Tax Reform 2.0’s “Universal Savings Account” Proposal Is Ill-Conceived
August 10, 2018
The “Universal Savings Accounts” (USA) proposal included in “Tax Reform 2.0” — the new tax framework from House Republican leaders — would drain federal revenues substantially, distribute its tax benefits overwhelmingly to the nation’s wealthiest households, and not likely boost private savings much at all.
USAs would be new tax-preferred accounts that are similar to Roth Individual Retirement Accounts (Roth IRAs), but they would have no income limits for participation and individuals would not have to wait until retirement to withdraw funds. As a result, they would create an incentive for individuals to shift savings that they now hold in taxable accounts to USAs in order to take advantage of the tax break that USAs would provide — namely, that unlike funds in existing taxable accounts, all of the interest, capital gains, or other earnings in USA accounts would be entirely tax free, forever.Unlike funds in existing taxable accounts, all of the interest, capital gains, or other earnings in USA accounts would be entirely tax free, forever.
Such a shift in savings from taxable accounts to USAs would not increase private savings, but it would lose significant amounts of federal revenue, enlarging deficits and debt. Moreover, wealthy households would receive the largest tax benefits by far because they already have substantial amounts of money in taxable accounts that they could shift to USAs. By contrast, households in the middle tend to hold only modest amounts of assets in taxable accounts, and they would receive much smaller tax benefits.
Although the Tax Reform 2.0 document, which House Ways and Means Committee Chairman Kevin Brady released in late July, provides few details on its USA proposal, USAs are an idea that Rep. Dave Brat and Sen. Jeff Flake have proposed in legislation, and House Speaker Paul Ryan’s “Better Way” plan discussed the Brat-Flake proposal favorably. Consequently, to analyze the likely impact of the USA provision of Tax Reform 2.0, which House leaders plan to bring to the House floor in September, we focus on the details of the Brat-Flake proposal. Our general findings, however, would pertain to any proposal that includes the basic elements of Brat-Flake.
USAs Would Do Little to Boost Savings
Under Brat-Flake, adults could place $5,500 each year into a USA. They would contribute such funds with after-tax dollars, but all earnings on the funds would be tax free, and USA account holders would owe no tax when they withdraw money. In that way, USAs would be similar to Roth IRAs — under which contributions are not tax deductible, but account holders don’t pay tax on the earnings as they accrue or when account holders withdraw them.
Nevertheless, USAs are more generous than Roth IRAs in at least two important ways. First, under a Roth IRA, an individual can withdraw funds only after retirement (with limited exceptions). Under Brat-Flake, however, USA account holders could withdraw funds at any time and for any reason. Second, Roth IRA contributions are limited to married couples that make less than $200,000 a year. USAs, however, would have no income limits on participation.
Rep. Brat has promoted USAs as a “more streamlined and flexible saving account option that will truly encourage savings.” In reality, they won’t increase savings much at all since, according to high-quality research, this type of tax-based savings incentive is largely ineffective, mostly subsidizing saving that households would have undertaken anyway. Harvard economist Raj Chetty and his coauthors examined the savings effects of tax-preferred accounts and found that 85 percent of individuals do not respond to such tax incentives. The other 15 percent respond mostly by shifting savings from taxable to tax-preferred accounts. Overall, the authors estimated that every dollar of tax-cut benefit only increased savings by one cent.
More than anything else, based on that research, USAs would likely mean that households with funds in taxable investment and savings vehicles — including regular savings accounts, the interest income from which is taxable — would get a tax cut simply by shifting their existing savings from taxable accounts to tax-preferred USAs. Despite providing a tax cut, that shift wouldn’t increase overall saving.
In fact, USAs could reduce savings for some households. That’s because households in the middle of the wealth spectrum hold 57 percent of their savings in retirement accounts, and USAs would give them much broader flexibility to withdraw funds before retirement, such as to pay regular bills. If some households in the middle (most of whom do not fully take advantage of existing tax-preferred accounts) shifted some of their contributions from retirement accounts to USAs, it could leave them less prepared for retirement since they may withdraw funds from their USAs. Even if they later replenished their USA funds after withdrawing them, they may have less savings for retirement since the temporarily removed funds wouldn’t generate returns when they were outside the account.
USAs Would Drain Federal Revenue
USAs would drain federal revenues because — as with Roth IRAs — investment earnings would escape taxation. Those who already have Roth IRAs would have an additional vehicle through which to contribute and then pay no tax on their investment earnings. And those with incomes above the Roth IRA’s $200,000 income limit would have a new opportunity to secure these tax benefits.
Neither the Congressional Budget Office nor any other agency has conducted a cost estimate for USAs. Nevertheless, their cost could be substantial — especially in the years beyond Congress’ traditional ten-year “budget window” for evaluating costs — as households shift increasing amounts of savings from taxable accounts into USAs.
