Corrections

M2 Index logs a correction here the day a page states something wrong — a mislabeled series, a stale claim, a data artefact — and never edits the entry away. Below are the first six, found while building this site and published on launch day.

How a correction gets logged

Our revision policy is simple: we catch it ourselves, or a reader emails hello@m2index.com — either way it becomes a dated entry, not a quiet edit. Each entry names what was wrong and what changed. The same date discipline runs the whole site: every figure carries the date we observed it, and rows older than 24 months carry a visible staleness badge. If a ranking page calls a competitor’s number stale, this log is why the claim holds.

Log

2026-08-15 — A growth rate was measured straight across a redenomination

Zimbabwe’s series drops from 671,372,604,617 (2022-04) to 388,815,894 (2022-05) — a factor of 1,727 in one month. That is a redenomination inside the series, not economics. Our ten-year window ran straight through it. We published 38.2%. The rate since the break is 338.2%, annualised over the 47 months to 2026-04, and Zimbabwe now tops the growth ranking.

Six series carry breaks. Equatorial Guinea moved from −3.9% to +12.5%. Grenada and Gambia lost their rates outright: 13 months and 1 month survive their breaks, too little to annualise. We withdrew those numbers rather than restate them. Sierra Leone and Guinea-Bissau were unaffected; their breaks predate the window. The collector now starts each window at the break, so a rate can begin decades after the series’ coverage start year. The growth table carries a Window column and a † on every shortened rate, because 338.2% over 47 months and South Sudan’s 69.6% over ten years are not the same kind of number.

2026-08-15 — US broad money was labelled “M2”

It isn’t the Fed’s M2. It’s the IMF’s harmonised broad money series (MFS_MA/BM_MAI, delivered over IMF SDMX), $30.68T as of 2025-12. The Fed’s own M2 (FRED M2SL) reads roughly $23T for the same period — a $7-8T gap that’s definitional: the IMF measure covers more instruments than the post-2020 Fed M2. Pages now name the series.

2026-08-15 — The US series was marked monthly

It’s quarterly — 138 observations since 1960, not the IMF’s usual monthly frequency. Specs blocks read a default, not the series’ own interval. Frequency now comes from the data itself.

2026-08-15 — Somalia ranked in the USD level table at $71,961

That’s four orders of magnitude below its neighbours. Our USD conversion at period-average rates (World Bank WDI) turned a residual-currency series in an effectively dollarised economy into an artefact, not a real figure. It is also one of only two rows sourced from MFS_DC/DCORP_L_BM — MFS_MA / MFS_DC are different IMF datasets, and the depository-corporations one covers only the shilling that remains. Somalia is now excluded from every level ranking. It stays in the growth rankings, which need no currency conversion.

2026-08-15 — The downloadable CSV counted the euro area twice

One row came from the ECB-sourced series, a second from the IMF’s own euro-area row, G163 — summing the file double-counted roughly $19.9T. G163 now carries exclude_from_rankings and drops from the export. One euro-area row ships.

2026-08-15 — “No national currency since 1999” was false for 9 of 20 members

Eleven founders adopted the euro in 1999; the other nine didn’t — Greece in 2001, Croatia as late as 2023. Pages now give each country its own adoption year, not one date that only fit the founders.