Pretax vs Roth: The Decision Most People Get Backwards

The Question Sounds Simple Until You Try to Answer It

Should you put your retirement contributions into the pretax bucket and pay tax on the withdrawals, or into the Roth bucket and pay the tax now? The conventional answer — "Roth if you expect to be in a higher tax bracket in retirement, pretax if you expect to be lower" — is technically correct and almost completely unhelpful. Nobody knows what their tax bracket will be in thirty years. The more honest version of the decision involves at least five other variables, and the way most people weigh them is backwards.

This post walks through the actual mechanics, the assumptions that change the answer, and the specific situations where the rule of thumb breaks. By the end you should be able to look at your own paycheck and make a reasoned choice rather than a vibes-based one.

The Mechanical Difference, Stated Cleanly

A pretax contribution (traditional 401(k), traditional IRA, traditional 403(b), traditional TSP) comes out of your paycheck before federal income tax is calculated. If you make $100,000 and contribute $10,000 pretax, the IRS treats your wages as $90,000 for income tax purposes that year. You do not pay tax on the contribution or its growth until you withdraw it in retirement, at which point every dollar — contributions and gains — is taxed as ordinary income.

A Roth contribution (Roth 401(k), Roth IRA, Roth 403(b), Roth TSP) is made with money that has already been taxed. Your paycheck shows the full income, you pay tax on it now, and the contribution goes into the account. Inside the account, growth is not taxed. When you withdraw in retirement, qualified distributions — contributions and gains — come out completely tax-free.

One critical thing to internalize: pretax does not reduce FICA. Social Security and Medicare withholding (the 6.2% + 1.45% on your paycheck) is calculated on gross wages before retirement deductions. Whether you go pretax or Roth, your FICA tax is identical. The only thing that moves is federal income tax — and, in most states, state income tax. If you plug numbers into our paycheck calculator and watch the "Federal Tax" line drop while "FICA" stays flat as you increase the pretax deduction field, that is exactly what you are seeing.

The Math, With Real Numbers

The core argument for pretax-vs-Roth equivalence rests on a piece of algebra that is worth understanding before you try to second-guess it. Assume your marginal tax rate now is t_now and your marginal rate in retirement will be t_future. You have $10,000 of pretax dollars to allocate. Investments grow at a multiple of g over the holding period.

If you go pretax: $10,000 grows to $10,000 × g, then gets taxed on the way out at t_future. You end up with $10,000 × g × (1 − t_future).

If you go Roth: you pay tax now, so only $10,000 × (1 − t_now) actually goes into the account. It grows to $10,000 × (1 − t_now) × g, and you owe nothing on the way out.

The two ending balances are equal when t_now = t_future. If your future rate is higher, Roth wins. If your future rate is lower, pretax wins. That is the entire conventional analysis, and it is mathematically airtight given its assumptions. The problem is that almost none of those assumptions hold cleanly in real life.

Where the Simple Model Breaks

The model assumes you have the same number of dollars going into either account. In practice that is rarely how the decision is framed. Most people decide on a contribution amount — "I want to put 10% of my paycheck in" — and then choose the bucket. If you put 10% of your paycheck into Roth, you are actually saving more in real economic terms than putting 10% pretax, because the Roth dollars are post-tax dollars. To make an apples-to-apples comparison, the pretax contributor would need to invest the tax savings in a taxable brokerage account. Most do not, which means in practice Roth often results in a higher effective savings rate. That is a behavioral argument for Roth that has nothing to do with future tax brackets.

The model also assumes a single marginal rate in retirement. Real retirees pull from multiple income sources — Social Security, pensions, pretax accounts, Roth accounts, taxable brokerage, sometimes a part-time job. The "marginal rate in retirement" is actually a question about the marginal rate on the next dollar of pretax withdrawal, which depends on everything else stacked under it. If you already have a substantial Social Security check and a small pension filling up the lower brackets, your incremental pretax withdrawals can land squarely in the 22% or 24% bracket even on a modest retirement income.

And the model assumes tax brackets stay constant in real terms, which is a heroic assumption over a thirty-year horizon. The 2017 Tax Cuts and Jobs Act rates are currently scheduled to sunset at the end of 2025 unless Congress acts. The historical norm for the top US bracket over the last seventy years has been substantially higher than today's 37%. Betting that future rates will be lower than current rates is, historically, a contrarian position.

Required Minimum Distributions Are the Quiet Asymmetry

This is the variable that most retirement-savings advice underweights. Pretax accounts are subject to required minimum distributions (RMDs) starting at age 73 (rising to 75 in 2033 under SECURE 2.0). The IRS forces you to withdraw a calculated percentage of the account every year whether you need the money or not, and the entire withdrawal is taxable as ordinary income.

