A New Take on The Greater Fool Theory

No, I'm not actually going to delve into the greater fool theory per se, nor am I going to talk about meme stocks, or any other investment.

Much has been made of the weak July jobs report, the significant revisions to the prior two months’ reports, and Pres. Trump’s subsequent announcement that he’s firing the Commissioner of the Bureau of Labor Statistics (BLS).

Those of you on the right believe the firing is a good move, and those of you on the left believe he's wrong to do so.

But I’ll bet all of you hold those opinions merely because of your partisan biases. I’ll bet none of you understand where the jobs report data come from, how they’re calculated, why there are revisions, how the revisions are made, etc. In other words, you’re unaware of the whole process, but you have a strong opinion about the process.

Because I hate to see people make themselves look silly by espousing uninformed opinions, I’m going to provide you with a free education. Let’s jump right in.

Here’s how the nonfarm payroll data (NFP) are actually collected and published each month by the Bureau of Labor Statistics (BLS).

1.     The data come from a survey, not from administrative records.

The BLS does not track every paycheck or tax record, or walk into businesses and count heads. Instead, it conducts a monthly survey of about 119,000 businesses and government agencies, covering over 600,000 individual worksites. This is called the Current Employment Statistics (CES) survey, generally referred to as the “establishment survey.”

These employers report the number of jobs (not people; an important distinction), along with hours and wages. The establishment survey is separate from the household survey, which is used to determine the unemployment rate.

2.     What do employers report?

Each participating employer provides the following information for a specific “reference week” each month (usually the week including the 12^th of the month):

·       Number of employees on the payroll, full-time and part-time

·       Total hours worked (for production/nonsupervisory employees)

·       Earnings for the week (average hourly and weekly earnings)

·       Industry classification (based on NAICS codes)

These are counts of jobs, so if one person has two jobs with two different employers that get included in the survey, both jobs get counted.

3.     Initial estimates involve modeling and imputation (don’t let the fancy words scare ya – this is just statistics, folks, like you use in your fantasy football league)

Because not all 119,000 employers respond on time (and the response rate has deteriorated since covid), the BLS has to impute (estimate) the missing data based on:

·       Historical patterns for late responders

·       Trends from respondents that did submit on time

·       Seasonal adjustment factors (think school years, Christmas holidays, summer retooling for auto manufacturers, etc.)

So when you hear that “payrolls rose by 205,000 in June,” that number is a modeled estimate based on partial survey data and historical assumptions – not a complete census.

4.     Two monthly revisions follow the initial release

Each month’s initial jobs estimate is revised twice:

·       First revision: published one month later, when more survey data have come in

·       Second revision: published two months later, using nearly full survey response

These revisions can be large or small, depending on:

·       Response rates (typically 70-75% at the initial release)

·       Trends in newly reported or corrected data

·       Seasonal adjustment quirks (especially during economic turning points or shocks like covid)

The revision process is not some nefarious, political book-cooking exercise. It happens no matter who is in office, and it happens no matter who runs the BLS. And if Pres. Trump appoints a hand-picked conservative lackey to run the bureau, the revisions will be calculated the exact same way they’ve always been calculated, and if those revisions don’t turn out the way the President likes, he’ll probably fire that person, too.

5.     Annual Benchmarking (Major Revisions)

Once a year – usually in February – the BLS makes a massive adjustment called the benchmark revision, which recalibrates the whole payroll series using actual employment counts from unemployment insurance tax records. (Lots of economic data are subject to annual revisions, btw.) This gives a more accurate picture than the survey-based estimates and can result in large upward or downward revisions going back up to 21 months.

6.     Limitations and Common Misunderstandings

·       The NFP report is not a literal count of all jobs in the country, it’s a statistical sample.

·       Jobs do not equal workers. People with multiple jobs get counted multiple times.

·       It excludes farm workers, the self-employed, and gig workers (unless they’re on a payroll).

·       Response rate matters: early estimates are vulnerable to errors during major trend changes (e.g., recessions, covid rebound)

·       Seasonal adjustment can distort monthly changes, especially around holidays, weather shifts, and school calendars. (This is why the numbers are always skewed when there’s a major winter storm or a hurricane.)

One simple observation is that the cadence of initial release – first revision – second revision is very similar to the cadence of GDP releases, and for similar reasons. The initial GDP estimate released each quarter is based on estimates of the factors of production; then, as more data comes in, more reliable GDP numbers are released in the form of second and third estimates, and long-term adjustments are sometimes made.

And it’s noteworthy that the head of the Bureau of Economic Analysis, who oversees the GDP data, is also a Biden administration appointee. Yet, because Pres. Trump loved the recent second quarter GDP growth figure of 3.0%, that agency head isn’t being fired. Hmmm …

Now that we’ve defined how the NFP data is obtained and calculated, let me make a couple of other observations. First, I’m not a fan of survey-based data. I wish the annual unemployment insurance tax data could be obtained in a monthly manner, because it’s actual hard data, and I can rely on that.

