Topics in the Economics of Aging.

• Author: Gitter, Robert J.

This book is part of an ongoing study on the economics of aging conducted by the National Bureau of Economic Research. The volume, the third in the series, contains a series of papers that vary widely in topics covered as well as the models and methods used.

Robin Lumsdaine, James Stock and David Wise examine three models of retirement to see if increased computational complexity will add predictive validity. The models use data from a Fortune 500 firm to examine the effects of current wages and future retirement benefits on the decision to retire. The most computationally complex model, dynamic programming, predicts better than the most simple probit model, but no better than the model of intermediate complexity, the option value model. As pointed out by Sylvester Schieber in his discussion of the paper, the models are not of retirement, but rather of voluntary exit from the firm, i.e., the employee may go on to work for another firm and not actually retire from the labor force. Hence the model has limited use in predicting retirement, per se. Further, from the firm's perspective, they would probably wish to know which employees will leave, not just the total number of exits.

The paper that should produce the most discussion is by Thomas MaCurdy and John Shoven. They show that over periods of 25 years or longer, common stocks have outperformed bonds. Although the result has been shown elsewhere, the authors present data to show that currently fewer than five percent of faculty allocate all of their retirement funds to CREF, the common stock fund. Further, less than one in ten faculty members allocate at least 75 percent of their retirement funds to CREF rather than the alternative bond fund, TIAA. These rates do not change even when controlling for age. There is a follow-up study for someone to determine if faculty are very highly risk-averse or just lack information.(1)

Axel Borsch-Supan, Vassilis Hajivassilou, Laurence Kotlikoff and John Morris use a new simulation method to examine the effects of health, functional ability and demographic factors on the living arrangements of the elderly. Their methods produce a better fit of the data and reconfirm that increased age and decreased functional ability as measured by an activity-of-daily living scale result in an increased probability of an elderly person residing in a nursing home. Borsch-Supan, Kotlikoff and Morris along with Jagadeesh Gokhale estimated a model of the provision of time...