Contrary to what the mainstream media is saying, the November jobs report from the Department of Labor is not a blowout, but a disappointment. The headline number of 263,000 additional non-farm payrolls sounds like robust growth, but that is not what most Americans are feeling right now. Even a cursory look under the hood of this report shows things are not as rosy as that headline number makes them appear.
The seemingly robust headline number of 263,000 jobs comes from the establishment survey of businesses, while the unemployment rate comes from the household survey. But the household survey also has a measure of employment, and that fell by 138,000. In fact, the household survey has been flat since March of this year, with essentially no jobs added over that time.
In the last eight months, there has been an unprecedentedly large divergence between the two surveys of 2.7 million jobs. One reason for this is that the establishment survey allows for double-counting, and the household survey does not. Over the last year, approximately one-fifth of the additional non-farm payrolls have been double-counting, not additional people employed.
The double-counting can occur for different reasons, such as people holding multiple jobs. Many people are getting second (or third) jobs to deal with inflation, but the household survey counts those people as employed -- just like when they had one job. However, the establishment survey counts each additional job that a person gets as another payroll.
Conversely, unincorporated self-employed workers are counted in the household survey, but not the establishment survey. But inflation has caused many small businesses to close over the last two years, and there has been a trend for the self-employed to go work for large corporations. There is no change in the number of people employed, but the headline jobs number actually increases because the person merely shifts from an uncounted category to one which is counted.
In fact, the double-counting in November was about 454,000 jobs – substantially more than the headline jobs number. It also explains the entire monthly divergence between the two surveys, which was 401,000. It turns out the economy might not have been employing any new people at all in November.
But there are more chinks in the labor market’s armor. Labor force participation remains weak and got worse in November. The low labor force participation rate is artificially depressing the unemployment rate so that even though the number of people employed declined in the household survey, the unemployment rate remained unchanged.
Meanwhile, the economy has been hemorrhaging full-time jobs, losing almost half a million since May. That equates to an average loss of over 2,600 full-time jobs per day. This is particularly ominous because businesses’ preferences often shift from full-time to part-time hiring before transitioning to layoffs in times of uncertainty.
The manufacturing sector offers further evidence of this tumult. Average weekly hours and average overtime in the sector declined, a common forerunner to layoffs, while the industry added just 14,000 jobs in November. That is less than half the number added in October, and the slowest growth in a year and a half. While employment is a lagging indicator, manufacturing often moves before the rest of the labor market, which is clearly running on fumes.
For the average person who is working, their paycheck is bigger, yet they are demonstrably poorer. That is because nominal earnings continue to trail inflation so that workers can buy less, despite a larger income. People are resorting to credit cards to maintain their standard of living – an unsustainable path.
Yet another concern is the seasonal adjustment made by the Department of Labor to the monthly jobs numbers. Because certain months have predictable seasonal variations, the Bureau of Labor Statistics adjusts the raw survey data to account for these seasonal changes, such as large increases in temporary holiday workers who are then let go in January. That adjustment gives an estimate for the monthly change in jobs which was not due to any seasonal variation.
However, the adjustments this year have been unusually large to the upside or adjusted up much more than usual. But what goes up must come down, since the seasonal adjustments for a year essentially zero out; if one month is revised up by 1 million jobs, another month must be revised down by roughly that amount. December’s downward revision will need to be 30 percent larger than normal to account for the unusually large upward revisions earlier in the year. In other words, this year’s jobs numbers have been front-loaded.
The devil is truly in the details of this jobs report. The labor market is clearly turning from growth to contraction. Considering that employment is a lagging indicator, the economy may be closer to recession than many pundits think.