Household Separation and Child Well-Being

The purpose of this Chapter is to help design a research agenda on risk and child well-being that is suitable for the Young Lives project. A large literature exists that documents the effect of various shocks on the material welfare of children. This literature has documented the negative effects that bad rainfall and other low income shocks has on child schooling (e.g. Jacoby and Skoufias 1997, Sawada 1997, Case, Fertig and Paxson 2005), health (e.g. Banerjee, Duflo, Postel-Vinay and Watts 2007, Maccini and Yang 2009), and nutrition (e.g. Glewwe, Jacoby and King 2001, Hoddinott and Kinsey 2001, Alderman, Hoddinott and Kinsey 2006, Leight 2008, Porter 2008). The death of parents often leads children to be placed in the foster care of relatives. While child fostering often increases the child's material welfare relative to no fostering, orphans and abandoned children often fare less well than the biological children of the foster family (e.g. Evans 2004, Akresh 2004a). If blood relatives take better care of foster children, having a large family is a form of insurance against orphanhood (e.g. Akresh 2004b, Ksoll 2007). The literature has also delved into the issue of sibling rivalry, particularly between boys and girls. Rose (1999), for instance, shows that negative shocks to the household affect girl mortality more than that of boys. What the economic literature in developing countries has failed to do till now is investigate the effect of the dissolution of the household on the material and psychological welfare of children. I suspect the main reason for this is the lack of suitable data. The Young Lives project can help fill this lacuna because it proposes to follow a panel of children over a long period of time. This stands in contrast with conventional panel datasets which follow households instead. When the composition of a household changes, a conventional panel survey typically looks for a household that occupies the same farm or dwelling and has members in common with the original household. For farming households, this often means following the farm — and hence following whoever is entitled to continue farming the land upon the dissolution — or recomposition — of the original household. Given that this person is nearly always an adult — often male — conventional panel surveys are not appropriate to investigate how the well-being of children is affected by shocks that force them to leave the household — such as the separation of the parents.1 How successful Young Lives (hereafter YL) can be in fulfilling this research objective depends on how well it keeps track of children over time. One particularly vulnerable group, runaway children, are notoriously difficult to follow so I expect YL not to be particularly successful at locating children who run away from home. This is of course unfortunate because this group is probably the most vulnerable, even if running away from home may be preferable to enduring abuse at home.2 We can hope that YL will find it easier to follow children who leave the original household/dwelling as the result of the separation or death of their parents. Separation is of particular interest given the paucity of available data. It has long been documented that female-headed households have lower per capital consumption than male-headed households (e.g. Handa 1996, De Graff and Bilsborrow 1993.). From this evidence one may worry that children raised in female-headed households may be worse off. The truth of the matter is, however, that female and male headed households are not directly comparable: while the majority of single-parent households headed by females are the result of separation from — or death of — a spouse, many male-headed households with children have both parents living together. It is well known that living together generates important material gains (e.g. Fafchamps and Quisumbing 2007, Browning et al. 2009). Using data from a developed country, Browning, Chiappori, Lewbel and Lewbel (2003) compare the material welfare that a couple can achieve on the same income by living together or separately. They show that, compared to living together, living separately generates a loss of material welfare equivalent to a 41% fall in income. Browning et al. (2009) argue this is probably a lower bound estimate. Given this, it is not surprising that female-headed households have lower material welfare. But to the extent that child custody often befalls to the mother, these observations also suggest that the material welfare of children is likely to deteriorate when their parents separate. Whether this is indeed the case remains to be shown rigorously. This is what I propose for the Young Lives project to do. This Chapter is organized as follows. In Section 2 I briefly summarize the literature on household formation, discussing the various material gains generated by cohabitation. This gives a idea of the magnitude of the potential losses resulting from separation, and it illustrates how these losses vary depending on the economic context. Next I introduced marriage market considerations as they provide a framework within which marriage and its converse, separation, can be modeled and analyzed. In particular, this framework makes testable predictions regarding which unions are most likely to be formed and which are most likely to fail. In Section 4 I examine the reasons why human beings elicit to have children, drawing from economics and evolutionary psychology. Combining these two stands of literature into a single framework generates testable predictions regarding which children are most likely to be neglected or abandoned. Section 5 discusses the possible fate of children when their parents separate and the role of legal and social norms aimed directly or indirectly at protecting children. Section 6 offers a brief discussion of the difficulty of empirically identifying the causal effect of separation on child welfare.