SAGA Progress Report - Briefing Report, November 2003 - SURVEY AND DATA COLLECTION EFFORTS

SAGA
B16 MVR Hall
Ithaca, NY 14853
(607) 255-8931
Fax (607) 255-0178
saga@cornell.edu

SAGA Briefing Report
November, 2003

VI. RESEARCH OUTPUT

We selected our geographical focus (Ghana, Kenya, Madagascar, South Africa, Uganda, and the West African region) from countries with USAID country missions and groups of countries comprising regional missions, and with commitments to working with the local members of The Secretariat for Institutional Support for Economic Research in Africa (SISERA). Below is a listing of research output from the work in these various countries.

MULTI-COUNTRY:
COMPLETED

Social Identity and Manipulative Interhousehold Transfers Among East African Pastoralists
October 2002
Huysentruyt, Marieke, Christoper B. Barrett, and John G. McPeak
We model interhousehold transfers between nomadic livestock herders as the state dependent consequence of individuals’ strategic interdependence resulting from the existence of multiple, opposing externalities. A public good security externality among individuals sharing a social (e.g., ethnic) identity in a potentially hostile environment creates incentives to band together. Self-interested interhousehold wealth transfers from wealthier herders to poorer ones may emerge endogenously within a limited wealth space as a means to motivate accompanying migration by the recipient. The distributional reach and size of the transfer are limited, however, by a resource appropriation externality related to the use of common property grazing lands. When this effect dominates, it can induce distributionally regressive transfers from ex ante poor households who want to relieve grazing pressures caused by larger herds. As compared to the extant literature on transfers, our model appears more consistent with the limited available empirical evidence on heterogeneous and changing transfers patterns among east African pastoralists.

Urban-Rural Inequality in Africa
September 2002
Sahn, David E. and David C. Stifel
In this paper we examine the relative importance of rural versus urban areas in terms of monetary poverty and seven other related living standards indicators. We present the levels of urban-rural differences for several African countries for which we have data. Then we examine the relative and absolute rates of change for urban and rural areas. Next, we employ simple cross-country regression analysis to examine how some potential covariates (e.g. openness, PPP GDP per capita, urbanization) affect urban-rural disparities in well-being. And finally, we conduct urban-rural decompositions of inequality, examining the within versus between (urban and rural) group inequality for asset inequality, education inequality, and health (height) inequality.

Bayesian Herders: Updating of Rainfall Beliefs In Response To External Forecasts
Revised March 2006
Lybbert, Travis J., Christopher Barrett, John G. McPeak, and Winnie K. Luseno
Temporal climate risk weighs heavily on many of the world’s poor. Model-based climate forecasts could benefit such populations, provided recipients use forecast information to update climate expectations. We test whether pastoralists in southern Ethiopia and northern Kenya update their expectations in response to forecast information. The minority of herders who received these climate forecasts updated their expectations for below normal rainfall, but not for above normal rainfall. This revealed preoccupation with downside risk highlights the potential value of better climate forecasts in averting drought-related losses, but realizing any welfare gains requires that recipients strategically react to these updated expectations.
Forthcoming in World Development

Smallholder Identities and Social Networks: The Challenge of Improving Productivity and Welfare
July 2003
Barrett, Christopher B.
This paper proposes a general framework for resolving the puzzle of how to reconcile the mass of recent evidence on the salutary effects of social capital at the individual level with the casual, larger-scale observation that social embeddedness appears negatively correlated with productivity and material measures of welfare. It advances an analytical framework that not only explains individual productivity or technology adoption behavior as a function of the characteristics or behaviors of others, but that also explains the aggregate properties of social systems characterized by persistently low productivity. Examples from Kenya and Madagascar are used to illustrate the phenomena discussed.

