BACK IN
MY PAY.
Data status ↗

Our standards / methodology

The fine print is the point.

Four different questions

Nominal change compares source-dollar prices. Real change adjusts the starting price by the CPI ratio. Price/income divides compatible annual price and gross household-income statistics. Work hours divide a price by either an explicitly supplied hourly wage or a verified wage benchmark included in the selected release. These metrics are not interchangeable.

Matched periods, honest gaps

Annual new-home medians come from the publisher’s Annual sheet, never an average of monthly medians. Household income uses H-5 All Races, median current dollars. Monthly prices do not receive an annual-income affordability verdict. Missing observations remain missing.

Source estimates and scenarios

Survey medians are estimates, not exact counts. Income duplicate-year rows use the first row, representing newer processing or population controls; row labels and footnotes are preserved. Rates are derived arithmetic means of all published weekly observations in a complete year. Personal inputs are scenarios, never source estimates.

Wage benchmarks and coverage

When included in an approved release, wage benchmarks match the price’s month or year. Average hourly earnings describe private production and nonsupervisory employees: mean gross payroll per paid hour, not median individual pay. Changes in industry and worker mix can change that mean. The general federal minimum starts in 1978 and uses calendar-day weighting of effective rates for period comparisons; it is not an observed wage or a legal-applicability calculator. State and local rates, tipped wages, youth and other exceptions, and territory-specific rules are outside that benchmark. Gross work hours exclude taxes and living costs. Missing wages remain unavailable; household income is never divided into a supposed observed hourly wage.

Precision and reproducibility

Method 1.0.0 uses IEEE-754 double arithmetic, stable log1p/expm1 mortgage math, and rounding only for display. It is an educational comparison, not a settlement ledger. JSON retains unrounded numeric results. Public receipts pin release, method and benchmark selections; old versions must be retained.

What the data cannot establish

Separate population medians do not describe actual buyers. National new homes are not local starter homes. CPI is not a dollar price. None of these tools identifies political causation, individual effort or the fortunes of a birth cohort.

Review still required

The data library includes annual CPI from 1913, Census home prices from 1963 and household income from 1967. Individual BLS price series begin later; monthly coverage extends through October 2026 and completed annual coverage through 2025. All 730 overlapping BLS values match the earlier API archive. Historical average-price methodology caveats are retained in the source notes. Freddie Mac PMMS is excluded pending reuse clearance; mortgage calculations use rates you enter. The October 2025 CPI gap is preserved.