FIRE Number and the 4% Rule

FIRE Number and the 4% Rule: How to Calculate It and How Safe It Is

Your FIRE number is your annual spending divided by the withdrawal rate you are willing to live with. At the classic 4 percent rate that means 25 times spending: a household spending $40,000 a year needs $1,000,000 invested. The 4 percent rule held up across historical 30-year retirements in US data, but it is a planning anchor, not a guarantee, and it gets thinner as the retirement gets longer.

By Finance Editorial Desk | October 22, 2026

Educational only, not financial, tax, or legal advice. Every household below is a labelled illustration with its inputs stated. Historical findings are taken from the original papers, and tax rules from IRS and Canada Revenue Agency pages, each linked where it is used and listed under Sources.

1. The formula, in one line

FIRE number = annual spending ÷ withdrawal rate.

Written as a multiple, it is spending × (1 ÷ withdrawal rate). At 4 percent the multiple is 1 ÷ 0.04 = 25. At 3.5 percent it is 28.57. At 3 percent it is 33.33.

Two details in that line do most of the work, and both are easy to get wrong.

Spending, not income. The input is what you spend in a year once you stop working, not your salary. Payroll deductions for retirement savings disappear in retirement, a mortgage may be paid off, commuting costs fall, while health insurance and travel may rise. A household earning $110,000 and spending $52,000 builds its number on $52,000.

Spending before or after tax. The 4 percent studies measured gross withdrawals and ignored taxes. If your money sits in pre-tax accounts, every dollar you withdraw is taxable income, so the spending figure you divide should include the tax you will owe on the withdrawal. Section 9 shows how that grosses up the number.

The FIRE Number Calculator applies this exact formula. Its default is $40,000 of annual expenses at a 4 percent safe withdrawal rate, which returns a target of $1,000,000. The separate Can I FIRE? calculator uses the same division but lets you change the withdrawal rate directly, which is the most useful single input to experiment with.

2. Where the 4 percent figure comes from

The number has two sources, and they measured slightly different things.

Bengen, 1994

William P. Bengen, a financial planner, published "Determining Withdrawal Rates Using Historical Data" in the October 1994 issue of the Journal of Financial Planning. He used Ibbotson Associates historical returns for common stocks and intermediate-term Treasury notes, starting a hypothetical retiree in every year from 1926 onward. Each retiree withdrew a fixed percentage of the starting portfolio in year one, then raised that dollar amount for inflation every year after, regardless of how markets moved.

His findings, in his own terms, using a 50/50 stock and bond mix as the base case:

• A 3 percent first-year withdrawal never produced a portfolio life under 50 years in the historical record, and he noted the same held up to about 3.5 percent.
• At 4 percent, "in no past case" had the portfolio been exhausted before 33 years, and in most cases it lasted 50 years or longer.
• A 4.25 percent first-year withdrawal could exhaust a portfolio in as little as 28 years.
• He described an initial 5 percent withdrawal as risky and 6 percent or more as "gambling".
• He recommended a stock allocation between 50 and 75 percent, calling allocations outside that band counterproductive.

He also assumed all money sat in tax-deferred accounts, so capital gains taxes did not enter the analysis. The paper is available as the Financial Planning Association's 2004 reprint (Bengen 1994 reprint, FPA, checked October 2026).

The Trinity University study, 1998

Philip Cooley, Carl Hubbard, and Daniel Walz, professors at Trinity University in San Antonio, published "Retirement Savings: Choosing a Withdrawal Rate That Is Sustainable" in the February 1998 AAII Journal. They used the S&P 500 for stocks and long-term high-grade corporate bonds for bonds, from 1926 to 1995, tested payout periods of 15, 20, 25, and 30 years, and called a portfolio a success if its ending value was above zero. Like Bengen, they did not adjust for taxes or transaction costs. Their text says withdrawal rates of 3 to 4 percent "continue to produce high portfolio success rates for stock-dominated portfolios" when withdrawals rise with inflation (AAII Journal, February 1998, checked October 2026).

The authors later published a follow-up using monthly data in Financial Counseling and Planning (1999). We could not find a copy of that paper on the journal's or the university's own site, so its success-rate table is not reproduced here; the AAII article above is the primary source this page relies on.

