Your Retirement Number and How to Compute It
A working professional can usually recite their salary, their bonus, and their tax bracket without hesitation, yet ask the same person how much total savings they need before they can stop working and the answer is often a shrug or a round guess. A retirement number replaces that guess with an actual calculation built from the person's own spending and income facts.
The core principle: turning a spending target into a savings target
A retirement number is the total portfolio value a person needs to accumulate to sustain their desired spending for the rest of their life without running out of money. The most common shortcut for calculating it relies on the concept of a safe withdrawal rate, a percentage of a portfolio that can reasonably be withdrawn each year, adjusted upward for inflation annually, without a high probability of depleting the portfolio over a multi decade retirement. The widely cited starting point is a 4% withdrawal rate, based on historical analysis of how various stock and bond portfolio mixes held up across rolling 30 year retirement periods in market history, though many planners today lean toward a somewhat more conservative 3.25% to 3.5% given lower expected future returns and longer life expectancies for early retirees in particular.
Rearranging the withdrawal rate relationship produces the retirement number directly: retirement number = annual spending needed from the portfolio / withdrawal rate. Because dividing by a percentage is mathematically the same as multiplying by its reciprocal, a 4% withdrawal rate is exactly equivalent to needing 25 times annual spending, since 1 / 0.04 = 25. A more conservative 3.5% withdrawal rate is equivalent to roughly 28.6 times annual spending, since 1 / 0.035 ≈ 28.6, and a 3% rate is equivalent to roughly 33.3 times spending.
The critical input that most people get wrong is not the withdrawal rate itself but the spending figure it is applied to. Retirement spending is rarely identical to current working years spending: a mortgage may be paid off by retirement, work related costs, commuting, professional wardrobe, childcare during working hours, typically disappear, and payroll taxes for Social Security and Medicare no longer apply to withdrawal income the same way they applied to a paycheck. At the same time, healthcare costs before Medicare eligibility and increased travel or leisure spending in early retirement years can push spending upward. The retirement number calculation is only as accurate as the spending estimate feeding into it, which is why estimating that figure carefully deserves at least as much attention as choosing between a 3.5% and 4% withdrawal rate.
The math: two worked examples of computing a retirement number
Worked example 1: a straightforward retirement number with guaranteed income subtracted. Suppose a professional projects $140,000 a year in retirement spending, a figure built from current spending adjusted for a paid off mortgage and reduced work related costs. At a 4% withdrawal rate, the gross retirement number is $140,000 / 0.04 = $3,500,000, equivalently 25 x $140,000 = $3,500,000. Suppose this professional also expects $32,000 a year in Social Security benefits starting at full retirement age. Because that income is guaranteed and does not need to be funded from the portfolio, it should be subtracted from the spending figure before applying the multiplier, not subtracted from the final dollar target. The portfolio only needs to cover $140,000 − $32,000 = $108,000 a year, producing an adjusted retirement number of $108,000 / 0.04 = $2,700,000, a full $800,000 lower than the unadjusted figure, purely from correctly accounting for a guaranteed income source.
Worked example 2: comparing withdrawal rate assumptions for an early retiree. Consider a professional planning to retire at 55 rather than 65, meaningfully extending the number of years the portfolio must last, potentially 35 to 40 years rather than the 30 year period the classic 4% research was built around. This professional projects $120,000 a year in spending with no guaranteed income until Social Security begins at 67. Using a 4% rate, the naive retirement number is $120,000 / 0.04 = $3,000,000. Given the longer retirement horizon, this professional instead applies a more conservative 3.25% rate, producing $120,000 / 0.0325 ≈ $3,692,000, an increase of roughly $692,000, or about 23% more than the 4% figure, reflecting the additional years the portfolio needs to sustain withdrawals before any guaranteed income arrives to share the load.
It is worth extending example 1 to show how a partial pension or a spouse's guaranteed income changes the calculation further. If this same professional's household also expects a modest employer pension of $15,000 a year on top of the $32,000 in Social Security, the portfolio only needs to cover $140,000 − $32,000 − $15,000 = $93,000 a year, producing a further adjusted retirement number of $93,000 / 0.04 = $2,325,000, now more than a million dollars below the original unadjusted $3,500,000 figure. This illustrates why accurately inventorying every source of guaranteed retirement income, not just the largest one, materially changes the target a person is actually saving toward.
What the research on withdrawal rates actually supports
The 4% figure traces back to historical simulation research testing how various stock and bond allocations would have performed across many overlapping 30 year retirement periods drawn from market history, seeking the highest withdrawal rate that would have survived the worst historical starting points without depleting the portfolio. The research supports 4% as a reasonable, historically tested starting point for a roughly 30 year retirement with a balanced stock and bond allocation, not as a guarantee that applies identically to every portfolio composition, time horizon, or future market environment.
