The role of aspirations and identities in decisions to invest in children’s schooling

I demonstrate that household investments in children's education in Ethiopia are affected by parents' self-beliefs (such as their locus of control), parents' aspirations for children's educational attainment, children's conceptions of their roles and identities in the household and at school, and children’s own preferences, all concepts not widely studied in development economics. Two empirical chapters report on a field experiment in which randomly selected adults watched documentaries about role models who were poor but succeeded in agriculture or small business. Six months later, parents' self-beliefs and aspirations for children's education were higher in the treatment than in the placebo and control groups. Enrolment of children in school, spending on education, saving and use of credit also increased. A third empirical chapter draws on longitudinal qualitative research to argue that children's preferences for their time allocation between work and school are strongly influenced by the desire to comply with valued identities as students and as independent earners and contributors to the household. The fourth chapter suggests that understanding children's preferences might improve predictions about their reaction to education policies. The literature predicts an increase in time in school will not improve test scores: children will reduce effort because they desire a limited amount of learning. I find a reform to lengthen the Ethiopian primary school day improves test scores. Although this could occur through many mechanisms, one possibility is that children do not prefer to limit their desired amount of learning. This suggests that better evidence on children's preferences might improve prediction of the effects of policies to alter school inputs. The conclusion reflects on whether the empirical relevance of concepts of self-beliefs, aspirations and identities implies that assumptions in standard models of decision-making in economics about the characteristics of beliefs and preferences ought to be rejected. I argue that these ideas can be captured by existing economic concepts of beliefs and preferences and by standard assumptions about these concepts. I suggest that, contrary to recent accounts building on human capital theory, self-beliefs should be viewed as beliefs, not non-cognitive skills. I consider aspirations as a type of preference, shaped by both objective constraints and self-beliefs. I consider identity as a preference for complying with a social role, but highlight that such preferences are often altruistic, rather than self-interested. In conclusion, I argue that economics should draw further on other social sciences, including psychology, to develop substantive theories of the formation and characteristics of beliefs and preferences. Doing so will suggest when it is appropriate to apply standard models and how their assumptions can be modified if their predictions do not hold.