Raising the Amounts That Can Be Contributed to Roth IRAs: The Dangers in the Short Run and the Long Run -- 6/21/00

Raising the Amounts That Can Be Contributed to Roth IRAs:
The Dangers in the Short Run and the Long Run
by Peter R. Orszag⁽¹⁾

Over the next few weeks, the House of Representatives is expected to consider a proposal to raise the amount that can be contributed to a Roth Individual Retirement Account (IRA) from $2,000 to $5,000. This proposal is contained in the "Small Savers Retirement Enhancement Act" (H.R. 1322), which was introduced by Rep. Elton Gallegly and had 191 co-sponsors as of June 9, 2000.

To many policymakers, the Gallegly legislation appears attractive: It has little short-run budgetary cost and seems aimed, as its title indicates, at "small savers." In fact, the impact of this legislation would be quite different than such initial impressions may suggest. This paper explores three key problems that the Gallegly bill raises:

1. The bill would disproportionally benefit higher-income earners, while doing little for middle-class families. Pension and IRA tax preferences already are substantially skewed toward higher-income individuals. Raising the amount that could be contributed to a Roth IRA from $2,000 to $5,000 would provide additional tax benefits disproportionately to those in the higher parts of the income distribution, but do little for middle-class families.

2. The bill would produce substantial tax losses in the long run. The bill's short-run cost estimate, which is small, is misleading. Due to the nature of Roth IRAs, increasing the amount that can be contributed to them would have little short-term cost. But it would significantly reduce revenue over the longer term. These longer-term revenue losses would mount as the baby boomers retired in increasing numbers, reducing revenue at the same time that other pressures on the budget were growing more intense.

3. The bill would endanger pension coverage for low- and moderate-income workers in small businesses. Increasing the amount that can be contributed to a Roth IRA would be likely, over time, to reduce pension coverage for workers in small businesses, because some small business owners would be able to meet their own pension needs through IRAs and consequently would no longer need to provide an employer-sponsored pension plan to meet those needs. Low- and moderate-income workers already have relatively low pension coverage; policymakers should be working to expand their coverage rates, not to provide business owners incentives to reduce them.

The Clinton Administration's proposal to establish "retirement savings accounts" (RSAs) represents an alternative way to boost retirement saving. The RSA proposal does not pose the problems that the Gallegly proposal does. Instead of benefitting only a very small number of middle-income families — those few that already deposit the full $2,000 in an IRA — the RSA proposal would be heavily focused on middle- and lower-income families and individuals. Nearly all married families with incomes of up to $80,000 would benefit more — in most cases, much more — from the RSA proposal than from raising the IRA contribution limit to $5,000. The RSA proposal also constitutes more responsible fiscal policy because it does not camouflage its costs in the initial years and result in burgeoning costs outside the five-year and 10-year budget periods that Congress employs. In addition, the RSA proposal does not provide similarly strong inducements to business owners not to offer employer sponsored pension plans.

Under current law, a taxpayer and spouse may each contribute up to $2,000 to a conventional IRA, a Roth IRA, or a non-deductible IRA. A couple thus may contribute a total of $4,000.

Problem #1: Disproportionate Benefits for Higher-income Individuals, with Little Effect on Middle-income Taxpayers

The tax subsidies for retirement saving that the federal government currently provides are skewed heavily toward more-affluent individuals. Treasury data show that two-thirds of the existing tax subsidies for retirement saving (including both private pensions and IRAs) accrue to the top 20 percent of the population. Only 12 percent of these tax subsidies accrue to the bottom 60 percent of the population. This suggests that any new retirement saving subsidies should be focused primarily on lower- and middle-income families.

Raising the Roth IRA contribution limit to $5,000, however, would provide additional and disproportionate benefits to higher-income earners. By its nature, the Gallegly proposal would benefit only those who are at the $2,000 IRA contribution limit under current law. It would have virtually no effect on families and individuals who either do not make any IRA contributions under current law or who deposit less than the current $2,000 IRA limit.

The logic is clear. Those who can not afford to deposit $2,000 in a IRA cannot deposit $5,000. They would not be affected by an increase in the Roth IRA contribution limit. This proposal would directly benefit only those who already make the $2,000 maximum IRA contribution.

