Safe Withdrawal Rate — Why 4%?

Financial planner William Bengen asked the right question: what is the maximum withdrawal rate such that the portfolio would have survived every rolling 30-year period in modern market history? He determined that this maximum Safe Withdrawal Rate (SWR) was 4%.

Allocation matters less than you would guess, but not as little as the rule’s popularizers suggest. Updated through 2018, the analysis finds a broad plateau — roughly 35% to 80% stocks — across which higher equity exposure has no discernible effect on the safe rate.⁴⁴ Step outside that band and the rule breaks: an all-bond portfolio supports well under 2.5%. The lesson is not that allocation is irrelevant but that there is a wide plateau of adequate answers and a cliff on the conservative side. Bengen himself recommended 50–75% equities — not because it raised the safe rate, but because within the plateau the higher allocation produced dramatically more upside in the typical case, and dying with too much money is a better problem than the alternative.

The mechanics: in the first year of retirement you withdraw 4% of the portfolio’s value, and in every later year you withdraw the same dollar amount adjusted for inflation. Follow the rule and, on the historical record, a 30-year retirement survives.

Two essential guardrails qualify the 4% figure, and both are worked properly with the withdrawal machinery in section “Withdrawal Strategies in Retirement”. The safe rate is valuation-sensitive — retire at the top of an expensive market and the historically safe rate has been closer to 3–3.5% — and the rule is deliberately rigid, ignoring how the portfolio actually performs, which is what dynamic withdrawal strategies like the Guyton–Klinger guardrails exist to fix. Use 4% as the anchor for the savings target below and as an initial planning baseline, not a static guarantee for spending; revisit the rate as the actual retirement date comes into view.

This gives a savings target: multiply the portion of your retirement spending that must come from investments by 25 ($\frac{100}{4}$). Only that portion — the rest is covered by Social Security, pensions, or other guaranteed sources, and you do not need to fund it twice.

The $25\times$ rule is stated pre-tax, and almost nobody adjusts for it. “Spend $150,000 a year, so save $3,750,000” is only right if the $150,000 represents a gross portfolio withdrawal, not net take-home spending. Where the portfolio sits decides the difference:

Roth accounts

Qualified withdrawals are tax-free. Here $25\times$ genuinely means $25\times$, and this is the one case the rule of thumb states correctly.

Traditional 401(k) and IRA

Every dollar comes out as ordinary income. To spend $150,000 at an effective rate of, say, 18% across the brackets, you must withdraw about $183,000 — so the target is $25\times \$183,000\approx \$4.6$ million, roughly 22% more than the headline.

Taxable brokerage

Better than traditional, worse than Roth. Only the gain portion is taxed, at long-term rates, so the drag depends on your basis. A position with 50% embedded gain taxed at 15% costs you about 7.5% of each withdrawal.

The general form: if $\tau$ is your blended effective rate on withdrawals,

Which makes the case for Roth conversions in low-income years, and for arriving at retirement with money in all three buckets so you can choose which one to draw from each year (chapter “Tax-Efficient Decumulation”).

Two structural constraints bind the rule: it was calibrated to a 30-year horizon — retire at 45 and you need a lower rate, since the failure modes compound over a longer window. And it rests on US market history, which is the most successful equity market of the twentieth century; studies extending the analysis to other developed markets find safe rates materially below 4%. Survivorship bias is not a reason to ignore the rule, but it is a reason not to treat 4% as a law of physics.