Slam the Door: Why Congress Should End the Backdoor Roth IRA

In September 2016, U.S. Senator Ron Wyden released a discussion draft of a “Retirement Improvements and Savings Enhancements Act” that would reform retirement planning policy.[1] Among the proposed reforms is ending “backdoor” Roth individual retirement account (IRA) conversions.[2] Backdoor Roth IRA conversion is a method for higher-income taxpayers[3] to avoid the income limits for Roth IRA contributions by rolling funds from a traditional IRA into a Roth IRA. This method provides a variety of tax advantages for taxpayers, including paying no tax on investment gains and avoiding required lifetime distributions.[4]

Senator Wyden is not the first politician to target this method. In his Fiscal Year 2016 and Fiscal Year 2017 budget proposals, then-President Obama advocated limiting Roth conversions to pre-tax dollars, curtailing the benefits that higher-income individuals receive from the conversion.[5] These proposals have yet to gain traction in Congress, but policymakers should continue to advocate for the elimination of the backdoor Roth IRA. In short, it is poor tax policy. Originally a legislative maneuver to facilitate reduced rates on capital gains and dividends, a decade later it is a costly tax expenditure that only benefits higher-income taxpayers and is removed from the policy goal of encouraging Americans, particularly lower- and middle-income Americans, to save for retirement.

This Essay discusses the Roth IRA within the current landscape of retirement policies, outlines the legislative history of the backdoor Roth IRA, and analyzes the mechanics and impact of the policy. The Essay then concludes that the backdoor Roth IRA is poor tax policy because it cuts revenue without furthering long-term legislative goals and argues that closing the backdoor Roth would be a small but easy step toward a more efficient and effective retirement tax policy.

Facilitating a Sound Retirement: Tax Expenditures and Retirement Planning

Tax expenditures like the backdoor Roth are a pervasive and much-debated component of tax policy.[6] They depart from the traditional tax policy goals of measuring income accurately and allocating tax burdens based on income and ability to pay. Instead, they attempt to guide taxpayers toward behaviors and activities that society has judged to be positive. Tax expenditures are government spending initiatives that are effected through special deductions, credits, deferred taxation, and other provisions of the tax code rather than through direct spending programs. They also function as alternatives to direct spending programs. Expenditures encourage a range of activities including home ownership,[7] charitable giving,[8] and investment in education.[9]

The government uses tax expenditures to encourage retirement savings through initiatives such as defined benefit and defined contribution plans,[10] the Saver’s Credit,[11] traditional IRAs,[12] and Roth IRAs.[13] These expenditures are a major component of the federal budget. In 2015, the Joint Committee on Taxation estimated that retirement expenditures would cost approximately $1 trillion from 2015 to 2019.[14] Traditional and Roth IRAs account for $116.7 billion of those expenditures.[15]

Tax expenditures encouraging retirement savings are arguably more important now than ever before. With the fate of Social Security at risk[16] and the number of Americans receiving pensions declining,[17] sound retirement policy is essential to prevent the elderly from slipping back into their status as a disproportionately impoverished age group.[18] These tax expenditures are particularly important for lower- and middle-income families because these groups have the most difficulty saving for retirement and are most vulnerable to the consequences of undersaving.[19] The legislative history of the backdoor Roth IRA reveals that facilitating secure retirements for these families was not the primary motivation for the policy.

Building the Backdoor: Legislative History and Congressional Intent

The backdoor Roth IRA came about as part of the Tax Increase Prevention and Reconciliation Act of 2005 (TIPRA),[20] serving to make the legislation revenue-neutral and thus smoothing its passage.[21] Because it was a reconciliation bill, TIPRA could not create a long-term deficit or it would trigger a point of order under Senate rules,[22] which would then require a sixty-vote override.[23] Senate Democrats were critical of TIPRA,[24] and Republicans held fifty-five seats in the Senate at the time the bill was under consideration by the Senate, not enough to override the point of order without bipartisan support.[25]

