ABSTRACT

heories of intergenerational mobility have set out to explain the well documented finding that a child's income is an increasing function of that of his parents. One simple mechanism through which this pattern may occur resides in parental ability to finance their children's human capital investments (BECKER and TOMES, 1986). On its own, however, this economic mechanism cannot confirm, or rule out, the possibility that in the long run incomes of individuals may be converging. Recent theories of distribution have highlighted some institutional mechanisms which may prevent convergence of incomes to occur in the long run. For instance, an imperfect credit market for human capital loans may reproduce inequalities in the long run (GALOR and ZEIRA, 1993). Alternatively, if education is financed through community based taxes, richer families may wish to be segregated from poor families in order to prevent the latter from paying little (if taxation is progressive) in return for enjoying high quality education for their children (DURLAUF, 1996). These are only two amongst the many possible explanations provided in this emerging literature surveyed in DURLAUF (1994), ATKINSON (1997), and PIKETTY (1998). While at the present time we do not possess data that cover sufficiently long time horizons in order to test whether in the long run life-cycle incomes of individuals are converging1, it is possible however to exploit limited information on the incomes of parents and children in order to test some predictions which follow from such theories. One proposition which emerges from the recent theories of distribution is that a child's expected utility is an increasing function of his/her parents' income. If this were indeed the case, then we would expect to observe that children raised in richer families are on aver-