Economy · labor marketLatest available releases through 2026-08-20
US Jobs Tracker
The labor market is stable, but hiring is weak: the Jobs Score is 63/100. Layoff pressure remains the strongest part of the picture at 80/100, while labor demand is weakest at 33/100. The latest payroll report lost 23k jobs and the three-month average slowed to +20k, but weekly claims and the JOLTS layoffs rate do not show broad firing. The evidence describes a low-hire, low-fire market—not a hiring boom and not yet a generalized layoff cycle.
Latest payroll change
-23k
July 2026 · 3m avg +20k
Unemployment
4.1%
U-3 headline · U-6 7.9%
Openings per unemployed
1.06×
7.36M openings · 6.92M unemployed
Sources, methodology & freshnessLast updated 2026-08-20 · Open ↓Close ↑
Source
US Bureau of Labor Statistics and Department of Labor via BLS and FRED; all scored histories and major-sector payrolls are committed and downloadable
Methodology
Thirteen de-duplicated inputs across four equal-weight dimensions. Fixed weak/strong bounds; JOLTS inputs use rates; claims are normalized per 100,000 payroll jobs. Pay and productivity are context only. History is release-aligned and current-vintage.
Updates
Daily pipeline; changes on weekly claims, monthly Employment Situation and monthly JOLTS releases|Last: 2026-08-20
Maintained & reviewed by Yuriy Matso — methodology shown on the page.
The Jobs Score through time — release-aligned since 2008
Every line segment holds the labor evidence actually available on that date. A monthly observation enters on its publication date; weekly claims update between jobs reports. Red bands mark final NBER recessions for context. The scoring function and bounds are identical throughout the replay.
4,863 business-day snapshots. Historical source values are today's revised vintages; release timing is preserved, complete pre-launch revision history is not.
02
The four labor dimensions — where the disagreement lives
Monthly snapshots show the sequence behind the headline. The current split—contained layoffs but weak hiring demand—is exactly the kind of condition a single payroll number hides.
contraction (<30)weak 30–45cooling 45–60stable 60–75broad strength (75+)no data
03
Every scored input and its fixed bound
Inputs map linearly between the red weak bound and green healthy bound, then clamp at 0 or 100. JOLTS turnover enters as rates; initial and continued claims are normalized per 100,000 payroll jobs. Inputs average within their dimension and the four dimensions receive equal weight. Open any input name for its complete chart and methodology.
The latest three labor releases — one joined answer
Employment Situation · 2026-08-07
July 2026 jobs report
Payroll change
-23k
Three-month average
+20k
U-3 / U-6
4.1% / 7.9%
Prime-age employed
80.4%
Weekly hours
34.3
JOLTS · 2026-08-04
June 2026 labor flows
Openings
7.36M
Openings / unemployed
1.06×
Openings rate
4.4%
Hires / quits
3.4% / 2.0%
Layoffs rate
1.1%
Weekly claims · 2026-08-20
The fastest layoff check
Initial claims, 4-week
204,000
Initial reference week
2026-08-15
Continued, 4-week
1,789,000
Continued reference week
2026-08-08
The score normalizes both counts by payroll employment; the display keeps the familiar headline levels.
05
Where the July 2026 jobs went — sector contributions
06
Pay and productivity — important context, not score votes
Nominal wage growth
3.2%
Average hourly earnings, YoY
Real wage growth
-0.2%
Hourly earnings minus CPI inflation
Labor productivity
+2.2%
Nonfarm business output per hour, YoY
Unit labor costs
+1.4%
Labor cost per unit of output, YoY
These measures describe job quality and inflation pressure, but they do not vote on whether employment is expanding. Faster wages can improve household income while also keeping services inflation sticky; productivity determines how much wage growth the economy can absorb without comparable cost pressure.
07
What changed over the last 30 days
Headline
+1
Jobs Score points since 2026-07-21.
Largest deterioration
Employment growth · -12
The dimension subtracting most from the latest month-over-month comparison.
Largest improvement
Worker utilization · +11
The dimension adding most to the latest comparison.
08
Historical alignment with recessions and labor contraction
One observation per month prevents weekly carry-forward from inflating the sample. “In recession” uses final NBER dates. “Labor contraction within 6m” is independent of the score: either the three-month payroll average turned negative or U-3 rose at least 0.5 percentage point during the following six months. Because the inputs are current revised histories, this validation is descriptive. It is not a vintage-pure forecast test.
Jobs band
Months
In recession
Labor contraction within 6m
Broad strength
40
0.0%
12.5%
Stabletoday
71
1.4%
11.3%
Cooling
70
0.0%
24.3%
Weak
14
28.6%
35.7%
Contraction
29
51.7%
89.7%
Broad strength
Months
40
In recession
0.0%
Labor contraction within 6m
12.5%
Stable
Today
Months
71
In recession
1.4%
Labor contraction within 6m
11.3%
Cooling
Months
70
In recession
0.0%
Labor contraction within 6m
24.3%
Weak
Months
14
In recession
28.6%
Labor contraction within 6m
35.7%
Contraction
Months
29
In recession
51.7%
Labor contraction within 6m
89.7%
09
What would change the current read
The fixed bands sit at 75/60/45/30. Today's 63 is 12 points below broad strength and 4 points above cooling. The bullish path is a recovery in the hires rate, payroll breadth and the three-month payroll average while claims stay contained. The bearish path is current weak demand turning into firing: rising normalized initial and continued claims, a higher JOLTS layoffs rate and U-3/U-6 moving above their twelve-month trends.