Fractal Poverty Traps
Published January 2006
Barrett, Christopher B. and Brent M. Swallow
This paper offers an informal theory of a special sort of poverty trap, one in which multiple dynamic equilibria exist simultaneously at multiple (micro, meso and/or macro) scales of analysis and are self-reinforcing through feedback effects. Small adjustments at any one of these levels are unlikely to move the system away from its dominant, stable dynamic equilibrium. Governments, markets and communities are simultaneously weak in places characterized by fractal poverty traps. No unit operates at a high-level equilibrium in such a system. All seem simultaneously trapped in low-level equilibria. The fractal poverty traps formulation suggests four interrelated strategic emphases for poverty reduction strategies.

Rural Poverty Dynamics: Development Policy Implications
September 2003
Barrett, Christopher B.
This paper summarizes a few key findings from a rich and growing body of research on the nature of rural poverty and, especially, the development policy implications of relatively recent findings and ongoing work. Perhaps the most fundamental lesson of recent research on rural poverty is the need to distinguish transitory from chronic poverty. The existence of widespread chronic poverty also raises the possibility of poverty traps. I discuss some of the empirical and theoretical challenges of identifying and explaining poverty traps. In policy terms, the distinction between transitory and chronic poverty implies a need to distinguish between "cargo net" and "safety net" interventions and a central role for effective targeting of interventions. Prepared for invited presentation to the 25th International Conference of Agricultural Economists, August 17, 2003, Durban, South Africa.

FORTHCOMING

A Comparative Analysis of the Determinants of HIV/AIDS Related Knowledge and Behavior
Peter J. Glick and David E. Sahn
In this paper we explore the determinants of HIV/AIDS related knowledge and behavior in four sub-Saharan African countries: Burkina Faso, Tanzania, Uganda and Zimbabwe. We use multiple point in time Demographic Health Surveys to explain both women’s knowledge of ways of preventing the spread of HIV/AIDS, as well as the utilization of condoms and nature of sex partners. In addition, we examine the determinants of testing behavior. A set of household, individual, and community explanatory variables are employed. Given that we have data from more than one survey in each country, we are able to both examine the changes in levels of knowledge and behavioral outcomes, and additionally, we can determine the extent to which they are driven by changes in the role of various covariates, for example, education or geographic location.

GHANA:

FORTHCOMING

The following papers to be published in a volume “Understanding Poverty in Ghana,” edited by Ernest Aryeetey and Ravi Kanbur:

Land and Rural Institutions Transformation
Ernest Aryeetey and Dzodzi Tsikata

Decentralization
Felix Asante and Joseph Ayee

Qual-Quant
Ellen Bortei-Doku Aryeetey and Ravi Kanbur

Public Expenditure Monitoring
Charles Jebuni and Anthony Tsekpo

Health Insurance
Kwadwo Asenso Okyere and K. Osei-Akoto

Education for HIV/AIDS
John Anarfi and Ernest Appiah

Household Asset Choice
Ernest Aryeetey

Skills Acquisition by the Poor
William Ahadzie and George Botchie

Inflation and the Poor
Nii Kwaku Sowa

Poverty Dynamics - Risk and Vulnerabilty
Abena Oduro and Kojo Appiah-Kubi

Non-traditional Exports and Poverty Reduction
Victor Nyanteng and Wayo Seini

Trade and Poverty
Charles Jebuni

KENYA:

COMPLETED

Decomposing Producer Price Risk: A Policy Analysis Tool With An Application to Northern Kenyan Livestock Markets
October 2002
Barrett, Christopher B., and Winnie K. Luseno
This paper introduces a simple method of price risk decomposition that determines the extent to which producer price risk is attributable to volatile inter-market margins, intraday variation, intra-week (day of week) variation, or terminal market price variability. We find that large, variable inter-market basis is the most important factor in explaining producer price risk in animals typically traded between markets. Local market conditions explain most price risk in other markets, in which traded animals rarely exit the region. Variability in terminal market prices accounts for relatively little price risk faced by pastoralists in the dry lands of northern Kenya although this is the focus of most present policy prescriptions under discussion.

Poverty Traps and Safety Nets
September 2003
Barrett, Christopher B. and John G. McPeak
This paper uses data from northern Kenya to argue that the concept of poverty traps needs to be taken seriously, and that if poverty traps indeed exist, then safety nets become all the more important. However, as presently practiced, safety nets based on food aid appear to be failing in northern Kenya.

