The 401k Math Most People Skip
Two arguments against maxing your 401k. Both wrong. Here's why.
What’s the number one piece of personal finance advice you hear constantly? No, not the one about living below your means. Right past that - yes, max out your 401k at work, that one. Why? Of course there’s the employer match, which is free money. But equally important: you save on taxes. If you’re in the top bracket, that’s 37% on every dollar.
Recently I read articles challenging that. Two arguments, both reasonable on the surface.
First: the 401k only works because of tax bracket arbitrage - your rate when contributing is higher than when you retire. But that assumes you’ll be poor in retirement. Nobody wants to be poor in retirement.
Second: you’re only deferring tax on $24,500 a year today, but after 30 years of compounding that becomes real money. Real money at any tax rate is a real tax bill.
Both sounded reasonable. I sat with an uncomfortable question: did I make a huge mistake maxing my pre-tax 401k year after year?
I did what I always do. I ran the numbers.
The math behind the fear
Take a household earning $800k filing jointly - a round number just above the 37% federal bracket threshold, useful for the illustration. Effective federal tax rate at that income: about 27%. Two choices with the $24,500 401k contribution:
Choice A (Roth or taxable): Pay 27% tax now, invest the remaining $17,885 in the S&P 500 for 30 years at 10% annual return. You end up with $312,083.
Choice B (pre-tax): Contribute the full $24,500, let it compound for 30 years. You have $427,510 — then pay 27% tax. Drum roll... $312,083.
Identical. It’s the commutative property of multiplication:
$24,500 × (1 − 0.27) × (1.10)^30 = $24,500 × (1.10)^30 × (1 − 0.27)
Here’s the intuition. In Choice A, you give the IRS their share upfront and invest the rest. In Choice B, you put both your share and the IRS’s share into the same fund - and pay them back 30 years later. The IRS’s slice compounds right alongside yours, and you hand it back at the end.
This is the key insight: assuming the same effective tax rate, it doesn’t matter when you pay the tax. The compounding doesn’t create a bigger tax burden. Thirty years of growth doesn’t make the bill larger in real terms. The fear that deferring tax on a small amount today creates a massive liability tomorrow is mathematically unfounded - as long as the rate is the same going in and coming out.
That assumption - same rate - is the entire game.
Where the math breaks open
At $800k of household income, the last $24,500 sits entirely above the 37% federal bracket threshold. Every single dollar of that $24,500 - if not contributed to the pre-tax 401k - is taxed at 37%. Not 27% effective. 37% marginal. Flat.
So Choice A doesn’t give you $17,885 to invest. It gives you $15,435. Thirty years later at 10% returns: $269,332.
Now Choice B. Contribute $24,500 pre-tax, let it compound to $427,510. In retirement with no other income, that $427,510 is taxed from the bottom up - 10% on the first $23,850, 12% on the next slice, 22% on the next, and so on. There’s one more detail worth noting: in working years at $800k, the standard deduction is already consumed by W-2 income - the 401k contribution gets no additional benefit from it. In retirement with no other income, the full $32,200 standard deduction applies before any brackets kick in, reducing the taxable amount further. The effective rate on $427,510 lands well below the top bracket - a meaningfully lower blended rate than what you paid going in. You keep significantly more. And $427,510 doesn’t have to come out in a single year. If your actual retirement spending is $150,000 a year, that balance stretches across nearly three years of withdrawals - each year’s taxable income sitting comfortably in the lower brackets, pushing the blended rate down further still.
Choice B beats Choice A by 25%. Not because of financial magic. Because you deducted at 37% marginal on the way in and paid a much lower blended rate on the way out. That gap - guaranteed by the structure of progressive taxation - is the actual argument for pre-tax contributions at high incomes.
This is what the standard advice misses. It isn’t about deferral. It isn’t about time value. It’s about where in the bracket stack your contribution sits today - at the top - versus where your withdrawal sits in retirement - starting at the bottom.
This is also the answer to the first concern. You don’t need to be poor in retirement for the math to work. You just need your retirement income - whatever it is - to be taxed from the bottom of the bracket stack up, rather than stacked on top of a high working income. A household living comfortably in retirement, with no W-2 pushing them into the top bracket, pays a far lower blended rate than the same household earning $800k at work. That’s not poverty. That’s the structure of progressive taxation working in your favor.
The one scenario where it doesn’t work
If your retirement income is high enough that every dollar of your 401k distribution stacks on top of $800k of other income, you pay the top rate on the way out too. The arbitrage disappears. That’s an enviable problem - and if that’s you, Roth is the right call.
But how often does this actually happen? A business generating top-bracket passive income indefinitely after the owner stops working. Rental income at scale. A portfolio large enough that dividends alone exceed $750k. For most high earners - even very successful ones - retirement income drops meaningfully from peak working years. The commutative property holds in theory; the rate gap holds in practice.
Two things worth knowing
The core argument stands on its own - but two details push the case further in this illustration.
If the alternative to pre-tax is a taxable brokerage rather than a Roth 401k, the pre-tax advantage widens further. Taxable accounts add capital gains tax on top - 20% plus the 3.8% net investment income tax at this income level - on top of whatever ordinary income tax applies at withdrawal.
And tax brackets are inflation-adjusted annually. The thresholds creep upward each year. The same nominal retirement income faces a slightly lower effective rate every year - meaning today’s blended rate estimate is a ceiling, not a floor.
What I decided
I still max my pre-tax 401k. Not because of a rule of thumb. Because I ran the math for my own situation: whatever I contribute is deducted at the top federal marginal rate today - and if retirement brings a meaningfully lower blended rate with no W-2 income stacking against it, the gap is real and structural. It doesn’t require market predictions or guesses about future tax policy. It comes from the structure of the tax code itself.
The articles that rattled me were making a real point - the 401k arbitrage is not automatic. It depends entirely on the gap between your marginal rate today and your blended rate in retirement. For most high earners, that gap is substantial and works in your favor.
Run your own numbers. The math is not complicated. And it is worth knowing before you make a decision that compounds for thirty years.
I am not a financial advisor. Nothing here is investment or tax advice. Run the numbers for your own situation and consult a CPA before making decisions. Fine Print Investing publishes weekly.