Roth vs. Traditional 401(k) for High Earners: A Worked Example

This article draws on work published at Kitces.com: Michael Kitces’s writing on the tax-equivalency principle, what tilts the Roth decision, and the limits of tax diversification (2009, 2012, and 2021), and Ben Henry-Moreland’s 2024 article on pre-tax contributions in peak earning years. We’d recommend reading them in full. We also credit Jeff Levine and Ed Slott where their ideas appear. The worked household, the figures, and the arithmetic below are our own.

Open a 401(k) enrollment screen and one of the first choices is the one most people make once and never revisit: traditional or Roth. A traditional, or pre-tax, deferral comes out of the paycheck before income tax, grows untaxed, and is taxed as ordinary income when it is withdrawn. A Roth deferral is taxed now, grows untaxed, and comes out tax-free in retirement, once the account is five years old and the saver is 59½. The investments inside can be identical. The only thing being chosen is when the tax is paid.

For a high earner that choice moves real money every year. In 2026 a worker can defer $24,500 into a 401(k), so a two-earner couple can defer $49,000, before catch-up contributions. In the 32% bracket, sending that $49,000 in pre-tax lowers this year’s federal income tax by $15,680. Sending it in as Roth gives that up in exchange for tax-free withdrawals decades from now. Unlike a Roth IRA, which a married couple cannot fund directly once income reaches $252,000 in 2026, a Roth 401(k) has no income limit, so the choice is open to nearly everyone whose plan offers both.

Most of what gets written about this decision argues from a prediction: tax rates are going up, so pay now; or you will be in a lower bracket later, so pay later. This article works the question the way we work it for a client, by running one household through fifteen working years and twenty-six retirement years and letting the retirement tax rate come out of the model instead of going into it as a guess. Then it asks the question most comparisons skip: what happens to the money the pre-tax saver does not send to the IRS?

The Whole Decision Is One Comparison

Start with the arithmetic that sits under everything else. Put $10,000 of salary toward retirement. As a Roth deferral in the 32% bracket, $6,800 goes in after tax. As a pre-tax deferral, all $10,000 goes in. Let both triple. The Roth account holds $20,400, all of it spendable. The pre-tax account holds $30,000, and if it comes out at 32%, $20,400 is left. The two are identical. Growth does not favor either account, and neither does the number of years, because multiplying by one minus the tax rate at the start gives the same answer as multiplying at the end.

Michael Kitces calls this the tax-equivalency principle, and it clears away a good deal of folklore. A Roth is not better because the growth is tax-free; the pre-tax account’s growth on the government’s share was never the saver’s to keep. A pre-tax account is not better because the deduction arrives sooner. If the rate going in equals the rate coming out, the two routes tie.

So the decision turns on one comparison: the rate the deferral saves now against the rate the same dollars will face later. Change the example so the pre-tax money comes out at 22%, and the pre-tax saver keeps $23,400 against the Roth saver’s $20,400. Bring it out at 37% and the pre-tax saver keeps $18,900. Every other consideration in this article is a refinement of that one comparison, or a reason the comparison is harder to make than it looks.

It is harder than it looks for two reasons. The first is that the rate coming out is not a single number. The dollars from this year’s deferral do not come out alone; they come out stacked on top of everything else the household will have in retirement: Social Security, the balances already saved, any pension or rental income. A high earner’s next pre-tax dollar comes out at the top of that stack, which is why the average rate a retiree pays can understate what the last dollars cost. The second is that the rate has to be judged over a period that may run fifty years, across a retirement, the death of a spouse, and an inheritance. That is a job for a projection, not a rule of thumb.

What We Want to Know First

When a client in a high bracket asks which box to check, the conversation starts with questions, because each answer moves one side of the comparison.

What is saved already, and where? A household with most of its savings in pre-tax accounts has already filled a good part of its future brackets. One with large taxable or Roth balances has room.

What is income likely to do from here? A pre-tax deferral made while income is still climbing saves tax at a lower rate than one made at the peak. A planned sabbatical, a job change, a move to part-time work, or a large liquidity event such as a business sale each change the rate on one side of the comparison or the other.

What will retirement cost, and what will pay for it? Spending drives withdrawals, and withdrawals drive the rate. A portfolio alone produces a different tax picture than a portfolio alongside Social Security, a pension, deferred compensation, or rents.

