US Payrolls—Is It All Coming Home to Roost?

• Nonfarm payrolls fell by 23k, household survey by –87k

• Revisions took out 103k jobs

• 2026 is tracking the weakest year for jobs since 2020

• The unemployment rate dipped as a quarter million dropped out of the workforce

• The plunging participation rate is hitting the youngest and oldest

• There remains no material World Cup effect

• Hours worked signal caution around Q3 GDP growth

• Wage growth cooled

• Hawks shouldn’t be encouraged by a falling UR

• Markets trimmed Fed hike bets

• US nonfarm payrolls m/m 000s / UR %, SA, July:
• Actual: -23 / 4.1
• Scotia: 50 / 4.3
• Consensus: 80 / 4.1
• Prior: 20 / 4.2 (revised from 57 / 4.2)

The US job market is clearly grinding to a halt after a brief spurt had sparked earlier optimism in March and April payrolls. Nonfarm payrolls fell by 23k in July over June and revisions took down June (now +20k instead of 57k) and May (now +63k instead of 129k). The disappointment was generally in line with Scotiabank Economics’ call which was at the bottom of consensus alongside communicated downside risk including multiple negative prints in our scenarios.

September hike pricing fell 3bps to 11bps post-data. Advice to receive Sept OIS has worked well as pricing has fallen from over a quarter point.

It’s not just payrolls either. The household survey registered another drop of 87k jobs in July after a drop of 507k in June. Whereas nonfarm payrolls are up by 426k ytd, the household survey is down by 1.8 million jobs ytd. Both readings are tracking the weakest year for jobs since the onset of the pandemic in 2020 (chart 1). The household survey is noisier, but also captures off-payroll jobs particularly at small businesses that may be struggling more than larger employers.

At the heart of the matter is that private sector jobs ex-health care have entirely lost previously brief momentum (chart 2).

Chart 3 shows that the weakness was fairly widespread on a sector basis. The goods sector added 25k payroll positions in July almost entirely due to construction (22k). Services added just 5k jobs with the main gainers being education/health (+25k) as education hiring flattened out (2.8k) and health sector hiring slowed to 22.6k. It’s possible that the expiration of ESSER funding is starting to hit the education sector amid expectations for outright losses over the next couple of reports. Health hiring may be dampening as the lagging effects of ACA subsidy reductions work through. Other service sectors that were up included professional and business services (+18k) with a minor assist from low quality temp help (3k), IT jobs (+13k after –11k previously) and mild gains in transportation and warehousing and communications/IT services.

The administration tried to pin weakness on the World Cup. That makes little sense. Yes, leisure and hospitality sector hiring was down 40k in July after a 43k drop in June. If the World Cup had any effect on this category, then it may have been through front-loaded hiring in May (+42k) but the declines over the past two months are double that prior gain. It’s a similar argument for retail trade where there were 19k fewer jobs in July and 4k fewer in June after only a 5k gain in May. There is no discernible World Cup effect in these numbers whether up or down.

So what gives? Multiple quarters of no material growth in inflation-adjusted personal disposable income, a negative housing wealth effect marked by falling real house prices, real wage compression and uncertainty’s effects may be dampening activity in discretionary spending categories. Spend more on groceries and gas, less on everything else while running out of room to deplete the saving rate as a shock absorber. The administration’s affordability sore points on energy, food, housing and tariffs as a tax hike on Americans may all be coming home to roost.

Government shed 53k jobs—almost all (-50k) at the state and local levels where budgets are being squeezed. A job is a job so I don’t like the administration’s data mining that removes this category.

Chart 4 shows the magnitude of the revisions. Chart 5 shows that perhaps next month’s revisions will be less acute given that the initial data collection rates were higher in July than June’s low reading.

Hours worked are stumbling and that’s not a good sign for Q3 GDP in isolation of everything else. Hours are tracking at only +0.2% q/q SAAR in Q3 after a 1.2% Q2 expansion (chart 6). To keep up growth in Q3 requires an activity-driven surge in productivity. Nowcasts pointing to strong Q3 GDP growth should be ignored at this point since there is a paucity of data upon which to base them and they tend to get massively revised over the quarter.

Why are folks leaving the US labour force in droves? The participation rate fell again to 61.4% (chart 7). It has dropped by a whopping 1.1 ppts since November which given the enormity of the US labour market translates into a big drop of 2.4 million in the size of the labour force over this period. All age groups are seeing declines of late (chart 8). Because it’s the biggest cohort with the highest part rate we should be concerned about the wavering participation rate for 25–54 year-olds led by 25–34 year-olds. The 55+ category has seen sharp declines in participation rates. The administration has tried to pin that on demographics and the boomers, yet the sudden plunge this year is hard to justify with that narrative. It’s doubtful that a two-decade long age cohort suddenly accelerated exits due to aging in the past few months. Further, since 2023, the declines have been across the younger and older workers (chart 9). That suggests they may be struggling the most with considerations such as rapid technological change. Put that high on your list for public policy purposes as it’s not a great state of the world if those worker categories go down.

In addition to technological change, cost pressures impacting employers and uncertainty effects, another possibility is that tighter immigration policy is hitting older workers the hardest. Direct effects of ICE detentions may capture part of this (chart 10) but so may indirect effects that are causing a fear-driven avoidance of workplaces. The role of technological change could also be impacting older workers harder.

Enter the unemployment rate that fell again (chart 11). Hawks will point to this as a sign of a healthy job market that perhaps merits tightening. But is the job market really healthy without job growth? Jobs drive incomes which drive consumption and GDP. There are feedback loop effects on the dual mandate that could be misjudged by focusing upon the unemployment rate. Practice some ratio decomposition here folks. The UR is distorted today versus the past by out-of-sample immigration policy changes.

The seasonal adjustment factor was low which shaved job growth (chart 12). Still, if we went with last July’s SA factor, then payrolls still would have only been up by 44k and bang on my 50k estimate; hence, still weak in line with the alternate scenario I applied using a lower SA. The seasonally unadjusted change in payrolls was weaker than last year but still within the bounds of a ‘normal’ July (chart 13).

Wage growth fell to just +0.6% m/m SAAR in July (chart 14). That drops the three-month moving average to 2.3% m/m SAAR. That should encourage the Fed from an inflation transmission standpoint. Knock present inflation out of that and real wages are falling which may portend disinflationary second-round effects of spikes in some prices.

Charts 15–19 break out a few of the key job categories by sector.