A recent Treasury study shows that in 1995, only a tiny percentage of taxpayers contributed the maximum amount to a conventional IRA. Comparable data are not yet available for Roth IRAs, which did not exist in 1995, but the results are likely to be similar. The reason for the similar effect is that the lifetime tax benefit from a dollar deposited in a Roth IRA is generally the same as the lifetime tax benefit from a dollar deposited in a traditional IRA. Participation rates thus are unlikely to be substantially higher in Roth IRAs than in traditional IRAs. (One could argue that relatively more Roth IRA depositors may be at the $2,000 maximum contribution level because the income limits for making Roth IRA contributions are significantly higher than the income limits for making deductible deposits to a conventional IRA. But it also is possible that the number of contributors to Roth IRAs who are at the $2,000 level may be smaller than the number of depositors to conventional IRAs at this level, because very high-income taxpayers who are not covered by an employer-provided pension plan are eligible to make contributions to a conventional IRA, but not to a Roth IRA. IRS data show that in 1995, roughly one-third of all taxpayers who made deductible contributions to conventional IRAs had incomes above the income limits for those IRAs.)

The Treasury study found that only four percent of all taxpayers who were eligible for conventional IRAs in 1995 made the maximum allowable $2,000 contribution.⁽³⁾ The analysis also found that only seven percent of taxpayers eligible to make deductible contributions to a conventional IRA made any IRA contribution in 1995.

The Treasury paper concluded: "Taxpayers who do not contribute at the $2,000 maximum would be unlikely to increase their IRA contributions if the contribution limits were increased whether directly or indirectly through a backloaded [Roth] IRA."⁽⁴⁾ It is only the very small minority of eligible taxpayers contributing the maximum $2,000 to an IRA who would likely benefit from raising the maximum contribution amount on Roth IRAs above $2,000.⁽⁵⁾

Raising the Roth IRA contribution limit consequently would not do anything to increase the amounts that all but a very small fraction of taxpayers save for retirement. Furthermore, those the proposal would benefit — individuals who already make the maximum IRA contribution — are likely to be higher-income earners.

In tax year 2000, eligibility for making a Roth IRA contribution phases out at $160,000 in adjusted gross income for joint filers and $110,000 for single filers. Couples that earn more than $100,000 and are eligible for Roth IRAs are much more likely to be able to set aside $10,000 ($5,000 for each spouse) than couples that earn $40,000 or $50,000. The tax subsidies the Gallegly bill would create consequently are likely to be skewed toward higher-income earners. They would add to the subsidies such individuals already enjoy under current pension and IRA tax laws.

The revenue loss from increasing the amount that can be contributed to a Roth IRA would be much smaller in the short term — i.e., over the years the Congressional budget resolution covers — than the revenue loss from an equivalent increase in the amount that can be contributed to a conventional IRA. This is because funds deposited in a Roth IRA do not result in an immediate tax deduction. But while an increase in the contribution limit for Roth IRAs would lose much less revenue in the years immediately ahead than raising the contribution limit for conventional IRAs, it would lose much more revenue in the years after that. The backloading of the tax subsidy under Roth IRAs is essentially a tax device that alters the timing of the revenue losses without reducing the overall amount of the loss; this device aggravates the drain that the tax break places on government finances in later years. Indeed, Roth IRAs were created specifically for the purpose of maneuvering around budget limits by structuring the IRA tax breaks so most of their cost occurs outside the budget window. Henry Aaron, a senior fellow at the Brookings Institution and a noted expert in tax policy, has described Roth IRAs as essentially a device to conceal the true cost of expanding IRA tax breaks.⁽⁶⁾

Fact and Fiction in Merrill Lynch's Promotion of the IRA Proposal

Merrill Lynch recently ran a full-page advertisement in Congressional Quarterly and National Journal promoting proposals to raise the IRA contribution limit to $5,000. The Merrill Lynch advertisement correctly notes that despite the economic boom, many Americans do not appear to be saving enough for retirement. But it includes two misleading statements.

First, the advertisement claims that "raising the IRA limit to $5,000 would encourage more savings and would especially benefit working women, whose savings needs are more acute because they are often in and out of the workforce to raise children." As this paper demonstrates, however, the increase in the limit to $5,000 would have no direct effect on the vast majority of taxpayers, including both working men and working women. Only a very small percentage of working women would receive any benefit from the proposal.