TIPRA’s headline provisions extended reduced tax rates on capital gains and dividends, resulting in long-term reductions in revenue.[26] Therefore, to ease its passage, TIPRA needed to include provisions that would raise revenue in order to offset the revenue reductions. However, because the Joint Committee on Taxation did not produce cost estimates for periods of more than ten years, the drafters of the legislation only needed to make TIPRA budget-neutral through 2015 to avoid the Senate point of order.[27]

Thus entered the backdoor Roth IRA. In 2006, the Joint Committee on Taxation estimated that over the first four years, the backdoor Roth IRA would increase tax revenues by approximately $5 billion as higher-income taxpayers took advantage of—and paid tax bills on—lump-sum rollovers.[28] The drafters of TIPRA slated the backdoor Roth provision not to kick in until 2010, thus allowing its revenue increase to fall within the ten-year revenue analysis window while avoiding the consideration of its long-term revenue losses.[29] The legislative strategy worked, and TIPRA passed without a Senate point of order.[30]

Congress was aware at the time of the backdoor Roth IRA’s passage that it would not facilitate greater retirement savings, particularly for those households for which increasing savings is most critical. As Brookings Fellow Peter Orszag warned Congress in 2005, “[r]ather than bolstering retirement security among middle- and lower-earners, proposals to increase income and contribution limits would generate significant asset shifting and be of primary benefit to households who are already disproportionately well-prepared for retirement.”[31] Instead, the driving force behind the backdoor Roth IRA was the need to facilitate the extension of capital gains and dividends rate cuts.[32]

Tax expenditures for retirement are well-established policy, designed to encourage retirement savings.[33] But retirement policy, and by extension tax expenditures for retirement savings, should not be designed to focus disproportionately on higher earners whose stable retirements are not in question, but should also target lower- and middle-income individuals whose retirements are more at risk.[34] As Michael Graetz explained almost three decades ago:

[I]t seems clear that national retirement policy includes two elements: (1) the maintenance of an adequate retirement income that will protect the elderly from widespread poverty, and (2) and income supplement to help ensure against an abrupt decline in a retiree’s lifestyle. From a public policy perspective, this implies that a higher percentage of preretirement wages must be replaced for low-and moderate-income workers to ensure that retirement not produce a less than adequate income . . . . It would be indefensible to consciously construct a national retirement security program that replaces a greater percentage of wages for high than for moderate and lower earners.[35]

An analysis of the mechanics and impact of the backdoor Roth IRA reveals that it does just that, taking away tax revenue while only benefitting higher-income taxpayers.

The Backdoor Roth IRA: Mechanics and Impact

The IRA is one category of tax expenditures encouraging retirement savings. IRAs are accounts in which a taxpayer deposits and invests funds that she can withdraw and spend in retirement. There are two broad categories of IRAs: traditional IRAs (both deductible and nondeductible) and Roth IRAs. The primary difference between traditional IRAs and Roth IRAs is the timing of the tax. Traditional IRAs defer tax. With deductible traditional IRAs, the taxpayer deducts the amount of the contribution from his income for the year of contributions and then pays tax on both the contribution and the gains when it is dispersed in retirement. With nondeductible traditional IRAs, the taxpayer pays tax on the contribution amount in the year of the contribution but does not pay tax on the gains until the income is dispersed in retirement. All taxpayers can contribute to a nondeductible traditional IRA, while income limits apply to deductible traditional IRAs.[36] The maximum annual contribution is $5,500 per person for both traditional and Roth IRAs combined.[37]

In contrast, Roth IRAs frontload tax payment. Individuals must fund Roth IRAs with post-tax dollars. But in retirement, the taxpayer receives both the contributions and the gains tax-free. Individuals with incomes under $133,000 and married couples with incomes under $196,000 can contribute directly to a Roth IRA.[38] While taxpayers must take required minimum distributions from traditional IRAs beginning at the age of seventy years and six months, there is no requirement to withdraw distributions from a Roth IRA.[39]

A backdoor Roth IRA allows a taxpayer whose income exceeds the threshold for a Roth IRA contribution to obtain the advantages of a Roth IRA by “rolling over” a traditional IRA into a Roth IRA. Many higher-income taxpayers have very large balances in their traditional IRAs that can be rolled over in one lump sum. Higher-income taxpayers can also take advantage of backdoor Roths by undertaking a sequence of “serial inversions” each year. Serial inversions involve a two-step process. First, the taxpayer can make a non-deductible contribution to a traditional IRA. Then, because there is no income limit to a Roth IRA conversion, the taxpayer can roll over the amount in the traditional IRA to a Roth IRA.[40] The taxpayer will owe tax on any amount in the traditional IRA that was deductible at the time of contribution as well as on any gain.

