Updated: Oct. 5, 11:05 am ET to make a correction from 35.5 to 34.5: “It has since fallen back to 34.5 hours in August, a level often seen in months before the pandemic.”
(The monthly Employment Situation is scheduled for release by the Bureau of Labor Statistics on Friday, October 7th, at 8:30 AM Eastern Time.)
The Federal Reserve Board has been aggressively raising interest rates to combat inflation, with a labor market it argues is overheated being in the center of its sights. Despite past hikes, the US continued to see strong job growth through August, although there was a modest uptick in the unemployment rate from a 50-year low.
Wage growth is erratic on a monthly basis, but the 0.3 percent monthly change reported for August was good news on the inflation front. However, more months of moderate wage growth are needed for the Fed to be confident that inflationary wage increases are not a problem.
Analysts have paid insufficient attention to hours growth. This matters for two reasons. First, aggregate hours are the best measure of labor demand. Hiring workers and increases in workweeks are alternative mechanisms for getting more labor. In August, the aggregate hours index was flat even though we added 315,000 jobs.
The other part of this story is that the level of average weekly hours is telling us about the difficulty of finding workers. The length of the average workweek increased sharply in the recovery, rising from an average of 34.4 hours in 2019 to a peak of 35.0 hours in January of 2021. It has since fallen back to 34.5 hours in August, a level often seen in months before the pandemic.
The private sector regained jobs at a very impressive pace. We are now 1,456,000 above the pre-pandemic level, including the preliminary benchmark revision. However, the economy cannot keep adding jobs at its recent pace with a labor market near full employment.
The overall jobs number will likely be close to 200,000 for September, which is still very strong. The fact that unemployment insurance claims remain at very low levels and the number of insured unemployed as a percent of the total workforce is near record lows suggests that demand for labor remains high.
We may see some shift from gains in the private sector to state and local government employment. Total government employment is still 754,000 (3.3 percent) below pre-pandemic levels, factoring in the benchmark revision. State and local governments have had difficulty attracting workers, as pay in the public sector has not nearly kept pace with wage gains in the private sector.
The Fed has been explicit in its concern that wage growth has been too fast to be consistent with its 2.0 percent inflation target. As mentioned earlier, the 0.3 percent growth in the hourly wage reported for August would be pretty much consistent with a 2.0 percent inflation rate (wage growth averaged 3.3 percent for 2019). There would have to be several more months of comparably restrained wage growth to provide clear evidence that wage growth had moderated.
The annual rate of increase for the most recent three months (June, July, August) compared with the prior three (March, April, May) was 4.9 percent. This is undoubtedly faster than would be consistent with 2.0 inflation, but it is noteworthy that it is slower than the 6.1 percent rate we saw at the start of 2022.
Much of the recent discussion of inflation has focused on Beveridge Curve predictions based on the ratio of job openings to unemployment, which has been at record highs. These predictions would show that wage growth is increasing, not slowing. This should raise questions as to the extent these models describe the extraordinary situation we are facing with the recovery from the pandemic.
While the unchanged number of hours worked in August was surprising, given the big rise in jobs, it is consistent with recent GDP numbers. The negative GDP numbers reported for the first two quarters, coupled with very rapid job growth, implied a collapse in productivity. It reportedly declined at a 7.1 percent rate in the first quarter and a 4.1 percent rate in the second quarter.
These sorts of declines would be consistent with supply chain disruptions making it difficult for workplaces to maintain normal operations. The high rates of job turnover we have been seeing would also depress productivity.
However, if the economy and labor market getting back to normal, these disruptions should be less substantial. Given the moderate growth in GDP expected for the quarter, it would not be surprising to see the index of hours again remain flat or even fall slightly in September. Also, if finding workers is less of a problem, employers will have little need to force workers to put in more hours as an alternative to hiring more workers.
Unemployment and Labor Force Participation Rates
Many people (including me) often make too much of month-to-month movements in the household survey. There is substantial error in the household data, and often month to month movements are simply driven by errors in the survey rather than real changes in the economy.
In this vein, the 0.3 percentage point jump in the labor force participation rate (LFPR) reported for August (0.4 percentage points for prime age workers) was most likely an error in the data, not a story of hundreds of thousands of people suddenly deciding that they felt like working. In the four months from March to July, the establishment survey showed the economy adding 1.6 million jobs, while the household survey showed employment actually declined by 168,000.
That story does not make sense. When the two surveys conflict, the establishment survey is likely closer to the mark. Weird movements in the household survey are not uncommon. For example, from March 2019 to July 2019, the prime-age LFPR fell by 0.9 percentage points. The economy grew at a 3.2 percent rate in the second quarter and a 2.8 percent rate in the third quarter. The establishment survey showed a gain of almost 600,000 jobs for this four-month period.
This problem in the household data was short-lived. The LFPR for prime age workers jumped in the next two months, and by September it was 0.2 percentage points above its March level.
The unemployment rate is likely to remain near its August level, while my bet is for another rise in LFPR in September.
Share of Unemployment Due to Quits
One measure of labor market strength is the willingness of workers to leave their job without having a new job lined up. This rose to 15.2 percent in August, tying the record high set in 2000. This figure is erratic, but if we are returning to a more normal labor market, it may edge down somewhat in September.
Construction Employment May Remain Strong
Construction is the major sector most impacted by the Fed’s rate hikes, however employment has continued to grow at a healthy pace. This is true even for residential construction, which added 10.9k jobs in August. Part of this story is that we have a huge backlog of houses that have been started but not completed due to supply chain problems. This means that even as the number of housing starts declines, workers are kept on to finish already started houses.
Overall Picture—Strong but Stabilizing Labor Market
There is no reason to believe that we will see any substantial weakening of the labor market in the September report. In a best-case scenario, we will see a picture of a labor market that looks like 2019, with the economy near full employment, but wages increasing at a more moderate pace.