How long might each spouse live? A long retirement gives required distributions more years to build. The years after the first death are taxed at single-filer brackets on much of the same income, a problem we cover in the widow’s tax penalty.

Where is the money going in the end? Pre-tax dollars given to charity are never taxed at all. Pre-tax dollars left to children are taxed at the children’s rates, usually within ten years and often in their own peak earning years.

None of these has a universal answer, which is the point. Two households with the same salary can reasonably check different boxes.

A Worked Household: 32 Cents In, About 25 Cents Out

Consider a hypothetical couple, Mark and Laura, both 50 in 2026 and living in Houston. Together they earn $520,000 and take the standard deduction, which puts their taxable income inside the 32% federal bracket whether they defer pre-tax or Roth. Their pay keeps pace with inflation, so they stay there. They have $900,000 in pre-tax retirement accounts and $300,000 in a taxable brokerage account. Each defers the full $24,500, and they keep doing so, with the limit rising with inflation, for fifteen years. At 65 they retire, spend $240,000 a year after tax in today’s dollars, and claim a combined $84,000 of Social Security at 70.

The model runs every year from 50 to 90. It assumes a 10% annual return while the couple is working, 7% in retirement, and 3% inflation. It uses the 2026 contribution limit and 2026 federal tax law, carries both forward with brackets and limits rising with inflation, and leaves unindexed the thresholds the law itself does not index, such as the ones that decide how much of Social Security is taxed. The temporary senior deduction has expired by the time they retire. In retirement the couple follows the approach we describe in the order of retirement withdrawals: spending from the taxable account first, with Roth conversions through age 74 sized each year to keep the household’s capital gains in the 0% rate, and required distributions beginning at 75. Once the taxable account is spent, our model draws pre-tax money only up to the top of the 12% bracket and takes the rest from Roth while there is Roth to take. For this couple the sizing rule produces no conversions in any of the four versions: while the taxable account is paying the bills, the gains realized to raise $240,000 a year already fill the 0% range, and once it is spent there is nothing outside the retirement accounts to pay a conversion’s tax. Catch-up contributions are left out, because for a spouse whose prior-year Social Security wages from the plan’s employer exceeded $150,000 the law requires those to be Roth either way. Any employer match is left out too, because it is the same in every version. Every figure is in 2026 dollars. The returns are assumptions for an illustration, not a forecast, and the result moves with them, as a later section shows.

Four versions of the same couple go through it. In the first, every deferral is pre-tax, and the $15,680 of tax saved each year is invested in the taxable account. In the second, every deferral is Roth. In the third, deferrals are split half and half, with the tax saved on the pre-tax half invested. The fourth is the first version with one change: the tax saving is spent.

At 65 the pre-tax couple holds $3,707,600 in pre-tax accounts and $1,177,100 in the taxable account. The Roth couple holds $2,413,100 pre-tax, $1,294,500 in Roth accounts, and $777,800 in the taxable account. The pre-tax couple looks richer by about $399,300, but the comparison is not yet fair, because more of their money still has a tax bill attached.

Retirement settles it. With everything pre-tax, the first required distribution at 75 is about $179,600, and ordinary taxable income climbs to about $302,800 by 90, well inside the 24% bracket, which runs from $211,400 to $403,550 for a married couple in 2026. Over the whole retirement, 9.8% of their ordinary taxable income is taxed at 10%, 30.1% at 12%, 42.6% at 22%, and 17.5% at 24%. None of it is taxed at the 32% they deducted against. The Roth couple’s first required distribution is about $114,600, their ordinary taxable income never passes $208,300, and their federal income tax and Medicare premium surcharges over retirement total about $589,100, against about $1,072,700 for the pre-tax couple.

The pre-tax couple pays nearly twice the tax in retirement and still comes out ahead. Counting what is left at the end of age 90, with any remaining pre-tax money taxed to heirs at 22%, the pre-tax version finishes with about $2,781,900, the half-and-half version with about $2,718,500, and the Roth version with about $2,656,500. The pre-tax route leads the Roth route by about $125,400, or about 4.7%.

The model also answers the question the whole decision turns on: what rate did the deferred dollars come out at? The Roth account the couple built by 65 would be worth about $3,486,000 at the end of age 90 if nothing touched it. Holding the same money pre-tax instead costs about $856,300 by that date, in tax paid by the couple along the way and by their heirs at the end. That is 24.6% of the money. The deferrals went in at 32 cents on the dollar and came out at about 25.

