Retirement Budget Benchmark Calculator

Updated August 6, 2026. Quick answer: the honest benchmark is narrower than most “retirement budget” tools pretend. We computed this from the Bureau of Labor Statistics Consumer Expenditure microdata ourselves, so we know exactly what it does and does not support: three age bands, three housing situations, and three spending categories that we can stand behind. Everything else is shown as a residual, and food is not in it — for a reason we explain below rather than hide.

What the tool is actually telling you

It answers one question: what do households like this one actually spend? It is a benchmark, not a budget. An average is a description of other people, not a recommendation for you, and the gap between your figure and this one is information rather than a verdict.

The median leads, and that is deliberate. For 65–74 the average total is $64,232 while the median is $50,068 — a gap of about $14,164. Averages in spending data are dragged upward by the highest spenders, so a tool that leads with the mean quietly tells most readers they are behind. The median is where the middle household actually is.

Housing situation changes the answer more than age does. That is the finding worth taking away: among 65–74 households, those still carrying a mortgage spend $25,927 on housing against $18,928 for outright owners — and their total spending differs by roughly $11,671 a year. It is also the one lever on this page that a household can actually pull.

What this tool deliberately will not do

Most calculators in this class offer more dimensions than their data supports. Ours offers exactly what survived the check, and here is the whole list of what was refused:

• No income tier. The study does not compute spending by income tier, so we do not offer it. A dropdown we could not populate honestly is worse than a missing one.
• No household size. Not a computed dimension — only an average family size per band (it falls from 2.18 at 55–64 to 1.587 at 75+, which is itself part of why spending falls).
• No food line, and this is the important one. Food is not a category we can publish from this source. In the Interview microdata it reconciles at only 0.55 to 0.62 against the published totals, because food is predominantly collected in the CE Diary survey. We would rather show you three categories we trust than four with one that is silently wrong by a third. Food sits inside the residual line.
• Two housing situations are missing. Households occupying without paying cash rent, and student housing, are in the data but too thin to publish — between one and forty-eight records per band. One of those cells contained a single household whose spending was more than three times its age group’s average. Publishing it as a benchmark would have been indefensible.
• Health and transport do not vary by housing situation here. They are computed for the whole age band, so the tool holds them constant while housing and the total change. The note under the table says so on every combination.

Methods

Source. BLS CE PUMD Interview 2024 — 23,176 consumer-unit records, weighted, with Fay-adjusted balanced repeated replication (k=0.5) for standard errors. Computed by us from the public microdata, not copied from a published table.

Reconciliation. Against BLS Table 1300, our totals land at 0.978 to 0.983 of the published figures across the three bands, with health at 0.946–0.973 and transport around 0.96. Housing sits near 0.915, which is the largest single gap and is disclosed rather than smoothed.

Housing-situation codes are the CE’s own CUTENURE values, read from the BLS data dictionary: owned with a mortgage, owned without one, and rented. We did not infer them from the spending pattern.

Standard errors are published beside every total. A benchmark that hides its error bars is asking to be read with more confidence than it has earned.

Sources

Bureau of Labor Statistics Consumer Expenditure Public-Use Microdata, Interview 2024; the CE PUMD data dictionary for the housing-situation codes; and BLS Table 1300 for reconciliation. All read 2026-08-06. The computed study and its permanent identifier are on our research page, free to reuse with attribution.

Honest gaps. Food is excluded, as explained above. The three published categories cover roughly 55% to 63% of total outlay depending on age band; the rest is a residual we do not break down. These are 2024 survey figures. And spending is not the same as needing to spend — households spend what they have, so a low benchmark in an older band partly reflects lower resources rather than lower needs.

The narrative companion to this tool is what retirement actually costs. To test a balance against a spending level, the can-I-retire calculator, and for where a given balance sits in the distribution, the amount series.

See methodology and corrections. General information, not financial advice. No advertising appears on this page.