SAGA PROGRESS REPORT (11/05-11/06) &
UPCOMING WORKPLAN (12/06-12/07)

II.1.1 Schooling Attainment and Cognitive Ability

• Analysis of the EPSPAM data reveal that rates of current enrollment are high overall, reflecting recent policies of the government such as fee elimination and provision of free books and school supplies to families. Still, the data point to clear urban rural gaps as well as differences by level of household resources. Almost all children in the sample enter school, so the differences show up as earlier school leaving for children in rural households or poorer households. Children in the subsample of smaller and more remote rural communities do quite poorly along all these dimensions compared with other rural areas and urban areas—a consistent pattern in these data.
  
  
• Econometric analysis of the determinants of cognitive achievement of 8-10 years olds suggests that (conditional on effects on enrollment and grade level) household wealth and parental schooling have little effect on test scores in four subjects; the only exception was a positive effect of mother’s education for several subjects. Several school quality factors do matter, however.
  

• Descriptive and econometric analysis of the EBMS data indicate that schooling matters strongly for cognitive skills, even for knowledge of ‘life skills’ that are presumably imparted largely outside of the classroom.
  
  
• Conditional on a child’s level of schooling, having better educated parents or enjoying the advantages of being in a wealthier household have only modest or inconsistent (across tests) benefits for academic performance. This is the case whether or not the regressions control for the possible correlation of parental education with unmeasured factors also affecting performance.
  

• Determinants of scholastic achievement in Madagascar (test performance): This work, in progress and noted above, is a multivariate regression analysis of the determinants of children’s achievement on standard math and French tests as well as oral math and ‘life skills’. Methodologically, this study is unusual in the use of instrumental variable techniques to deal with unobserved factors associated with both parental schooling and student performance.
  
  
• Determinants of school progression and dropout in Madagascar: This study will consider the dynamics of educational attainment, focusing on the factors determining primary completion and transition to lower secondary school. The effects of both school and household characteristics, as well as economic and health shocks recorded in retrospective interviews, will be modeled.
  
  
• Effects of early cognitive abilities on school attainment in Senegal: In this paper, we consider the impact of early cognitive ability, measured by testing after first grade, on subsequent attainment and academic achievement.
  
  

II.1.2. Community Schools in Senegal (http://www.saga.cornell.edu/images/nal-senegal.pdf)

• Profiles and Learning of the Pupils
  
  
  • Gender Distribution: Among Communitiy School (CS) pupils, 63 percent are femaleswhile 60 percent among the enrolled pupils at the time of the survey were females. Considering they have less chance to progress into secondary and higher education levels, the higher proportion of the female pupils suggests that the unequal educational opportunity for girls remains an issue. This suggests a need to focus on integrating CS with the regular school system to provide opportunities for girls.
    
    
  • Acquisition of Cognitive and Practical Skills: While 67.7 percent of the pupils in the CS were illiterate when they enrolled, the study found that 93.7 percent of enrolled pupils and 92.3 percent of former CS pupils are able to read fluently French and other national languages.
    
    
• Infrastructure and Equipment
  
  
  • Infrastructure and Equipment and Physical Conditions: By and large, the community schools are in dire physical and learning conditions. While nearly all the schools (92.7 percent) are equipped with seats, up to 50 percent of the sampled schools are in temporary locations (abris provisoires). Only 45 percent have the necessary teaching/learning tools, and less than half of the pupils (48.8 percent) have textbooks.
    
    
• Teaching Staff and Management
  
  
  • Teaching Staff: The mere fact that the teachers in these schools are referred to, as “volunteers” indicates the type and magnitude of the problems. Teaching in these schools is not a professional choice, but rather “à défaut de mieux” or a stepping-stone to better work and/or educational opportunities. In addition to the lack of initial professional training, many do not benefit from proper supervision. While the majority of the CS students are females, women constitute only 36.4 percent of the teaching staff.
    
    
  • Management: Among the sampled schools, only 17.4 percent have received supportfrom the local institutions.
    
