Savings

Since 1926 the Internal Revenue Code has provided for deferral of taxes on retirement savings in plans that file with the government to be ‘‘qualified’’ for this special tax treatment. Contributions to such plans are not taxed as income when made, and interest and investment gains are not taxed when realized. Instead, both are taxed upon withdrawal from the plan. Many such incentives have been added to the law since 1926. Not all are for retirement savings, with incentives now offered for saving towards the purchase of a new home, saving for sending children to college, or saving for medical expenses.

The savings issue that arises when such incentives are discussed, is the impact of such incentives (in particular, IRAs and 401(k) plans) on personal saving rates. At the aggregate level, tax-qualified retirement plans represent a tremendous store of wealth in vehicles earmarked specifically for retirement—$10.5 trillion as of 1998, up from $3 trillion just one decade earlier. At the individual level, the latest data on 401(k) accumulations indicate the potential these vehicles have for generating retirement wealth. According to the EBRI/ICI Participant-Directed Retirement Plan Data Collection Project, the average 401(k) account balance was $55,000 at year-end 2000 (up 47 percent from the average account balance at year-end 1996). More significantly, the average balances of older workers with long tenure indicate that a mature 401(k) program will produce substantial account balances. For example, individuals in their 60s with at least 30 years of tenure have average account balances in excess of $185,000.

Some argue that the tax-preferred treatment and the implicit government subsidy of saving through such plans, along with the provision of a degree of self-discipline that results from automatic saving, results in higher levels of saving than would otherwise exist without such programs. Others maintain that such preferential tax treatment merely serves as an inducement to transfer existing savings into such vehicles or to use such vehicles for saving that would have occurred even without such programs.

Interest in this issue is spurred by the fact that individual tax deferrals for employer-based retirement plan contributions and earnings carry a high estimated cost to the federal government, relative to other programs. The U.S. Treasury Department estimates that in fiscal 2000, the net exclusion of pension contributions and earnings will result in a federal tax revenue loss of $99.8 billion, and for fiscal years 2000 through 2004, these provisions will result in a tax revenue loss of $527.2 billion over the five-year period (see Executive Office of the President 1999). One analysis (Salisbury 1993) found that about one-half of the retirement-tax preference was attributable to public-sector-defined benefit retirement plans, about one-third to private-sectordefined contribution plans, 15 percent to private-sector-defined benefit plans, and 2 percent to public-sector-defined contribution plans.

Impact on savings. James Poterba, Steve Venti, and David Wise (1996), based on a body of research that they conducted over time, concluded that contributions to tax-qualified personal-retirement savings plans (IRAs and 401(k) plans) represent largely new savings and thus such plans have a significant positive effect upon saving rates.

Steven Venti and David Wise (1995), utilizing the Survey of Income and Program Participation (SIPP), and based on a household longitudinal methodology, found no significant reduction in other saving when households begin contributing to IRAs.

What would households eligible for saving incentives have saved in the absence of these incentives? Eric Engen, William Gale, and Karl Scholz (1996) argue that empirical analysis of this question is difficult and subject to biases that generally lead to overestimation of the impact of saving incentives. They conclude, based on a body of empirical work, that controlling for these biases largely or completely eliminates estimated positive effects of saving incentives on saving. They do qualify this conclusion by stating that such incentives may increase saving for some people and that they may eventually increase saving in the long run.