Individuals who under-invest in education experience negative outcomes along most measurable dimensions. Although such under-investment is common, it is not adequately explained by existing economic theory. We disaggregate the canonical educational investment decision into a series of incremental educational opportunities, and thereby endogenously separate economic agents into high- and lowparticipation equilibria. We derive self-productivity in cognitive ability development, and we identify the effects of specific noncognitive skills. Our results suggest that early intervention should focus on children’s noncognitive skills, whilst later intervention should not target disadvantaged individuals directly — it should focus instead on specified aspects of their educational provision.