# Fiat-money lesson: chart-data handoff

Snapshot: 16 October 2026. Public U.S. government/FRED data only.

## Ready-to-use files
- fiat_money_monthly.json: chart-ready observations plus sources and methodology.
- fiat_money_monthly.csv: same 236 monthly rows, January 2007 through August 2026.
- raw_snapshot.json: original date/value strings parsed from the official FRED tables, including September 2026 wages and CPI to document their later availability.
- build.mjs: dependency-free offline rebuild and validation.
- validation.json: checks and endpoint statistics.
- sha256.json: hashes for reproducibility.

Reproduce with `node build.mjs` in this directory. No installation, credentials, network access, or paid service is needed. This rebuild uses the captured snapshot; it does not silently fetch revised data.

## Meaning of the chart
All four indices equal 100 in January 2007. Level-series formula: 100 × value / January 2007 value. Real hourly earnings: 100 × (hourly earnings / CPI) / (January 2007 hourly earnings / January 2007 CPI).

Latest common month is August 2026 because M2's latest published month is August. Wages and CPI extend to September, which is intentionally excluded from the common-range chart.

August 2026: M2 328.05; nominal hourly earnings 183.39; CPI 164.24; real hourly earnings 111.66. Thus, average real earnings are about 11.7% above January 2007; this entire window does not show a decline in average purchasing power per hour. Inflation did erode the larger nominal gain. A fixed dollar amount retains roughly 60.9% of its January 2007 purchasing power against this CPI measure.

## Required chart behavior
- Use dates as monthly labels, not release dates.
- Keep October 2025 CPI and real earnings null/blank; show a gap and a missing-data note. Do not bridge, interpolate, or convert it to zero.
- Use the same time range and base across series.
- Label nominal earnings and real earnings explicitly.
- Prefer a separate real-earnings panel so the much larger M2 increase does not flatten its movement.
- Include “Monthly, seasonally adjusted; January 2007=100. Retrieved 16 October 2026. Data subject to revision.”
- Do not use the M2/wage gap as a numerical estimate of lost purchasing power or as proof of causation.

## Sources and limitations
M2: Board of Governors of the Federal Reserve System, via FRED.
https://fred.stlouisfed.org/series/M2SL
https://fred.stlouisfed.org/data/M2SL

Hourly earnings: U.S. Bureau of Labor Statistics, via FRED. All employees, total private, dollars per hour. This is an average of a changing group of jobs, not median wages or the experience of the same worker over time. Composition changes are especially important in 2020.
https://fred.stlouisfed.org/series/CES0500000003
https://fred.stlouisfed.org/data/CES0500000003

Consumer prices: BLS CPI-U all items, U.S. city average, via FRED. The basket does not represent every household, and CPI is not an asset-price index.
https://fred.stlouisfed.org/series/CPIAUCSL
https://fred.stlouisfed.org/data/CPIAUCSL

October 2025 CPI was not published because of the lapse in appropriations. BLS approximation used for seasonal-factor computation does not supply a published observation for this chart.
https://www.bls.gov/cpi/additional-resources/2025-federal-government-shutdown-impact-cpi-faq.htm

These are latest revised values retrieved on the stated date, not a real-time vintage for each historical month. CES undergoes routine monthly and seasonal revisions, and CPI seasonal data also revise. August 2026 earnings are still preliminary at this snapshot.
https://www.bls.gov/web/empsit/cesfaq.htm
https://www.bls.gov/cpi/seasonal-adjustment/

M2 incorporates annual seasonal revisions. The July 2026 change in IRA/Keogh netting methodology applies historically; do not splice this snapshot with earlier vintages. The September 2026 H.6 release independently confirms the endpoint.
https://www.federalreserve.gov/releases/h6/20260324/
https://www.federalreserve.gov/releases/h6/20260728/
https://www.federalreserve.gov/releases/h6/20260922/

Official FRED graph CSV downloads were not accessible through the available web tool. The official HTML data tables were accessible, so date/value rows were parsed directly without manual transcription. All 236 dates were checked. There are 235 complete three-series observations and one published missing CPI value. No forecasts or imputations were added. A comprehensive historical-vintage audit was not performed.