What the studies did not test

Read these studies for what they are: a record of how often a fixed rule survived the past in one country's markets, for retirements up to 30 years.

• No period longer than 30 years. A 40-year-old retiree may need 50 years.
• No fees. A 1 percent annual fund or advisory fee is a quarter of a 4 percent withdrawal.
• No taxes. Both studies say so directly.
• US data only, and a strong century for US stocks. Other countries' markets had worse stretches.
• Success means "did not hit zero". A portfolio that ended at $1 counts the same as one that tripled.

None of this makes 4 percent wrong. It makes 4 percent a reasonable starting point for a roughly 30-year retirement with a stock-heavy, low-fee portfolio and some flexibility. Early retirees usually adjust downward, and the next table shows what that costs.

3. Withdrawal-rate sensitivity: what each half-point costs

The FIRE number moves inversely with the withdrawal rate, so small changes in the rate move the target by large dollar amounts. This table is pure arithmetic on stated spending levels:

Two readings matter more than any single cell.

Moving from 4 percent to 3.5 percent raises the target by 14.3 percent. For a $60,000 household that is $214,286 more capital. Moving to 3 percent raises it by a third.

The cost of caution is time, not just money. Section 5 converts those extra dollars into extra working years. For a household saving 50 percent of take-home pay, the step from 4 percent to 3.5 percent adds roughly 1.6 years of work at the return assumption used there.

A practical decision rule follows from the historical tables and the length-of-retirement problem:

This is a framing device for testing, not a recommendation. Your own rate depends on how much of your spending is fixed, how much income arrives later from pensions or government programs, and how willing you are to cut spending after a bad year.

4. Worked example: running the calculator defaults by hand

The FIRE Number Calculator projects a balance month by month. It converts the annual return into an equivalent monthly rate, (1 + annual return)^(1/12) − 1, grows the balance by that rate, adds the monthly contribution, and stops in the first month the balance reaches the target. Here are its default inputs traced through:

• Age 30, current investments $50,000
• Annual expenses $40,000, so the target at 4 percent is $40,000 ÷ 0.04 = $1,000,000
• Monthly contribution $1,500 ($18,000 a year)
• Annual return 7 percent

Run forward, the balance first reaches $1,000,000 in month 247, which is 20.6 years, at about age 50.

Now change one assumption at a time:

The calculator's own FAQ notes that it uses nominal returns and suggests a real return when you want inflation-adjusted results. That matters: at 7 percent nominal with 2 to 3 percent inflation, the $1,000,000 reached at 50 buys less than $40,000 of today's spending. Entering a real return, such as the 5 percent row, keeps the target and the projection in the same year's dollars. The difference in this example is 3.75 years of work.

5. Years to financial independence by savings rate (original table)

If the target is a multiple of spending, and spending is whatever is left after saving, then the savings rate sets both the size of the target and the speed of the deposits. Income level drops out of the arithmetic entirely. A household saving 50 percent of take-home pay on $60,000 reaches FI at the same time as one saving 50 percent on $300,000, if their returns match.

Method and assumptions. Starting from zero invested. Take-home pay is constant in real terms. Spending is take-home pay minus savings, and stays the same in retirement. Savings are invested at the end of each year. The target is 25x spending (4 percent rate) unless the column says otherwise. Years solve the future-value-of-an-annuity equation, t = ln((target × r + savings) ÷ savings) ÷ ln(1 + r), which is the same closed-form method the Can I FIRE? calculator uses. The returns are real (after inflation) and are illustrations, not forecasts.

Calculated by the Finance Editorial Desk, October 2026. Illustration only.

What the table says

The first gains are the largest. Going from 10 to 20 percent cuts about 15 years at 5 percent real. Going from 60 to 70 percent cuts under 4. The reason is that each extra point of savings rate both adds to deposits and shrinks the target.

Returns matter less at high savings rates. At a 70 percent savings rate, the gap between a 5 percent and a 0 percent real return is under 2 years, because most of the target is built from deposits. At 20 percent it is more than 60 years. Lower savers are far more exposed to a decade of weak returns.

The 0 percent column is a stress test, not a worst case. It shows how much of the outcome is your own saving. At 50 percent, deposits alone get you there in 25 years; compounding at 5 percent real shortens that to 16.6.

