Life-Cycle Portfolio Choice: The Role of Heterogeneity and Under-diversification

The empirical work on household portfolio choice documents two facts. One, the stock market participation rate is low and hump-shaped over the life-cycle, two, the conditional share of stocks is also low but does not appear to change much during the life-cycle. In contrast the standard life-cycle portfolio choice model predicts high and monotonically increasing participation rates and conditional stock shares that are low and exhibit dramatic changes with age. In this paper I consider a number of extensions to this basic framework. I find that a small per period participation cost is needed to generate a hump shaped life-cycle profile of participation rates. Under a realistic calibration the quantitative effect is minor. Progressive social security and the assumption that the risk of stock portfolios is declining in household wealth as a consequence of better diversification opportunities 'an assumption that has some empirical support' though provide substantial amplification and significantly improve the ability of the model to match the data. Under-diversification also reduces the average portfolio share of stocks conditional on participation and, together with the intergenerational transmission of wealth makes it insensitive to age, consistent with the empirical evidence.