    
• Output and Outcome toward the National Goal of Universal Education, Fighting Poverty, and Sustainable Development
  
  
  • Universal Basic Education: It is worth noting that according the sampled pupils, 38.6 percent of those who leave early primary school from the regular schools are enrolled in the CS.
    
    
  • Fighting Poverty: CS alumni are often among the poorest segments of society. Mostare engaged as in agro-pastoral or agricultural activities.
    
    
  • Toward Sustainable Development: One remarkable outcome is the consciousness of responsible citizenry and commitment toward the contribution development of local knowledge as an investment in permanent source for future development.
    
    

II.2.1 Institutional Analysis and Health Delivery Systems

• The survey of health facilities reveals severe inadequacies in infrastructure: for example, only 53% had electricity, only 60% had an adequate source of water (tap or pump), and less than 38% of facilities have supplies of drugs adequate to their needs.
  
  
• The effects of the 2002 crisis on health care utilization were severe but apparently short=lived: consultations fell by about 10% but since then have rebounded strongly.
  
  
• Direct observation of health practitioners (by doctors carrying out this part of the survey) suggests that standard treatment protocols are often, even typically, not followed completely. For example, in only about one fifth of the centers did practitioners note lethargy in their patients.
  
  
• Research on improving health care quality frequently uses client satisfaction surveys conducted outside health facilities (user exit surveys). However, econometric analysis using the linked household, facility, and exit surveys in the EEEFS indicates that responses in user exit surveys are often substantially biased toward indicating higher satisfaction than is actually the case, likely reflecting the phenomenon of ‘courtesy bias’. The results thus raise doubts about the reliance on such surveys to understand consumer perceptions and health care demand.
  
  
• Several aspects of care that patients observe and respond favorably to are uncorrelated with measures of provider skills (with respect to diagnoses and treatment) obtained from direct clinical observation. This suggests the need for direct observation of facilities and practitioners to understand quality. It also suggests that providers interested in increasing consumer satisfaction, hence demand for their services, potentially would try to achieve this objective by altering their behavior in ways that do not improve actual process quality.
  

II.2.2 Infant and Under-five Mortality

II.2.3 HIV/AIDS

• Analysis of survey compatibility over time presents a mixed picture. There is evidence in several countries, of the nine used for the analysis, of differences in sampling, as indicated, for example, by differences in mean education level or adult women’s heights. These can be controlled in regression analysis. More difficult for understanding changes in behavior over time are cases where responses themselves change due, for example, to increasing concern to give socially appropriate answers about one’s sexual behaviors. There is evidence of this but only in a minority of our country/region samples.
  
  
• Overall, some reductions in risk behaviors are seen for each of the behaviors studied, for both men and women. Particularly noteworthy is that for the case of condom use among men and women with persons other than co-habitating partners.
  
  
• Among the behaviors examined, the least progress has been made in terms of increasing the share of abstinent women.
  
  
• The one group and country where negative signs predominate is for women in Nigeria. Although this may reflect in part issues of survey compatibility, the apparent trend in Africa’s largest country is still worrisome. Likewise, the most recent signals in terms of behavior change are decidedly mixed for women in Uganda. The country with the most dramatic changes in behavior, at least in percentage terms, is Mozambique.
  
  

II.2.4. Poverty and Reproductive Health

• Reproductive Health, HIV/AIDS and Poverty
  This involves the analysis of the interactions among the following: the supply of reproductive health services; household demand for reproductive health services; HIV incidence and HIV-related risk behaviors and knowledge; and poverty. With respect to reproductive health services, given the HIV focus of this framework paper, special attention will be given to aspects of these serviced that are closely related to HIV risk, such as contraceptive provision, HIV education and counseling, and HIV testing. (See http://www.saga.cornell.edu/images/wp219.pdf)
  
  
• Poverty, Maternal Morbidity and Mortality in Africa
  This work explores the impact of poverty on morbidity and mortality, taking a health production function approach. We would also look at issues of the deleterious consequences of negative shocks in the form of a mother’s illness or death on the household and individual family members. In this regard, we would examine the mechanisms through which these shocks contribute to households falling into poverty and the related deleterious consequences such as children leaving school or suffering health consequences as a result of the loss of their primary care giver. (See Working Paper 213)
  