Second, the advertisement states that for "years the IRA has been an invaluable savings tool for over 30 million households of working Americans." This is likely to create the impression that 30 million Americans a year make deposits into IRAs. In fact, Treasury data show that the number of taxpayers making IRA contributions is about five million a year, one-sixth of the 30 million number. (The most recent reliable data available show that 5.3 million taxpayers made contributions to IRAs in 1995.) The 30 million figure appears to reflect the total number of people who have contributed to IRAs over a period of nearly 20 years, including large numbers of higher-income individuals who contributed in the early 1980s when there were no income limits on IRAs.

Under Roth IRAs, no revenue losses occur upfront when funds are deposited in the IRA because no tax deductions are provided on those deposits. Instead, the revenue losses occur in later years when individuals withdraw funds from their Roth IRA accounts; rather than being taxable, the withdrawals are tax-free. The revenue losses from raising the Roth IRA limit thus are hidden, because they are largely pushed outside the 5-year and 10-year budget windows. But they are real nonetheless.

The point at which the revenue losses caused by raising the Roth IRA contribution limit would become large is about the same time that the baby boom generation will begin to retire. The nation may face new and potentially serious fiscal problems at that time. The proposal to increase the contribution limits for Roth IRAs would aggravate these long-term fiscal problems. The mounting revenue losses that the proposed Roth IRA expansion would engender would eventually have to be paid for through higher taxes, lower spending, or higher levels of government debt than would otherwise be the case. If the revenue losses are offset by deeper reductions in benefits provided under various government programs, middle- and low-income families would be likely to finance most of the cost of providing what is primarily a new tax benefit for more affluent individuals.

The proposal could endanger pension coverage for workers at some small businesses, because it could create incentives for small-business owners not to establish an employer pension plan and instead to meet their own retirement saving needs through the substantially enlarged IRA contributions the proposal would permit.

Currently, a small-business owner with, say, $125,000 in income can deposit $4,000 a year in a Roth IRA ($2,000 for the owner and $2,000 for the owner's spouse). If the owner wants to set aside a larger amount, say $10,000, in tax-favored retirement savings, the owner must establish an employer pension plan and make contributions through the plan. If the owner does so, the firm must make pension contributions for its employees as well.

If the Roth IRA contribution limits are raised to $5,000, however, the owner will be able to use the Roth IRAs to put away $10,000 a year in tax-advantaged retirement saving without having to incur the expense of operating and making contributions to an employer-sponsored pension plan for the firm's employees. As a result, the legislation would provide a strong incentive for new businesses, as well as businesses that have reached a level of stability at which they otherwise might institute a pension plan, not to establish one. Given the high rate of small-business creation and expansion in the U.S. economy, the effect over time could be substantial.

The Clinton Administration's proposal to establish "retirement savings accounts" (RSAs) represents an alternative way to boost retirement saving. The RSA proposal does not pose the array of problems that the Gallegly proposal does. Under the RSA approach, the Treasury would provide tax credits to match contributions that married couples with incomes up to $80,000 a year (and individuals with incomes up to $40,000) make to retirement saving accounts. The matching rate would be highest for lower-income families and gradually phase down as income rose.

Compared to the proposal to raise the Roth IRA contribution limit, the RSA proposal represents a far more efficacious way to increase retirement saving among middle- and lower-income working families. While the Roth IRA proposal would directly benefit only a very small number of middle-income families — those few that already deposit the full $2,000 in an IRA — the RSA tax subsidies would be focused heavily on middle- and lower-income families and individuals.

As noted, raising the Roth IRA contribution limit would have virtually no effect in increasing participation rates in IRAs among those not participating under current IRA rules and also would not be likely to affect those who participate in IRAs but contribute less than $2,000. By contrast, RSAs would provide subsidies to modest savers to increase the amounts they save by matching contributions they make, starting with the first dollar they save. As a recent Center analysis showed, for most couples with incomes below $80,000 and most individuals with incomes below $40,000, RSAs would provide a substantially larger subsidy for retirement saving — and a much more powerful inducement for such saving — than raising the IRA contribution limit would.⁽⁷⁾

