The Roth 401(k) contribution was introduced by the Economic Growth and Tax Relief Reconciliation Act of 2001 (EGTRRA). It was then (and has been since) largely a budget gimmick – a way to reduce the size of the ‘tax expenditure’ on retirement savings.
Further ‘Rothification’ of Tax Code retirement savings provisions is currently being considered as a way to finance other tax policy objectives, like reducing rates. This strategy is clearly behind Congressman Dave Camp’s (R-MI), Chairman of the House Ways and Means Committee, proposal to mandate that 401(k) plans include a Roth option and provide that, after a participant has ‘used up’ half her 401(k) contribution limit with ‘regular’ 401(k) contributions, additional contributions have to be Roth. (See our article Congressman Camp’s comprehensive tax reform proposal.)
Minor Roth changes have been used (and continue to be proposed) as a way to raise revenues to pay for policy initiatives unrelated to retirement savings. For instance, the Small Business Jobs Act of 2010 (SBJA10) included, as a revenue-raiser, a provision allowing in-the-plan Roth conversions in certain circumstances. Proposals to further liberalize Roth conversion rules often come up when policymakers are searching for revenue.
From the point of view of plan participants and plan sponsors, Roth math can be a little confusing. For some participants, Roth contributions will produce greater benefits (net of taxes) than regular contributions. For others, they produce smaller benefits. Which outcome applies often depends on the participant’s marginal tax rate when the contribution is made and when it is distributed.
In this article we begin with a discussion of Roth tax benefit math (that is, Roth math from the participant’s point of view) – the relative value of Roth tax benefits in different scenarios. We’ll then briefly describe Congress’s Roth budget math – which is what drives Roth policy generally. We’ll conclude with a brief review of possible Roth policy changes.
Roth vs. regular 401(k) contributions
The basics of participant Roth vs. regular contribution math can be described in the following three propositions:
1. Assuming identical tax rates at contribution and distribution, $1,000 saved as a Roth contribution produces the same tax benefit as it would if saved as a regular 401(k) contribution.
While not particularly intuitive, this equivalence is a commonplace. Here’s a simple version of the math: Assume two taxpayers in the highest marginal bracket (39.6%) at the time of contribution (year 1) and the time of distribution (year 10) and a 5% earnings rate. They both want to save $1,000. One makes a regular 401(k) contribution; one makes a Roth contribution. The following table shows what they will get at distribution.
Bottom line: the regular contributor doesn’t pay taxes on the way in, but does pay on the way out; the Roth contributor pays taxes on the way in but not on the way out. They both wind up with the same amount at distribution, net of taxes.
2. Assuming identical tax rates at contribution and distribution, the Roth maximum contribution produces greater tax benefits than the regular contribution.
Again, this proposition isn’t very intuitive. Because taxes have already been paid on the Roth contribution, it is, effectively, ‘bigger’ than a regular contribution. Thus, a $17,500 (the current limit on 401(k) contributions) Roth contribution produces a bigger tax benefit than a $17,500 regular contribution. Here’s the math: Assume two taxpayers in the highest marginal bracket (39.6%) at the time of contribution (year 1) and the time of distribution (year 10) and a 5% earnings rate. One makes a maximum $17,500 regular 401(k) contribution, one makes a maximum $17,500 Roth contribution. The following table shows what they will get at distribution.
The reason for this outcome is that the amounts going in aren’t both ‘apples’ (that is, they aren’t comparable). The Roth contributor started with $28,974, paid $11,474 in taxes (39.6%), and then contributed the $17,500 balance to the Roth 401(k). The regular contributor simply contributed $17,500. Taking account of tax effects, the Roth contributor contributed more.
3. If tax rates are higher in the year of contribution than in the year of distribution, regular contributions produce a bigger ultimate benefit net of taxes; if tax rates are lower in the year of contribution than in the year of distribution, Roth contributions produce a bigger benefit.
This is a relatively intuitive proposition. Evaluating it, however, is difficult. There are at least two major variables. The first is personal: An individual may expect to be in a higher tax bracket when she retires – a young worker may legitimately have this expectation. Or she may expect to be in a lower one, e.g., a ‘peak earner’ who doesn’t expect a rich retirement.
The second depends on federal tax policy: There are a lot of proposals to reduce marginal tax rates (in connection with the elimination or reduction of certain tax preferences). Congressman Camp’s proposal (noted above) would do so.
Let’s go back to our first table and compare outcomes where the highest marginal rate, for the year of distribution, is reduced to 25% (as it is, for instance (and with a lot of qualifications), in the Camp proposal).
Now let’s consider the opposite possibility, an increase in marginal tax rates. There are certainly those who argue for one. Here’s the same table with the highest marginal rate for the year of distribution set at 50%.
As noted, these results are pretty intuitive. In a sense, the Roth contributor is speculating on the size of future (that is, year-of-distribution) tax rates. If those rates remain the same as the year-of-contribution rates, his choice is tax neutral. If they go down, he ‘loses.’ If they go up, he ‘wins.’
The challenge for participants and sponsors
In a plan that offers both regular and Roth 401(k) contribution options, these complexities – the non-intuitive nature of the outcomes and their dependence on future unknowns – make participant decision-making difficult and, for some, daunting. In addition, there are other considerations – the 5-year holding period and qualified distribution rules may reduce the liquidity of Roth 401(k) contributions. And it’s generally thought that there is a behavior bias towards getting the 401(k) tax deduction up front.
For sponsors, in addition, Roth 401(k)s present administrative, recordkeeping and communications challenges. The result: some sponsors have been reluctant to include a Roth 401(k) option in their plans.
