The 4% Rule Explained: How Much You Actually Need to Retire — and Why the Answer Is "25× Expenses, Roughly"

The rule originates in 1990s research on a precise question: what fraction of a portfolio could a US retiree withdraw in year one — then adjust for inflation annually — without running out of money over 30 years, in the worst historical case?

Financial planner William Bengen (1994) tested withdrawal rates against every rolling 30-year period of US stock and bond history, including retirements starting into the Great Depression and the brutal 1966–1982 stagflation. The worst-case survivable rate landed near 4%. The Trinity study (1998, updated repeatedly) confirmed it: a 50–75% stock portfolio withdrawing an inflation-adjusted 4% survived 30 years in roughly 95–100% of historical periods.

Inverting 4% gives the famous shortcut:

Annual spending ÷ 0.04  =  Annual spending × 25  =  "your number"

Spend 40,000/year  →  need ≈ 1,000,000
Spend 60,000/year  →  need ≈ 1,500,000

Three fine-print items that the popular version drops:

• It is a worst-case rule, not an average. In most historical periods, 4% withdrawals left retirees with more money after 30 years than they started with — the rule is calibrated to the unlucky minority of start dates.
• It assumed a 30-year horizon. Retire at 40 and plan for 50+ years, and the safe rate drops (research suggests ~3.25–3.5% for very long horizons).
• It ignores fees and taxes. A 1% advisory fee effectively turns your 4% into 3% of spendable withdrawal; taxes on withdrawals reduce it further. The rule describes portfolio math, not take-home income.

Here is the risk that makes retirement math genuinely different from accumulation math. Two retirees can earn the identical average return over 30 years and have opposite outcomes — because withdrawals make the order of returns matter.

Both start with 1,000,000, withdraw 40,000/year (inflation-adjusted), and average 7% returns. Retiree A hits a 40% crash in years 1–2; Retiree B hits the same crash in years 25–26:

• Retiree A (early crash) is forced to sell depressed assets to fund withdrawals — each 40,000 withdrawal at the bottom sells nearly 7% of the shrunken portfolio. The portfolio never recovers its footing and can be exhausted by year 20–25.
• Retiree B (late crash) spent 24 years compounding first; the same crash, arriving when withdrawals are a tiny share of a grown portfolio, barely dents the plan.

Same average, same withdrawals — ruin versus abundance. That asymmetry is sequence-of-returns risk, and it is why the first ~5–10 years around retirement are called the "fragile decade". Standard defenses, each with trade-offs:

• Cash/bond buffer: 1–3 years of expenses in stable assets, drawn during crashes so equities are never sold at the bottom.
• Flexible spending ("guardrails"): cutting withdrawals 10–20% in bad years dramatically raises survival odds — rigid inflation-adjusted spending is the rule's most unrealistic assumption, and its biggest hidden safety valve.
• Glidepath: entering retirement bond-heavy and re-raising equity exposure over the first decade (a "bond tent") specifically blunts the fragile-decade risk.
• Part-time income early in retirement: even modest earnings in years 1–5 disproportionately protect the portfolio, because they offset withdrawals exactly when sequence risk is highest.

The rule was mined from US market history — the single best-performing major equity market of the 20th century. Applying it elsewhere requires honesty about that:

• Global research says "closer to 3–3.5%". Studies replicating Bengen's method on non-US developed markets (and on globally diversified portfolios with international return data) generally find historical safe rates below 4% — some markets (Japan post-1990, pre-war European markets) saw sustained periods where 4% failed badly. A globally diversified portfolio with a ~3.5% initial rate is the conservative reading of the evidence.
• State pensions change the arithmetic entirely. The 25× target applies to spending not covered by other income. A retiree with a full UK State Pension, German Rente, or Canada CPP+OAS covering half their spending needs 25× of only the other half — most non-US planning underweights this offset.
• Inflation and currency risk differ. The rule assumes withdrawals grow with a moderate, mostly-anchored inflation rate. In economies with historically volatile inflation (or for anyone whose spending currency differs from their portfolio currency), the real-value guarantee at the rule's core is weaker — favoring inflation-linked bonds, real assets, and lower initial rates.
• Tax wrappers shift the net rate. Whether withdrawals come from pre-tax (401(k), RRSP, EPF), post-tax (Roth, ISA, TFSA), or taxable accounts changes spendable income at the same gross withdrawal — the 25× target should be computed on gross spending including expected taxes.

A defensible global default: 25× is the optimistic floor, 28–33× (a 3–3.5% rate) is the conservative planning band — tightened or loosened by your pension coverage, horizon, and flexibility.

The full calculation, front to back, for someone 35 years old planning to retire at 60:

Step 1 — Retirement spending today:        45,000/year
Step 2 — Minus pension/other income:      −12,000/year (estimated state pension)
         Portfolio must cover:             33,000/year (today's money)
Step 3 — Inflate 25 years @ 3%:            33,000 × 1.03^25 ≈ 69,100/year
Step 4 — Target multiple (3.5% rate ⇒ ~28.6×):
         69,100 × 28.6 ≈ 1,975,000  ← the number
Step 5 — Monthly saving needed, 25 yrs @ 7% nominal:
         ≈ 2,440/month (from zero); less with existing savings,
         less again if contributions step up with salary

Observations that generalize from every such calculation:

• The spending estimate dominates everything. A 20% error in step 1 moves the target by 20%; no investment decision moves it that much. Retirees who track actual spending for a year plan on solid ground; everyone else is extrapolating a guess.
• Cutting spending counts double: each 1,000/year of permanent spending reduction both shrinks the target by ~25,000–30,000 and frees 1,000/year of extra saving.
• Starting age is brutal arithmetic: the same target from age 25 (35 years of compounding) needs roughly half the monthly saving of starting at 35. The FIRE movement's core insight is just this equation read aggressively.
• Run it in real terms as a sanity check: using real returns (nominal minus inflation) with today's spending avoids the inflation step and typically produces the same answer with fewer chances to fool yourself.

The 4% rule is a planning anchor, not an autopilot. The factors that routinely matter more than the second decimal of the withdrawal rate:

• Healthcare and long-term care. The largest unmodeled expense in most plans, and the most country-dependent — a US retiree's exposure differs by an order of magnitude from a German or Canadian one. Whatever your system, it deserves an explicit line, not an implicit hope.
• Spending is not flat. Real retiree spending typically follows a "smile": higher in the active early years (travel), lower in the middle, rising again late (care). Rigid inflation-adjusted withdrawals model none of this.
• Flexibility is worth more than precision. The research is unambiguous: retirees able to cut 10–15% in bad markets can safely start at meaningfully higher rates than the rigid-withdrawal math suggests. Building a budget with a defined "can cut" layer is a better use of effort than refining the rate's third decimal.
• Longevity is a distribution, not a number. Plan for the 90th percentile of your lifespan, not the average — half of couples reaching 65 see at least one partner pass 90 in most developed countries. Annuities and state pensions are longevity insurance precisely because they cannot be outlived.
• The plan will be revised. A retirement plan is not a 30-year prophecy; it is this year's best estimate, re-run annually with real balances and real spending. The savers who succeed are rarely the best forecasters — they are the most consistent revisers.

All figures are illustrative and historical patterns are not guarantees. For decisions of this size, a fee-only fiduciary advisor familiar with your country's pension and tax system is worth the cost.