The debate is usually framed as a prediction: will your tax rate be higher in retirement than it is today? If yes, go Roth. If no, go pre-tax. Guess right and you picked the better account.
That framing is incomplete. There is a mathematical proof that most financial advice skips entirely: under one specific condition, pre-tax and Roth accounts produce exactly the same after-tax wealth. Not approximately. Algebraically identical.
What follows from that proof is more useful than any rule of thumb. The choice between account types is entirely a tax rate arbitrage problem, and most of the real complexity lives in correctly estimating that rate differential — and in the structural variables the simple algebra ignores.
The Math of Each Path
Start with the same dollar amount in pre-tax terms: call it $C. Using pre-tax dollars as the common unit matters; it is the only way to make the comparison apples-to-apples.
Pre-tax path: $C goes in with no immediate tax consequence. It compounds for n years at annual rate r. At withdrawal, you pay your retirement marginal tax rate, t_r, on the full balance.
After-tax result = C × (1 + r)^n × (1 − t_r)
Roth path: You pay your current marginal rate, t_c, upfront. That leaves C × (1 − t_c) to invest. It compounds for the same n years at the same rate r and comes out tax-free at withdrawal.
After-tax result = C × (1 − t_c) × (1 + r)^n
Multiplication is commutative. Both expressions produce the same result when t_c = t_r.
When t_c is lower than t_r: Roth wins. Tax was paid at the lower rate.When t_c is higher than t_r: Pre-tax wins. Tax was deferred until the rate was lower.When t_c equals t_r: the account type is mathematically irrelevant.
The practical implication follows directly: the only variable that determines which account produces more after-tax wealth is the relationship between your current marginal rate and your retirement marginal rate. Everything else in the debate is downstream of that.
Where the Equivalence Breaks
The algebra assumes a single tax rate at each time point. Several real-world variables complicate that assumption.
Marginal vs. effective rate. The proof uses one rate, but tax brackets are progressive. Not all retirement income is taxed at the top marginal rate. Your effective rate on withdrawals is typically lower than the marginal rate paid during peak earning years. This creates a systematic bias toward pre-tax accounts for engineers contributing at 24-37% marginal rates and withdrawing across a progressively structured income stack in retirement.
Required Minimum Distributions. Traditional pre-tax accounts require minimum annual withdrawals starting at age 73. RMDs are taxable income and can push you into higher brackets, trigger Medicare premium surcharges (IRMAA), and eliminate flexibility to manage your taxable income year to year. Roth accounts have no RMDs during the owner's lifetime, which matters for both tax planning and estate planning.
State taxes. Some states exempt retirement income from state tax; others tax it at full ordinary rates. Contributing in a high-tax state and retiring in a no-income-tax state creates a real rate differential that the federal-only analysis misses entirely.
Tax diversification. Holding both pre-tax and Roth balances in retirement gives you the flexibility to draw from whichever account creates less taxable income in a given year. In years with large medical expenses, reduced spending, or Roth conversion opportunities, that control has real economic value that no single-scenario comparison captures.
Run the comparison with your own rate assumptions. When the two lines on the chart overlap, the current and retirement tax rates are equal and both accounts produce the same after-tax wealth.
What This Means In Practice
For most engineers at peak earning years, the systematic starting point is pre-tax contributions now and Roth conversions later. Marginal rates during peak working years tend to land at 24-37% federal. Effective retirement withdrawal rates tend to be lower: progressive brackets, the standard deduction, and income flexibility work together to reduce the effective rate on a diversified retirement income stream.
The tax environment over the next 20-30 years is genuinely uncertain. The Tax Cuts and Jobs Act's individual provisions are scheduled to sunset after 2025, which would push the 22% and 24% brackets back toward 25% and 28%. Roth contributions lock in today's rates regardless of future legislative changes, and that optionality has value even if the rate environment turns out to be stable.
A practical framework: contribute pre-tax up to the employer match, which is always the right answer regardless of account type. After the match, run your own numbers. Engineers in the 22% bracket or below often find Roth competitive. Engineers at 32% or above usually favor pre-tax. In the 24% bracket — where much of a typical engineering career is spent — the answer often depends on state tax situation and expected retirement income structure. The tool above is the right place to start that calculation.
Takeaways
• Pre-tax and Roth produce identical after-tax wealth when the contribution and withdrawal tax rates are equal; that equivalence is the starting point for all analysis
• The account choice is a tax rate arbitrage problem, not a product preference
• Marginal vs. effective rate, RMDs, and state taxes all introduce real-world complexity that the simple equivalence proof ignores
• Tax diversification across both account types gives flexibility in retirement that no single-scenario comparison can capture
• The prime Roth conversion window is typically the gap between retirement and age 73 when Required Minimum Distributions begin