Automatic Data Processing (ADP)'s 38K print contradicts steady JOLTS—this is the signal that decides whether September is really a “two-hike” month

On Sep 2, ADP’s National Employment Report said private-sector employment increased by just 38,000 jobs in August. In the same release, ADP also reported base pay up 3.2% year-over-year and gross pay up 4.7%, but the key investor takeaway was labor demand itself slowing sharply versus prior months and against typical expectations.

The macro problem for markets is that this comes while JOLTS for July (released separately by the BLS) showed job openings and hiring that were largely unchanged—so the labor-market tape is split. That split is exactly what matters for the decision tree leading into Friday’s BLS Employment Situation print (scheduled for Aug. 7, 2026 per BLS’s Employment Situation page).

JOLTS showed that in July job openings were little changed at 7.3 million, and hires were little changed at 5.1 million. But hires in JOLTS are a rate of realized hiring activity; ADP is a private payroll growth estimate based on payroll processing data.

So when ADP prints 38K while JOLTS stays stable, the plausible interpretation is not “labor is fine” or “labor is collapsing” in a single stroke—it’s that the conversion from openings into actual payroll additions is slowing, or job additions are becoming more concentrated in segments/establishments not captured the same way. ADP’s own segment detail also supports that the weakness is not uniform: goods-producing employment fell by 10,000, while service-providing rose by 48,000.

When ADP points to weaker payroll additions, the market immediately asks whether Friday’s BLS Employment Situation will confirm the same cooling. If the payroll print also under-delivers, rate-hike odds tend to compress; if it surprises to the upside, “two-hike” paths can reassert.

What’s tricky for investors is that the ADP number is not the official payrolls report—so the short-term market reaction can be muted if other labor indicators (like JOLTS) stay calm. But if the next datapoint (Friday’s BLS) aligns with ADP, the repricing tends to be sharper because traders have fewer “holdout” indicators to lean on.

Labor-market weakening tends to show up first in hiring-intensity businesses (staffing and recruitment), then in risk/rating and financing-adjacent data providers through credit and default assumptions, and finally in payroll-adjacent software and outsourcing through mix (more employer compliance work, fewer net-new hires).

The constraint for investors is that ADP’s read is private payroll growth, so the cleanest equity translation is to companies whose revenues track hiring velocity or job-cycle turnover.

In the next 1–3 days, the question is simple: does Friday’s BLS payrolls release land closer to ADP’s cooling signal or closer to JOLTS’s steadier hiring picture? That determines whether “two-hike” odds fade quickly or persist.

Over 1–3 quarters, the question becomes structural: does labor demand remain uneven (as ADP’s goods vs. services mix suggests) or does it broaden into a sustained hiring downcycle? If the pattern repeats—weak payroll additions with steadier openings—investors should treat labor-market conversion (openings → hires/payrolls) as the key risk factor rather than the headline level of vacancies.