Moreover, any official cost estimate of USAs within the 10-year budget window would likely be artificially low. That’s because if some people shift new contributions from an existing tax-preferred account, such as a 401(k) or traditional Individual Retirement Account (IRA), to a USA account, that would change the timing of their tax benefits and, in turn, make the costs much smaller in the first ten years than in subsequent decades.
With a 401(k) or traditional IRA, individuals receive a tax break up front, because they either contribute to their 401(k) contributions with pre-tax dollars or receive a tax deduction when they contribute to a traditional IRA. But these individuals then pay tax when they withdraw funds, typically in retirement. With USAs, account holders would make contributions with after-tax dollars, so USAs wouldn’t provide an upfront tax benefit. But USA account holders would enjoy the tax benefit later, paying no tax on their investment earnings.
Thus, if some account holders shifted contributions from 401(k)s and IRAs to USAs, they would pay more taxes up front, raising federal revenue in the ten-year budget window. But the federal government would then lose revenue after ten years because it wouldn’t later collect taxes on withdrawals of funds that would have been in 401(k)s and IRAs.
USAs might even show an initial net revenue gain if lots of households shifted contributions from 401(k)s or traditional IRAs and paid the upfront taxes in exchange for future tax benefits. In a political sense, that could make it easier for lawmakers to vote for this proposal. In the past, policymakers have on various occasions sought to use timing shifts in tax payments to mask the long-term costs of tax-cut proposals and thereby facilitate their passage. Just last year, House Republicans reportedly considered changing the tax benefits of 401(k)s to more closely resemble Roth 401(k)s because that would have made their tax bill appear less costly.
USAs Would Mostly Help the Wealthy, Do Little for the Middle
USAs would deliver most of their tax benefits to wealthy Americans while doing little for those in the middle or at the bottom. In large part, that’s because the wealthy are best positioned to take advantage of the main tax benefit of USAs: they can reduce their taxes by rearranging their portfolios, shifting their savings from taxable accounts to tax-preferred USAs.
Currently, wealthy households have substantial assets in taxable accounts that they could shift into USAs. Households in the top 1 percent of wealth have, on average, $9.4 million in assets that they could shift into USAs, according to our analysis of the Federal Reserve’s Survey of Consumer Finances, using the methodology of New York University economist Edward Wolff. Outside the top 1 percent, the other households in the top 20 percent have an average of $470,000 in taxable accounts that they could shift into USAs. Thus, wealthy households could contribute to USAs year after year simply by shifting funds from their existing taxable accounts; they wouldn’t have to save more to take advantage of the tax break.
For the middle 60 percent on the wealth scale, the average household has just $16,000 that it can shift into USAs. Given the $5,500 per-adult contribution limit for USAs, a married couple with this amount of USA-eligible savings would need roughly two years to shift all of its existing eligible assets from taxable accounts into USAs. The median (as opposed to the average) amount of USA-eligible assets among the middle 60 percent is far lower — just $900 — so more than half of these households couldn’t shift any additional assets into USAs after one year. Once they had shifted those assets, these households could get an additional tax break only by boosting their savings rate. The research cited above, however, strongly suggests that they wouldn’t do so to a significant degree. In the end, wealthy households would get a hefty tax cut that grows over time, while those in the middle would get far less.
Moreover, most households seem poorly positioned to take advantage of USAs by the sheer fact that only a small share of households takes full advantage of existing tax-preferred accounts today. Contribution limits on these accounts are already generous: two married workers can contribute a combined $48,000 each year to their 401(k)s and IRAs (not including employer contributions), rising to $62,000 if they are 50 or older. But just 5 percent of 401(k) participants contributed the maximum allowable amount to their 401(k) in 2006, a Congressional Budget Office analysis found, and those that did were disproportionately households with incomes above $160,000.
USA benefits are tilted to the top in another important way as well: the tax savings would be proportional to a household’s tax rate on investments. Interest income, such as from a savings account or a bond portfolio, is taxed at ordinary income tax rates. A married couple with total income of over $1 million a year is in the 37 percent tax bracket and, consequently, would save 37 cents on the dollar if its interest income was earned tax free in a USA as opposed to a taxable account.  By contrast, a couple making under $100,000 a year would save at most 12 cents on the dollar because it’s in the 10 or 12 percent tax bracket.
Nor would USAs help middle-income savers when it comes to other types of investments. Take, for instance, capital gains — the gains from the sale of stocks or other assets. The top tax rate on long-term capital gains is 20 percent, but the rate is 0 for a married couple making up to about $100,000. Thus, that married couple would receive no additional tax benefit from holding such assets in a USA as opposed to a taxable account; either way, the couple would pay no tax on the gains. More than 70 percent of filers would fall into this group, according to IRS data. (See Figure 1.) By contrast, a married couple making $1 million a year would enjoy a significant tax advantage by shifting its assets into a USA. Rather than pay a 20 percent tax on any capital gain, it would pay no tax at all.