Some other examples of economic surveys are the various consumer confidence or sentiment surveys, and manufacturing surveys like the ISM (Institute of Supply Management) survey. Regarding the former, it’s been demonstrated (by me, among others) that there’s little correlation between consumer sentiment and actual spending. So to gauge how consumers really feel about the economy, I’d rather look at what they’re actually spending than what they think about economic conditions (in part because the average consumer being surveyed doesn’t know jack about economics).

Similarly, rather than ask a bunch of manufacturers how they feel about business conditions in their industry, I’d rather look at how much they’re producing, how much of their capacity they’re utilizing, etc.

In other words, I’ll always prefer hard data over soft data. But sometimes, you’ve gotta go with what you’ve got, and in the case of jobs data, the NFP survey is pretty much what we’ve got; it’s simply impossible to go in and count every employee at every business in the country every single month.

Now, let’s look at some of the reasons the magnitude of the revisions has increased since covid. First, the survey response rate has declined somewhat since then, so the initial estimate is somewhat less reliable. Second, business churn in the covid aftermath (companies going out of business and new businesses starting – referred to as the birth/death rate) has increased, distorting the models, because that rate is a significant factor in the revision calculations.

However, by late 2023 and early 2024, the adjustments made to the models to compensate for the greater birth/death rate may have overcompensated, especially as job growth began to slow. That led to some of the overstated initial numbers during that period, which later saw massive downward revisions like the one made to the March 2024 data (bookmark that point – I’ll return to it shortly).

More recently, the DOGE cuts have resulted in large job losses at several government agencies (remember that nonfarm payrolls include government employees), but there may not be enough, or the right, personnel left to respond to the CES survey, at least timely. So there may some distortion there. And immigration trends could play a role (we don’t know how much, because we really don’t know how many employers would have actually reported illegal immigrants on their payrolls in 2021-24, when the illegal immigrant population soared). Of course, that trend is reversing in 2025.

Now, I’ll speak to the notion that there’s political bias in the data, and I’ll do it in one word:

Bullshit.

I’ve looked at the monthly revision data trends going back to 2017, and they’re consistent. In other words, even before this “Biden appointee” took over the BLS (and she isn’t the one crunching the numbers, people), the revisions – including those during Trump’s first term – have been substantially similar, especially in light of job growth trends over that period (job growth has slowed gradually post-covid as the labor market has matured and we’ve recovered the jobs lost due to the covid shutdown, including accounting for normal growth that would have occurred had the shutdown not taken place).

And the notion that the revisions are one-sided is patently absurd: the latest annual revision was a massive downward adjustment to the March 2024 data of more than 800,000 jobs. In other words, the BLS determined that jobs in the post-covid years while Biden just happened to be President were over-estimated by that amount, and needed to be adjusted downward. So the revision didn’t “help” Biden, it reduced "his" job growth numbers.

Again, those revisions aren’t manipulated, they come from actual tax records from the Quarterly Census of Employment and Wages (QCEW), which employers are legally required to submit – hard data. The revisions cut both ways, under administrations of both parties, regardless of whether the head of the BLS was appointed by the President in office or a President from the opposition party. There is no trend indicating bias.

I’m going to say that again: there is no trend indicating bias in either the initial NFP releases or the revisions, whether we’re looking at the monthly revisions or the annual adjustments. That’s a factual statement based on an understanding of the methodology, which is fixed and politically agnostic, and a review of the data.

Anyone who knows me, knows two things about me: first, I lean conservative. And second, regardless of my leanings, when it comes to economics and statistics, I call balls and strikes.

In that spirit, Pres. Trump was foolish to fire the Commissioner of the BLS because he doesn’t like the jobs numbers. In fact, those numbers provide him with the strongest argument he could want for Jerome Powell – who’s almost equally foolish – to cut interest rates, because the data indicate that job weakness has gotten ahead of the Fed, just as inflation did in 2021.

The President is the bigger fool because firing any agency head whose agency releases statistically sound data just because the President doesn’t like the way the data makes him look erodes market confidence in all economic data, because it leads market participants to believe that now the data really will be manipulated by Presidential fiat. Just like making up total BS tariff rates that other countries were supposedly charging us caused the markets to lose confidence in April, nearly leading to a bear market. (And no, the market isn't made up of some nebulous "they" who manipulate stock prices to make Trump look bad. The "they" is actually you and me.)

But those who, after reading this, continue to insist that the labor market data is biased, just because they heard it from their favorite media source, despite the fact that they’ve never researched the methodology or the data itself, are the biggest fools of all.

P.S. If you’re reading this, and your own political bias is making you feel smug because you believed all along that Pres. Trump was wrong to fire the BLS Commish, and now you think you’ve been vindicated by what you've just read, remember these two things: first, your belief was no less based on a lack of information than that of those who believed the firing was righteous.

And second, you’re the fools who still have your collective hair on fire because Sydney Sweeney made a jeans ad.