  
• Labor Market Opportunities and Fertility This aspect of our work focuses on a key aspect of the demographic transition: women’s labor market activities or, more generally, women’s time allocation decisions. Women in poor countries tend to work at home, in agriculture, or in informal self-employment activities. These characteristics are related to the other features of the demographic transition in ways that are important to our thinking about population dynamics, reproductive health, and poverty in Africa. (See http://www.saga.cornell.edu/images/wp218.pdf)
  
  

II.3.1 Q-Squared

• Economists’ analysis of poverty is almost always quantitative, based on survey data, while anthropologists and sociologists are more likely to rely on qualitative poverty assessments. That these different methods often arrive at different conclusions about poverty changes is disturbing, and has begun to attract the attention of scholars in many social sciences. SAGA researchers have been at the forefront of efforts to bring together researchers from different disciplines to reconcile the apparent contradictions of quantitative and qualitative approaches to poverty analysis. Activities in this area during the last year consisted of editing of the World Development Special Issue. It is due to be published in mid-2007. See Appendix I for a Table of Contents for this issue, showing the article titles and authors who are included.
  

II.3.2 Access to Social Services

• There is evidence, though less universal than is commonly assumed, that the demand for girls’ schooling is more responsive than boys’ to gender neutral changes in school cost or distance as well as quality. Among these policies, increasing the physical accessibility of schools emerges as a measure that is likely to result in disproportionate enrollment gains for girls.
  
  
• Where gender gaps are large or persistent, however, direct targeting of girls is probably necessary. Formal evidence from a number of demand or supply side interventions, including subsidies to households and to schools to enroll girls and the provision of girlsonly schools, suggests the potential for targeted measures to yield substantial gains for girls.
  
  
• Many other policies, such as subsidized childcare or flexible school scheduling that address the opportunity costs of girls’ time, hold promise but for the most part have yet to be subjected to rigorous assessment.
  

II.3.3 Land Tenure

II.3.4 Political Liberalization, Decentralization and the Social Economics of Development

II.4. Risk, Vulnerability and Poverty Dynamics

• We completed editing the Analytical Perspectives on the Economy of Ghana volume, and it will be published by James Currey in mid-2007.
  
  
• The study of risk management, insurance, and social networks in Ghana examines individual-level risk sharing networks and tests for risk pooling. We define and identify the socially invisible as those with few or no relationships with other members of the village. We then use a general equilibrium framework to test for full insurance using individual-level panel data. Preliminary results show that younger farmers and recent migrants are most likely to be socially invisible. In addition they fail to achieve even partial insurance. On the other hand, visible individuals achieve full risk pooling with other visible individuals both within the village and at the network level. Because a village is the aggregate of the socially visible and invisible, the hypothesis of full risk pooling is rejected at the village level. (Vanderpuye-Orgle, J. and Barrett, C. (2006). Risk Management, Insurance and Social Networks in Ghana. Mimeo, Cornell University.)
  
  
• We also find that efficient allocation of idiosyncratic risk among members of social networks suggests that given binding financial constraints, policy should distinguish between idiosyncratic versus covariate risk: policy targeting for idiosyncratic shocks could focus on the socially invisible whereas interventions ex poste of covariate shocks could target everyone. Given the evidence that the socially invisible are farmers, contract interlinkages may be used as a means of quasi- insurance: interlinking credit and product sales or input delivery to provide alternate means of insuring this subpopulation.
  

• Using a multi-market model for Madagascar that focuses on income-generating activities in an agricultural sector that is characterized by seasonal variability, we find evidence that investments in rural infrastructure and commercial food storage have both direct and indirect benefits on poor households. Stifel, D., J.C. Randrianarisoa (2006) “Agricultural policy in Madagascar: A seasonal multi-market model,” Journal of Policy Modeling, December 2006 (28): 1023–1027.
  