The tax benefit for 401(k) retirement savings – regular or Roth contributions – reduces tax revenues. This tax benefit is (along with, e.g., Child Tax Credit, the Earned Income Tax Credit, the mortgage deduction and health care tax benefits) on a short list of tax preferences that tax reformers have targeted. The strategy of the reformers is to reduce some or all of these tax preferences and use the resulting revenue increase to fund a decrease marginal tax rates.
Given the theoretical equivalence of the tax benefits for either regular or Roth contributions, you would think it wouldn’t matter which style of contribution was preferred. But it turns out that, because of the way the cost of tax preferences is calculated for Congressional budgeting, regular contributions ‘cost’ more than Roth contributions.
That cost, in lost tax revenues, of 401(k) plan tax benefit is generally calculated on a ‘cash flow’ basis. Estimates of the revenue loss generally include: (1) deductions/exclusions for contributions (revenue lost); (2) untaxed trust earnings (revenue lost); and (3) taxes paid on distributions (revenue gained). They may also include certain taxpayer behavioral responses to any proposed change. The element of this process that is critical for understanding the budget ‘magic’ of Roth 401(k) contributions is that these estimates are made for the current year and over a budget window of, generally, 10 years.
This ‘budget window cash flow’ approach may work well for certain tax expenditures, but many have criticized its application to 401(k) tax incentives. Assuming constant contributions, distributions, and tax rates, it is possible that a cash flow approach would capture, over time, the ‘true cost’ of 401(k) tax benefits. But given what has actually happened over the past 20 years, cash flow numbers are likely to overestimate the cost of regular (as opposed to Roth) 401(k) contributions. Because 401(k) plans are relatively new, and baby boomers have only just begun to retire, deductions/exclusions for contributions are relatively high, and taxes paid on distributions are relatively low.
Again, because of this budget window cash flow methodology, the cost of Roth 401(k) contributions is underestimated. Because Roth contributions are taxed upfront, more revenue comes in in the early period, that is, inside the budget window. Because the tax benefit is only ‘realized’ on distribution, more of the tax loss occurs in the later period, outside the budget window.
So for policymakers, simply changing regular 401(k) contributions to Roth contributions ‘increases tax revenues’ without affecting the real tax benefit (assuming constant tax rates.)
There are basically two types of Roth proposals being considered in Congress: (1) major, comprehensive changes to the 401(k) contribution rules designed to generate significant revenue to help fund cuts in marginal rates; and (2) less major (or even minor) changes designed to produce revenue for specific, lower cost policy proposals.
The Camp proposal
An example of the first type of proposal was included in Congressman Camp’s comprehensive tax reform proposal released in February 2014. Here are the Camp Roth provisions:
401(k) plan sponsors would generally be required to include a Roth contribution option in the plan (there would be an exception for certain small employer plans).
Oversimplified version: sponsors would have to include a Roth 401(k) option in their plans; and contribution maximizers would have to make half their contributions as Roth contributions.
For some participants, these changes might actually improve retirement outcomes (net of taxes). For others, e.g., those whose highest marginal tax rate in retirement is lower than it was when the contribution was made, they might worsen outcomes. It would, as we discussed above, make the entire process of figuring out what to contribute more difficult (and conceivably less appealing). And, because of the preference (bias towards) ‘deductions now, taxes later,’ participants would probably perceive it as a reduction in tax benefits, whatever the long-term outcome.
Sponsors would have to hire staff, build systems and develop communications to implement this Roth program. And it’s possible participant dissatisfaction and frustration with and bias against Roth contributions might reduce overall 401(k) plan participation. That would present a long-term retirement policy challenge: if your participants are not saving enough in the 401(k), how will you help them prepare for an adequately funded retirement?
Might something like this pass? Unlikely, but maybe, if the House, the Senate and the Administration return to the idea of a ‘grand bargain’ on the budget, entitlements and tax reform, or even if they just try to do a stand alone bi-partisan tax reform bill – it has happened before. These possibilities will depend in part on the results of the November 2014 elections.
With respect to the second type of proposal – a less comprehensive Roth 401(k) change designed to raise revenues for a specific policy initiative – an example from the past is the (as noted above) change included in SBJA10. That change generally allowed participants to convert money in their regular 401(k) account to a Roth 401(k). Generally, only money that could have been distributed was eligible for this treatment. In connection with the conversion, the participant would pay taxes (and then pay no taxes on actual distribution) – hence the short-term revenue-raising appeal. This was a very minor change – the participant could, before SBJA10 passed, get the same tax result by simply taking a distribution, rolling it into an IRA, and then converting the IRA to a Roth IRA.
A bigger change, that would produce more revenue, would be to allow the conversion of non-distributable regular 401(k) accounts to Roth accounts. Proposals to do this have (informally) been considered and are likely to continue to come up when Congress is hunting for revenue to fund proposals (generally bi-partisan and generally popular) that cost money.
These more targeted Roth proposals are generally voluntary both at the sponsor level (e.g., the sponsor generally does not have to implement a Roth conversion program if it doesn’t want to) and at the participant level (e.g., the participant doesn’t have to convert). One can imagine a scenario in which they are made mandatory.
Revenue policy not retirement policy
Finally, let’s note that none of these policy initiatives have anything to do with retirement savings policy. They are only being considered because they raise revenues that (generally) will be used for other, non-retirement policy purposes. Nevertheless, they produce real world consequences for participants and sponsors.
401(k) plan sponsors will want to consider familiarizing themselves generally with the ‘Roth option’ and related issues – we may be dealing with more of them in the relatively near future.