruction Model for Resettling Displaced Populations (forthcoming) 22 A Framework for World Bank Involvement in Post-Conflict Reconstruction (forthcoming) 23 Bioengineering of Crops: Report of the World Bank Panel on Transgenic Crops Related ESSD Publications Monitoring Environmental Progress: A Report on Work in Progress Nurturing Development: Aid and Cooperation in Today's Changing World Toward Sustainable Management of Water Resources Water Supply, Sanitation, and Environmental Sustainability: The Financing Challenge The World Bank Participation Sourcebook THE WORLD BANK Washington, I).C. 20433 l'SAX Telephone: 202-477-1234 Facsimile: 202-477-6391 Tele\: AICI 64145 W)R ID)K\Nk NICA 248423 WORII,M\Nk Cable \ddress: INTItIFRAD WAS 11l NGODC WVorld WVide Veb: http:/www.no