Reading your own row. Savings rate here is saved dollars ÷ take-home pay. If you count employer 401(k) matches or RRSP contributions that come off your pay before it reaches your account, add them to both the savings and the take-home figures before dividing, or your rate will look lower than the deposits really are. For a quick look at doubling times at different returns, the Rule of 72 calculator is a fast cross-check, and the compound interest calculator shows the year-by-year balance behind any row.

If your savings rate is below the level where any of this feels reachable, the first job is a cash buffer, not a FIRE plan. The resilience floor guide works out the minimum savings rate and emergency fund that keep one bad month from becoming debt.

6. Lean, regular, fat, coast, and barista FIRE, with the arithmetic

The labels are community shorthand, not defined terms. Each one changes a different input in the same formula, which is the useful way to think about them. All figures below are illustrations.

Lean FIRE: lower spending

A household spending $30,000 a year needs $30,000 × 25 = $750,000 at 4 percent. The catch is that a lean budget has less room to cut after a bad market year, because more of it is rent, food, and insurance. Bengen's and Trinity's rules assume you keep withdrawing the inflation-adjusted amount no matter what, and a lean budget is the case where that assumption is most literal. Many lean planners pair a lower number with a lower rate, which pulls the target back up: $30,000 at 3.5 percent is $857,143.

Regular FIRE

$50,000 a year × 25 = $1,250,000.

Fat FIRE: higher spending, usually a lower rate

$120,000 a year at 4 percent is $3,000,000. A fat budget has more discretionary spending that can be trimmed in a downturn, which argues for the 4 percent end, but planners targeting fat budgets over long retirements often test 3.5 percent: $120,000 ÷ 0.035 = $3,428,571. The Can I FIRE? calculator's "fat" preset uses exactly that pairing, $120,000 at 3.5 percent.

Coast FIRE: stop contributing, let time finish the job

Coast FIRE asks: how much do I need invested today so that, with no further contributions, it grows to my full FIRE number by a conventional retirement age?

Coast number = FIRE number ÷ (1 + real return)^(years until retirement).

A 35-year-old with a $1,250,000 target at 65 and a 5 percent real return needs $1,250,000 ÷ 1.05^30 = $289,222 today. At a 4 percent real return the same target needs $385,398. Once coast is reached, work only has to cover current spending, not future savings.

Check the other direction too. Someone with $200,000 at 35 who stops contributing reaches $200,000 × 1.05^30 = $864,388 at 65, short of $1,250,000. At 5 percent it would take 37.6 years from 35 for $200,000 to grow into $1,250,000 alone, so on its own that balance reaches the target at about 72 and a half, not 65.

Barista FIRE: part-time income covers part of spending

If part-time work brings in $20,000 a year net and spending is $50,000, the portfolio only has to fund the $30,000 gap: $30,000 × 25 = $750,000, versus $1,250,000 for full FIRE. The $500,000 difference is exactly $20,000 × 25, which is a useful way to value part-time income: every $1,000 a year of dependable side income is worth $25,000 of portfolio at a 4 percent rate. In the US, part-time work that comes with health coverage can be worth more than its wage, for reasons covered in section 10.

Variant summary

7. Sequence-of-returns risk: why the order of returns matters

Two retirees can earn exactly the same average return and end up far apart, because withdrawals turn the order of returns into a permanent difference. Losses early in retirement force you to sell more shares to fund the same withdrawal, and those shares are not there for the recovery.

Illustration. Start with $1,000,000, withdraw $40,000 at the start of each year (in real terms, so no separate inflation step), and apply these ten real annual returns: −20%, −12%, +4%, +10%, +8%, +12%, +6%, +15%, +9%, +7%. Retiree A gets them in that order. Retiree B gets the same ten returns in reverse. Both sequences compound to the same 3.31 percent a year, and with no withdrawals both would end at $1,385,015.

After two years, Retiree A's $40,000 withdrawal is 6.24 percent of the remaining balance, a rate Bengen would have called gambling. After ten years A is $220,060 behind B despite identical average returns, and A has a smaller base for the next twenty years.

