But this sort of analysis ignores the endogenous nonconvexities brought about by the price system. For instance, even if there are many different education levels, the wage payoff to such level will generally be determined by the market. There is good reason to argue (see, e.g., Ljungqvist (1993), Freeman (1996) and Mookherjee and Ray (2002b, 2003)) that the price system will sort individuals into different occupational choices, and that there will be persistent inequality across dynasties located at each of these occupational slots. Thus an augmented theory of history dependence might predict a particular proportion of the ultrapoor trapped by physical nonconvexities (low nutrition, ill-health, debt, lack of access to primary education), as well as a persistently unequal dispersion of dynasties across different occupational choices, induced by the pecuniary externalities of relative prices.