Prime-age employment falls below 50% in an OECD country

Where things stand (data pulled 2026-10-14)

I downloaded the resolving OECD dataflow (DSD_LFS@DF_IALFS_EMP_WAP_Q, ages 25–54, both sexes, quarterly, adjusted and unadjusted) from https://sdmx.oecd.org/public/rest/data/OECD.SDD.TPS,DSD_LFS@DF_IALFS_EMP_WAP_Q,1.0/.EMP_WAP...._T.Y25T54..Q?startPeriod=1990-Q1&format=csvfilewithlabels. All 38 members are in it. The latest values (2026-Q2, seasonally adjusted) are:

• Türkiye 63.3% (unadjusted 63.5%; the seasonal low in Q1 2026 was 62.1%)
• Mexico 74.0%, Costa Rica 74.1%, Italy 75.4%, Chile 76.5%, Colombia 76.9%
• Every other member is above about 79%. The US is at 80.5% for 2026-Q3.

So the lowest member would need to fall about 13 points, and the next-lowest about 24 points. Nothing points to a collapse. TurkStat's August 2026 release had unemployment at 7.8%, employment of 32.5 million and a stable 15+ employment rate (Gazete Oksijen, 2026-09-30). That puts Türkiye's 2026-Q3 figure for ages 25–54 at roughly 63–64%. Türkiye's 15+ headline rate (48.4%) is already below 50%, but it covers a different age range and does not qualify. In July 2026 the OECD's Employment Outlook said OECD employment was at an all-time high. Its September 23, 2026 interim outlook raised global growth to 2.9% despite the Iran-war energy shock.

Reference class and base rate

Using the OECD's annual 25–54 series (OECD.ELS.SAE, DSD_LFS@DF_LFS_INDIC, EMP_RATIO, back to 1960 for some countries) and the quarterly series, no OECD member has ever published a 25–54 rate below 50%:

• Türkiye's lowest annual value is 52.6% (series from 1988). Its lowest quarterly value is exactly 50.0% (unadjusted, 2009-Q1), which would not count under the "strictly below" rule.
• The biggest quarterly drops on record, from 2019-Q4 to the COVID low (adjusted), were Colombia 15.6 points, Chile 13.3, Costa Rica 12.5, Mexico 9.7 and the US 9.0. Türkiye fell only 4.2.
• The biggest multi-year declines were Greece and Spain, about 12–13 points during the euro crisis.

To resolve YES, Türkiye would need about three times its COVID drop, or a Latin American member would need about 1.6–2 times the largest drop ever recorded.

New members offer no route. The latest 25–54 rates of the accession candidates are all high: Brazil 76%, Argentina 81%, Indonesia 76%, Thailand 88%, Peru 81%, Bulgaria 84%, Croatia 85%, Romania 78%. South Africa is at about 51% and was below 50% in 2020–22, but it is neither a member nor a candidate.

Ways to YES (rough probabilities to end-2031)

• A pandemic worse than COVID, with deeper lockdowns: about 0.6%. Roughly 1.5% a year for an event that disruptive, times about 10% that it pushes a Latin American member below 50% for a whole quarter.
• AI-driven displacement: about 0.6%, mostly in 2029–31. Mainstream projections are tiny: Metaculus's labor hub expects AI to cut employment by about 0.7% by 2030. Kalshi prices only about 10% on US unemployment above 10% in December 2030, which corresponds to a 25–54 rate around 75%, far from 50%.
• A Turkish depression-scale collapse, or a war involving a member: about 0.4%. A large war could also stop statistics being published at all.
• A publication or data error: about 0.2%.

Together that is about 1.8%. I spread it across quarters as a hazard rising from 0.05% to 0.11% per quarter. The near quarters are close to zero because the Q3 2026 figures are effectively known, and a qualifying number needs a quarter that ends after the opening date plus a reporting lag of 1–3 months.

Strongest case against my estimate

AI progress could be discontinuous, and the question family seems built to catch that. Türkiye starts only about 13 points above the line, and unadjusted Q1 readings sit about 1.5 points below the adjusted series. If I underrate fast-takeoff scenarios, 3–5% would be defensible. But even extreme AI forecasts rarely imply a fall of 13–30 points in employment for this age group by 2031, so I stay near 2%.

Sources

• OECD quarterly employment-rate dataflow (SDMX link above), retrieved 2026-10-14.
• OECD annual LFS indicators: https://sdmx.oecd.org/public/rest/data/OECD.ELS.SAE,DSD_LFS@DF_LFS_INDIC,/.EMP_RATIO.._T.Y25T54.?startPeriod=1960&format=csvfilewithlabels
• TurkStat August 2026 labour-force figures via Gazete Oksijen, 2026-09-30.
• OECD Interim Economic Outlook press release, 2026-09-23 (oecd.org).
• OECD Employment Outlook 2026 coverage (Personnel Today, 2026-07-08).
• Kalshi and Metaculus AI-unemployment pricing via tech-insider.org, 2026-09-06.

Question. By the last day of each quarter, has the OECD or a national statistical agency reported an employment rate below 50% for people aged 25 to 54 in any OECD member country, for any quarter or year?

Definitions.

• Employment rate: employed persons aged 25–54 inclusive divided by the total population of the same age covered by the official national labour-force statistics, expressed as a percentage. The denominator includes people outside the labour force. Use the source's standard employment definition, including self-employment and any contributing family workers it classifies as employed, and its standard population coverage.
• National rate: the country-wide, all-sexes combined estimate for the entire 25–54 age group, not a regional, sex-specific, occupational, educational, or other subgroup rate.
• OECD member country: a country that is a full OECD member on the publication date of the qualifying estimate, as established by the OECD's official membership record. Accession candidates and OECD-wide aggregates do not count.
• Quarter or year: an official estimate for a calendar quarter or calendar year, not a single month, a quarter-end snapshot, or a forecast. Either a seasonally adjusted or unadjusted estimate qualifies.
• Below 50%: strictly less than 50 using the most precise numeric value published by the qualifying source for that estimate. A value published only as 50.0% does not qualify; a confidence interval extending below 50% does not qualify unless the point estimate is below 50%.

Resolves YES if at least one qualifying official publication or database release reports such a rate below 50%. Preliminary estimates and subsequent revisions qualify. A qualifying revision resolves on its own publication date; a later upward revision does not undo a YES. No attribution to AI is required.

The Resolution window and Observation period conventions apply: the publication must be dated on or after 2026-09-16, and the quarter or year measured must end after that date.

Resolves NO if no qualifying publication has appeared by the horizon. Exclude employment counts, labour-force participation rates, unemployment rates, rates for other age ranges, estimates for subgroups, forecasts, and calculations made by the resolver from separate counts or monthly observations. An officially published 25–54 employment rate derived by the OECD or the national statistical agency from such data does count. Republishing or revising observations for periods ending on or before the opening date does not count.

Resolving source. The OECD's Employment Database, including OECD Data Explorer's Employment rate quarterly dataflow OECD.SDD.TPS, DSD_LFS@DF_IALFS_EMP_WAP_Q (measure EMP_WAP, age Y25T54, sex _T, frequency Q), its annual labour-force statistics and OECD Employment Outlook statistical tables, or their official successors; alternatively, the relevant country's national statistical agency's official labour-force survey release, statistical table, or public database reporting the specified national rate. The OECD and the national statistical agency are independently sufficient; neither must corroborate the other. This is a named-source requirement, not the broader Public confirmation convention A2. Mirrors, news reports, and third-party analyses cannot independently resolve the question. Use publication date under Resolution date (A3), including the dated release or update that first makes a qualifying database value public.