Notice what the couple’s rate did not do. It did not fall to the low rate an average-rate calculation might suggest, because these were the top dollars on a stack that already included Social Security and distributions from the $900,000 they had saved before. And it did not stay at 32%, because this household, even with $3.7 million in pre-tax accounts, never reaches the $403,550 of taxable income where the 32% bracket begins.

Two assumptions carry weight in that result, and a third matters less than it might seem. The couple’s taxable account is tax-efficient: a 2% dividend yield, no trading along the way, and a stepped-up basis on whatever is left at death. A less tax-efficient account would narrow the pre-tax lead. Most of what the pre-tax couple has left at 90, about $3,105,600 before the heirs’ tax, is still pre-tax, so the rate their heirs pay matters a great deal, as the sections below show. The third is the retirement approach itself, which does not decide the comparison for this couple. We reran it with conversions large enough to fill the 12% bracket, then the 22% bracket, and with the plain rule of thumb of spending pre-tax money before Roth. Pre-tax finished ahead of Roth each time, by about $92,500, $175,500, and $216,300. None of those conversion plans left the pre-tax couple itself better off with heirs at 22%. Raising $240,000 a year from the taxable account already fills their first retirement years with capital gains, so this couple has no truly low-income years to convert in. Conversions earned their place only when the heirs’ rate was higher, as a later section shows.

The Part Most Comparisons Skip: What Happens to the Tax Saving

Look again at the two paychecks. A pre-tax deferral of $49,000 lowers the couple’s take-home pay by $33,320. A Roth deferral of $49,000 lowers it by the full $49,000. The pre-tax couple has $15,680 a year that the Roth couple does not, and over fifteen years that comes to $235,200 in today’s dollars.

Every comparison that shows pre-tax ahead, ours included, assumes that money is invested. In the model it is the entire advantage, and more. Investing the tax saving adds about $981,700 to what the pre-tax couple has at 90. Carrying a tax bill on the account costs about $856,300. The $125,400 lead is the difference between those two numbers.

Now run the fourth version, where the tax saving is absorbed into ordinary spending. The couple makes the same deferrals and retires on the same $240,000. They finish with about $1,800,200, which is $981,700 less than the couple who saved the difference and $856,300 less than the couple who chose Roth.

This is the part of the conversation we think gets too little attention. Many people talk about pre-tax and Roth and never address what happens to what was saved in taxes. In our experience many households simply spend it. No one decides to; the money arrives as a slightly larger paycheck and disappears into the month. For that household the Roth would have been better, and the reason is behavior, not tax planning: choosing Roth forces a higher effective savings rate. The same $24,500 deferral is a bigger commitment when it is made after tax, and the commitment happens automatically.

Jeff Levine has made a related point about Roth conversions: a conversion paid for from outside savings works like a forced savings account for the future tax bill, because people rarely connect a taxable account with the tax they owe on an IRA and may spend it. Our observation is the same idea one step earlier, at the contribution.

There is a second cost, and it is the one that compounds. Money that is spent every year becomes the household’s normal, which is what planners mean by lifestyle creep. A couple who lets the tax saving flow into the monthly budget for fifteen years has not only saved less. They have built a lifestyle that costs $15,680 a year more, and they will want to keep it in retirement. If they do, spending $255,680 a year, they finish with about $958,200. That is a little over a third of what the Roth couple has and about $1,823,700 less than the couple who saved the difference. The Roth couple never had a lifestyle to scale back, because that money never reached their checking account.

Put the results side by side and the pattern is hard to miss. Choosing between pre-tax and Roth moved the outcome by about $125,400. What the couple did with the tax saving moved it by $981,700 or more. A lower savings rate and a higher spending rate pull in the same direction, which is why this is a conversation about habits as much as taxes.

The couple who spent the saving did get something for their money. They spent an extra $235,200 over fifteen years, and spent intentionally, on things a household has decided matter to it, that can be the right use of it. Spending is not the mistake. The mistake is choosing the pre-tax box for its tax advantage and then letting the advantage disappear, because that leaves the household with the pre-tax account’s tax bill and without the savings meant to pay it.

