The employment market is losing momentum
Signs of a slowing US employment market are increasing, adding further conviction to my expectation that the US will enter a recession in 2023.
While January’s bumper jobs report drew plenty of headlines and led to market expectations for future Fed rate hikes drastically increasing, February’s jobs report was largely lost amidst the collapse of Silicon Valley Bank.
While a deeper look at the latest employment data reveals some positive news, there’s also a host of items that suggest growing labour market weakness.
Let’s dive in.
I hope you enjoy this research piece. To help support my work, please consider liking & sharing my work.
Q3 QCEW data reverses — now in-line with the monthly CES jobs report
Before analysing February’s jobs report, I will revisit a topic that I have previously spoken about — the difference between the comprehensive Quarterly Census of Employment and Wages (QCEW) employment report vs the monthly Current Employment Statistics (CES) jobs report.
In a nutshell, the QCEW survey includes data from more than 95% of employers, while the CES establishment survey includes data from ~6% of all employers. In order to improve the accuracy of the CES survey, it is benchmarked annually.
Though this benchmark is significantly lagged. It is done each February, to QCEW data from Q1 of the PRIOR year — i.e. January’s monthly CES report, which was released in February, was benchmarked to QCEW data up to March 2022.
In order to gain a more up-to-date picture of the comprehensive employment situation, the Federal Reserve Bank of Philadelphia (Philly Fed), releases its Early Benchmark Revisions (EBR) report approximately one week after the latest QCEW report becomes available, which is released a little over five months after the end of a quarter. While this is still a material lag, it is less lagging than the CES benchmark revision.
While the QCEW and CES reports don’t generally see huge divergences, one such divergence opened up in Q2 2022, with the Philly Fed’s review of QCEW data suggesting that just 10,500 new jobs were added, versus the monthly CES surveys now revised estimate of 988,000. Given that the monthly CES jobs report had only been benchmarked to data up to Q1 2022, this divergence suggested that the monthly CES jobs report could be overstating job growth.
While one divergent quarter doesn’t make a trend, the divergence was corroborated by the household survey, which is not subject to the same annual benchmarking process, and which had also began to record materially less job growth than the establishment survey since Q2 2022.
Given this combination, the Q3 QCEW data had outsized importance. If it again showed that the monthly CES survey was overstating job growth, there would be a major problem, as the next benchmark is not due until February 2024. On the other hand, if it reversed, reasons to doubt the accuracy of monthly establishment survey are significantly reduced.
Now that the Q3 QCEW data has been released, what did it say? That the divergence seen in Q2 2022 REVERSED. With revisions also made to prior QCEW data, there is no longer a significant difference between the QCEW and CES establishment data.
Over the year to September 2022, the Philly Fed’s EBR report based on QCEW data, estimates that 6,072,000 jobs were created vs 5,904,000 estimated by the monthly CES establishment survey.
Though interestingly, the divergence between the household and establishment survey continues. In February, the household survey again recorded significantly lower job growth (177k) versus the establishment survey (311k). While I will now move to more deeply analyse the establishment survey data, whose accuracy has been supported by the latest QCEW data, the ongoing divergence between it, and the household survey, remains a point that continues to raise questions.
Headline employment growth remains strong…
With 311,000 nonfarm payroll jobs added in February, headline employment growth continues to paint a relatively strong picture, with 3-month average job growth of 351k continuing to remain well above its pre-COVID level.
but this is being supported by government jobs — private sector growth is decelerating
Though an important detail is that in recent months, this headline number has been supported significantly by government jobs — January and February saw 118,000 and 57,000 new government jobs created respectively, versus a monthly average of just 4.6k since 2016.
3-month rolling average government job growth thus continues to rise, and outside of the peak COVID affected re-hiring period, is around peak levels.
While 3-month rolling average private sector employment has been supported by January’s gain of 386,000, all other months since October have recorded job gains below 300,000.
While job gains of over 200,000 a month are still relatively strong versus historical levels, the broader trend is one of ongoing deceleration.
Employment diffusion index further highlights the ongoing deceleration in private job growth
In-line with the gradual slowing that has been seen in private sector employment growth, is the tapering of growth in the private employment diffusion index produced by the BLS.
In February, the total private employment diffusion index recorded a value of 56. While this remains above the key level of 50, which indicates that the majority of private industry sectors saw employment growth during February, it represented a major 12 point decline from the 68 that was recorded in January, and the lowest result since April 2020. This suggests that February’s private payroll growth was supported by a smaller number of private industries.
The more cyclical manufacturing sector is now shedding jobs
Focusing on the more cyclical manufacturing component of the employment market, reveals a much weaker picture.
The establishment survey estimates that manufacturing employment declined in February for the first time since April 2021.
On a three-month rolling average basis, employment growth remains positive, but only at a modest average of 5,000 jobs per month.
Likewise, after falling close towards the 50 level in recent months, the manufacturing diffusion index fell into negative territory in February, falling to a level of 47.2.
Falling manufacturing employment corroborated by weakening Fed manufacturing surveys
Corroborating the decline in manufacturing employment that was recorded by the BLS’ monthly survey in January, are the manufacturing surveys conducted by the various Fed branches.
For many months now, key indicators from the average of all Fed branch manufacturing sector surveys have been moderating significantly.
