The Impact of Own Children on Retirement Portfolio Composition in the United States

We estimate models of retirement portfolio compositions using Wave 1 of the Health and Retirement Survey. The sample consists largely of pre-retirement individuals in their early 50s. Treating number of own children as endogenous, we find little firm evidence that either levels of specific asset holdings or their shares in a portfolio of assets are affected by number of own children. The exception is housing wealth, which also carries a significant consumption component. While some parents may well come to rely on their children for retirement support, we cannot support the notion that it typically is intentional.