A separate, mechanical point leans the same way, and it builds on what Kitces wrote in 2009 about contribution limits. Applied to a 401(k), the deferral limit is the same number for both kinds of contribution, but a Roth dollar is a whole dollar and a pre-tax dollar is partly the government’s. A saver who hits the limit shelters more after-tax wealth with Roth, and the pre-tax saver’s tax saving has to live in a taxable account, where dividends and gains are taxed along the way. That drag is visible in our numbers. The couple’s deferrals came out at about 25 cents, but pre-tax needed to save about 27.6 cents going in just to tie the Roth route. For this household, with its tax-efficient side account, pre-tax has to win by about three points before it wins at all, and the hurdle grows with a less efficient account and the longer it compounds.

The practical test is simple. If the tax saving has a destination, such as an automatic transfer to a brokerage account that matches it, the pre-tax math holds. If it does not, the math describes a different household.

What Moves the Answer

Thirty-two cents in and about 25 cents out sounds like a comfortable margin. After the hurdle it is a lead of under 5%, and it does not take much to close it. Each line below changes one thing about Mark and Laura and reruns all forty-one years.

The rate saved today. If their deferrals saved 35 cents on the dollar instead of 32, the pre-tax lead grows to about $220,100. If they saved 24 cents, Roth finishes about $104,500 ahead, and the half-and-half mix lands between the two. For this household the tie sits near 27.6 cents, between the 24% and 32% brackets, whose boundary is $403,550 of taxable income for a married couple in 2026, a jump Kitces has described as one of the natural break points in the decision.

What is already saved. If the couple reached 50 with $2,500,000 in pre-tax accounts instead of $900,000, their required distributions alone carry them into higher brackets, and Roth finishes about $57,400 ahead on a far larger base, or about $107,900 ahead with each couple on the best of the four retirement approaches we tested. On balances that size either figure is close to a tie. Ed Slott, one of the best-known advocates for Roth accounts, makes this point often: large pre-tax balances grow into large required distributions.

Other income in retirement. Add $100,000 a year of pension, deferred compensation, or rental income, and it fills the low brackets before the first 401(k) dollar comes out. The pre-tax lead disappears: Roth finishes about $7,800 ahead, another tie.

The survivor. If one spouse dies at the end of age 80, the survivor files as a single taxpayer, keeps two thirds of the Social Security, and spends 80% as much. The lead shrinks to about $75,200. Because we model a first death in every plan, that is closer to a real plan’s starting point than $125,400, and in our reruns a first death together with any one of the changes that favor Roth (a 24-cent rate, $2,500,000 already pre-tax, $100,000 of other income, or heirs at 32%) put Roth ahead, by about $34,800 to $149,100.

Who inherits. If the children who inherit the leftover pre-tax money pay 32% on it instead of 22%, the lead falls to about $15,500. If they pay 12%, it grows to about $235,400. Under the ten-year rule for inherited retirement accounts, an adult child’s own salary often decides which of those it is. Each version of the couple has a response to high-bracket heirs. In our reruns with heirs at 32%, the pre-tax couple did better converting in retirement, enough to fill the 22% bracket, than converting nothing, and the Roth couple did better spending its pre-tax money first and leaving the Roth to the children. With each couple doing that, Roth finishes about $55,300 ahead.

Charity. If the leftover pre-tax money goes to charity, no one ever pays tax on it, and the pre-tax lead grows to about $367,400. A household with real charitable intent, during life through qualified charitable distributions from an IRA or at death through a beneficiary designation, has a use for pre-tax dollars that a Roth cannot improve on. Slott himself, in making the case for Roth, advises keeping some pre-tax money for charitable giving.

Three of those single changes are enough to erase the pre-tax lead for this couple: a 24-cent rate, $2,500,000 already pre-tax, and $100,000 of other income, the last two by margins too small to call. Two changes together leave less doubt. With $2,500,000 already pre-tax and heirs at 32%, Roth finishes about $171,200 ahead. With $2,500,000 already pre-tax and $100,000 of other income, Roth leads by about $157,800. With a 24% rate today and heirs at 32%, Roth leads by about $199,200, and with a 24% rate and $2,500,000 already pre-tax, by about $272,700.