The most leading component, new orders, was the first to record outright declines. This was followed by the hours worked component, which began falling in December. The most lagging indicator (and particularly so on this occasion given the difficulty firms have had in finding enough staff to meet an artificial surge in demand), being employment, has also seen a significant moderation over many months.
Some Fed branches have begun to record declines in the employment subcomponent (including the long-running Philly Fed survey), but the average of all Fed branch surveys continues to just keep its head above water, recording a reading of 0.3 in February, and 0.6 in March.
With the values so close to an overall contraction, it’s not surprising to see the BLS record a decline in manufacturing employment in February.
The more leading indicators of new orders and hours worked suggest that further weakness in manufacturing employment is likely to lie ahead.
Fed surveys of services sector activity show the broader employment market is also weakening
Given its more cyclical nature, the manufacturing sector often precedes broader economic weakness.
The various Fed branch surveys of the services sector, are indicating that yet again, the weakness that has been seen in the manufacturing sector, is spreading.
Sales based indicators of services sector activity has been negative for 3 of the past 4 months. The hours worked component is teetering around 0. Part-time employment has been negative for four of the past six months, and recorded its largest negative value since July 2020 in March (-3.8). The full-time employment category fell in March and is now relatively close to 0, recording a reading of 1.7 in March.
Weakening demand for labour is also showing up in decelerating wage growth
Decelerating wage growth is providing another indicator of an employment market whose strength is decelerating.
Average hourly earnings growth of 0.2% in February was the slowest growth recorded since February 2022, and marked the third consecutive month of decelerating MoM wage growth.
On a 3-month annualised basis, average hourly earnings growth has fallen from 4.9% in December 2022, to 3.6% in February, the lowest rate of growth since March 2021.
On a 6-month annualised basis, average hourly earnings growth fell to 4.1% in February, a level not seen since June 2021.
Quits have declined to the lowest rate since March 2021
While from elevated levels, the rate of people quitting their jobs has declined materially over recent months, falling to a level not seen since March 2021. This indicates that individuals are either less confident of finding better employment elsewhere, or are receiving fewer offers from other firms competing for their labour.
Initial unemployment claims have continued to be relatively low, but this isn’t surprising
While the signs of a decelerating employment market continue to grow, unemployment claims continue to remain relatively low.
This shouldn’t be overly surprising, as the above mentioned indicators show a jobs market that is weakening, but which is not yet experiencing outright declines, and which has instead so far continued to generate enough new jobs to keep the unemployment rate relatively low.
Given my expectation that the employment market will weaken further over the months ahead, eventually leading to a material increase in the unemployment rate, I expect unemployment claims to rise over the months ahead.
The broader picture of a moderating jobs market shouldn’t come as a surprise: Fed tightening generally leads to higher unemployment & recessions
While the overall level of job creation has remained relatively strong, keeping initial unemployment claims and the unemployment rate relatively low, the number of indicators pointing to a jobs market that is losing momentum, continues to grow, and as discussed above, includes:
moderating private sector job growth;
a deceleration in the BLS’ private sector employment diffusion index, including a particularly significant drop in February from the level recorded a month earlier;
falling manufacturing employment in February;
a continued moderation in the BLS’ manufacturing employment diffusion index, which turned negative in February;
Fed branch surveys showing an ongoing weakening of key manufacturing and services sector indicators;
an ongoing moderation in average hourly earnings growth; and
an ongoing moderation in the rate of quits.
With the Fed aggressively tightening, this shouldn’t come as a surprise.
When unemployment rates are high, the Fed loosens monetary policy, encouraging a boom and lower unemployment. When unemployment rates are relatively low, the Fed tightens monetary policy, encouraging a bust, and higher unemployment.
The Fed’s perpetuation of boom-bust economic cycles via its artificial altering of interest rates, can clearly be seen in the historical data.
Don’t expect this time round to be different.
With the money supply declining, current tightening is extreme — history shows that as opposed to high inflation, the greater concern should be that of a deflationary bust
Indeed, instead of being different (i.e. expecting a soft or no landing), there’s a strong argument that this time round, there will be a particularly negative economic shock.
The reason for this is simple: the Fed’s current tightening is relatively extreme.
By focusing on the lagging indicators of services prices and employment, the Fed has continued to tighten whilst the money supply has not only moderated, but is experiencing NEGATIVE YoY growth.
Not since the Great Depression has the Fed raised rates while the M2 money supply was declining. Not since 1938 has the M2 money supply declined on an annual average basis.
An inflationary, credit driven economic system does NOT react well to a falling money supply. As such, they are rare events that policymakers generally seek to avoid. History provides a clear lesson as to why: such periods are generally associated with economic depressions and busts.
No longer should high inflation be the primary concern surrounding the US economy — a declining money supply will resolve this problem. Instead, the recent trouble in the US banking sector is providing a warning sign of the greater danger that now lies ahead for the US economy — a deflationary bust.
With employment generally a lagging economic indicator, the increasing signs of a weakening employment market, thus suggest that a more significant economic downturn may not be too far away.
As such, I continue to expect that the US will see a recession in 2023.
Thank you for reading!
Thank you for continuing to read and support my work — I hope you enjoyed this latest piece.
In order to help support my independent economics research, which aims to democratise access to institutional grade insights & analysis (as opposed to it being the exclusive domain of hedge funds & asset managers), please like and share this latest research piece.
If you haven’t already subscribed to Economics Uncovered, subscribe below so that you don’t miss an update.