Tom Bundervoet

Health and Civil War in Rural Burundi

We combine household survey data with event data on the timing and location of armed conflicts to examine the impact of Burundi’s civil war on children’s health status. The identification strategy exploits exogenous variation in the war’s timing across provinces and the exposure of children’s birth cohorts to the fighting. After controlling for province of residence, birth cohort, individual and household characteristics, and province-specific time trends, we find an additional month of war exposure decreases children’s height for age z-scores by 0.047 standard deviations compared to nonexposed children. The effect is robust to specifications exploiting alternative sources of exogenous variation.

Civil War, Crop Failure, and Child Stunting in Rwanda

We combine Rwandan household survey data with event data on the timing and location of localized crop failure and armed conflict to examine the impact of these distinct shocks on children’s health status. The identification strategy exploits exogenous variation in the shocks’ geographic extent and the exposure of children’s birth cohorts to the shock. We find that in poor and nonpoor households, boys and girls born during the conflict in regions experiencing fighting are negatively affected, with height-for-age z-scores 1.05 standard deviations lower. Conversely, only girls are negatively affected by crop failure, with these girls exhibiting 0.86 standard deviations lower height-for-age z-scores, and the impact is worse for girls in poor households. Results are robust to using alternative shock exposure measures, different geographic boundary definitions for the affected regions, and household-level production and rainfall shocks as alternative measures of crop failure.

Livestock, Land and Political Power: The 1993 Killings in Burundi

This article examines the characteristics of the victims of the October 1993 massacres in Burundi. Using information on parents of the respondents of a 2002 demographic household survey, the author finds that the probability of a parent being killed during the events is significantly affected by age, sex and wealth, in the sense that older, wealthier and male persons were more likely to be killed. Using the median level of education of a parents offspring to proxy the parental investment in human capital, the author finds that people who invested more in the human capital of their children ran a higher risk of being killed. The analysis also shows important spatial disparities in the killings. In trying to explain these locational effects, the author focus on two key hypotheses set forth with respect to the October 1993 events in Burundi: the land crisis and the questionable role played by the Front Démocratique du Burundi (FRODEBU), the dominant political actor at that time. The author finds that communal land pressure significantly increases the probability of being killed and that communal popular support for FRODEBU increases, in a non-linear fashion, the risk of being affected by the killings. The results are interpreted in light of the prevailing political economy of 1993 Burundi.

What explains Rwanda's drop in fertility between 2005 and 2010 ?

Following a decade-and-a-half stall, fertility in Rwanda dropped sharply between 2005 and 2010. Using a hierarchical age-period-cohort model, this paper finds that the drop in fertility is largely driven by cohort effects, with younger cohorts having substantially fewer children than older cohorts observed at the same age. An Oaxaca-Blinder decomposition is applied on two successive rounds of the Demographic and Health Survey. The findings show that improved female education levels account for the largest part of the fertility decline, with improving household living standards and the progressive move toward non-agricultural employment being important secondary drivers. The drop in fertility has been particularly salient for the younger cohorts, for whom the fertility decline can be fully explained by changes in underlying determinants, most notably the large increase in educational attainment between 2005 and 2010.

The Impact of Violent Conflict on Child Health: What are the Channels?

Child health during and after violent conflicts has been a priority for both policymakers and academics, as ill-health in early life can be impossible to make up for in later life, and has important effects on education and adult wages. In order for policy interventions to mitigate health impacts, it is essential to understand the channels through which conflict impacts on child health. This briefing uses empirical results of research in Burundi and Rwanda to identify these channels. It outlines the policy implications of these findings, arguing that policymakers should prioritise interventions to rebuild agricultural capacity and improve sanitation conditions and nutrition in displacement camps.

Exploring Psychological Distress in Burundi During and After the Armed Conflict

We assessed symptoms of psychological distress among a population-based sample of 9000-plus adults in Burundi during (1998) and after (2007) armed conflict. After exploratory and confirmatory factor analysis to an 8-item, self-report measure, we identified two domains of psychological distress “Depression/Anxiety” and “Functioning” with good fit to data. The questionnaire was invariant in males and females. Depression and Anxiety symptoms during conflict were more frequently reported than Functioning symptoms; all symptoms were more frequently reported by women. Psychological distress was found in 44xa0% of individuals during conflict and in 29xa0% 2xa0years after the conflict. Results call for further research in Burundi that can inform the development of mental health interventions.

The educational impact of shocks in utero: Evidence from Rwanda

&NA; Research on the impact of violence and conflict on education typically focuses on exposure among a cohort of school‐aged children. In line with the fetal origins hypothesis, this paper studies the long‐run effect of exposure to adverse maternal health shocks while still in the womb. Exploiting the sudden and discrete nature of the Rwandan genocide and an identification strategy based on temporal and spatial variation, we find that the cohort in utero during the genocide reported on average 0.3 fewer years of schooling in the 2012 Rwanda. Population and Housing Census and was 8% points less likely to finish primary school relative to the cohort in utero just a couple of months later. HighlightsChildren who were in utero during the Rwanda 1994 genocide have lower educational attainment, eighteen years later, than children who were not exposed to violence while in the womb.Pre‐natal genocide exposure did not affect enrolment in school, but did affect the educational performances of those affected. Children who were exposed to genocide while in utero had completed on average 0.3 fewer years of education and were 8% points less likely to finish primary school, relative to children born a couple of months later.The effects of in‐utero exposure to genocide were most apparent in the first trimester of gestation, and to a lesser extent in the second trimester.The adverse education impact of genocide, or indeed of any other severe episode of violence, goes beyond the direct impact on school‐aged children. Individuals exposed before they were even born are affected too, and a role of public policy in post‐conflict settings should be to remediate the potential adverse impact of prenatal exposure as much as possible.

Bright Lights, Big Cities: Measuring National and Subnational Economic Growth in Africa from Outer Space, with an Application to Kenya and Rwanda

This paper uses the night lights (satellite imagery from outer space) approach to estimate growth in and levels of subnational 2013 gross domestic product for 47 counties in Kenya and 30 districts in Rwanda. Estimating subnational gross domestic product is consequential for three reasons. First, there is strong policy interest in how growth can occur in different parts of countries, so that communities can share in national prosperity and not get left behind. Second, subnational entities want to understand how they stack up against their neighbors and competitors, and how much they contribute to national gross domestic product. Third, such information could help private investors to assess where to undertake investments. Using night lights has the advantage of seeing a new and more accurate estimation of informal activity, and being independent of official data. However, the approach may underestimate economic activity in sectors that are largely unlit notably agriculture. For Kenya, the results of the analysis affirm that Nairobi County is the largest contributor to national gross domestic product. However, at 13 percent, this contribution is lower than commonly thought. For Rwanda, the three districts of Kigali account for 40 percent of national gross domestic product, underscoring the lower scale of economic activity in the rest of the country. To get a composite picture of subnational economic activity, especially in the context of rapidly improving official statistics in Kenya and Rwanda, it is important to estimate subnational gross domestic product using standard approaches (production, expenditure, income).