The number of individuals displaced worldwide from their homes is at a level not seen since World War II. In Afghanistan, decades of conflict have resulted in one of the worst cases of displacement, which is likely to persist into the future. With this in mind, this chapter investigated the long-term consequences of conflict and displacement in Afghanistan by inquiring how the younger generations within displaced households may fare in their particular circumstances. More specifically, the study compared the differences in child-related outcomes between households that have experienced internal displacement and those that have not. As a way to infer the future livelihood prospects of the child, these outcomes relate to human capital formation, concentrating on both education and nutrition.
The analysis finds strong evidence that displacement leads to inferior nutritional outcomes, with displaced households 17 % less likely to have eaten meat in the week prior to the survey. For those households that had eaten meat, it was consumed 30 % less in comparison to nondisplaced households. As for education, there is an indication that displacement has resulted in worse school attendance. These last estimates, however, are not robust to the inclusion of location-based fixed effects, suggesting that the dynamics within the communities of displacement are more likely to influence educational outcomes regardless of whether the household is displaced. This is likely a result of the lack of local services (e.g., schools) in communities throughout Afghanistan, but especially in rural areas where many of the displaced are located. Indeed, the NRVA report for 2011-2012 makes note of the low-absorption capacity of the educational system overall, while also highlighting general impediments to school attendance such as insecurity and distance to schools (CSO 2014). Otherwise, it may be that areas commonly receiving the displaced are generally poorer, which results in a substitution of schooling for incomegenerating activities not only for children of displaced households but for those of nondisplaced households as well.
An evident limitation of this analysis is the inability to fully account for endogeneity—namely, selection bias of displacement. Efforts are made to minimize any potential bias by considering only those households involuntarily displaced because of conflict, insecurity, persecution, or natural disaster—arguably exogenous shocks—and who chose their destination location without consideration for the outcomes in question (e.g., services including education, food, and health). Nonetheless, it is still possible that some households may be systematically more exposed to violence, and therefore displacement, given inherent characteristics leading to inconsistent estimates. With this in mind, the estimates should still be regarded as lower bounds for two reasons that make it probable displaced households are better off a priori relative to nondisplaced ones: local violence, if targeted, is likely targeted towards wealthier members of the community; and the migration journey itself, even internally, is costly. As such, these estimates, if in fact imprecise, should in principle be underestimating any true effect that shows a negative difference between displaced and nondisplaced households, making them still fundamentally meaningful.
Another clear drawback is the fact that the analysis is done entirely at the household level. Although certain questions from the original survey focused on individuals within the household (e.g., adult respondent’s level of education), a level of insight is lost by not having all information at the individual level. For instance, the results for food insecurity pertain to the whole households. Nevertheless, it is not difficult to imagine a case where intrahousehold differences in food consumption exist, especially between children and adults or even by gender. Even though cost limitations frequently hamper more detailed data collection at the individual level, future surveys in this still understudied context are much needed.