What reduces the damage

• Spending flexibility. Bengen's paper follows a 1929 retiree who cut his 1930 withdrawal by 5 percent and kept it at that reduced level: 30 years in, his wealth was 25 percent greater than without the cut. A plan that can cut 5 to 10 percent of spending in a bad year is materially safer than one that cannot.
• A lower starting rate. Bengen found 3 to 3.5 percent lasted 50 years or more in his data, which is why early retirees often plan at those rates.
• A cash or bond reserve for the first years. Holding one to three years of spending outside stocks lets you avoid selling shares in a crash. It reduces long-run returns somewhat in exchange.
• Working a little in the first decade. Barista income in years one to five has an outsized effect, because that is when sequence risk bites hardest.

The drawdown strategies themselves, such as guardrails and withdrawal ordering, are a topic of their own; the RRIF minimum withdrawals guide shows how Canada's forced minimums interact with any withdrawal plan once accounts convert.

8. United States: getting money out before 59½

The FIRE number assumes you can spend the portfolio. Before age 59½, money in a 401(k) or traditional IRA usually comes out with an extra cost. The IRS states that early distributions from qualified plans and IRAs face an additional tax equal to 10 percent of the portion that is included in gross income, on top of ordinary income tax (IRS Topic 558, checked October 2026).

The exceptions are where early-retirement plans are built. The IRS exceptions chart separates which apply to employer plans and which apply to IRAs (IRS exceptions to tax on early distributions, checked October 2026):

The rule of 55 is plan-specific. It applies to the plan of the employer you leave in or after the year you turn 55. Roll that 401(k) into an IRA and the exception no longer applies to the rolled money, because the rule of 55 is not an IRA exception.

72(t) payments are a commitment. Substantially equal periodic payments are calculated by one of three IRS methods (required minimum distribution, fixed amortization, or fixed annuitization). The interest rate used cannot exceed the greater of 5 percent or 120 percent of the federal mid-term rate for either of the two months before the first payment. Payments must continue unchanged until the later of five years after the first payment or age 59½. Modify them early and the 10 percent tax applies to that year, plus a recapture of the 10 percent tax that would have applied to every earlier year, plus interest (IRS substantially equal periodic payments, checked October 2026).

Illustration only: a $500,000 IRA amortized over a 40-year period at 5 percent produces a fixed payment of $29,139 a year. At 4 percent the same balance produces $25,262. The real period comes from IRS life-expectancy tables for your age, so your figure will differ; the point is the scale. A 72(t) schedule converts a pre-tax account into a fixed income stream, but you cannot raise it when you need more or cut it when markets fall without triggering recapture.

A layered US plan often funds the years before 59½ from taxable brokerage accounts first, uses the rule of 55 if the last employer's plan allows it, and treats 72(t) as a tool for a specific gap rather than the default. Model the pre-tax balance you are building on the 401(k) calculator.

9. Canada: RRSP withholding, the TFSA, and when CPP and OAS start

Canada has no age-based penalty on RRSP withdrawals. The friction is tax, and how it is collected.

RRSP withdrawals

Every RRSP withdrawal is included in income for the year. The issuer withholds tax at source based on the size of the withdrawal: 10 percent on amounts up to $5,000, 20 percent on more than $5,000 up to $15,000, and 30 percent on more than $15,000. In Quebec the rates are 5, 10, and 15 percent, with provincial tax withheld as well (CRA, tax rates on RRSP withdrawals, checked October 2026).

Withholding is a prepayment, not your tax bill. A $12,000 withdrawal outside Quebec has $2,400 withheld (20 percent). Taken as three $4,000 withdrawals, $1,200 is withheld (10 percent each). Your final tax is identical either way, because the full $12,000 is income on your return. Smaller withdrawals only move more of the payment to filing time, which surprises people who budget from the net deposit.

For FIRE planning the useful feature is that an early retiree with little other income can draw RRSP money in low-income years at a low marginal rate. The RRSP must be wound up in the year you turn 71: withdraw it, transfer it to a RRIF, or buy an annuity (CRA, options for your RRSPs at 71, checked October 2026).

TFSA

Earnings inside a TFSA are generally not taxed, and you can withdraw whenever you want without penalty (CRA tax tip on the TFSA, checked October 2026). The annual dollar limit for 2026 is $7,000 (CRA limits table, checked October 2026), and amounts withdrawn are added back to your room on January 1 of the following year (CRA, withdrawing from a TFSA, checked October 2026). That makes the TFSA the natural account for the pre-60 bridge years: no withholding, no age rule, and the room is not lost. The TFSA guide has the room arithmetic, and the RRSP vs TFSA calculator compares the two for a given tax rate.

