Traditional versus Roth is one comparison of two tax rates
The claim
For the same cost to today's take-home pay, a traditional and a Roth account differ only by the ratio (1 - retirement rate) / (1 - current rate); growth rate, horizon and contribution size cancel exactly.
Why we rate it strong
The core result is algebra, not an empirical finding, and it holds for any return and any horizon. The behavioural corollary — that people do not adjust their contribution when the tax treatment changes — is measured separately in plan records.
Put the same amount of after-tax money into each account. In a traditional 401(k) the deduction means that net cost buys a larger gross contribution, which grows and is taxed on withdrawal. In a Roth it buys a smaller gross contribution, which grows and is not. Write both out and the growth multiple appears on both sides and cancels, as does the contribution. What survives is (1 - rate in retirement) divided by (1 - rate today).
So the question is only whether your marginal rate will be lower when you draw the money than it is when you earn it. For most people mid-career it will be: retirement income is usually lower than working income, and the standard deduction plus the shape of the brackets means the first tranche of withdrawals is taxed very lightly. Early in a career, in a low bracket, or in a year with unusually little income, the comparison flips.
The argument people actually hear — that a Roth grows tax-free — is true and not decisive, because the traditional account also grows untaxed. Tax is levied once in each case; the only question is at which end, and at what rate.
One real asymmetry survives the algebra, and it only binds at the ceiling. The statutory limit caps the gross contribution, not its after-tax value. Someone contributing the maximum shelters more real money in a Roth than in a traditional account — and pays correspondingly more for it out of this year's pay. Beshears and co-authors found that when employers added a Roth option, employees did not lower their contribution percentage to compensate, so choosing Roth quietly raised their real saving rate.
Two rules of thumb follow, and both are weaker than the arithmetic. Rate uncertainty is symmetric, so splitting between the two is a hedge rather than a mistake. And tax law changes over thirty years in ways nobody forecasts, which is an argument for holding some of each rather than for a confident bet on either.
Where this breaks down
- The comparison assumes you actually invest the traditional account's tax saving rather than spending it. If it is spent, the traditional side is smaller than the algebra says.
- Marginal rate in retirement is not a single number. Withdrawals interact with Social Security taxation, Medicare premium brackets and required minimum distributions, and those interactions can push the effective rate above the headline bracket.
- State income tax can dominate. Earning in a high-tax state and retiring in a no-tax one favours traditional by more than any federal comparison, and the reverse is also true.
- Employer matching contributions are traditional money regardless of which side you choose, so nobody ends up entirely Roth in a workplace plan.
Sources
Follow these rather than taking our word for the summary.
John Beshears, James J. Choi, David Laibson and Brigitte C. Madrian (2017). Does front-loading taxation increase savings? Evidence from Roth 401(k) introductions
Journal of Public Economics, 151, 84-95
Finding: Contribution rates were essentially unchanged after employers introduced a Roth option, so employees choosing Roth increased their after-tax saving without intending to.
Internal Revenue Service (2026). Retirement topics - 401(k) and profit-sharing plan contribution limits
irs.gov
Finding: The 2026 elective deferral limit is $24,500, with an $8,000 catch-up from age 50 and $11,250 for ages 60 to 63; total annual additions are capped at $72,000.