Updated August 7, 2026. Quick answer: the number is close to right against one denominator and badly wrong against the one most people assume. Households headed by someone 65–74 spend 76.9% of what 55–64 households spend — near enough to 80%. But they spend only 53.8% of what 55–64 households earn. The rule almost never says which denominator it means. Aim at 80% of income and you are aiming at $97,257 a year; households 65–74 actually spend $65,354 — a gap of $31,903.
The claim: “You need 80% of your pre-retirement income to retire”
Verdict: MUTATED — Right figure, wrong meaning.
Evidence tier: primary. Verdict class chosen from the taxonomy fixed before the evidence.
Where the number comes from
The 80% figure circulates as a planning rule with no attribution, so we went to the one federal survey that measures what households actually spend: the Bureau of Labor Statistics Consumer Expenditure Survey, Table 1300, Age of reference person: Annual expenditure means, 2024 — the most recent published year.
Read from bls.gov — CE published tables, Table 1300 (2024) on 7 August 2026.
We could not verify the rule’s origin, and we are saying so rather than implying one. The 80% rule is widely attributed to mid-century replacement-rate research and, separately, to the 4% withdrawal literature. We attempted the Social Security Administration’s own replacement-rate publications and were refused at the door: ssa.gov returned HTTP 403 on every path we tried, including its homepage. We did not spoof a browser identity to get around it. So this page audits what the number does against real data, not where it was born.
What that source actually measured
Here is what the CES actually measured in 2024. These are means, not medians, and each column is a different set of households:
| Age of reference person | Income before taxes | Annual expenditures |
|---|---|---|
| 55–64 | $121,571 | $84,946 |
| 65–74 | $75,460 | $65,354 |
| 65 and older | $67,462 | $61,432 |
| 75 and older | $56,028 | $55,834 |
Run the rule against each denominator and it gives two different answers:
- Spending against spending. $65,354 ÷ $84,946 = 76.9%. The rule is approximately right.
- Spending against income. $65,354 ÷ $121,571 = 53.8%. The rule overshoots by more than 26 points.
- Take all households 65 and older and both fall further: 72.3% of pre-retirement spending and 50.5% of pre-retirement income.
That is the mutation. “80% of pre-retirement income” states the income denominator explicitly and is wrong by that measure. It is roughly correct only if you silently swap in spending — which is a different quantity, lower than income by every saved and taxed dollar.
The honest current figure
Retiree spending falls with age, and the rule has no mechanism for that. Households 75 and older spend $55,834 — 65.7% of the 55–64 figure. A single fixed percentage applied at 65 is describing the most expensive year of a declining series as though it were the average.
The gap is also narrowing from the other end. At 75 and older, income ($56,028) and spending ($55,834) are within $200 of each other. Whatever the rule is doing at 65, it is not describing later retirement at all.
The “$1 million” sibling — reported, deliberately not given a verdict
The same advice usually travels with a dollar version: you need $1 million to retire. We are not issuing a verdict on it, and the reason is a rule of this desk. We could not verify its origin: four independent routes to the 1994 withdrawal-rate paper it is usually traced to failed on 7 August 2026 — a TLS certificate error, a dead redirect, a 404, and one archive host our tools will not fetch. Judging a rule of thumb against a statistic it never claimed to describe would be the same error this desk audits others for.
What we can report is what American households actually hold. The Federal Reserve’s Survey of Consumer Finances (2022 wave, 2022 dollars) puts retirement-account balances for families headed by someone 65–74 at a median of $200,000 against a mean of $609,230 — the mean is 3.0 times the median, because a small number of very large balances drag it upward.
Correction, 7 August 2026. This page first published these figures as a median of $207,100 and a mean of $945,300, taken from a summary of the SCF Bulletin rather than from the survey’s own tables. Both were wrong. The figures above are read directly from the Federal Reserve’s published Table 6 (“Family holdings of financial assets”), retirement-accounts column, for the 65–74 age row. The error overstated the mean by more than $336,000 and the mean-to-median ratio as 4.6× instead of 3.0×. Corrected here rather than edited silently, because a page that audits other people’s sourcing has no business hiding its own.
Read from federalreserve.gov — 2022 SCF, Table 6 (Family holdings of financial assets) on 7 August 2026.
Two things that figure does not mean. It counts only families who have a retirement account at all — the same Fed table puts that at 54.3% of all families, so the 45.7% with no retirement account whatsoever are outside the median entirely. And a median is not a target: it describes what people have, never what anyone needs.
The Fed attaches its own caution to the means in this table: “Because estimates in this table have not been reviewed for robustness, they may be sensitive to outliers.”
What to cite instead
“Households headed by someone 65-74 spend 76.9% of what 55-64 households spend, but only 53.8% of what they earn (BLS Consumer Expenditure Survey, Table 1300, 2024). The “80% of pre-retirement income” rule is approximately right against pre-retirement SPENDING and overshoots by more than 26 points against pre-retirement INCOME – and it rarely says which it means.”
Source: Clear Money Guide, The Provenance File, 2026. Free to reuse with attribution under CC BY 4.0.
What this audit does not establish
🔴 These are different households, not the same households followed over time. The CES is cross-sectional: the 65–74 group in 2024 is not the 55–64 group of ten years earlier, grown older. They differ in lifetime earnings, pension coverage and survival. A true replacement rate requires following the same household across retirement, and this table cannot do that. Every ratio on this page is a comparison between age groups in one year, and should be read as such.
They are means, and means are pulled by the top. The CES publishes means in Table 1300. Medians would very likely be lower on both income and spending, so the spending-to-income ratio in particular is not necessarily what a typical household experiences. BLS publishes standard errors alongside: $4,133 on 55–64 income and $1,648 on 65–74 spending, among others.
This audit does not say what any individual needs. It tests a population rule against population data. Housing tenure alone can move a household’s answer more than the rule’s entire margin, and health costs are the acknowledged tail risk the averages hide.
The rule’s origin remains unestablished, for both the 80% and the $1 million variants, and ssa.gov was unreachable to us on the day we looked.
Related
The decision itself is owned elsewhere on this site and this page does not restate it: can I retire with $1 million · with $1.5 million · with $2 million · retiring on Social Security alone.
This is an audit in the Provenance File. Our method, the six verdict classes and the desk laws are published at how we check money claims. If you think this is wrong, tell us — with a source. No advertising appears on this page and we earn nothing from it.