CPP and OAS timing

Government pensions reduce the portfolio's job after they start, which can shrink a Canadian FIRE number considerably. The timing rules, from Canada.ca:

• CPP can start as early as 60 or as late as 70. Payments are reduced by 0.6 percent for each month before 65 (36 percent at 60) and increased by 0.7 percent for each month after 65 (42 percent at 70) (Canada.ca, when to start CPP, checked October 2026).
• OAS can start at 65 or as late as 70, rising 0.6 percent for each month of deferral, up to 36 percent at 70. A full OAS pension needs 40 years of residence in Canada after age 18 (Canada.ca, when to start OAS, checked October 2026).

Two FIRE-specific points. First, CPP is based on your contributions over your working life, so leaving work at 40 leaves many zero or low years in your record, and your benefit will be lower than a career-length estimate. Use your own Service Canada estimate rather than an average. Second, high RRIF withdrawals after 65 can trigger the OAS recovery tax, so the order you draw accounts in matters; the OAS recovery tax guide works the threshold math and the CPP timing guide compares start ages in dollars.

Illustration with hypothetical figures: if your CPP estimate at 65 were $10,000 a year, starting at 60 would pay $6,400 (36 percent less) and starting at 70 would pay $14,200 (42 percent more). Use the CPP calculator with your own record.

A two-stage FIRE number

Because benefits start later, a flat 25x number overstates what an early retiree with pensions needs. Illustration: a 45-year-old spends $50,000 a year. Suppose government benefits from 65 would cover $18,000 a year in today's dollars (a hypothetical figure, not an estimate for any person).

• After 65, the portfolio covers $32,000: $32,000 × 25 = $800,000.
• From 45 to 65, it also covers the $18,000 the benefits will later replace: 20 years × $18,000 = $360,000, assumed held conservatively at a 0 percent real return.
• Total: $1,160,000, versus $1,250,000 for a flat $50,000 × 25.

The structure works the same in the US with Social Security. The trade-off is that it depends on benefit estimates that may change and on a bridge account you must not raid. And a 45-year-old faces a horizon well beyond 30 years, so testing the $800,000 part at 3.5 percent ($914,286) is a reasonable sensitivity check.

10. Taxes and healthcare change the spending input

Gross up for tax

If most of your portfolio is pre-tax (401(k), traditional IRA, RRSP), withdrawals are taxable. Illustration: assume an average tax rate of 15 percent on withdrawals. To spend $40,000 you would withdraw $40,000 ÷ 0.85 = $47,059, and the FIRE number becomes $47,059 × 25 = $1,176,471, not $1,000,000. Your average rate depends on your jurisdiction, account mix, and other income; the point is that ignoring it understates the target by roughly the tax rate. A mix of pre-tax, tax-free (Roth, TFSA), and taxable accounts lowers the effective rate by letting you choose which bucket funds each year.

Healthcare before 65

In the US, Medicare is health insurance for people 65 or older who meet citizenship or residency requirements (Medicare.gov, checked October 2026). A US retiree at 45 has twenty years of private coverage to buy, usually through an employer plan continuation or the health insurance marketplace, where premiums and subsidies depend on income and change by year. Price the premium and the out-of-pocket maximum for your area and add both to annual spending before dividing. A gap here can be larger than the gap between a 4 percent and a 3.5 percent rate.

In Canada, provincial plans cover medically necessary physician and hospital care, but coverage for prescription drugs, dental care, and vision differs by province and often came from an employer plan. Leaving work early usually means pricing those privately and adding them to the spending figure.

