00:00Economists polled by Reuters expected 90,000 jobs.
00:04America added 29,000.
00:07Stocks rose anyway, and yields fell,
00:10because markets don't trade the headline alone.
00:12They trade the surprise and what it changes about expectations.
00:17Here is how the Bureau of Labor Statistics actually wrote it.
00:21Both non-farm payroll employment and the unemployment rate, it says,
00:26changed little in September.
00:28However, the headline everyone quotes is one figure lifted out of that sentence.
00:33The headline jobs number comes from the current Employment Statistics Survey,
00:37which most people call the Establishment Survey.
00:40Every month, the Bureau of Labor Statistics collects payroll records
00:45from about 119,000 businesses and government agencies.
00:49Those employers cover roughly 622,000 worksites,
00:54and the active sample represents about 26% of all non-farm payroll employees.
01:00From those records, BLS estimates employment, hours, and earnings.
01:05And here is the detail almost everyone gets wrong.
01:09Non-farm payrolls counts jobs on payrolls.
01:12It does not simply count individual people.
01:15So why non-farm?
01:16It's tempting to say the survey excludes farmers and stop there,
01:20but it excludes more than that.
01:22Agricultural workers sit outside the survey's scope,
01:25and so do self-employed people whose businesses are unincorporated,
01:30unpaid family workers, and private household workers.
01:33The payroll survey is a picture of payroll jobs at businesses and government agencies,
01:39a large part of the economy, but not all of it.
01:42Which brings us to the most important idea in this entire video.
01:46The jobs report is built from two completely different surveys.
01:50The establishment survey asks employers about their payrolls.
01:54The household survey asks people about their own work status.
01:57The establishment survey produces non-farm payrolls.
02:01The household survey produces the unemployment rate.
02:04Two samples, two questions, two different concepts.
02:08And there is a reason both exist.
02:10An employer knows exactly how many people it paid last month,
02:15but it has no idea whether someone it didn't hire is out looking for work.
02:19A household knows that, but cannot tell you the payroll of the company down the road.
02:24Each survey can answer a question the other structurally cannot.
02:28So picture one person.
02:30He works at a coffee shop, and he drives for a delivery company.
02:34On the payroll side, that's two jobs reported by two different employers.
02:38On the household side, he is one employed person.
02:41Same human being, two different numbers, and both are correct.
02:45Neither survey made a mistake.
02:47They were counting different things.
02:49That is why payrolls and the unemployment rate can move together,
02:52or in opposite directions, without either one being wrong.
02:56Now, why does the payroll number keep changing after it's published?
03:00Three different things are going on.
03:02First, monthly revisions.
03:04The first estimate is published before all the responses are in,
03:08so BLS revises it in each of the next two months as more arrive.
03:12That is exactly what happened in this release.
03:14July was revised down to negative 10,000, and August down to 133,000.
03:21Together, those two months hold 60,000 fewer jobs than previously reported.
03:27Second, the annual benchmark.
03:28Once a year, BLS re-anchors the survey estimates to far more complete unemployment insurance tax records.
03:36Third, the birth-death model.
03:38New businesses and closures cannot all be observed immediately,
03:42so BLS estimates their net effect with a published statistical model.
03:47That is an estimate of real firm formation, not a number pulled out of the air.
03:51Which leads to something worth internalizing.
03:55The payroll number is an estimate from a sample, and BLS publishes how precise it is.
04:00The 90% confidence interval on the monthly change in non-farm employment
04:05is on the order of plus or minus 122,000.
04:09For the monthly change in the unemployment rate,
04:12it is about plus or minus three-tenths of a percentage point.
04:16That does not make a smaller number meaningless.
04:18It means you should not read a monthly estimate, as if it were an exact headcount.
04:24So when the next report lands, don't just read the headline.
04:27There are five numbers worth checking.
04:29One, payrolls against what was expected.
04:32Two, the unemployment rate.
04:34Three, wage growth month over month and year over year.
04:38Four, revisions to previous months.
04:41Five, labor force participation,
04:43because the unemployment rate can fall simply because people stopped looking for work.
04:47Once those five are second nature, two more add real texture.
04:52Which industries added or shed jobs tells you whether growth is broad
04:56or concentrated in one corner of the economy.
05:00And average weekly hours often move before headcount does,
05:03because employers tend to adjust shifts before they adjust payrolls.
05:08So why does the Federal Reserve care?
05:10Congress gave the Fed two goals, maximum employment and price stability.
05:15Notice what that does not include.
05:17A target number for monthly payrolls.
05:20Maximum employment is not directly measurable, and it changes over time.
05:24So the Fed reads a wide range of labor indicators alongside inflation and the rest of the economy.
05:31Labor data is one input into an assessment, not a trigger wired to a rate decision.
05:37Which is why a single payroll print almost never settles anything on its own.
05:42Which finally explains the market reaction.
05:45Markets don't trade the number, they trade the surprise.
05:48The surprise is the gap between what arrived and what was already expected and priced in.
05:53That surprise is then read through the current environment,
05:56where inflation is, where growth is, and what investors already assume about policy.
06:01Only then do you get repricing, across yields, the dollar, equities, and gold.
06:06And every arrow in that chain is conditional.
06:09That is also why the reaction looks instant.
06:12Nobody is reading the report in the first second.
06:15Prices are simply moving to where the new expectation sits, and the reading comes afterward.
06:21Which is why the most popular rule about this report doesn't hold up.
06:25You've heard it.
06:26Soft payrolls bring rate cuts, so equities rally.
06:29Watch what actually happened on October 2nd.
06:32Payrolls missed badly, stocks rose, and treasury yields fell.
06:36And none of that was about a rate cut.
06:38Going into this report, markets were pricing the possibility of a rate increase.
06:43The weak number pushed the odds of a hike at this month's meeting down to around 12%.
06:49Stocks did not rally because easing arrived.
06:52They rallied because tightening got less likely.
06:54Same direction, completely different mechanism.
06:57Change the backdrop, and the sign can flip.
07:01Soft payrolls alongside falling inflation can read as room for easier policy.
07:07Alongside stubborn inflation, the same number can read as stagflation instead.
07:12And a very weak report can read as recession risk rather than relief.
07:17So here is the session itself.
07:20The report lands at 8.30, an hour before the stock market opens,
07:24so the first reaction happens in futures.
07:27By the close, the S and P tracker finished the day higher.
07:31And the two-year yield, the dollar and gold had all moved on the wire within minutes of the release.
07:37Next time the jobs report drops, you have a framework.
07:41Payrolls versus expectations, unemployment, wages, revisions, participation.
07:45And then the question that decides what any of it means, what is the inflation and policy backdrop right now?
07:52The headline gets the attention, the context decides what it's worth.
07:56Payrolls counts jobs, not people, the unemployment rate comes from a different survey entirely,
08:02and the first print is an estimate that will be revised.
08:05Get those three straight, and you are reading the report better than most of the commentary about it.
