|Cato Policy Analysis No. 73||June 3, 1986|
by Peter J. Ferrara
Peter J. Ferrara, a Washington attorney, is author of Social Security: The Inherent Contradiction and editor of Social Secu- rity: Prospects for Real Reform, both published by the Cato Institute.
The tax-reform bill currently before the Senate is one of those rare pieces of legislation that are truly historic and revolutionary. It deserves the wide support it is receiving. But there is one major flaw in the bill that is so serious it must be corrected before passage by the full Senate. That is the provision sharply restricting the deduction for Individual Retirement Accounts (IRAs).
Unfortunately, the debate over IRAs has been marred by widely trumpeted but flawed and inaccurate analysis. These inaccuracies must be cleared up. The fact is that workers would receive much higher retirement benefits from their IRAs with the deductible IRA under current law than with the non- deductible IRA under the Senate Finance Committee bill. Restoring the full deductibility of IRAs under the Committee bill is necessary if workers are to receive about the same benefits in retirement from their IRAs as under current law.
Moreover, it should be recognized that the policy justifications for fully deductible IRAs are still applicable and compelling under the Committee's sweeping tax-reform proposal. Opinion polls also show that the public strongly supports restoring IRA deductibility. Doing so would only strengthen the appeal of tax-reform legislation.
The IRA System and the Proposed Restriction
Under current law, all workers earning $2, UUO or more per year may contribute up to $2,000 per year to an IRA. Non- working spouses may contribute up to $250 per year. All IRA contributions are currently tax-deductible. Returns on IRA investments also accumulate tax-free. All withdrawals from an IRA are counted as income subject to tax in the year of withdrawal. For withdrawals made before age 59 1/2, the taxpayer must in addition pay a tax penalty equal to 10 percent of the amount withdrawn. This is the same tax treatment accorded all pension plans, except for varying annual contribution limits and withdrawal policies. These current IRA provisions were adopted in the 1981 tax legislation.
The Senate Finance Committee tax-reform bill would allow full IRAs as under current law only to workers not covered by employer pensions, except that the early-withdrawal penalty for these workers would be increased to 15 percent. Workers covered by employer pensions would be allowed to make IRA contributions up to the current limit but would receive no deduction for them. For these workers, returns to IRA investments would continue to be tax-free. Contributions would not be included in taxable income when withdrawn, though returns earned over the years would be. The early withdrawal penalty for these workers would remain at 10 percent. With these changes, about 60 percent of all workers, and 80 percent of IRA holders, would no longer be eligible for the IRA deduction.
EBRI's Flawed Analysis
Recently, the Employee Benefits Research Institute (EBRI) released a study purporting to show that the nondeductible IRA under the Senate Finance Committee tax-reform bill would, due to the lower tax rates proposed in the bill, provide equal or greater value in benefits for most workers than IRAs under current law, with a deduction for contributions but higher tax rates. This conclusion is erroneous, resulting from demonstrably incorrect analysis and methodology in calculating the relative values of the IRA under current law and under the Senate Finance Committee bill.
One major flaw in the EBRI analysis is that it emphasized cases where workers have the same marginal tax rates in retirement as in their working years. Under current law, most workers have lower marginal tax rates in retirement, when their incomes are usually lower. This disparity results in a significant advantage for IRAs under current law, since contributions are deducted at the higher rates during working years but are taxed as income at the lower rates prevailing in retirement.
