Assessing the “orphan effect” in determining development outcomes for children in 11 eastern and southern African countries

There are more than 45 million orphaned children in sub-Saharan Africa, 11.4 million because of AIDS – representing approximately 80% of all such orphans worldwide. Programming for orphans is therefore a major policy issue. This article uses the latest publicly available Demographic Health Survey data from 11 eastern and southern African countries to provide a snapshot of the distribution and demographic characteristics of orphans, and to test whether orphans have worse outcomes along a series of socioeconomic outcomes. While the analysis is primarily descriptive, multivariate analysis is undertaken to determine whether orphan/non-orphan differences persist after controlling for age, region of residence and household wealth. Results show, among other things, that while orphan prevalence is higher in urban areas, the majority of orphans live in rural areas. Furthermore, while urban orphans appear to be more vulnerable than their urban peers the most vulnerable children in the region are, in fact, living in rural areas. Multivariate regression analysis also revealed that after controlling for other cofounders, orphan status per se is not the most important negative determinant of children's wellbeing. Gender and region of residence are much more important predictors of poor schooling outcomes, and for all outcomes household wealth is the single most important correlate of better outcomes, with the threshold falling typically between the second and third quintiles. While orphanhood is clearly one important dimension of child vulnerability, other factors are not only important but in some cases have a much stronger quantitative association with child development (e.g. household poverty). Overall social policy as well as the targeting of specific interventions should recognize this fact, and approach child development in a holistic and integrated manner; for example combining specific orphan services such as psychosocial support within broader poverty alleviation and social protection programming such as social cash transfers, school bursaries and health service fee waivers.

Keywords:

• orphans
• non-orphans
• eastern and southern Africa
• developmental outcomes
• children

Notes

1. Population-level studies typically use Demographic Health Surveys (DHS) or National Household Budget Surveys. Examples of these include Case, Paxson, and Ableidinger (2004), Stewart (2008) and Palermo & Peterman (2008).

2. In all countries orphan prevalence increases monotonically with age, with extremely low prevalence rates at young ages (under age 5). For this reason, we collapse our age analysis into the two groups shown in Table 3.

3. In Kenya and Uganda approximately 7% of births per year are from HIV positive mothers. Antiretroviral therapy (ARV) among HIV exposed babies is around 5% and 7% in Kenya and Uganda, respectively (UNICEF ESARO estimates, 2007).

4. With regard to frequency of illness among young children, only Rwanda collected this data in its DHS. There, more orphans reported being sick in the last two weeks among the under-five-year age group than non-orphans (33% vs. 26%).

5. According to the Education For All (EFA) Global Monitoring Report (2007), upper secondary school enrolment is 22% and 115.5% in Lesotho and Malawi, respectively.

6. The calculations shown here will differ from those reported in the respective country DHS reports because those reports use different subsets of questions, and sometimes report responses for individual questions. Our purpose here is to define a standard composite response so that we can make meaningful comparisons across countries.