Pre-registration
Elevated US JOLTS job openings are a leading indicator of broad private-sector wage growth; when job openings accelerate, average hourly earnings growth follows within six months.
Falsification criterion — what would disprove this
This hypothesis is considered falsified if:
SUPPORTED if monthly YoY log growth in JOLTS job openings predicts private average hourly earnings YoY six months later with Pearson r >= 0.45, and months in the top quartile of openings growth show median six-month-ahead wage-growth acceleration of at least +0.25 percentage points. REFUTED if r <= 0.15 or median acceleration <= 0. PARTIAL otherwise. Exclude March 2020 through December 2021 from the primary correlation because pandemic labour-market measurement and reopening distort the signal.
formal test & threshold
test: fred_jolts_openings_lead_private_ahe_yoy_6m threshold: SUPPORTED: lead correlation >= 0.45 and top-quartile median wage-growth acceleration >= +0.25pp; REFUTED: correlation <= 0.15 or acceleration <= 0.
Method
- Template
descriptive- Clustering
none- Sample
- 1 countries · 2007 – 2026
- Evidence type
- associational
Monthly lead-lag event-window diagnostic; openings YoY at t is compared with wage YoY at t+6, and wage acceleration is wage_yoy(t+6) minus wage_yoy(t).
Data
| Variable | Source | Transform |
|---|---|---|
private_average_hourly_earnings_yoy outcome | fred:CES0500000003tier 1 | 100 * log difference from 12 months earlier |
jolts_job_openings_yoy treatment | fred:JTSJOLtier 1 | 100 * log difference from 12 months earlier |
● ready · ● pending · ● reconstruct-needed
Detailed result card
Result card - jolts_openings_wage_pressure_us_2007_2026
Verdict: refuted - lead r=-0.219; top-quartile median 6m wage acceleration=0.07pp; n=200
Predeclared Test
Support requires openings YoY to correlate at least 0.45 with private AHE YoY six months later, and top-quartile openings-growth months to show at least +0.25pp median wage-growth acceleration six months later.
Primary sample excludes 2020-03 through 2021-12. Vintages: data/vintages/fred/JTSJOL@2026-04-30T114542Z.parquet, data/vintages/fred/CES0500000003@2026-04-30T114914Z.parquet.
Strongest opposing argument
Every hypothesis ships with its charitable opposing argument. The framework earns credibility by handling objections at their strongest, not weakest.
Notes
FRED-runnable labour-market tightness test using JOLTS job openings rather than quits because the on-disk BLS quits vintage is too short for a serious event-window test.