 U.S. House of Representatives Committee on Ways and Means, “Brady Releases Tax Reform 2.0 Listening Session Framework,” July 24, 2018, https://waysandmeans.house.gov/brady-releases-tax-reform-2-0-listening-session-framework/.
 The document says only the following; “Creating a new Universal Savings Account to offer a fully flexible savings tool for families.”
 Office of U.S. Congressman David Brat, “H.R. 937 - The Universal Savings Account Act,” https://brat.house.gov/legislation/the-universal-savings-account-act.htm.
 Office of the Speaker of the House of Representatives, “A Better Way,” June 2016, https://abetterway.speaker.gov/_assets/pdf/ABetterWay-Tax-PolicyPaper.pdf
 Internal Revenue Service, “Amount of Roth IRA Contributions That You Can Make for 2018,” https://www.irs.gov/retirement-plans/plan-participant-employee/amount-of-roth-ira-contributions-that-you-can-make-for-2018. Note there are some ways individuals can skirt the income limits. See Keith H. Clark, "How to Leverage a Backdoor Roth IRA," Kiplinger, December 27, 2017, https://www.kiplinger.com/article/retirement/T046-C032-S014-how-to-leverage-a-backdoor-roth-ira.html.
 Office of U.S. Congressman David Brat.
 Raj Chetty et al., “Active vs. Passive Decisions and Crowd-Out in Retirement Savings Accounts: Evidence from Denmark,” The Quarterly Journal of Economics, 129 (3) (2014): 1141-1219. Chetty et al. also find that while 85 percent do not respond to tax subsidies, these “passive savers” can be convinced to save by measures such as automatic enrollment in savings accounts. See also Eric M. Engen, William G. Gale, and John Karl Scholz, “The Illusory Effects of Savings Incentives,” Journal of Economic Perspectives, Vol. 10 (4) (1996): 113-138.
 “Savings” is defined as financial and retirement assets. CBPP analysis of Survey of Consumer Finances (SCF) data. Methodology based on Edward N. Wolff, “Household Wealth Trends in the United States, 1962 to 2016: Has Middle Class Wealth Recovered?” NBER Working Paper No. 24085, November 2017, http://www.nber.org/papers/w24085.
 Joseph Lawler, “Republicans could target 401(k)s in tax reform gimmick,” Washington Examiner, September 25, 2017, https://www.washingtonexaminer.com/republicans-could-target-401-k-s-in-tax-reform-gimmick.
 This section contains data both on the income and wealth distribution depending on the data source. The general points about the distribution of the benefits from USAs mostly hold regardless of whether one focuses on the income or wealth distribution, given the strong correlation between the two. See Jonathan Fisher et al., “Inequality and Mobility using Income, Consumption, and Wealth for the Same Individuals,” National Poverty Center Working Paper Series 16-02, April 2016, http://www.npc.umich.edu/publications/u/2016-02-npc-working-paper.pdf.
 CBPP analysis of Survey of Consumer Finances (SCF) data. We assumed that assets identified in the SCF data that can be shifted into USAs include directly held pooled investment funds, directly held stocks, directly held bonds, other managed assets, savings bonds, certificates of deposit, and all liquid assets other than checking accounts and prepaid cards.
 Wolff, 2017.
 Workers under 50 can contribute each year up to $18,500 to their 401(k)s and $5,500 to their IRAs. Workers over 50 can contribute an additional $6,000 to their 401(k)s and $1,000 to their IRAs. Internal Revenue Service, “Retirement Topics - IRA Contribution Limits,” https://www.irs.gov/retirement-plans/plan-participant-employee/retirement-topics-ira-contribution-limits; and Internal Revenue Service, “IRS Announces 2018 Pension Plan Limitations; 401(k) Contribution Limit Increases to $18,500 for 2018,” https://www.irs.gov/newsroom/irs-announces-2018-pension-plan-limitations-401k-contribution-limit-increases-to-18500-for-2018.
 CBO, 2011.
 Calculations in this section do not include the effects of the Net Investment Income Tax for simplicity. If they did, they would show an even larger benefit for high-income taxpayers (an additional 3.8 cents on the dollar).
 This assumes the household takes the standard deduction.
 Internal Revenue Service, “Table 3.4: All Returns: Tax Classified by Both the Marginal Rate and Each Rate at Which Tax Was Computed, by Marital Status, Tax Year 2015 (Filing Year 2016),” https://www.irs.gov/statistics/soi-tax-stats-individual-statistical-tables-by-tax-rate-and-income-percentile.