That pattern is how we read the decision in practice. For Mark and Laura the lean toward pre-tax held with none of these features, came close to a toss-up with most of them taken one at a time, and turned toward Roth with two. A real household’s count comes from its own projection. A physician couple in their early forties with modest balances and a steep income curve ahead is in a different position than a couple at 58 with $3 million in rollover IRAs, a pension, and children who are attorneys, even if this year’s W-2s match.

Two more things move the answer that the model above holds fixed. Deductions and credits that phase out with income can make a pre-tax deferral worth more than the bracket rate. One example is the cap on the state and local tax deduction, which begins shrinking above $505,000 of modified adjusted gross income; for a household that itemizes and pays enough state and local tax to feel it, a deferral inside that range saves more than 32 cents on the dollar, though most Texas households do not pay enough for it to matter until well above that line. And a low-income year, between jobs or early in a career, reverses the comparison for that year alone, which is a reason to revisit the election whenever income changes instead of setting it once.

The last two things that move the answer are inputs to the illustration: the return it assumes and what the couple spends.

The assumed return. Every figure so far assumes 10% a year while the couple is working. At 8% the pre-tax lead is about $156,500, at 9% about $150,400, and at 11% about $65,500. The direction matters more than any one of those numbers. The faster the accounts grow, the larger the pre-tax balance at 65 (about $2,929,100 at 8% and $4,172,000 at 11%), the larger the required distributions, and the higher the brackets they reach. A Roth balance grows just as fast and has no bracket to climb. An 8% return in retirement in place of 7% leaves the lead at about $120,200. Lower returns also leave far less at 90 on the same spending, which is a different problem from the one this article is about.

Retirement spending. The lead depends on what the couple spends, too. At $220,000 a year it is about $87,800, at $200,000 about $54,000, and at $180,000 about $35,100: the less they spend, the more the pre-tax balance and its required distributions grow. At $260,000 a year it falls back to about $57,300, because the pre-tax couple’s larger withdrawals raise their tax and push them into a higher Medicare surcharge tier, while the Roth couple covers the extra spending from its Roth account. More spending does not simply mean a wider lead. In every one of these reruns pre-tax still finishes ahead. The size of the lead is what moves.

What If Tax Rates Go Up?

The strongest argument for Roth is about uncertainty. Today’s rates are known and tomorrow’s are not, the federal debt is large, and a Roth takes one variable off the table. Slott makes the first of those points often, and the argument deserves a number.

Our projections use current law, and then we stress the result. The 2025 tax law removed the scheduled expiration of today’s brackets, which means they have no sunset date, not that they cannot change. So suppose every ordinary income tax rate is one quarter higher throughout the couple’s retirement, with capital gains rates left where they are: 12% becomes 15%, 22% becomes 27.5%, 24% becomes 30%. Their deferrals were still deducted at 32%. The pre-tax lead of about $125,400 turns into a Roth lead of about $28,700, which leaves the two routes within about 1% of each other. That test leaves the heirs at 22%; with the higher rates and heirs in the 32% bracket together, Roth leads by about $123,800.

Even the higher rates stay under the 32% the couple saved, so why does the lead go? Because the margin was thin to begin with. Under today’s rates the deferrals came out at about 25 cents and needed to go in at about 27.6 just to tie, which left less than five of the 32 cents as a cushion. Rates a quarter higher on the couple’s own withdrawals use up that cushion. The result is a near tie, not a clear win for Roth. Ben Henry-Moreland’s point is that a household’s own rate tends to move more across its lifetime than the law moves it, and for many households that is the stronger force. In our example it is a weaker force, because this couple’s savings grow to nearly $5 million: retiring from a $520,000 income to a $240,000 lifestyle takes them from the 32% bracket only as far as the 22% and 24% brackets.

A move is the other stress worth running. A household that earns in Texas and retires to a state with an income tax generally pays that state’s tax on pre-tax withdrawals (some states exempt part or all of retirement income), having deducted against no state tax at all. With a flat 5% state tax throughout retirement, the couple’s pre-tax lead becomes a Roth lead of about $15,300, a tie for practical purposes. The reverse case, earning in a high-tax state and retiring to Texas, widens it, and the same logic is why Roth conversions look better after such a move, as we note in our checklist for moving to Texas.

Henry-Moreland adds a point about flexibility that we find persuasive. A pre-tax balance can be converted to Roth later, in a year of the household’s choosing, at that year’s rate. A Roth balance cannot be converted back. Choosing pre-tax in a high-income year keeps the option to pay the tax in a lower-income year, such as the years between retirement and required distributions that we describe in the retirement gap years. Levine’s answer for a top-bracket earner asking about a Roth conversion is often “not yet,” and the option to convert later fits that answer.