As explained above, at the time that the legislation creating the backdoor Roth IRA was passed in 2006, experts estimated that it would initially raise revenue because taxpayers would make large, lump-sum rollovers and pay a hefty tax bill. As noted above, the Joint Committee on Taxation predicted that backdoor Roth IRA conversions would raise about $5 billion in revenue during their first four years from 2011 to 2014.[41] However, the Joint Committee also predicted that beginning in 2015, five years after the introduction of the backdoor Roth IRA, the backdoor Roth IRA would cause the U.S. Treasury to lose revenue. The Joint Committee predicted lost revenue of $9 billion over the six years from 2015 to 2020.[42]

The advantages of the Roth IRA versus the traditional IRA for higher-income taxpayers are vast and varied. In a commentary written following the introduction of the backdoor Roth, David Halperin outlined the benefits that Roth IRA conversions provide to higher-income taxpayers.[43] Some of these advantages apply primarily to large, lump-sum conversions while others apply to both lump-sum and serial conversions.[44] First and most obviously, the conversion removes the income cap on eligibility for tax-preferred Roth IRAs.[45] Second, assuming that some tax was due at the time of the conversion (for example, tax owed on gains within the traditional IRA), the taxpayer increases his relative amount of tax-preferred savings when he uses other savings to pay that tax.[46] Third, because there is no requirement for distributions on a Roth IRA during the taxpayer’s lifetime, the tax preferences can be prolonged.[47] Fourth, assuming that some tax was due at the time of the conversion, the size of the taxpayer’s taxable estate is reduced by the amount of the payment.[48] Finally, the taxpayer is able to lock in current tax rates, eliminating the risk of future rate increases.[49]

While much of the contemporary commentary has focused on the impact of lump-sum backdoor Roth conversions,[50] serial conversions also have the ability to substantially reduce revenue. Imagine a higher-income Taxpayer Y who places $5,500 in a nondeductible traditional IRA each year from the age of twenty-five until sixty-five. Without a Roth IRA conversion, at the end of forty years, assuming a seven percent return, Taxpayer Y will have approximately $1,098,000 in her traditional IRA, of which $878,000 is attributable to gains. When the funds are distributed, she will pay ordinary income rates on the gains, leaving a post-tax value of $746,800.[51]

Now assume the same fact pattern, except that Taxpayer Y rolls over her traditional IRA contributions to a Roth IRA each year following her contribution. After forty years, Taxpayer Y will have $1,098,000 in her Roth IRA, just as in the traditional IRA. However, in this case any distributions will be tax-free—meaning that she is able to keep the full $1,098,000, even post-tax. Thus, the U.S. Treasury loses $351,200 in revenue from this one taxpayer because of serial conversion from a traditional to a Roth IRA. And there is evidence that this practice is widespread: Vanguard estimated that in 2013, more than twenty thousand of its customers were completing Roth IRA conversions.[52]

In consequence, backdoor Roth IRA conversions have the potential to produce significant declines in U.S. Treasury revenues in coming decades, creating a lasting and damaging impact on the budget.

Recommendation

The backdoor Roth IRA bleeds revenue without encouraging retirement savings among the taxpayers for whom it is most important: lower- and middle-income households who have the most difficulty saving and are most vulnerable to the consequences of undersaving. Sound retirement policy is a critical goal, and the United States should focus its tax expenditures on policies that actually change behaviors. In our current polarized political environment, it may not be feasible to divert funds currently expended on the backdoor Roth IRA to programs targeted toward lower-income taxpayers, such as the Saver’s Credit. However, both sides of the political spectrum should be able to support the notion that we should not continue a tax expenditure that is unlikely to change taxpayer behaviors. Closing the backdoor Roth IRA is a simple first step toward a more efficient and effective retirement policy.


