Life Cycle Theories of Savings and Consumption

Economists have developed three major theories of consumption and saving behavior: (1) The life-cycle hypothesis (Modigliani and Brumberg, 1954; Modigliani and Ando, 1957; Ando and Modigliani, 1963); (2) the permanent income hypothesis (Friedman, 1957); and (3) the relative income hypothesis (Dusenberry, 1949). All three theories have their conceptual roots in the microeconomic theory of consumer choice. However, the life-cycle and permanent income hypotheses are the most similar; both theories assume that individuals attempt to maximize their utility or personal well-being by balancing a lifetime stream of earnings with a lifetime pattern of consumption. The relative income hypothesis is quite different. Dusenberry theorized that individuals are less concerned with their absolute level of consumption than with their relative level—the idea of "keeping up with the Joneses."

The life-cycle hypothesis has been utilized extensively to examine savings and retirement behavior of older persons. This hypothesis begins with the observation that consumption needs and income are often unequal at various points in the life cycle. Younger people tend to have consumption needs that exceed their income. Their needs tend to be mainly for housing and education, and therefore they have little savings. In middle age, earnings generally rise, enabling debts accumulated earlier in life to be paid off and savings to be accumulated. Finally, in retirement, incomes decline and individuals consume out of previously accumulated savings.

Empirical studies of the life-cycle hypothesis have generated a large literature. Studies that have focused on the savings behavior of older persons, however, have been inconclusive regarding the correspondence between observed savings behavior and the pattern of saving and dissaving predicted by the life-cycle hypothesis. Many studies seemingly in conflict with the life-cycle hypothesis, have found that older persons continue to save in retirement. Several explanations have been offered for this. King (1985), for example, notes that saving in retirement is not necessarily inconsistent with the life-cycle hypothesis, if one accounts for the aversion of individuals to uncertainty about the future (e.g., how long they will live and future inflation). Another explanation is that the generosity of pensions reduces the need to save in preparation for retirement and to dissave while in retirement. Life-cycle savings patterns in some European countries that have generous pension systems such as France, Germany, and Italy appear to be consistent with this explanation. Another related explanation for lack of dissaving in retirement is that deteriorating health may limit the ability of individuals to consume at levels that are higher than their pension income. Moreover, the pension wealth that retired persons hold is not liquid and they are not able to draw down their pension wealth any faster than the annuity payments that they receive. This health aspect of life cycle savings and consumption patterns raises an interesting question: Should payments for health insurance also be viewed as a form of savings, and receipt of health care services as drawing down one's "health insurance wealth"?

A number of other studies, however, have found evidence of a hump-shaped pattern of savings that is consistent with the life-cycle hypothesis. It is important to note that most studies have tended to underestimate the degree of dissaving among older persons, because these studies have not generally accounted for the decumulation of pension wealth associated with Social Security and private pension payments.

Pension payments are probably the best example of decumulation of savings in the latter stages of the life cycle. Under Social Security and defined benefit pension plans, older persons have established a claim on a future stream of income payments that is generally some function of each person's earnings history and life expectancy. The expected total value of this stream of income payments in current dollars over their remaining lifetime is known as their pension wealth. Thus, as retirees receive pension payments, they draw down their pension wealth. This factor has generally not been taken into account in studies that have examined whether older persons dissave in retirement, as would be predicted by the life-cycle hypothesis. After accounting for personal contributions and withdrawal of benefits from pensions, Jappelli and Modigliani (1998) find evidence for the expected hump-shaped savings profile.

In an analysis of the savings behavior of the baby boom generation, Gist et al. (1999) estimate that Social Security and pension wealth accounted for more than half of all wealth for 90 percent of the pre-retired population. As a consequence, it seems clear that failing to account for the reduction in pension wealth implied by the receipt of Social Security or other defined benefit pension payments leads to a substantial underestimation of dissaving in retirement. Moreover, in countries such as the United States, where outof-pocket health care costs are rising more rapidly than the value of pension payments, one might expect to observe dissaving in the retirement years. In particular, among the old-old population, rising medical and long-term care expenditures are likely to occur at a point in the life cycle where, for many, the real value of their pension income has eroded over time and may be inadequate to cover out-of-pocket health care costs.