Understanding heterogeneity in the effects of parental separation on educational attainment in Britain : do children from lower educational backgrounds have less to lose?

Abstract

We use the British Cohort Study 1970 to show that the proportion of children achieving a tertiary education degree is 8 percentage points lower for the offspring of separated parents than for children from intact families. Moreover, the children of highly educated parents experience a two times larger 'separation penalty' than the children of less educated parents. We find a similar pattern of heterogeneity in effects for the likelihood of participation in academic education (A-Levels) beyond school leaving age but not for school grades at age 16. We test three different explanations for heterogeneity in the parental separation penalty: changes in family relations, changes in income, and negative selection into separation based on unobserved characteristics. We address the potential endogeneity of parental separation by including pre-separation observable characteristics, individual fixed effects models, and a placebo test. Our key finding is that changes in family income, but not those in family relations or selection, explain a large part of heterogeneity in the effects of parental separation. Children with more highly educated parents face a larger decline in family income if parents separate and, in addition, declines in family income of equal amounts entail more negative consequences for their educational attainment.