[1].   Press Release, U.S. Senate Comm. on Fin., Wyden Proposal Would Crack Down on Tax Avoidance in Retirement Plans, Create New Opportunities for Working Americans To Save (Sept. 8, 2016), http://www.finance.senate.gov/ranking-members-news/wyden-proposal-would-… [http://perma.cc/2EH6-APAS].

[2].   See Senator Ron Wyden, Summary of Discussion Draft: Retirement Improvements and Savings Enhancements (RISE) Act of 2016, Senate Fin. Committee 2 (Sept. 8, 2016), http://www.finance.senate.gov/imo/media/doc/RISE%20Act%20discussion%20dr… [http://perma.cc/AV3U-V979].

[3].   For the purposes of this Essay, “higher-income taxpayers” are defined as individuals with incomes over $133,000 per year and married couples with incomes over $196,000 per year. These figures represent the income thresholds at which taxpayers can no longer contribute directly to a Roth IRA. See U.S. Dep’t Treasury, Internal Revenue Serv., Publication 590-A, Contributions to Individual Retirement Arrangements (IRAs) 2 (2016) [hereinafter Publication 590-A], http://www.irs.gov/pub/irs-pdf/p590a.pdf [http://perma.cc/RB98-WTWR]. 

[4].   See infra text accompanying notes 39, 43–49.

[5].   Excerpts Available of Obama Budget Analytical Perspectives, Tax Notes Today, Feb. 10, 2016; JCT Examines Revenue Provisions in Obama’s 2016 Budget Proposal, Tax Notes Today, Sept. 30, 2015.

[6].   A full discussion of the wisdom, efficiency, and impact of tax expenditures on behavior and economic equality is beyond the scope of this Essay. For insight on some of those issues, see generally Stanley S. Surrey, Pathways to Tax Reform: The Concept of Tax Expenditures (1973) (characterizing, in a seminal work, certain tax code provisions as expenditures rather than taxes); Daniel N. Shaviro, Rethinking Tax Expenditures and Fiscal Language, 57 Tax L. Rev. 187 (2004) (conceptualizing tax expenditures within the context of fiscal language issues more generally); Victor Thuronyi, Tax Expenditures: A Reassessment, 1988 Duke L.J. 1155 (presenting the definition of substitutable tax provisions as an alternative to tax expenditures); and Edward A. Zelinsky, Efficiency and Income Taxes: The Rehabilitation of Tax Incentives, 64 Tex. L. Rev. 973 (1986) (discussing why tax expenditures may lead to economic efficiencies).

[7].   See, e.g., I.R.C. § 163(h)(2)(D) (2012) (creating home mortgage interest deduction).

[8].   See id. § 170 (creating deduction for charitable giving).

[9].   See, e.g., id. §§ 221, 529 (creating deductions for student loan interest payments and 529 savings plans).

[10].  See id. §§ 401, 412.

[11].  See id. § 25B.

[12].  See id. § 408.

[13].  See id. § 408A.

[14] Joint Comm. on Taxation, 114th Cong., Estimates of Federal Tax Expenditures for Fiscal Years 2015–2019, at 39 (Comm. Print 2015), http://www.jct.gov/publications.html?func=startdown&id=4857 [http://perma.cc/8A6V-6C9Z].

[15].  Id.

[16].  Stephen C. Goss, The Future Financial Status of the Social Security Program, 70 Soc. Security Bull. 111, 111 (2010), http://www.ssa.gov/policy/docs/ssb/v70n3/v70n3p111.html [http://perma.cc/CX39-E9Q5] (estimating that Social Security costs will increase by 2035 so that current taxes will only be able to cover seventy-five percent of scheduled benefits).

[17].  See, e.g., Barbara A. Butrica et al., The Disappearing Defined Benefit Pension and Its Potential Impact on the Retirement Incomes of Baby Boomers, 69 Soc. Security Bull. 1, 1 (2009), http://www.ssa.gov/policy/docs/ssb/v69n3/v69n3p1.html [http://perma.cc/4X2F-25Z8] (noting the steady decline in defined benefit pension plans from 1980 to 2008). 