Subsequent research and updated market data have generally supported using a somewhat lower rate for retirements expected to last meaningfully longer than 30 years, an increasingly common situation for professionals retiring in their fifties, and for portfolios with a more conservative asset allocation than the roughly 50% to 75% equity mix the original research emphasized. This is the basis for the increasingly common practice of using 3.25% to 3.5% rather than a flat 4% for early retirement planning, a more conservative buffer rather than a rejection of the underlying research.
It is also worth noting what the historical research does and does not claim about market conditions going forward. The 4% figure was derived from past data and represents a rate that survived the worst historical starting points observed, not a rate mathematically guaranteed to survive every conceivable future scenario. Treating it as a well tested, reasonable planning assumption, subject to periodic review as new market data accumulates, is a more accurate reading of the research than treating it as an immutable law of retirement finance.
Applying this to a real professional's plan
For a high earning professional, the practical starting point is building an honest, itemized estimate of retirement spending rather than reflexively assuming current spending will simply continue unchanged. This means walking through major categories explicitly: housing, expected to drop meaningfully if a mortgage is paid off by retirement age; healthcare, expected to rise, particularly in the years before Medicare eligibility begins; taxes, which shift in structure once income comes from withdrawals and Social Security rather than wages; and discretionary spending, travel and hobbies in particular, which often rises in the early, more active years of retirement before tapering later.
The next step is inventorying every source of guaranteed income the household will actually receive, Social Security for each spouse, any pension, rental income from a property intended to be kept rather than sold, and subtracting the total from projected spending before applying the withdrawal rate multiplier, exactly as shown in the worked examples above. Skipping this step and applying the multiplier to gross spending instead of net spending is one of the most common errors in a do it yourself retirement number calculation, and it can overstate the required portfolio by hundreds of thousands of dollars for a household with a meaningful pension or dual Social Security benefit.
Professionals planning an earlier than typical retirement, a real possibility for high earning fields with strong savings capacity during the working years, should also lean toward the more conservative end of the withdrawal rate range, since a longer retirement horizon is the single factor most likely to expose an aggressive withdrawal assumption to a genuinely bad, multi decade sequence of poor early returns, a risk covered in more detail under sequence of returns risk elsewhere on this site.
Professionals with variable income, a physician with locum tenens or side consulting income, an attorney with an irregular bonus structure, should also stress test their retirement number against a range of spending assumptions rather than a single point estimate. Calculating the number under both a base case and a somewhat more conservative spending assumption gives a sense of how sensitive the target actually is to the underlying estimate, which is useful information when deciding how aggressively to save in years when income allows for it.
Finally, a retirement number is not a figure calculated once and left untouched. Spending expectations, health status, family circumstances, and market conditions all evolve, and a number calculated at 40 deserves to be recalculated at 50 and again at 55 as actual circumstances, and actual portfolio performance, become clearer than they could be a decade or two out.
Actionable breakdown
- Estimating retirement spending
- Build spending estimates from categories, not a single guess.
- Account for a paid off mortgage reducing housing costs.
- Budget separately for healthcare before Medicare eligibility.
- Calculating the number
- Subtract Social Security and pension income before multiplying.
- Use 3.25% to 3.5% for retirements expected to exceed 30 years.
- Use 4% as a reasonable baseline for a standard length retirement.
- Keeping it current
- Recalculate the number every few years, not once permanently.
- Update guaranteed income estimates as benefits are confirmed.
- Adjust the withdrawal rate if planned retirement length changes.
Common pitfalls
Applying the multiplier to gross spending instead of spending net of guaranteed income: this overstates the required portfolio, sometimes by hundreds of thousands of dollars.
Using current working years spending instead of a projected retirement spending estimate: mortgage payoff, work costs, and healthcare needs typically shift the real figure meaningfully.
Treating the 4% rule as a guarantee rather than a historically tested starting point: it performed well across most historical periods but is not immune to a genuinely unfavorable future sequence of returns.
Ignoring how a longer than typical retirement changes the required withdrawal rate: an early retirement calls for a more conservative rate than the 30 year period the original research emphasized.
The bottom line
Divide expected annual retirement spending, net of Social Security and any pension, by a withdrawal rate appropriate to your expected retirement length, and recalculate the resulting number every few years as your actual circumstances become clearer.
The 20 percent savings rate for late starters · Why savings rate beats returns early on · Withdrawal strategies · Financial independence guide · Withdrawal rate (glossary) · All articles