If you have saved aggressively into pretax accounts and your portfolio has grown well, your RMDs can be large enough to push you into a higher marginal bracket than you were in during your working years. A retired couple with $2 million in a traditional IRA is looking at a starting RMD around $75,000 a year, on top of Social Security, on top of any other taxable income. The pretax-vs-Roth math that ignored this scenario gets the answer badly wrong.

Roth IRAs are not subject to RMDs during the original owner's lifetime. Roth 401(k)s used to require RMDs but no longer do as of 2024 under SECURE 2.0. That asymmetry — pretax forces taxable distributions, Roth does not — gets larger in dollar terms the more successfully you save.

The Honest Decision Framework

Stripping out the hand-wavy versions, the practical decision comes down to a small number of questions.

What is your marginal federal rate right now? If you are in the 10% or 12% bracket, Roth is almost certainly the right answer. Paying 12% now to avoid an unknown future rate that is unlikely to be lower than 12% is a strong bet. The pretax tax break is not large enough in absolute dollars to matter much. If you are in the 32%, 35%, or 37% bracket, the immediate tax deduction from pretax contributions is substantial and the case for at least some pretax allocation is real.

What is your state income tax situation, both now and probably later? A California resident in the 9.3% state bracket who plans to retire in Texas or Florida gets a state-tax break twice over by going pretax — they avoid California tax on the contribution now and pay zero state tax on the withdrawal later. The reverse pattern (low-tax state now, planning to retire in a high-tax state) tilts toward Roth.

How much do you already have in pretax accounts? If your traditional 401(k) is already at $500,000 in your forties, you are setting up RMD pressure later. Diversifying into Roth at that point is a hedge against bracket creep at age 73, not a bet on tax rates.

Do you have employer match? Employer matches go into a pretax bucket regardless of where your own contribution lands. If you are entirely Roth on your side, the match still builds your pretax balance. That naturally creates some pretax exposure without effort and is a reason not to feel obligated to allocate some of your own contribution to pretax for "diversification."

What does your household income look like over the cycle? A two-earner household with one spouse in residency or law school for three years has a window of unusually low marginal rates. That window is screaming for Roth contributions and Roth conversions. The opposite — peak earning years with a working spouse — is where pretax has its strongest case.

The Cases Where the Rule of Thumb Is Just Wrong

A few specific situations where the conventional advice misleads.

High-income earners assuming they "must" go pretax. The intuition is that the immediate tax deduction is too valuable to pass up at a 35% marginal rate. But if you are at 35% now and your accumulated retirement balance will throw off $300,000 a year in RMDs plus Social Security plus pension, your retirement marginal rate is plausibly in the same neighborhood. The deduction is not as decisive as it looks.

Young earners going entirely pretax because "I'm in a high bracket now." A 26-year-old in the 24% bracket has forty years for compounding. The tax-free growth on a Roth contribution made now is enormous. The 24% rate is high relative to a probable retirement rate of 12% only if you assume tax brackets stay constant for forty years and your retirement income lands in the 12% band. That is a lot of assumption.

Anyone above the Roth IRA income limits assuming they "can't" go Roth. The Roth IRA direct-contribution phaseout in 2024 starts at $146,000 single and $230,000 married filing jointly. Above that, direct Roth IRA contributions are not allowed. The backdoor Roth (contribute non-deductible to a traditional IRA, convert to Roth) remains legal as of mid-2026 and is widely used by high earners. Roth 401(k) contributions have no income limit at all.

The catch-up contribution rule change. SECURE 2.0 requires that catch-up contributions (the extra amount workers 50+ can contribute) be made on a Roth basis for employees who earned more than $145,000 in the prior year. This rule took administrative effect in 2026. If you are 50+ and high-earning, part of your catch-up is now Roth by law, which simplifies the decision for that slice of dollars.

What Actually Helps

The single most useful thing you can do before making this decision is to look at a concrete paycheck and see what the numbers do under each scenario. The paycheck calculator lets you plug in salary, filing status, and pretax deduction amount and see federal tax withholding update line by line. Comparing the pretax case (deduction filled in) against the Roth case (deduction zero) shows you exactly how many dollars of tax you would defer by going pretax — which is the immediate, concrete number the abstract argument tends to obscure.

From there, the right question is not "which is mathematically optimal." It is whether the certain, present-value tax break from pretax is worth more to you than the uncertain, future-value tax flexibility of Roth, given the realistic range of futures you might end up in. For most people in the middle income brackets, with most of a career ahead of them, a split — some pretax, some Roth — is a defensible answer that does not require you to predict the future correctly. The mistake is choosing the bucket once at age 25 and never revisiting it for the next thirty years as your income, your balances, and your tax law all change underneath you.