FORTHCOMING

Enhancing Access, Accountability and Empowerment for the Poor through Decentralization and Participation: The case of agricultural extension services
Omiti, John, et al.
This study explores the extent to which decentralization of agricultural extension strengthens popular participation, local accountability and empowerment of the poor by enhancing access to agricultural extension information, as well as how this affects efficacy in serving the poor.

The Role of Rural Factor Markets in Reducing Poverty, Risks and Vulnerability in Rural Kenya
Karugia, Joseph, Willis Oluoch-Kosura, and Paswel P. Marenya
This study investigates the role of access to rural factor markets in influencing poverty, risk and vulnerability among the rural poor in western Kenya.

Effects of Market Price Volatility On Production Patterns and Apparent Retreat Into Subsistence Farming by Kenyan Smallholders
Nyangito, Hezron, et al.
This study analyzes the effects of the changing food market structure on price distributions and crop production patterns and marketing strategies among small-scale farmers in Makueni and Kakamega districts in Kenya.

A Study of Producer Organisations in the Liberalised Kenyan Economy
Nyoro, James, et al.
This study explores the factors that drives institutional change at the level of producer cooperatives, self-help societies and other grass-roots level organisations and the impact of those changes on market participation, incomes, risk exposure and general welfare of farmers in Kenya.

Papers from the March 2004 Workshop on Mixing Qualitative and Quantitative Methods of Poverty Analysis in Kenya:

Quantitative Perspective on Poverty Dynamics in Kenya
Germano Mwabu
Social Aspects of Dynamic Poverty Traps
Nelson Mango
Qualitative and Quantitative Poverty Appraisal: Maximizing Complementarities and Minimizing Tradeoffs
Christopher B. Barrett
Pathways out of Poverty in Western Kenya
Anirudh Krishna, Patti Kristjanson, Maren Radeny and Wilson Nindo
Two further papers to be finalized.

MADAGASCAR:

COMPLETED

Determinants of HIV Knowledge and Behavior in Madagascar: An analysis using DHS Data
Randriamamonjy, Josee, Peter J. Glick, David E. Sahn
Madagascar’s low prevalence of HIV/AIDS, just 0.3%, suggests that it has so far escaped the epidemic that has plagued other countries in southern African. However, Madagascar is also characterized by a set of ideal conditions for the rapid spread of HIV to the general population: limited access to health and social services, the presence of high illiteracy, low condom use, widespread poverty, and high rates of sexually transmitted infections. To estimate the determinants of HIV/AIDS knowledge and sexual behavior among adolescent women aged 15 to 49 years old in Madagascar, we use data from the Malagasy Demographic and Health Surveys conducted in 1997 and the Commune Census conducted in 2001.

The Complex Dynamics of Smallholder Technology Adoption: The Case of SRI in Madagascar
Moser, Christine M. and Christopher B. Barrett
This paper explores the dynamics of smallholder technology adoption, with particular reference to a high-yielding, low-external input rice production method called the System of Rice Intensification (SRI) in Madagascar. We present a simple model of technology adoption by farm households in an environment of incomplete financial and land markets. We then use a probit model and a symmetrically trimmed least squares estimation of a dynamic Tobit model to analyze the decisions to adopt, expand and disadopt the method. We find that seasonal liquidity constraints discourage adoption by poorer farmers. Learning effects – both from extension agents and from other farmers – exert significant influence over adoption decisions.

Better Technology, Better Plots or Better Farmers? Identifying Changes In Productivity And Risk Among Malagasy Rice Farmers
Barrett, Christopher B., Christine M. Moser, Joeli Barison and Oloro V. McHugh
It is often difficult to determine the extent to which observed output gains are due to a new technology itself, rather than to the skill of the farmer or the quality of the plot on which the new technology is tried. This attribution problem becomes especially important when technologies are not embodied in purchased inputs but result instead from changed farmer cultivation practices. We introduce a method for properly attributing observed productivity and risk changes among new production methods, farmers and plots by controlling for farmer and plot heterogeneity using differential production and yield risk functions. Results from Madagascar show that the new system of rice intensification (SRI) is indeed a superior technology. Although most observed productivity gains appear due to farmer aptitude, the technology alone generates estimated average output gains of more than 37 percent. These findings also help resolve several outstanding puzzles associated with observed low and incomplete uptake and high rates of disadoption of SRI in spite of the technology’s manifest superiority.

