Source: Mariacristina De Nardi, National Bureau of Economic Research (NBER), NBER Working Paper No. w21106, April 2015
From the abstract:
In the data, wealth is very unequally distributed, even more so than labor earnings and income, and the saving rate of wealthy people is high. Many dynamic models used for quantitative policy evaluation imply that once households get rich, they dissave. As a result, these models generate too little wealth concentration in the hands of the wealthiest compared with the observed data. This raises the question of the robustness of the policy lessons that we learn from environments in which key aspects of saving behavior in the model are not consistent with those in the observed data. Mechanisms that raise the saving rate of richer people, and thus generate more realistic saving behavior and wealth concentration in dynamic quantitative models, have been proposed. These mechanisms include heterogeneity in patience, transmission of human capital and voluntary bequests across generations, entrepreneurship or high returns to capital coupled with borrowing constraints, and high earnings risk for the top earners. More work is needed to evaluate these explanations both individually and jointly and to quantitatively assess their importance. Additionally, more work to explore alternative or complementary mechanisms is warranted.