ADP jobs report vs. BLS jobs report: what's the difference?

On September 2, 2026, ADP reported that private employers added just 38,000 jobs in August, the slowest pace since January. Two days later, the Bureau of Labor Statistics releases its own official count for the same month, and it can show a meaningfully different number. Here is what each report actually measures, why they do not always agree, and which one actually moves the Fed and the stock market.

What is the ADP jobs report?

ADP, the payroll processing company, publishes its National Employment Report on the Wednesday before the BLS's Friday release, built from anonymized, real payroll records covering roughly 26 million US workers at businesses that use ADP to run payroll, according to ADP's own research on the report's methodology. It counts only private-sector jobs, not government employment, and counts a worker as long as they remain on a company's payroll, even if unpaid during a strike or disaster. August's report, released September 2, 2026, showed private employers added 38,000 jobs, down from 46,000 in July and below the 47,000 economists expected, according to ADP's August release. Large employers with 500 or more workers drove nearly all of the gain, while firms with under 50 workers added just 3,000.

What is the official BLS jobs report?

The Bureau of Labor Statistics' monthly Employment Situation report, usually released the first Friday of each month, is actually built from two separate surveys, according to the BLS's own Current Employment Statistics documentation. The establishment survey samples roughly 119,000 businesses and government agencies representing about 622,000 individual worksites, and produces the headline nonfarm payrolls number, which includes both private and government jobs. A separate household survey of about 60,000 households produces the unemployment rate, an entirely different sample, which is why the two figures inside a single BLS report can occasionally tell slightly different stories from each other. Because the BLS survey is far larger, includes government employment ADP does not track, and only counts workers actually paid during its reference week, it is the number the Fed and financial markets treat as the authoritative read on the labor market, not ADP's.

Why do ADP and BLS numbers diverge?

The two reports frequently tell different stories about the very same month because they are not measuring the same thing. In July 2026, ADP reported private payrolls rose 44,000, while the BLS's July report, released two days later on August 7, showed total nonfarm payrolls fell 23,000, a swing large enough to move Fed rate-hike odds and send the S&P 500 to a record high the same day. See why a weak jobs report sent stocks up, not down for what happened next. A 2019 Federal Reserve study comparing the two series found they work best as complements, not one predicting the other, with correlation running strong over long stretches but breaking down within any single month, especially around events like strikes, hurricanes, or a rapidly shifting labor market.

Which jobs report actually moves the stock market?

Both move markets, but differently. ADP's report lands first and moves futures and yields briefly on release morning as an early, imperfect signal, exactly what happened September 2, 2026, when the weak 38,000 print helped stocks and Treasury yields settle down after a rough start to the month. The BLS report typically produces the larger, more durable reaction, because it is the number Federal Reserve officials cite directly in their own policy statements, and it lands closest to FOMC meetings.

The honest counterargument: is ADP just noise?

After a case like July 2026's, where ADP and BLS pointed in opposite directions, it would be easy to write ADP off entirely. That overcorrects:

• ADP never claimed to forecast BLS. Since retooling its methodology with Stanford's Digital Economy Lab in 2022, ADP explicitly describes its report as an independent measure of the labor market, not a preview of the official number, so judging it by how well it predicts BLS misreads its own stated purpose.
• The two series move together over time. The Federal Reserve researchers who studied both found the optimal estimate of true payroll growth weighs ADP and BLS roughly equally, because combining them reduces measurement error in both. A single month's mismatch is noise inside a relationship that holds up over longer stretches.
• Being first and large still has value. A sample of 26 million workers, even an imperfect one, is the earliest real signal available each month, two full days before the number that actually moves Fed policy.

The honest read is that ADP is a useful, early, imperfect data point, not a preview of Friday's number and not something to dismiss because one month diverged.

What a beginner should actually do

1. Do not trade on ADP alone. Treat it as an early, private-sector-only signal, not a preview of Friday's official report.
2. Give more weight when both agree. A weak ADP print followed by a weak BLS print is a real signal. A weak ADP print followed by a strong BLS print, like July 2026, means the ADP number was mostly noise.
3. Watch the actual BLS release on September 4, and the August CPI report on September 11, the two data points still standing before the Fed's September 15 and 16 meeting. See the full rate-decision breakdown.
4. Keep contributions on schedule regardless of either report. See time in the market beats timing the market for why reacting to any single monthly data point tends to cost more than it saves.
5. If a weak jobs report worries you, check your cash position first. See how big an emergency fund you actually need before adjusting anything else.

The quick version

• ADP's National Employment Report tracks private-sector payroll data from about 26 million workers at companies that use ADP, released the Wednesday before the BLS report
• The BLS jobs report surveys about 119,000 businesses and government agencies covering roughly 622,000 worksites, includes government jobs, and is the number the Fed treats as authoritative
• ADP's August 2026 report, released September 2, showed private payrolls up just 38,000, the slowest pace since January
• The two reports can point in opposite directions for the same month: July 2026 saw ADP report a 44,000 gain while BLS reported a 23,000 loss
• A 2019 Federal Reserve study found ADP and BLS work best as complements, weighted roughly equally, not as one predicting the other
• ADP moves markets briefly on release morning; BLS produces the larger, more durable reaction because it is the number the Fed cites directly
• The official August BLS jobs report lands September 4, 2026, and the August CPI report lands September 11, both before the Fed's September 15 and 16 meeting

Neither report is wrong when they disagree. They are measuring overlapping but different slices of the same economy, at different sample sizes, with different rules for who counts. The useful habit is not picking a side. It is waiting for the one the Fed actually watches before treating either number as an answer.