FORTHCOMING

Characteristics of the Malagasy health care system.
A detailed descriptive overview of the data from the several surveys we are conducting.

Impact of the crisis and the suspension of user fees on the quality, efficiency and equity of the Malagasy health sector.
This study will examine the effects of the 2002 political and economic crisis, which led effectively to transportation blockade over much of the country, on the quality of and demand for health services. At the same time, it will examine the effects of the temporary suspension of the fee policy (PFU) at the end of the crisis. The effect of user fees, especially on health care utilization by the poor, is a crucial policy question in Madagascar, especially since the government has recently decided to reinstate the fees.

Explaining Efficiency in the Malagasy health sector.
This study will provide a statistical analysis of the efficiency of service delivery by public and private health care providers. It will apply recent techniques in stochastic frontier analysis to measure quality-adjusted efficiency in different facilities. It will then use econometric methods to explain the differences in efficiency between facilities, e.g. what policy factors seem to influence efficiency the most?

Determining policies to maximize incentives for efficiency and quality in public health care.
Lack of monitoring and incentives for good performance are often cited as key factors behind poor efficiency and quality in the public health care system. Related to this is said to be the inability of managers to exercise control over resources and personnel. This study will use the detailed information on incentives facing health care practitioners, facility managers, and health district commissioners to assess the impacts on quality and efficiency of incentives and management discretion in personnel policy and resource allocation, and to suggest ways to improve the performance of the health care system.

Equity in the Malagasy health care system.
This study will assess equity consideration in health care, addressing questions such as: Do the poor have less access to health care services? Do they have less access to high quality care? Are they able to benefit from private alternatives? What are the implications of the reinstatement of the fee policy and decentralization of the health sector for the poor’s use of and quality of health care services?

SOUTH AFRICA:

FORTHCOMING

We expect to publish a volume “Poverty and Policy in Post-Apartheid South Africa,” edited by Haroon Bhorat and Ravi Kanbur, that will include the following papers:

An assessment of Distributional Data in South Africa
Murray Leibbrandt and Ingrid Woolard

Evolution of Poverty and inequality, 1995-2000
Morne Oosthuizen and Laura Poswell

Evolution of the labor market using the 1995 and 2002 LFS/OHS
Haroon Bhorat

Human Capital Accumulation and Household Poverty
Harris Selod and Yves Zenou

Inflation and the Poor
Haroon Bhorat, Ravi Kanbur and Morne Oosthuizen

Is Public Expenditure Pro-poor?
Servaas van der Berg

Trade, Openness, Employment and Poverty
Rashad Cassim

Exchange Rate Volatility and Poverty
Lawrence Edwards

The earner/non-earner composition of poor and non-poor households, with implications for the impact of minimum wages on unemployment and poverty.
Ravi Kanbur

HIV/AIDS and Poverty using the 1998 DHS.
Haroon Bhorat

Crime and poverty
Berk Ozler

Mobility in the post-1994 period
Michael Carter and Julian May

Economic Growth, the Macroeconomic Environment and Poverty
Johannes Feddercke and Farah Pirouz

Internal Migration and Household Poverty post-1994
Dori Posel and Daniela Casale

Social Security and Poverty Alleviation
Charles Meth

UGANDA:

COMPLETED

Growth and Poverty Reduction in Uganda, 1992-1999: A Multidimensional Analysis of Changes in Living Standards.
October 2003
Younger, Stephen D.
This paper examines Uganda’s progress on poverty reduction when poverty is measured in multiple dimensions. In particular, I consider poverty measures that are defined across household expenditures per capita or household assets, children’s health status, and in some cases, mother’s literacy. The comparisons are robust to the choice of poverty line, poverty measure, and sampling error. In general, I find that multidimensional poverty declined significantly in Uganda during the 1990s, although results for the latter half of the decade are more ambiguous. While there was clear progress in the dimension of expenditures and assets, improvement in children’s height-for-age z-scores is less certain for the 1995-2000 period. I also make poverty comparisons for individual regions and urban and rural areas in the country. Rather surprisingly, progress on multivariate poverty reduction is less clear in Central region and in urban areas.