Splitting on Purpose

If the future is uncertain, why not do half of each? At times we lean that way, and the conversation is about hedging unknowns. It helps to be exact about what a split buys.

Kitces argued in 2021 that splitting is not diversification. Any change in future rates that helps the Roth half hurts the pre-tax half by a matching amount, so a split narrows the range of outcomes. Less regret if rates rise, less benefit if they do not. In our base case the half-and-half couple finishes at about $2,718,500, between the other two, giving up about $63,400 of the pre-tax lead in exchange for being less wrong if the future differs from the projection. In the version with rates a quarter higher, the same split finishes about $27,900 ahead of all pre-tax.

In principle a mix can also be more than a hedge, though our reruns give that only thin support. Because the tax code is progressive, the first pre-tax dollars a retiree withdraws fall in the low brackets and the last fall in the high ones, so a household whose projection sits near a bracket break can do better with some of each than with either alone. In our reruns the mix finished ahead of both pure choices once, in the version with the 5% state tax, by about $18,200 over all pre-tax and about $2,900 over all Roth. It also finished a little behind both in a few versions where the two were nearly tied, such as the one with heirs at 32%. Near a tie the mix can land on either side, which is a reason to find the answer in the household’s own numbers and not to default to 50/50.

So we lean toward a deliberate split when the direction is unclear, toward pre-tax in years that are clearly high-income, and toward Roth in years that are clearly low. The election can change every year, and over most careers it changes more than once.

Where the Catch-Up and the Mega Backdoor Fit

Two neighboring rules shape the mix whether a saver chooses or not. Starting in 2026, a worker 50 or older whose prior-year Social Security wages from the plan’s employer exceeded $150,000 must make catch-up contributions as Roth: the $8,000 catch-up, or $11,250 at ages 60 through 63. If Mark and Laura each earn more than that, as much as $16,000 a year of catch-up money, or $22,500 at ages 60 through 63, goes into Roth accounts no matter how they elect the base $24,500, which is one reason the base deferral can lean pre-tax without the household ending up with no Roth money at all. The mechanics are in the Roth catch-up rule.

Plans that allow after-tax contributions open a second door. The mega backdoor Roth uses space above the deferral limit, so it does not compete with the pre-tax deferral; it competes with the taxable account for money the household would save anyway. The order we usually see work for a high earner is the pre-tax deferral and the full employer match first, then the after-tax space, and it is also the cleanest answer to the behavioral problem above: a household that wants both the deduction and the forced saving can take the deduction on the first $24,500 and direct the tax saving into the after-tax layer.

The same rate comparison runs through decisions that look unrelated. It is why we lean toward pre-tax deferral first for most physicians in their peak years, as we describe in when a physician should start financial planning, and it is the whole of the Roth conversion question, where the household chooses the year instead of the paycheck.

How We Think About It

We lean toward pre-tax deferrals first for high earners in their peak earning years, and toward Roth conversions later, in the lower-income years when the household chooses the rate, sized each year from a multi-year projection. “Roth everything” is not our position. In the worked household the pre-tax dollars went in at 32 cents and came out at about 25, and the pre-tax route finished about $125,400 ahead. That is a lean and not a wide margin: rates a quarter higher in retirement, or a move to a state with a 5% income tax, were enough to bring the two routes within about 1% of each other.

That lean rests on a condition, and we treat the condition as part of the advice. The pre-tax answer assumes the tax saving is saved. When it would be spent, we lean toward Roth, because a Roth deferral works as a forced higher savings rate, and in our numbers the couple who chose Roth finished about $856,300 ahead of the couple who chose pre-tax and spent the difference. So one thing we check is where the tax saving goes. An automatic transfer that matches it is the simplest way to make sure it is saved.

In our plans the mix is modeled, not assumed. We start with the questions above: what is already saved and where, what income is likely to do, what retirement will cost and what will pay for it, how long each spouse might live, and where the money is meant to go. We project the household’s tax picture under current law, year by year, through retirement and through the survivor’s years. Then we talk through what higher rates would do to it and, where it applies, what a move to a state that taxes income would do. We lean toward revisiting the election each year, because a sabbatical, a job change, or a liquidity event changes the answer for that year.