Much more importantly, the EBRI analysis failed to account completely for the full value of the IRA deduction at the higher tax rates under current law. The correct way to evaluate the relative values of the IRA under current law and under the Senate Finance Committee bill is to compare the benefits each taxpayer would receive for contributions of equal net cost. Thus, a taxpayer in the 50 percent tax bracket under current law could contribute $2,000 to an IRA for a net cost of $1,000, since the taxpayer would receive back $1,000 in reduced taxes due to the deduction of his $2,000 contribution at the 50 percent tax rate. By contrast, with the nondeductible IRA under the Senate Finance Committee bill, the taxpayer would incur a net cost of $1,000 with a contribution of only $1,000. Consequently, to compare current law with the new tax bill properly, for a 50 percent bracket taxpayer one must compare a $2,000 contribution under current law for a net cost of $1,000 with a $1,000 contribution under the Senate Finance Committee bill for a net cost of $1,000. One would then calculate the interest accumulated on these contributions over the years and compare the ultimate after-tax benefits workers would receive at withdrawal under the higher tax rates of the current system and the lower rates of the Senate Finance Committee bill
The EBRI study did not follow this methodology, which is a standard analytical method used by tax and investment advisers. For example, for a 50 percent bracket taxpayer under current law, EBRI did not add in any of the tax benefits resulting from the deduction for an IRA contribution of $1, UUU. Instead, it simply compared an IRA contribution of $1,000 under current law with a contribution of $730 or $680 under the Senate Finance Committee bill, figuring that with nondeductibility $270 or $320 of the $1,000 available for contribution would be taxed away under the 27 percent or effective 32 percent tax rates of the Committee bill. This methodology clearly does not take into account the full value of the IRA deduction at the higher tax rates under current law. The net cost to the 50 percent bracket taxpayer under current law for a $1,000 IRA contribution is only $500, as the taxpayer receives back $500 in taxes due to the deduction of the $1,000 at the 50 percent tax rate. The EBRI study, however, compares this net-cost contribution of $500 under current law with contributions of $680 or $730 under the Senate Finance Committee bill. For the 50 percent bracket taxpayer, the correct comparison is rather a $1,000 contribution under current law and a $500 contribution under the tax-reform bill.
To show the significance of these errors in the EBRI study, The IRA Reporter, a national trade industry periodical, recalculated the relative values of the deductible IRA under current law and the nondeductible IRA under the Senate Finance Committee bill, using the proper, equal-net-cost methodology. The results are shown in the accompanying charts and tables.
Taxpayers at different marginal tax rates were assumed to make contributions involving equal net costs of $1,000 to an IRA under current law and under the Senate Finance Committee bill. As in the EBRI study, the calculations were done twice, first assuming an annual return of 10 percent interest, and, second, assuming an annual return of 7 percent interest. The data presented here include the value of each IRA contribution after 30, 20, 10, and 5 years, respectively, and the amount of after-tax benefits the worker would receive for withdrawing this value in retirement years or as an early withdrawal during working years. Examples of contributions to an IRA under the Senate Finance Committee bill with full deductibility restored are also presented.
The data show that all workers would receive much higher benefits in retirement with the deductible IRA under current law than with the nondeductible IRA under the Senate Finance Committee bill. Restoring the full deductibility of IRAs under the Committee bill is necessary just for workers to receive about the same retirement benefits from their IRAs under the bill as they would under current law.
For example, a contribution under current law with a net cost of $1,000 from a worker in the 50 percent tax bracket during his career and the 38 percent tax bracket in retirement would be worth $21,637 after tax when withdrawn in retirement after 30 years. But with the nondeductible IRA under the Senate Finance Committee bill, a contribution with net cost of $1,000 from a worker at the effective 32 percent tax rate during working years and the 27 percent tax rate in retirement would be worth only $13,008 after tax when withdrawn in retirement after 30 years. The IRA value under current law in this comparison is consequently 66 percent greater than the IRA value under the Senate Finance Committee bill, as shown in Chart 1.
Similarly, a contribution under current law with net cost of $1,000 from a worker in the 25 percent tax bracket during working years and the 16 percent tax bracket in retirement would be worth $19,543 after tax when withdrawn in retirement after 30 years. But with the nondeductible IRA under the Senate Finance Committee bill, a contribution with net cost of $1,000 from a worker in the 27 percent tax bracket during working years and the 15 percent tax bracket in retirement would be worth only $14,982 after tax when withdrawn in retirement after 30 years. The IRA value under current law in this comparison is consequently 30 percent greater than the IRA value under the Senate Finance Committee bill, as shown in Chart 2.
Restoring full IRA deductibility under the Senate Finance Committee bill would merely make the IRA values under the bill roughly equivalent to IRA values under current law in most cases. Thus, for the worker under the Committee bill at the 27 percent tax rate during working years and 15 percent tax rate during retirement years, a $1,000 net-cost contribution to a deductible IRA would be worth $20,318 after tax when withdrawn in retirement after 30 years, compared with $19,543 for a worker under current law in the 25 percent tax bracket during working years and 16 percent tax bracket in retirement. The deductible-IRA value under the Committee bill in this comparison would consequently be within 4 percent of the current-law value, as shown in Chart 2.
