Roth vs traditional: how to actually decide

A traditional 401(k) or IRA takes your money before tax, shelters everything it earns, and taxes each dollar you eventually withdraw as ordinary income. A Roth takes money that has already been taxed, shelters everything it earns, and taxes nothing on the way out. Same investments, same shelter, different collection point. The choice is almost never about the accounts themselves.

Which is why most of the arguments you hear are wrong in the same way. Tax-free withdrawals are not an advantage in isolation, and neither is an upfront deduction. If your marginal tax rate is the same in the year you contribute and the year you withdraw, the two accounts hand you an identical after-tax result, to the dollar. Everything that decides the question lives in the gap between those two rates.

The symmetry, in numbers

Take $8,000 of gross income, invested until it triples, at a marginal rate of 24% in both the contribution year and the withdrawal year.

Identical, and not by luck — multiplication does not care about order, so shaving 24% off the front and shaving 24% off the back are the same operation performed at different times. The refund a traditional contribution generates is not a bonus, and the tax-free withdrawal a Roth generates is not free money. Both are the same tax, collected once.

Change one rate at a time and the picture separates immediately:

Same $8,000 of gross income and the same tripling in every row; only the rates move. The rule falls out of the arithmetic without any judgment call: contribute where your rate is high, withdraw where it is low.

Your rate later is not simply your bracket later

The comparison above is honest but slightly rigged, because it treats the withdrawal rate as a single marginal number. In practice, contributions and withdrawals sit at opposite ends of your income stack.

A contribution comes off the top, so it saves you your marginal rate — the highest one you pay. Withdrawals fill your income from the bottom: the standard deduction absorbs the first slice at zero, the lowest bracket takes the next, and so on upward. A retiree whose income is mostly plan withdrawals therefore pays a blended rate noticeably below the marginal rate they were deducting at while working. Comparing marginal-now to marginal-later overstates the Roth case, sometimes badly, and the fix is to compare marginal-now to the blended rate on the whole withdrawal — which the income tax calculator will show you on a realistic retirement income.

The exception matters too. A pension, meaningful taxable brokerage income, or an already-large pre-tax balance fills those low brackets first, and every plan withdrawal then stacks on top at the margin. If that describes your likely retirement, the blended-rate discount disappears and the naive comparison is closer to right.

Who leans Roth

Anyone in an unusually low-rate year is the clearest case. A resident, a first-year associate, a graduate student, a parent taking reduced hours, someone between jobs — deducting at the bottom of the schedule against an unknown but plausibly higher future rate is the weak side of the trade, and it is exactly the position most people are in during their first working decade.

There is a second, quieter Roth argument the table above hides, because that table held gross income constant. Contribution caps are stated in dollars going in, and the cap is the same number for both account types. So a saver who genuinely maxes out shelters more real purchasing power in a Roth, because a dollar of after-tax money is worth more than a dollar of pre-tax money carrying a future bill. The traditional saver's equivalent tax savings have to be invested somewhere unsheltered, where dividends and rebalancing are taxable every year. If you are hitting the cap rather than choosing a percentage, this is a genuine and frequently decisive edge — see the 401(k) guide for how the caps and the match interact.

Long horizons also help Roth in a way that has nothing to do with returns. Forty years of compounding on a traditional balance compounds the future tax bill alongside the balance; the Roth version simply does not have one to grow.

Run your numbersRoth vs traditional calculator→

Who leans traditional

Peak earning years, plainly. If you are deducting near the top of the schedule and expect a retirement funded mostly by drawing down accounts, the spread between your two rates is wide, real, and in your favor — this is the position where a deduction is worth the most and the eventual withdrawal is worth the least.

It is worth being skeptical of the common counterargument, that rates will inevitably rise so everyone should choose Roth. That is a forecast about future legislation, and it is competing against a much more reliable forecast about your own income: a career normally ends with the highest earnings the taxpayer will ever have, followed by a retirement funded at a lower run rate. Your personal rate trajectory usually swamps any plausible change to the schedule itself.

The deduction also has one property the Roth benefit lacks — certainty. You know your rate this year exactly. Your retirement rate is a projection built on income you have not earned, spending you have not done, and rules that have not been written. When the two sides look close, the side you can measure deserves some weight.

The state-tax move

Federal law bars a state from taxing the retirement income of someone who no longer lives there, which makes state tax an unusually clean lever. A deduction taken against a high state rate, followed by withdrawals in a state with no income tax, adds a spread on top of the federal comparison that has nothing to do with brackets. The reverse works too: contributing to a Roth while living somewhere with no state income tax, then retiring to a high-tax state, locks in the good outcome permanently.

Most of this decision is a forecast. The state piece is closer to a plan, so if a move is genuinely likely, let it count.

RMDs and the flexibility tiebreakers

Traditional balances arrive with a schedule attached. Once you reach the age the law sets, a percentage of the balance must come out each year and be taxed, whether or not you want the income. That forced withdrawal is what turns a large pre-tax balance into a problem in its own right: it can lift the taxable share of Social Security, cross Medicare premium surcharge thresholds, and produce an effective marginal rate above the one you originally deducted at. Roth IRAs have never carried a lifetime withdrawal requirement, and Roth balances in workplace plans were brought into line by a recent change.

Two smaller tiebreakers point the same way. Roth IRA contributions can be withdrawn at any age, tax and penalty free, which makes a Roth IRA quietly serviceable as a backstop in a way a traditional account never is. And inheritance is asymmetric: an inherited traditional account hands a beneficiary a tax bill, usually compressed into ten years and landing in that person's own peak earning years, while an inherited Roth hands over the same money with the tax already settled.

Holding both beats getting it right

The real reason to split is not that you cannot predict your future bracket. It is that two buckets give you a control one bucket does not.

A retiree with balances on both sides chooses, every year, how much taxable income to realize: draw traditional money up to the top of a low bracket, then top up from the Roth without adding a dollar of income. That single dial governs the taxable share of Social Security, Medicare surcharge thresholds, the rate applied to capital gains, and marketplace subsidies before Medicare starts. Someone holding only traditional money has no dial at all. Someone holding only Roth money has low brackets going unused every year, which is its own kind of waste.

What to actually do

Take the full employer match before anything else — the match is worth more than any bracket argument, and its tax treatment is a detail by comparison. Then use the rate comparison: an unusually low-rate year goes Roth, a peak-earning year goes traditional, and the wide middle is best served by splitting rather than agonizing. A common and defensible default is traditional inside the workplace plan, where the balance grows large, and Roth in the IRA, where the flexibility is most useful.

If your income blocks a direct Roth IRA contribution, the backdoor route still gets Roth money in. And if the split ends up lopsided toward pre-tax — as it usually does for high earners — the low-income years between leaving work and claiming benefits are when a conversion ladder can rebalance it at rates you will never see again.