10
What this tracker cannot tell you
This is a national labor-market breadth score, not a probability, payroll forecast or measure of every worker's experience. It will miss regional and occupational pain; household and establishment surveys can diverge; JOLTS is lagged and revised; claims cover only eligible unemployment-insurance filers. Historical release timing is preserved, but old values use today's revisions. The append-only payroll-vintage archive begins 2026-08-16 and currently contains 1 snapshot; it does not invent unavailable pre-launch first estimates.
Employment growth asks whether jobs and hours are expanding; labor demand asks whether employers are hiring and workers are moving; layoff pressure asks whether job loss is spreading; worker utilization asks how much broad slack remains. Keeping the questions separate exposes low-hire, low-fire markets that one headline cannot describe.
2
Rates and normalized counts make history comparable
JOLTS hiring, quits and layoffs enter as rates rather than raw levels. Initial and continued claims are divided by total payroll employment and expressed per 100,000 jobs. This prevents a larger modern workforce from mechanically appearing weaker than the smaller workforce of earlier cycles.
3
Breadth and leading labor inputs sit beside payrolls
The three-month payroll average is joined by BLS payroll diffusion, average weekly hours and temporary-help employment. Payroll diffusion asks how many private industries are adding jobs; hours and temporary staffing can turn before employers change permanent headcount.
4
Dimensions receive equal weight
Each input maps linearly between fixed weak and strong bounds and clamps at 0 or 100. Inputs average inside their dimension, then the four dimensions average equally. Three different layoff series therefore cannot overpower employment growth or worker utilization merely because they update more often.
5
Release timing is preserved; revisions are disclosed
A June JOLTS observation enters history when it was released, not in June. Historical values use today’s revised source histories, however. An append-only jobs-report snapshot archive starts at launch so future payroll revisions can be shown honestly without pretending earlier first estimates were stored.
Who Uses Jobs Tracker
Macro investors
Distinguish a genuine layoff cycle from a quieter low-hire, low-fire slowdown.
Rate watchers
See whether wage pressure is occurring alongside strong demand or deteriorating employment.
Business owners
Compare hiring, hours and industry job breadth before making a national-demand conclusion.
Research writers
Use the dated score, sector reconciliation and linked source series as an auditable labor-market summary.
Pro Tips
01
Start with the dimension gap
Low layoffs do not mean strong hiring. When layoff pressure scores well but labor demand scores poorly, the correct description is a frozen or low-hire, low-fire market—not boom and not broad contraction.
02
Treat openings as intent
An opening becomes stronger evidence when hiring and quits confirm it. The openings-per-unemployed ratio is useful, but vacancies that remain unfilled can overstate effective demand.
03
Watch hours before payrolls
Employers can trim schedules faster and more cheaply than headcount. A sustained decline in weekly hours and temporary help can flag weakening labor input before total payroll employment contracts.
04
Separate worker pay from job health
Nominal wages, real wages, productivity and labor costs are displayed as context. Faster wages help workers but may increase inflation pressure, so they do not receive a simplistic healthy/unhealthy vote.
Common Issues & Solutions
Why not use payroll growth alone?▾
Payrolls are revised and can be concentrated in a few industries. The tracker adds breadth, hours, temporary staffing, hiring, quits, layoffs, claims and household-survey utilization so one estimate cannot dictate the verdict.
Why is strong layoff pressure wording associated with a high score?▾
The dimension score measures health. It is never the raw layoff rate. A high Layoff Pressure score means pressure is low: claims and employer-initiated separations remain contained.
Does 63 mean a 63% chance jobs will grow?▾
No. It is a fixed-bound breadth score. It means the four dimension scores currently average 63 on the published 0–100 labor-health scale.
Is this a real-time-vintage backtest?▾
Not fully. Release dates are respected, but old observations include later source revisions. The page labels its historical results descriptive. The project’s append-only payroll snapshot archive begins at launch and will make future revision analysis vintage-aware.
Frequently Asked Questions
Is the US job market healthy right now?▾
The Jobs Tracker answers from four separate dimensions: employment growth, labor demand, layoff pressure and worker utilization. The live verdict changes whenever a weekly claims release, monthly jobs report or JOLTS report changes the underlying evidence.
What indicators are in the Jobs Score?▾
The score uses the three-month payroll average, payroll diffusion, weekly hours, temporary-help growth, openings per unemployed worker, hiring and quits rates, workforce-normalized initial and continued claims, the JOLTS layoffs rate, U-3 and U-6 relative to their twelve-month averages, and prime-age employment relative to trend.
What is a low-hire, low-fire labor market?▾
It is a market where employers are reluctant to add workers but are also reluctant to lay off existing staff. Hiring, quits and openings cool while claims and layoffs remain low. That produces weak labor demand but healthy layoff pressure in this tracker.
Why use payroll diffusion?▾
Total payroll growth can be supported by only a few large industries. The BLS diffusion index counts whether roughly 250 private industries are increasing or decreasing employment. Above 50 means gains are broader than losses; below 50 means contraction is broader.
How often does the Jobs Tracker update?▾
The file rebuilds in the daily pipeline. Initial and continued claims can change weekly; the Employment Situation and JOLTS components change monthly on their release dates. The score holds the last known information between releases.
Does the Jobs Tracker predict recessions?▾
It measures current labor-market breadth rather than forecasting recessions. The historical table shows how score bands overlapped final NBER recessions and how often payroll contraction or a material unemployment increase followed within six months, using release-timed but currently revised data.