Alan Auerbach and Laurence Kotlikoff have done their assignment; they have predicted the impact of the upcoming demographic shift on saving and capital formation. They conclude that the significant projected increase in the elderly and the young relative to the working-age population will substantially depress the national saving rate beginning around the year 2010. Their conclusion rests on the standard life-cycle model where households accumulate wealth in their working years and then draw down their wealth to support themselves in retirement. According to this theory, the demographic shift should produce a dramatic increase in the number of dissavers relative to savers and this development should substantially reduce national saving. Who can argue with that logic?
While the logic may be flawless, the importance of demographics as opposed to other factors in determining national saving deserves some consideration. In fact, three things should make us somewhat cautious about incorporating a lower saving rate into our national strategic planning.
The first, which the authors acknowledge themselves, is the evidence from the 1980s. Their model, as well as several others, suggests that, based on demographics, the saving rate should have increased during the 1980s. The projected rise in private saving should have occurred because the large decline in the proportion of the population that was young swamped the increase in the proportion that was old. But instead of a projected rate of 12.8 percent, saving averaged 4.3 percent for the decade. This means either that demographic factors are not very important or that other influences overwhelmed the demographics. The difficulty is that economists have been at a loss to explain the behavior of national saving during the 1980s. Studies have explored the impact of capital gains from equities and housing, the effects of higher real interest rates, a reduction in the need for precautionary saving, the effect of slower income growth, and a host of other factors as possible reasons for the precipitous decline in saving. Unfortunately, to date the puzzle remains unsolved. However, the fact that models based on demographics were off by nearly 200 percent in projecting the 1980s saving rate suggests that forecasts for the future should be taken with a grain of salt.
The second argument for caution in interpreting the results is the fragility of the evidence documenting the importance of demographics. The strongest piece of evidence comes from a series of cross-sectional studies comparing international differences in saving rates ( Modigliani 1970, Feldstein 1980, Modigliani and Sterling 1983, and Horioka 1989). These studies, which were originally undertaken to determine the impact of social security on saving, generally found a statistically significant negative relationship in the late 1960s and early 1970s