Traditional vs Roth: It's a Marginal Rate Question, Not an Age Question

Start with the math that makes it a wash

Before the interesting complications, it helps to see the baseline case, because it clarifies what the decision is really about.

Suppose you have $10,000 of pre-tax income, your marginal rate is 24% today, the investment triples over your holding period, and your rate at withdrawal is also 24%.

• Traditional: the full $10,000 goes in untaxed. It triples to $30,000. You withdraw and pay 24%, leaving $22,800.
• Roth: you pay 24% first, so $7,600 goes in. It triples to $22,800. You withdraw tax-free, leaving $22,800.

Identical. This is not a coincidence — multiplication is commutative, so it makes no difference whether you apply the growth factor or the tax factor first. When the two rates are equal, the accounts are mathematically equivalent.

Which means the entire decision reduces to one question: is your marginal rate today higher or lower than the rate that will apply to those dollars when you withdraw them? Everything else is a refinement on that.

The asymmetry nobody mentions: which rate applies

Here is the subtlety that flips a lot of intuitions, and it is the single most useful thing in this article.

A traditional contribution is deducted from your top dollars. If you earn $120,000 as a single filer in 2026, your last dollars sit in the 24% bracket, so a deduction saves you 24 cents on the dollar. That is a genuine marginal rate.

But withdrawals in retirement do not come off the top of anything. They fill the brackets from the bottom. A retiree with no other income who withdraws $60,000 from a traditional IRA gets the standard deduction first — $16,100 for a single filer in 2026 — then fills the 10% bracket, then the 12% bracket, and only the remainder reaches 22%. Their blended rate on that withdrawal is far below 22%.

This is why comparing "my bracket now" to "my bracket then" systematically overstates the case for Roth. You are deducting at a marginal rate and withdrawing at something closer to an effective rate. For a great many people, traditional wins on this asymmetry alone.

You deduct off the top and withdraw from the bottom. Comparing one marginal rate to another marginal rate quietly stacks the deck toward Roth.

The 2026 brackets, for reference

You cannot run this comparison without knowing where the lines are. For tax year 2026, single filers face:

• 10% on taxable income up to $12,400
• 12% from $12,400 to $50,400
• 22% from $50,400 to $105,700
• 24% from $105,700 to $201,775
• 32% from $201,775 to $256,225
• 35% from $256,225 to $640,600, and 37% above that

For married filing jointly the brackets are exactly double at every step up to 32%, running 10% to $24,800, 12% to $100,800, 22% to $211,400, 24% to $403,550, then 32% to $512,450, 35% to $768,700 and 37% above. Standard deductions for 2026 are $16,100 single, $32,200 married filing jointly and $24,150 head of household.

Note the shape of it. The jump from 12% to 22% is the largest single step in the entire schedule — nearly doubling. Someone whose income sits near the top of the 12% bracket is in an unusually favorable position for Roth contributions, because they are paying a genuinely low rate now and any withdrawal in retirement is likely to be taxed at 12% or more. Someone in the 32% bracket is in the opposite position.

So why does "young means Roth" persist?

Because it is a reasonable proxy that people mistook for the rule itself. Three real effects sit underneath it.

Young workers often have low current rates. A 24-year-old in the 12% bracket genuinely is unlikely to face a lower rate in retirement. But a 24-year-old software engineer earning $160,000 is in the 24% bracket, and the logic does not apply to them at all.

Longer horizons mean more growth inside the shelter. This is often stated as a reason Roth wins, and it is usually stated wrong. As shown above, growth is rate-neutral when the two rates match. More growth does not favor Roth by itself; it amplifies whichever side the rate comparison already favored.

Contribution limits create a real, genuine asymmetry. This one is legitimate and underappreciated. The 2026 limit of $24,500 is a limit on dollars contributed, not on after-tax value. A maxed Roth 401(k) at $24,500 shelters $24,500 of after-tax money; a maxed traditional 401(k) at $24,500 shelters what is really about $18,600 of after-tax money at a 24% rate, with the remaining value sitting as a deduction you spend or invest in a taxable account. For someone who reliably maxes out, Roth effectively lets more real wealth into the shelter. This is the strongest argument for Roth that does not depend on predicting future rates.

The factors that sit outside the rate comparison

Several considerations do not show up in the two-rate math but change real outcomes.

• Required minimum distributions. Traditional balances must eventually be drawn down starting at age 73 or 75 depending on birth year, whether you need the money or not, and those distributions can push you into a higher bracket. Roth IRAs have no RMDs during the owner's lifetime, and Roth 401(k)s were exempted as of 2024.
• Medicare premium surcharges. IRMAA surcharges on Medicare Part B and D are based on modified AGI from two years prior. Large traditional withdrawals raise MAGI; qualified Roth withdrawals do not. Crossing an IRMAA threshold by a single dollar raises premiums for the entire year.
• State tax arbitrage. Contributing while living in a high-tax state and withdrawing while living in a state with no income tax is a real, sizable advantage for traditional — and it runs the other way if you plan to retire somewhere with higher taxes than where you work now.
• Heirs. Inherited traditional accounts generally must be emptied within ten years, and every dollar is taxable to the beneficiary at their rate — frequently their peak earning years. Inherited Roth accounts follow the same ten-year timeline but come out tax-free.
• Tax diversification. Holding both gives you a lever in retirement: you can draw from traditional up to the top of a low bracket and take anything further from Roth, managing your own taxable income year by year. That optionality has value that no single-rate projection captures.

A framework you can actually use

Lean traditional if you are in the 24% bracket or above, you expect materially lower income in retirement, you live in a high-tax state and may leave it, or you cannot max out the limit anyway — in which case the deduction is real money you can save elsewhere.

Lean Roth if you are in the 10% or 12% bracket, you are early in a career with a clear path to much higher earnings, you reliably max out contributions, you expect a pension or large Social Security benefit that will keep retirement income high, or you are specifically trying to leave tax-free assets to heirs.

And if you genuinely cannot tell — which is the honest answer for most people in the 22% bracket looking thirty years ahead — split the contribution. Dividing between the two is not indecision; it is buying optionality against a rate you cannot forecast. Nobody knows what the brackets will look like in 2056, and a strategy that only works if you guessed right is a fragile strategy.