We lean toward Roth as the facts stack up: a lower rate saved today, a pre-tax balance already large enough to fill the middle brackets, a pension or other income in retirement, or heirs likely to inherit in high brackets. In the worked household, one of those alone was enough to bring the two routes to a tie or to put Roth ahead. We lean toward a deliberate split when the direction is unclear, knowing it narrows the range of outcomes without raising the average. And we lean toward keeping enough pre-tax money for the giving a household intends, since dollars given to charity are never taxed.

The trade-off is that a projection is not a prediction. The pre-tax route bets that the household’s own income will fall more in retirement than tax law will rise, which has been a reasonable bet and is not a certain one. In our example the bet paid by a narrow margin, and the margin narrowed further as the assumed return rose. A Roth removes that uncertainty at a price, and for some households the certainty is worth paying for.

Common Questions

Is a Roth or a traditional 401(k) better for high earners?

For many high earners in their peak earning years, a traditional pre-tax 401(k) deferral comes out ahead of a Roth deferral, though often by a narrow margin, and only if the tax saved each year is invested and not spent. The deduction is taken at today’s high rate, and the withdrawals are usually taxed at lower rates in retirement. In our worked example, a couple in the 32% bracket saw their deferred dollars come out at about 25 cents each and finished about $125,400 ahead with pre-tax, a lead of under 5% that higher tax rates in retirement were enough to erase. We lean toward pre-tax first in peak years and toward Roth when the tax saving would be spent, when the bracket today is lower, or when large pre-tax balances, other retirement income, or high-bracket heirs are already in the picture.

Is there an income limit for Roth 401(k) contributions?

There is no income limit on Roth 401(k) contributions. Any employee whose plan offers a Roth option can use it, up to the $24,500 deferral limit for 2026, which pre-tax and Roth deferrals share. The income limit people remember belongs to the Roth IRA, which a married couple cannot fund directly once modified adjusted gross income reaches $252,000 in 2026.

What happens if tax rates go up after I choose a traditional 401(k)?

Higher tax rates narrow the advantage of a pre-tax deferral, and a large enough increase erases it. In our worked example, raising every ordinary income tax rate by one quarter throughout retirement, with capital gains rates left where they are, turned a pre-tax lead of about $125,400 into a Roth lead of about $28,700, leaving the two routes within about 1% of each other. A pre-tax balance can also be converted to Roth later, in a low-income year, and that option runs one way only: a Roth balance cannot be converted back.

Should I split my 401(k) contributions between Roth and traditional?

Splitting contributions between Roth and traditional is a hedge that narrows the range of outcomes; on its own it does not raise the expected result. Whatever helps the Roth half hurts the pre-tax half by a similar amount. In our worked example the half-and-half couple finished between the all-pre-tax and all-Roth versions, and in reruns where those two were nearly tied the mix landed slightly ahead of both once and slightly behind both in others. We lean toward a deliberate split when the direction of a household’s future tax rate is unclear, and toward one side or the other when it is not.

Does maxing out a Roth 401(k) save more than maxing out a traditional 401(k)?

A maxed-out Roth 401(k) shelters more after-tax wealth than a maxed-out traditional 401(k), because the $24,500 limit is the same for both and a Roth dollar has no tax bill attached. The pre-tax saver catches up only by investing the tax saved, in a taxable account where dividends and gains are taxed along the way. In our worked example, a couple who deferred pre-tax and spent the tax saving finished about $856,300 behind the couple who chose Roth.

Do high earners have to make catch-up contributions as Roth?

Beginning in 2026, workers 50 or older whose prior-year Social Security wages from the plan’s employer exceeded $150,000 must make 401(k) catch-up contributions as Roth. The rule covers the $8,000 catch-up and the $11,250 catch-up at ages 60 through 63. The base $24,500 deferral can still go in pre-tax, Roth, or a mix.

Can I choose a traditional 401(k) now and convert to Roth later?

Pre-tax 401(k) money can be converted to Roth later, usually after it is rolled to an IRA or through an in-plan conversion where the plan allows one. The converted amount is taxed as ordinary income in the year of the conversion, so the years between retirement and required distributions, when income is often lowest, are the usual window. That one-way option is a large part of why we lean toward pre-tax deferrals in peak earning years.