Development Economics
Labor Economics

Development Economics
Labor Economics
Applied Econometrics
Microeconomic Theory

May 1999 (Oral) Development Economics, Labor Economics
May 1998 (Written) Microeconomic and Macroeconomic Theory

Flexibility of Household Structure: Economic Motivations and Consequences of Child Fostering in Burkina Faso

Professor Christopher Udry
Professor Michael Boozer
Professor T. Paul Schultz

May 2004

Ph.D., Economics, Yale University, expected May 2004
M. Phil., Economics, Yale University, May 2000
M.A., Economics, Yale University, May 1999
B.A., Economics, University of California, Berkeley, with Highest Honors, May 1993
B.S., Political Economy of Natural Resources, U.C., Berkeley, with Highest Distinction, May 1993

Research Grants and Fellowships
     National Science Foundation Doctoral Dissertation Research Fellowship, 2000-2001
     Social Science Research Council International Dissertation Field Research Fellowship, 2000-2001
     J. William Fulbright Student Researcher Award for dissertation research in Burkina Faso, 2000-2001
     J. William Fulbright Supplemental Grant for additional dissertation research in Côte d’Ivoire, 2001
     Institute for the Study of World Politics Doctoral Dissertation Research Fellowship, 2000-2001
     David Boren National Security Education Program Graduate International Fellowship, 2000-2001
     Yale Center for International and Area Studies Dissertation Research Grant, 2000-2001
     Yale University, Economic Growth Center Travel Grant, 2000
     Yale Center for International and Area Studies Summer Pre-dissertation Grant, 1999
     Yale University Program in Agrarian Studies Summer Pre-dissertation Grant, 1999
Academic Fellowships and Honors
     Yale University, Dissertation Fellowship, Fall 2002
     Yale University, Cowles Foundation for Research in Economics Summer Fellowship, 2001, 2002
     Ryoichi Sasakawa Young Leaders Fellowship, 1999-2000
     Yale University Doctoral Fellowship, 1997-2001
     Phi Beta Kappa, 1993

Teaching Assistant, Economic Development in Africa (undergraduate), Yale University, Spring 2004
Teaching Assistant, Introductory Microeconomics (undergraduate), Yale University, Spring 2002

Unité d'Enseignement et de Recherche en Démographie, University of Ouagadougou, Burkina Faso
Visiting Researcher, Dissertation Field Research, September 2000-October 2001
     Received 6 grants totaling $90,000 to fund a quantitative household survey to understand families’
     motivations for fostering children. Hired, trained, and supervised research project team of 35 members
     (included 2 field supervisors, 5 field managers, 15 field enumerators, 12 data entry agents, and 1 data entry
     supervisor). The project team located and interviewed the sending and receiving households involved in 95%
     of the fostering exchanges for 316 foster children. The households were located in 109 different villages
     across Burkina Faso and 25 villages in Côte d’Ivoire.

U.S. Peace Corps, Togo, West Africa, Environmental Protection Specialist, June 1995-July 1997
     Conducted a participatory rural needs’ assessment in 20 villages. Organized workshops on improved
     agricultural practices for individuals and farmer cooperatives. Received a grant funding construction of a
     hand-dug well providing year-round potable water for 75 villagers. Coordinated the construction and
     supervised the technician and 30 village workers.

Resources for the Future, Washington, D.C. Research Assistant, 1994

U.C. Berkeley, Department of Agricultural and Resource Economics, Research Assistant, 1991-1993

"Risk, Network Quality, and Family Structure: Child Fostering Decisions in Burkina Faso." October 2003 (job market paper). Paper presented at:
     Bureau for Research in Economic Analysis of Development (BREAD) Conference, December 2003
     Northeast Universities Development Consortium (NEUDC) Conference, October 2003
     University of Chicago, Economic Research Center Seminar, June 2003

"School Enrollment Impacts of Child Fostering: Evidence from Burkina Faso." In progress.

"Are the ICRISAT Villages Representative of Burkina Faso? A Study of Pareto Inefficient Intra-Household Allocations." October 1999.

Discussant, Northeast Universities Development Consortium (NEUDC) Conference, October 2002, 2003

"U.S. National Security Impacts of Child Migration in Africa." Presented at National Security Education Program’s conference on U.S. National Security, Washington, D.C., September 2002.

"Conducting Fieldwork in West Africa." Presented at Fulbright training program, June 2002.

"Designing a Research Project: An Iterative Approach to Integrate Theory and Fieldwork." Presented at Social Science Research Council’s International Dissertation Field Research Workshop, February 2002.