  
• We found that poor infrastructure and low performance institutions restrain the convergence toward a stable high equilibrium of the rice yields in Madagascar. Increasing technical efficiency through better education, availability of basic health services and access to market, only pay-offs under good production environmental conditions. Randrianarisoa J. C., C. B. Barrett (2006) “Structural Rice Yield Dynamics and Technical Efficiency in the Highland of Madagascar,” mimeo.
  
  
• We explored the constraints on agricultural productivity and priorities in boosting productivity in rice. We pay particular attention on exploring why chemical fertilizer uptake rates appear so low. Fertilizer use on rice appears only marginally profitable and highly variable across years. Our willingness-to-pay estimates suggest that fertilizer demand is highly price sensitive, suggesting that low fertilizer uptake in rural Madagascar largely reflects prices beyond the reach of most farmers, especially poorer ones. Minten B., J. C. Randrianarisoa, C. B. Barrett (2006). “Productivity in Malagasy rice systems: wealth differentiated constraints and priorities,” Invited panel paper prepared for presentation at the International Association of Agricultural Economists Conference, Gold Coast, Australia, August 12-18, 2006. (See Productivity in Malagasy Rice Systems: Wealth-differentiated Constraints and Priorities ).
  
  
• Comparative qualitative and quantitative evidence from rural Kenya and Madagascar are examined in an attempt to untangle the causality behind persistent poverty. Our results suggest the existence of multiple dynamic asset and structural income equilibria, consistent with the poverty traps hypothesis. Christopher B., P. Marenya, G. McPeak, B. Minten, M. Murithi, O. Kosura, F. Place, J.C. Randrianarisoa, J. Rasambainarivo and J. Wangila, (2006) “Welfare Dynamics in Rural Kenya and Madagascar,” Journal of Development Studies, vol. 42, no. 2 (February 2006): pp. 248 - 277.
  

• Asset dynamics in southern Ethiopia exhibit patterns consistent with the notion of a poverty trap. Moreover, Boran pastoralists appear to recognize this in their stated subjective expectations of rainfall-conditional herd growth. Unpacking the overall dynamics, one finds that two factors account for the apparent existence of poverty traps: (i) adverse rainfall events – drought – that causes severe herd loss, and (ii) lower herding ability among a subpopulation of herders. These dynamics have strong implications for the design of herd restocking programs and also point to important holes in social safety nets within the Boran community, such that the likelihood of external transfers to poor households crowding out private transfers appears very low.
  (http://www.saga.cornell.edu/images/wp201.pdf and on http://www.saga.cornell.edu/images/wp206.pdf)
  
  
• Data being collected routinely in northern Kenya by Kenya’s Office of the President’s Arid Lands Resources Management Program (ALRMP) permit reasonably accurate forecasting of community-level child nutritional status at leads of 3-6 months. This suggests the possibility of both improved early warning for food crises and of using weather-based index insurance or catastrophe bonds to provide timely financing of drought interventions. (http://www.saga.cornell.edu/images/wp203.pdf).
  
  
• Panel data on pastoralists’ risk perceptions collected in northern Kenya and southern Ethiopia indicate considerable heterogeneity in subjects’ assessment of the near-term risks they face. Individual risk perceptions respond robustly to shocks experienced by others in the community, indicating active information flow and learning. Most variation in risk perceptions is between rather than within communities or within households, suggesting that geographic targeting is likely to prove relatively effective in addressing the proximate risks faced by vulnerable households in the arid and semi-arid lands of east Africa.
  
  
• A broader review of pastoral development strategies in east Africa and a synthesis of lessons learned from this body of work points to several key foci for reducing pastoralist poverty and vulnerability: firmer recognition of pastoralists’ land rights so as to maintain mobility, improved livestock productivity through genetic and nutrition improvements, improved marketing infrastructure, redesigned post-drought restocking programs, improved access to health and education services, more responsive safety net interventions, and political empowerment of pastoral communities.
  

Previous Section | Next Section

Return to 2006 SAGA Annual Report and Workplan for 2007