In some cases, however, even restoring full IRA deductibility would still leave IRA values under the Senate Finance Committee bill significantly lower than under current law, though much closer than without the deduction. Lowering the tax rates under the Committee bill depreciates the value of the IRA deduction, and this depreciation can, in a few instances, outweigh the beneficial impact of the lower tax rates on withdrawals. Thus, for a worker under the Committee bill at the effective 32 percent tax rate during working years and 27 percent tax rate during retirement years, a $1,000 net-cost contribution to a deductible IRA would be worth $18,733 after tax when withdrawn in retirement after 30 years, compared with the value of $21,637 for a worker under current law in the 50 percent tax bracket during working years and 38 percent tax bracket in retirement. The current-law value in this example would consequently still be about 15 percent greater than the deductible-IRA value under the Committee bill, as shown in Chart 1. This does not imply that rates should not be lowered, but it does clearly indicate the importance of restoring the full IRA deduction.
The IRA Success
IRAs have proven extremely effective in developing private resources devoted to retirement needs. In the last four years, total assets saved in IRAs have soared to more than $250 billion, already over 15 percent of the total assets saved in pensions. Over 28 million households--about one-third of all households-- and 40 million individuals have opened IRAs. The popularity of IRAs seems to be rapidly accelerating, with more and more individuals participating and contributing more and more in savings each year.
The huge amount of private resources accumulating in IRAs will substantially reduce pressure over the long run for government to provide increased retirement support through Social Security and other programs. Policymakers in Washington are already debating where additional resources should come from to finance long-term nursing-home care for the future aging population, to pay for retirement medical care with Medicare so badly underfunded, and to cover the enormous retirement demands of the baby-boom generation. Social Security and pensions do not appear adequate to meet all these needs. Without the highly successful IRA system building up private resources to meet these needs, the burden will probably fall on government, requiring greater and greater amounts of government spending.
Moreover, IRAs simply remove the discriminatory, multiple- tax burden on private savings that exists under our tax system and that will continue to exist under the Senate Finance Committee bill. Returns to savings and investment are subject to income taxation three times--once by the corporate income tax, once by the ordinary personal income tax, and again by the capital gains tax. Moreover, without expensing for the costs of capital investment, investors are taxed on supposed investment profits before they have recovered the cost of their investment. By increasing the corporate tax burden and eliminating the capital gains exclusion, the Senate Finance Committee bill probably worsens this discriminatory tax treatment.
In recognition oi the current tax discrimination against savings, the Treasury originally recommended in its 1984 tax-reform study that the IRA contribution limit be increased to $2,500 per person. The Treasury believed that this expanded IRA would provide the most equitable, pro-growth tax system. The White House agreed, finally proposing a lesser though still major increase in the contribution limit.
Our national retirement policy is supposed to be based on the concept of the "three-legged stool"--Social Security, pensions, and private savings. IRAs support the private-savings component. Without IRAs, how can we expect private savings to perform a major role in the face of the heavy, discriminatory, multiple-tax burden that otherwise applies to savings and capital under our tax system? The tremendous boom in private retirement savings that has occurred under the auspices of IRAs would collapse under the weight of this discriminatory tax burden.
IRAs have proven to be excellent retirement policy. They are the only vehicle available to virtually all workers in all circumstances. They avoid all vesting problems, since funds paid into an IRA immediately belong to the worker. They avoid all portability problems, since the IRA funds are under the worker's ownership and control wherever he goes. They offer workers greater freedom and self-reliance than Social Security or pensions, where the fate of the worker is in the hands of others and the worker's eligibility can be lost or cut off. IRAs get workers maximally involved in controlling and planning for their retirement and participating in the private economy and investment market. IRAs are clearly the most flexible, best-suited retirement vehicle for our modern, highly mobile, broadly diverse work force.
Restoring full IRA deductibility may also be justified on the grounds that IRAs produce additional, badly needed savings for our economy. IRAs may be expected to produce substantial additional savings because they reduce the discriminatory, multiple-taxation burden on the returns to savings. The rapidly growing amount of IRA assets provides strong encouragement that substantial new savings are being generated and that even more may be generated in the future.
Recent studies by Harvard professor David Wise indicate that IRAs have in fact substantially increased savings. Wise reviewed data regarding individual savings behavior and found that the amount most individuals are saving in IRAs is substantially greater than they could be expected to save otherwise. On the basis of statistical analysis of both individual and national savings data, Wise's analysis suggests that about 80 percent of IRA contributions are new savings that would not have been accumulated otherwise. This would mean that IRAs have increased national savings by about $200 billion, with most of this increase occurring in the last four years.