[18].  The poverty rate among the elderly (those sixty-five and older) was sixty-five percent in 1960, twice the rate of non-elderly adults. By 1995, the poverty rate among the elderly was ten percent and therefore lower than the rate among non-elderly adults. See Gary V. Engelhardt & Jonathan Gruber, Social Security and the Evolution of Elderly Poverty 2 (Nat’l Bureau of Econ. Research, Working Paper No. 10466, 2004), http://www.nber.org/papers/w10466.pdf [http://perma.cc/2AYL-94U5]. 

[19]Joel Slemrod & Jon Bakija, Tax Ourselves: A Citizen’s Guide to Debate over Taxes 286–87 (4th ed. 2008).

[20].  Tax Increase Prevention and Reconciliation Act of 2005, Pub. L. No. 109-222, 120 Stat. 345 (2006).

[21].  See Joel Friedman & Robert Greenstein, Joint Tax Committee Estimate Shows that Tax Gimmick Being Designed To Evade Senate Budget Rules Would Increase Long-Term Deficits, Ctr. on Budget & Pol’y Priorities 1–3 (2006), http://www.cbpp.org/sites/default/files/atoms/files/4-25-06tax.pdf [http://perma.cc/2AHX-XNWA].

[22].  2 U.S.C. § 641(d)(2) (2012) (preventing changes to reconciliation legislation that would result in a net increase in the deficit); see also James V. Saturno, Cong. Research Serv., 97-865, Points of Order in the Congressional Budget Process 6 (2015); Robert Keith & Bill Heniff Jr., Cong. Research Serv., RL33030, The Budget Reconciliation Process: House and Senate Procedures (2005).

[23].  2 U.S.C. § 621 note (allowing an override of various provisions of the Congressional Budget Act of 1974, Pub. L. No. 93-344, 88 Stat. 297, including the provision codified in 2 U.S.C. § 641(d)(2), with the vote of three-fifths of the Senate).

[24].  Discussing the Bill, Senate Democrat Max Baucus stated: “It’s time to drop the obsession with capital gains and dividends tax cuts that don’t expire for years, and pass the tax cuts that are important to American taxpayers right now.” Press Release, Office of Senator Max Baucus, Senate Parliamentarian Upholds Senate Rules Against House Capital Gains and Dividends Measure in Tax Reconciliation (Feb. 1, 2006) (on file with author); see also Capital Gains, Dividend Rate Cut Extensions May Need Offset for Reconciliation Protection, BNA Daily Tax Rep. (Feb. 2, 2006) (noting Democratic opposition to the capital gains and dividend rate cut extensions included in TIPRA).

[25]Mildred Amer, Cong. Research Serv., RS22007, Membership of the 109th Congress: A Profile 1 (2006).

[26].  Tax Increase Prevention and Reconciliation Act of 2005, Pub. L. No. 109-222,
§ 102, 120 Stat. 345, 346 (2006).

[27].  See Friedman & Greenstein, supra note 21, at 1–3.

[28].  See Joint Comm. on Taxation, 109th Cong., Estimated Revenue Effects of Various Proposals Related to Roth IRAs (Comm. Print 2006), cited in Friedman & Greenstein, supra note 21, at 4 n.6.

[29].  Kurt Ritterpusch & Jonathan Nicholson, IRA Conversion Proposal Draws Attention as Means To Stave Off Tax Bill Point of Order, BNA Daily Tax Rep., Mar. 29, 2006.

[30].  Edmund L. Andrews, Senate Approves Extension of Bush Tax Cuts, N.Y. Times (May 12, 2006), http://www.nytimes.com/2006/05/12/washington/12spend.html [http://perma.cc/V92E-Y7SX].

[31].  Retirement Policy Challenges and Opportunities for Our Aging Society: Hearing Before the H. Comm. on Ways & Means, 109th Cong. 48 (2005) (statement of Peter R. Orszag, Retirement Security Project Director, The Brookings Institution) [hereinafter Orszag Statement].

[32].  See Friedman & Greenstein, supra note 21; Ritterpusch & Nicholson, supra note 29.