Robust Multidimensional Spatial Poverty Comparisons in Uganda.
May 2003
Duclos, Jean-Yves, David E. Sahn, and Stephen D. Younger
We investigate spatial poverty comparisons in Uganda using multidimensional indicators of well-being. In contrast to earlier work, our methodology applies equally well to what can be defined as "union", "intersection," or "intermediate" approaches to dealing with multidimensional indicators of well-being. Further, unlike much of the stochastic dominance literature, we compute the sampling distributions of our poverty estimators in order to perform statistical tests of the difference in poverty measures. We apply our methods to two measures of well-being from the 1999 Uganda National Household Survey, the log of household expenditures per capita and children’s height-for-age zscores. Univariate comparisons based on expenditures alone rank the regions from richest to poorest as Central, Western, Eastern, and Northern. When comparing results for urban areas in one region with rural areas in another, however, we have a richer set of results.

FORTHCOMING

Determinants of Poverty Dynamics.
Mukunge, Ashie and Ibrahim Kasirye
This paper uses the 1992-1999 panel of households in the Integrated Household Survey (IHS) and National Household Survey (NHS) to model change in poverty status over time. A draft is circulating internally for comments, and the authors expect to release a working paper version before the end of the year.

Multidimensional Intertemporal Poverty Comparisons.
Younger, Stephen D.
This paper uses the 1992 IHS and 1999 NHS cross-sections to compare poverty over time in Uganda, where poverty is measured is multiple dimensions. In particular, the author considers household expenditures per capita, children’s nutritional status (height), and mother’s literacy. Results are less optimistic than univariate comparisons of expenditures (e.g., Appleton, 2001), with some regions and areas not showing multivariate improvement. A draft is circulating internally for comments, and the authors expect to release a working paper version before the end of the year.

Multidimensional Spatial Poverty Comparisons.
Younger, Stephen D., David E. Sahn, Jean-Yves Duclos
This paper builds on Duclos, Sahn, and Younger (2003a, 2003b) to make spatial poverty comparisons when poverty is measured in the dimensions of household expenditures per capita and children’s nutritional status (height) in Uganda and other African countries. Most regional comparisons are consistent with prior expectations based on univariate poverty comparisons based on expenditures alone. However, comparisons of rural areas in one region with urban areas in others are more nuanced, with rural areas in some regions actually appearing less poor than urban areas in others. The Uganda results are published in Duclos, Sahn, and Younger (2003b), and the authors are now adding results from other countries for this paper.

Modeling Infant Mortality over Time.
Ssewanyana, Sarah and Stephen D. Younger
This paper, like the previous two, will address the concern in Uganda that not all dimensions of well-being are improving as rapidly as incomes. It will use birth history recall data from the DHS to construct time series for infant mortality from the mid-1970s to 2000. It will then model infant mortality rates, attempting to understand how both macro and micro variables have influenced mortality rates over time. To date, the authors have a preliminary set of results. They expect to have a draft ready before the end of the year.

Modeling Behavior and HIV/AIDS.
Sahn, David E. and Peter J. Glick
This research will model a variety of behaviors that both determine and are affected by HIV/AIDS transmission in Uganda. Using DHS data, the authors will examine the impact of knowledge about HIV/AIDS and other public interventions on the probability of choosing to be tested for HIV, condom use, and sexual activity. The authors have preliminary results and expect to complete a draft in early 2004.