11. Common mistakes

1. Using income instead of spending. A $100,000 earner spending $55,000 who builds a target on $100,000 will work years longer than needed.
2. Mixing nominal and real numbers. A target in today's dollars projected with a 7 percent nominal return is reached "on time" in dollars that buy less. Use a real return, or inflate the target.
3. Treating 4 percent as a 50-year rule. The studies stop at 30 years. For longer horizons, test 3 to 3.5 percent.
4. Forgetting fees. Withdrawal rates in the studies are before costs. A 0.75 percent fee on a 4 percent plan leaves 3.25 percent of spending power.
5. Ignoring tax. Pre-tax money is partly owed to the government. Gross up spending first.
6. Leaving out lumpy costs. Car replacements, roofs, and dental work belong in average annual spending. Divide a $36,000 car every 12 years into $3,000 a year.
7. Planning on the rule of 55 after rolling to an IRA. The exception does not follow the money.
8. Counting on a full CPP or Social Security benefit after a short career. Early leavers have fewer contributory years.
9. No buffer for the first bad year. A sequence like Retiree A's in section 7 is when people return to work. Plan the response in advance.
10. Measuring net worth instead of investable assets. A paid-off home does not produce withdrawals unless you sell it. Use the net worth guide to separate the two.

12. A checklist before you trust your number

• Annual retirement spending, including lumpy costs, health coverage, and tax on withdrawals.
• Withdrawal rate matched to your horizon: 4 percent for about 30 years, lower for longer.
• Future income streams (pension, CPP, OAS, Social Security, part-time work) entered as a reduction from spending, with a bridge sleeve for the years before they start.
• A real return assumption, with the target in today's dollars.
• An account map: which balances can be spent before 59½ in the US, and which withdrawals are taxable or withheld in Canada.
• A downturn plan: how much spending you would cut, or how many years of spending sit outside stocks.
• An annual recheck: rerun the FIRE Number Calculator with updated spending and balances, and compare against the inflation calculator to keep the target in current dollars.

More retirement and savings tools sit on the guides and hubs index.

Sources

• Bengen, W. P., "Determining Withdrawal Rates Using Historical Data", Journal of Financial Planning, October 1994, FPA 2004 reprint: https://www.financialplanningassociation.org/sites/default/files/2021-04/MAR04%20Determining%20Withdrawal%20Rates%20Using%20Historical%20Data.pdf (checked October 2026)
• Cooley, Hubbard and Walz, "Retirement Savings: Choosing a Withdrawal Rate That Is Sustainable", AAII Journal, February 1998: https://www.aaii.com/journal/article/retirement-savings-choosing-a-withdrawal-rate-that-is-sustainable (checked October 2026)
• IRS Topic 558, additional tax on early distributions: https://www.irs.gov/taxtopics/tc558 (checked October 2026)
• IRS, exceptions to tax on early distributions: https://www.irs.gov/retirement-plans/plan-participant-employee/retirement-topics-exceptions-to-tax-on-early-distributions (checked October 2026)
• IRS, substantially equal periodic payments: https://www.irs.gov/retirement-plans/substantially-equal-periodic-payments (checked October 2026)
• Medicare.gov, get started with Medicare: https://www.medicare.gov/basics/get-started-with-medicare (checked October 2026)
• CRA, tax rates on RRSP withdrawals: https://www.canada.ca/en/revenue-agency/services/tax/individuals/topics/rrsps-related-plans/making-withdrawals/tax-rates-on-withdrawals.html (checked October 2026)
• CRA, options for your RRSPs in the year you turn 71: https://www.canada.ca/en/revenue-agency/services/tax/individuals/topics/rrsps-related-plans/rrsp-options-when-you-turn-71/options-your-rrsps.html (checked October 2026)
• CRA, TFSA and other registered plan limits: https://www.canada.ca/en/revenue-agency/services/tax/registered-plans-administrators/pspa/mp-rrsp-dpsp-tfsa-limits-ympe.html (checked October 2026)
• CRA, withdrawing from a TFSA: https://www.canada.ca/en/revenue-agency/services/tax/individuals/topics/tax-free-savings-account/withdraw.html (checked October 2026)
• CRA, understanding the TFSA: https://www.canada.ca/en/revenue-agency/news/newsroom/tax-tips/tax-tips-2021/understanding-tax-free-savings-account-tfsa.html (checked October 2026)
• Canada.ca, when to start your CPP: https://www.canada.ca/en/services/benefits/publicpensions/cpp/cpp-benefit/when-start.html (checked October 2026)
• Canada.ca, when to start your OAS: https://www.canada.ca/en/services/benefits/publicpensions/old-age-security/when-start.html (checked October 2026)