"Results from Qualitative Interviews on Child Fostering in West Africa." Presented at Unité d'Enseignement et de Recherche en Démographie, October 2000.

Journal of Political Economy

U.S. Forest Service and National Park Service, 1990-1993 (17 months experience)
Firefighter Wyoming Hotshot Crew; Forestry Technician, Colorado; Backcountry Ranger, Alaska

French (fluent), Moba (introductory)

In sub-Saharan Africa, child fostering is a prevalent and socially accepted institution where parents send their own biological children to live with a different family or take in a child from another family. Based on household survey data I collected in rural Burkina Faso, approximately twenty-seven percent of households either sent or received a foster child between 1998 and 2000. Almost ten percent of all children aged five to fifteen inclusive were sent to live away from their biological parents during this time period, and children spent, on average, two years and nine months away from their parents. Living away from their parents might put these children at risk of reduced educational attainment or worse health outcomes, which would further impair Africa’s economic growth and limit its ability to achieve sustainable development.

Two of the goals of this dissertation are to understand why households send and receive children and then to measure the welfare impacts of that decision. The analysis is based on a unique dataset I collected during eighteen months of fieldwork in Burkina Faso. The research methodology that involved locating the sending and receiving households participating in each fostering exchange, combined with the survey instrument, make these data particularly appropriate for achieving these goals.

In the first paper, I present a theoretical framework in which the allocation of children across households in a social network is Pareto efficient. This framework provides the motivation for three principal factors influencing the household decision to foster a child. First, households use child fostering as a risk-coping mechanism in response to exogenous income shocks. Second, households with better opportunities, measured in terms of the quality of their social network, are more likely to foster. Third, in these households, children perform chores and having too many or too few children in a given gender and age class may not optimize household production. Therefore, parents are more likely to foster children to offset these demographic imbalances. The data I collected provide empirical evidence that these three factors significantly affect the household's decision to send a child, but not the decision to receive a child. Increases of one standard deviation in a household’s shock, percentage of good network members, or number of older girls, would increase the probability of sending a child by twenty-eight, twenty, and thirty-four percent, respectively, above the current level of fostering.

This paper contributes to three distinct literatures: child fostering, risk-sharing, and social networks. First, the paper builds on Ainsworth’s (1990) seminal work on child fostering by confirming her results that child labor plays a role in fostering decisions. The paper also validates two new explanations, risk-coping and network quality, for why households foster children. Another reason for fostering, often cited by sociologists and demographers working in West Africa (Goody, 1982; Isiugo-Abanihe, 1985), is educational investment, but I do not find evidence in support of this. Second, the paper extends the risk-coping literature by providing evidence for another mechanism that households adopt to deal with adverse shocks and income fluctuations. None of the existing empirical economics research has tested whether households use child fostering as a risk-coping strategy. Third, the paper extends previous research on social networks by considering their impact on a new outcome, child fostering, and suggesting two new quantitative measures of network quality.

The second paper, in progress, attempts to measure the welfare impacts of child fostering. Most international development organizations and academic research claim that living away from one’s biological parents is detrimental to a child's welfare (UNICEF, 1999; Case, Paxson, Ableidinger, 2002; Bishai and Suliman, 2003). Yet several other researchers argue there is considerable heterogeneity in the schooling and health outcomes of foster children that varies across countries, wealth classes, and reasons for the child being sent (Castle, 1995; Lloyd and Blanc, 1996; Ainsworth and Filmer, 2002). However, all of these researchers use cross sectional data, which limits them to comparisons between foster children and their non-fostered host family siblings. Their results will be biased if there is some unobservable factor omitted from the analysis that is correlated with both fostering and the welfare measure.

This paper uses two approaches to consistently estimate the effect of fostering on school enrollment. First, I utilize the panel dimension of the data to calculate a fixed effects or difference-in-differences estimator. Limiting examination to the cross section indicates that only sixteen percent of foster children are enrolled compared with thirty-one percent of host family siblings. However, incorporating enrollment information prior to fostering and controlling for host sibling enrollment shows that foster children are no worse off, compared to their prior enrollment, after they are sent to live away from their biological parents. The difference-in-differences coefficient estimate indicates that being fostered actually increases enrollment by 1.5 percentage points. Second, I use a two stage least squares strategy to instrument for the endogenous variable, fostering status, using measures of household shocks, network quality, and family structure as instruments. The choice of instruments follows from the analysis in the first paper. Both the fixed effects and instrumental variables approaches indicate that not controlling for omitted variables in measuring the welfare impacts of child fostering yields biased and incorrect results.