Given these insights, in the second part of the paper the authors move to a much more general dynamic life-cycle model within the context of an overlapping generations economy, allowing for general equilibrium. This richer quantitative model is based on that of Abbott et al (2013). The generations are linked by altruism and the initial ability of the child is drawn from a distribution that depends on parental ability. This links the generations both in terms of initial endowments and in terms of financial transfers, generating two important channels for intergenerational persistence. First children’s ability is correlated (albeit not perfectly) with that of their parents. Second, altruism will imply that parents make unconditional transfers to children at the start of their life taking into account their ability and with a full understanding of the child’s optimal behavior. These transfers are of central importance because, in the absence of perfect credit markets, they improve funding for college. Another channel for alleviating LCs is student labor supply, which can react both to changes in policy and to the amount of transfers provided by the parents. The authors also allow students to borrow a fraction (estimated from the data) of their educational costs. In contrast to Abbott et al. (2013) they do not model the full structure of the student grant and loan system, which includes an element of means testing with grants and subsidized loans for those with low-income parents. This is potentially very important because it makes it easier for higher ability but low parental wealth children to attend college, despite the absence of any substantial parental transfers. Ignoring the existence of such loans at baseline may distort the resulting optimal policies.