Some argue that individuals borrow to make their IRA contributions, resulting in no savings increase. But this criticism overlooks the fact that the individuals who do borrow generally pay that borrowing off in the short term, leaving the long-term IRA savings intact. Some argue that individuals shift existing savings into IRAs. But this assertion ignores the lock-in effect of IRAs. With funds contributed to an IRA for retirement and with a penalty for early withdrawal, the worker is less likely to spend them earlier on major consumption items, such as a boat, a new car, or a vacation.
Moreover, workers generally have little fluid savings available for shifting into IRAs. Wise points out that apart from home equity the median savings for all families is only $1,200. For families earning $30,000 to $40,000 with a household head 45 to 54 years old, the median is only $4,600. While some shifting of existing savings may occur in the years just after an IRA option is allowed, over the years there will be less and less savings remaining to shift into an IRA. For this reason, many believe that the biggest net increase in savings from lRAs is just now starting to occur.
Wise also notes that the big rush to contribute to IRAs just before the tax deadline every year further suggests that individuals are not merely shifting existing long-term savings into IRAs. If individuals were in fact doing so, they would do it as early as possible to obtain the advantage of the tax-exempt interest returns.
Some doubt the impact of IRAs in increasing savings because the savings rate is down from 1981, when IRAs were first allowed for all workers. But IRAs are only a small part of the overall economy and overall savings. Many other factors could offset any beneficial impact from IRAs.
There is in fact a ready explanation for the decline in the savings rate since 1981. Asset values in the stock and bond markets have soared since that time, increasing the wealth held by investors. This increased wealth performs the same function for investors as increased savings, but it is not reported as an annual addition to savings. If investors reduce their reported savings because they are saving by retaining the increased wealth due to asset appreciation, the reported savings rate will decline, but real savings held by the public may increase dramatically.
To see this, consider a model that assumes that an individual's savings rate increases as his income increases and the return to savings increases. Assume that at a fixed return the individual will save 10 percent of his income when earning $50,000 per year, 15 percent when earning $100,000, 20 percent when earning $150,000, and 25 percent when earning $200,000. Assume that after years of saving $5,000 a year on a $50,000 salary, the worker holds a portfolio of $50,000, which jumps in value in one year to $200,000 due to asset appreciation. For the year, the worker will earn $200,000: 150,000 in asset appreciation and $50,000 in salary. The worker will consequently want to save $50,000 (25 percent of $200,000), instead of the usual $5,000. But he has added $150,000 to his net wealth in that year through asset appreciation. He may therefore decide to save nothing out of his salary that year, spend $100,000 of his accumulated funds, and still meet his goal of an effective $50,000 increase in savings held. While the worker would consequently add $50,000 to his portfolio, it would not be reported as a savings increase because increases in portfolio values are not reported as savings. Instead, reported savings will actually fall $5,000 because the worker did not save $5,000 out of his $50,000 salary as in prior years. Such an effect may have been reducing reported savings rates since 1981, even while actual savings held may have increased sharply.
Some challenge the fairness of IRAs by arguing that too much of the benefit goes to higher-income taxpayers. But 75 percent of taxpayers with an IRA earn less than $50,000. The income distribution of workers covered by IRAs is in fact not significantly different from the income distribution of those covered by pensions. The great majority of people who contribute to IRAs are from the middle class.
Any vehicle for private savings will involve somewhat greater participation by those with more income because they have more resources to make the savings. But that doesn't mean the vehicle is unfair, as long as the tax treatment involved simply amounts to more neutral, nondiscriminatory, fair taxation of savings, as with IRAs. And it does not mean that one should oppose encouraging greater reliance on private savings rather than government spending for those for whom such savings is feasible.
Moreover, it is ironic that the proposed IRA restriction would discriminatorily exclude blue-collar union workers, who are broadly covered by employer pensions, from eligibility for the IRA deduction. The exclusion would also fall most harshly on highly mobile young workers who may be eligible for an employer pension and therefore ineligible for the IRA deduction but who are likely to find it difficult to earn significant vested pension rig
© 1986 The Cato Institute
Please send comments to webmaster