Complexity in Retirement Savings Policy

The U.S. retirement system is often described as a three-legged stool in which the legs represent Social Security, employer pensions, and individual savings. This metaphor can be somewhat misleading, however, because it inaccurately suggests that the three sources of retirement income are approximately equal. In 2000, three-fifths of individuals age 65 and over received at least half their income from Social Security, and for nearly one-fifth Social Security was the only source (Social Security Administration, 2002a). The shares for the other "legs" are much smaller: only one-fifth of the population received as much as 20 percent of their income from individual savings, and just one-tenth received this much from employer pensions. Moreover, these figures vary dramatically with income: half of the over-65 population in the lowest income quintile received all of their income from Social Security, while a third of the population in the highest income quintile received at least half of their income from labor-market earnings. These differences illustrate an additional limitation of the stool metaphor--it is a static representation of a dynamic process. In particular, it does not consider the endogeneity of savings choices, and it is therefore of limited use in assessing the role of complexity in retirement savings behavior. For workers, retirement savings behavior essentially involves a trade-off between current and future consumption. Tax policies can be constructed to allow a more efficient trade-off, but overly complex rules will certainly discourage their use. Unfortunately, the sections of the U.S. tax code and regulations dealing with retirement savings policy are among the most arcane areas of our current tax system, for both individuals and businesses.