Contributions to RSAs also would be more likely to add to national saving than contributions to IRAs, an important issue since increasing national saving should be one of the nation's top priorities in preparing for the retirement of the baby-boom generation. Increasing the IRA limit would result in a large portion of the new IRA tax subsidies going to more-affluent taxpayers, a group that can shift funds from existing saving vehicles rather than increase the amount they save out of their income. By contrast, the RSA proposal would concentrate its subsidies on lower- and middle-income families, a group much less likely to have substantial financial assets to shift. Deposits made in a saving vehicle such as RSAs consequently are more likely to represent new saving than the increased amounts that would be deposited in IRA accounts if the IRA contribution limit is raised to $5,000. Indeed, new research by Eric Engen of the Federal Reserve Board and William Gale of the Brookings Institution suggests that tax-preferred retirement saving undertaken by lower-income workers is much more likely to represent new saving (rather than asset shifting) than tax-preferred retirement saving undertaken by higher-income workers.⁽⁸⁾

The RSA proposal also would pose less danger of creating incentives for small businesses not to offer employer pension plans. An increase in the IRA contribution limits could obviate the need for some small-business owners to offer employer plans to build substantial tax-advantaged retirement accounts for themselves. RSAs would have a much more limited effect in this regard because they would not be available to business owners (or anyone else) whose incomes exceed $80,000. (Roth IRAs, by contrast, are available to business owners and executives with incomes up to $160,000.) Since RSAs would not provide retirement tax subsidies for business owners or executives with incomes of more than $80,000, they could not replace employer pension contributions for most business owners and executives. As a result, the creation of RSAs would not provide much of an incentive to an owner to drop or fail to initiate an employer-sponsored pension plan.

The proposed expansions in Roth IRAs would cause substantial tax losses outside the budget window and provide additional and disproportionate benefits to higher-income earners, while doing very little for middle-class families. These expansions also could endanger pension coverage for some workers at small businesses. The proposed expansion seems an unwise course to follow.

The Administration's RSA proposal represents a more promising approach to increasing retirement saving and retirement income for lower- and middle-income working families. Although it costs more than the Gallegly proposal in the short run, the RSA proposal does not create longer-term liabilities through burgeoning tax losses as the Roth IRA proposal does. Policymakers who seek to boost retirement security among lower- and middle-income families would obtain much more beneficial results from the RSA proposal than the Gallegly approach.

1. Peter R. Orszag is President of Sebago Associates, Inc., an economics consulting firm, and lecturer in economics at the University of California, Berkeley. From 1995 to 1998, he served as Special Assistant to the President for Economic Policy and as Senior Economist on the Council of Economic Advisers.

2. If a taxpayer remains in the same tax bracket throughout his or her life, including after retirement, a dollar deposited in a conventional IRA will yield the same tax benefits over a lifetime as a dollar deposited in a Roth IRA. A non-deductible IRA offers much less in lifetime tax benefits, since neither deposits nor withdrawals are excluded from taxation.

3. Robert Carroll, "IRAs and the Tax Reform Act of 1997," Office of Tax Analysis, Department of the Treasury, January 2000.

5. Advocates for raising the $2,000 limit may argue that doing so would attract more workers to contribute to Roth IRAs in the first place. For example, one such argument would be that there are large fixed costs associated with learning about IRAs and investing in them, so that individuals will not find it worthwhile to do so in exchange for the opportunity to make a $2,000 contribution but will find it worthwhile in exchange for the opportunity to make a $5,000 contribution. Such an argument would strain credulity; the explanation for the low IRA participation rate among middle-income families is not likely to be the fixed costs of setting up an account. Similarly, advocates may argue that the $2,000 limit provides a psychological benchmark against which individuals judge their savings behavior, so that someone always saving "half the benchmark" would save more if the limit were $5,000 rather than $2,000. This argument is not especially persuasive either; those who contribute to Roth IRAs today but deposit less than $2,000 are likely to do so because they cannot afford to put $2,000 aside, not because they wish to save "half the benchmark" or some other such fraction of it. Moreover, even if this argument were valid, its effects would appear to be limited — only seven percent of eligible taxpayers made any contribution to a deductible IRA in 1995, and the figures are likely to be similar for contributions to Roth IRAs.

6. Statement of Henry J. Aaron before the House Committee on Ways and Means, March 13, 1991.

8. Eric Engen and William Gale, "The Effects of 401(k) Plans on Household Wealth," Paper prepared for the TAPES conference, Gerzensee, Switzerland, May 22-24, 2000.