Tax Incidence.
Matovu, John, and Margaret Banga.
This study will examine the incidence of taxes in Uganda in 1999, updating a previous study by Chen, Matovu, and Reinikka (2001) for 1992 data. A particular concern is to look at the graduated tax, which is a main source of revenue for districts and thus key to Uganda’s decentralization plans. Both participatory assessments in Uganda have found this tax to be extremely unpopular. The authors have begun their analysis and expect to complete a draft early in 2004.

Demand for Health Care Consultations.
Ssewanyana, Sarah, and Stephen D. Younger
The 2002 round of the National Household Survey has an unusually rich set of information on respondents’ access to health care and the quality of those services. This paper will use this information to estimate the demand for public and private health care. Given that user fees were recently abolished, understanding these demands is particularly relevant for policy makers in Uganda. The authors expect to begin work early in 2004.

Public Water Supply and Women’s Time Use.
Glick, Peter J., and Stephen D. Younger
This paper uses an econometric analysis to ask whether public investments in water supply will reduce the work burden on females relative to males. It considers the implications for time allocated to the following activities: water collection itself, all domestic activities, market oriented work, and leisure. The preliminary results suggest that, in Uganda and Madagascar, such investments can have at best only limited impacts on time use and the gender distribution of work and leisure. The authors have an extensive set of results, and they plan to prepare a draft before the end of the year.

Agricultural Commercialization and Children’s Nutritional Status.
Bahiigwa, Godfrey, and Stephen D. Younger
This paper responds to a direct request and concern of the Ministry of Agriculture in Uganda. The Plan for the Modernization of Agriculture (PMA) is a central feature of Uganda’s poverty reduction strategy. The PMA, in turn, aims to promote the transformation from subsistence to commercial farming. This strategy has raised the concern, however, that commercialization may have negative consequences for children’s nutrition. While available evidence casts doubt on this concern (e.g., von Braun and Kennedy, 1994), examining the issue for Uganda-specific data will be more persuasive for Uganda’s policy-makers. Authors plan to begin work in early 2004.

WEST AFRICA

SENEGAL:

COMPLETED

The impact of family literacy on the earnings of illiterates: Evidence from Senegal.
October 2003
Sarr, Leopold
The paper investigates the extent to which the sharing of literacy knowledge within the household affects the labor force participation and the earnings of illiterate workers in Senegal. The paper uses an intra-household model of literacy to a Senegalese household dataset and provides evidence that parental literacy and education do not capture all sources of external literacy benefits and that illiterate members also benefit from other literate members of the household.

FORTHCOMING

Determinants of primary enrollment of girls and boys.
Will address the reasons for low primary enrollment in Senegal, especially of girls, by examining the role of household factors, school factors (e.g., quality and distance), and community level factors. The choice among public and increasingly important private and community school alternatives will also be considered.

Dropout and Progression through school in Senegal.
This report will explore the reasons for the very high rates of repetition and dropout in the Senegal school system. The roles of gender, exam performance, school and teacher quality and practices, and household wealth and changes in household fortunes (e.g. though illness or bad harvest years) will be considered. The report will also specifically address the factors influencing the transition (or lack of transition) from primary to lower secondary school of girls and boys, and recommend policies to encourage or enable parents to enroll and keep their children in school.

Determinants of student learning.
Little is known about the school and household factors that affect learning, as measured by test scores. This report will address this gap by combining the test score information with the wide range of information from the complementary surveys on household factors (parental education, assets, etc) and school factors (teacher skills and practices; quality of school facilities; staff monitoring and management; use of multi grade teaching; and for girls, teacher gender and attitudes and gender-related aspects of the school environment). Specific school factors will be identified that can be used to boost student learning, and especially, girls’ learning.

Schooling and acquisition of ‘life skills’.
We will assess whether schools also impart crucial non-academic knowledge and skills, for example, basic health knowledge. This will be done using tests administered to school age children in their homes, thus allowing us to compare knowledge of life skills of children who have attended school and those who have not (or who have dropped out).

Next Section

Return to SAGA Briefing Report (November 2003)