[33].  See Michael J. Graetz, The Troubled Marriage of Retirement Security and Tax Policies, 135 U. Pa. L. Rev. 851, 852 (1987) (describing tax expenditures encouraging retirement saving as “generally applauded” while noting the problems of relying on such expenditures); Norman Stein, Slouching Towards a Consumption Tax and the End of Retirement Income Security, 9 Fla. Tax Rev. 119, 125 (2008) (describing the wide range of tax expenditures encouraging retirement savings and explaining that “the commonly accepted rationale for the income tax expenditure is to help as many Americans as possible create income security for that period of life when they are no longer supporting themselves with wage income”).

[34].  See Orszag Statement, supra note 31, at 36 (encouraging stronger incentives for retirement savings for lower- and middle-income families); Daniel I. Halperin, Special Tax Treatment for Employer-Based Retirement Programs: Is It ‘Still’ Viable as a Means of Increasing Retirement Income? Should It Continue?, 49 Tax L. Rev. 1, 8 (1993) (encouraging a refocus of tax expenditures for retirement policy on programs focused on enhancing the retirement security of low and moderate earners).

[35].  Michael J. Graetz, Retirement Security Policy: Toward a More Unified Viewin Social Security: Beyond the Rhetoric of Crisis 91, 93–94 (Theodore R. Marmor & Jerry L. Mashaw eds., 1988).

[36].  If a taxpayer is covered by an employee-sponsored plan, contributions to traditional IRAs are fully deductible in the year of the contribution for individuals with incomes below $62,000 and married couples with incomes below $99,000. There are phase-out provisions allowing partial deductions for incomes between $62,000 and $72,000 for an individual and $99,000 and $119,000 for a married couple. No deduction is allowed for individuals making $72,000 or more or married couples making $119,000 or more. More generous income limits are available for those who are not covered by an employer-sponsored plan. Publication 590-A, supra note 3, at 2, 13.

[37].  Id. at 5.

[38].  Similar to the traditional IRA, there are phase-out provisions allowing only partial Roth IRA contributions for incomes between $118,000 and $133,000 for an individual and $186,000 and $196,000 for a married couple. See id. at 2.

[39].  Compare Treas. Reg. § 1.408-8 Q&A (1)(a) (as amended in 2014) (discussing traditional IRAs), with Treas. Reg. § 1.408A-6 Q&A (14)(a) (as amended in 2014) (discussing Roth IRAs).

[40].  Some commentators believe that step-transaction doctrine should apply to the backdoor Roth IRA, thus invalidating it. That question is beyond the scope of this Essay. For discussions on the backdoor Roth IRA and step-transaction doctrine, see J. William Harden & David R. Upton, Backdoor Roth: Still the Standard of ‘Can I Get Away With It, Tax Notes Today, Oct. 6, 2016; J. William Harden & David R. Upton, Backdoor Roths: The Ethical Standard of Can I Get Away With It, Tax Notes Today, Nov. 24, 2015; and Kaye Thomas, Step Transactions and Backdoor Roth Contributions, Tax Notes, Jan. 7, 2016.

[41]Joint Comm. on Taxation, supra note 28.

[42].  Id.

[43].  See David Halperin, Fun and Games with the Roth IRA, Tax Notes, July 10, 2006, at 167–69.

[44].  It is likely that most well-advised taxpayers with large balances in traditional IRAs completed lump-sum rollovers when the backdoor Roth IRA was first introduced, so the primary impact of the backdoor Roth IRA moving forward will be the impact of serial conversions. However, because no data is available to support that assumption, this Essay addresses the continuing impact of both lump-sum and serial conversions. 

[45].  Halperin, supra note 43, at 167.

[46].  Id.

[47].  Id.

[48].  Id.

[49].  Id.

[50].  See, e.g., Friedman & Greenstein, supra note 21; Halperin, supra note 43.

[51].  For simplicity, this example assumes a forty percent marginal tax rate. Individuals in the top federal income bracket currently have a marginal rate of 39.6%. See I.R.C. § 1 (2012).

[52].  IRA Insights: The Benefits of a “Backdoor” Roth, Vanguard 1 (2014), http://www.vanguard.com/pdf/ISGIRA9.pdf [http://perma.cc/H4C4-CRAU], cited in Thomas, supra note 40, at 116.