August housing construction: even more recessionary than before

 

 - by New Deal democrat

A puzzling relationship this year has been that the housing data has been classically recessionary for a number of months, and yet the economy has not rolled over. And this morning’s dismal report on housing construction was even more recessionary. 

Let’s start with the most dismal number of all: permits (gold in the graph below) declined -50,000 to 1.312 million annualized. Excluding the immediate COVID lockdown months of April through June 2020, this was the lowest number since June 2019. The more noisy starts (blue) also declined by -122,000 to 1.307 million annualized. And the metric that is the least noisy of all and conveys the most signal, single family permits (red), declined -19,000 to a 3+ year low of 856,000 annualized:

From the post-pandemic peaks, starts are down 28.2% from their peak, permits 31.7%, and single family permits 31.1%. Although I won’t bother with the graph this month, all of those have been typical readings for the onset of most of the recessions of the past 50+ years, although in two cases - 1991 and the Great Recession - they were down by over 50%. 

On the other hand, on a YOY% basis, starts are down 6.0%, permits are down -11.1% and single family permits are down 11.5%. Typically all three have been down 20% or more at the onset of recessions in the past, although in the 1991 and 2001 recessions, they were only down about -10%:

Note that there have been a number of times, for example 1966, 1987, and 1995, where construction has been down -10% or more YoY without a recession occurring.

Let’s turn next to the number of housing units under construction. As I have written many times in the past several years, it is the best “real” measure of the economic impact of housing (blue in the graphs below). In August they declined -20,000 to a new four year low of 1.317 million annualized. They are also down 23.2% from their peak:

The above graph shows how they have followed single family permits (red), as expected. More often than not in the past by the time a decline in units under construction had declined by this much, a recession had already begun. The only two exceptions were the late 1980s, where the pre-recession decline was -28.2%, and 2007, where the pre-recession decline was -25.6%. 

Now let’s update housing units under construction with the typical final shoes to drop before recessions, houses for sale (gold) and residential construction employment (red), in comparison with units under construction, all normed to 100 as of their respective post-pandemic peaks. As I noted in the past month, after revisions both the number of employees in residential construction and new one family homes for sale peaked in March and have declined almost uniformly since:

Now here is the same data presented in YoY% change format:

Note that with the exception of 1974 and the COVID recession, houses for sale and (once available) employment in residential construction had turned down YoY before the recessions had begun. By contrast, at present these metrics are higher by 8.1% and 1.5% respectively. But at their present rates of decline, both could be negative YoY by January.

Finally, as I discussed last month, one reason why the steep decline in housing has not caused a recession yet is that other durable goods purchases, and in particular motor vehicle purchases, have not followed suit. Since then we did get an update on both passenger vehicle (gold in the graphs below) and heavy truck sales (red). Here’s the historical pre-pandemic record, averaged quarterly to cut down on noise:

Note that both types of vehicle sales were lower YoY, with truck sales typically down over 10% YoY.

Here is the monthly post-pandemic view:

While truck sales are down -15.8% YoY, passenger vehicle sales are higher by 6.2%. But as the graph below shows, in their present range passenger vehicle sales (gold) could easily turn negative YoY as early as next month:

Meanwhile, even with yesterday’s increase, nominal retail sales of motor vehicles remain in their range since last November.

To sum up, today’s housing construction report for August was very much recessionary, although in some YoY comparisons, I would expect further damage before the actual onset of one. But that could easily occur within the next four to six months. The next big datapoint to watch for will be the update on persona spending on durable and consumer goods.

August industrial production: overall neutral trend continues

 

 - by New Deal democrat

So much is imported that industrial production is much less central to the US economic picture than it was before the “China shock,” but it remains an important if diminished economic indicator. It has been trending generally sideways this year, and that trend continued in August.

Headline industrial production (blue in the graph below) rose 0.1%% in August, but after revisions to prior months, the net was a decline of -0.1% compared with the initial reading last month for July. Manufacturing production (red) increased 0.3%, but after revisions was up 0.2% compared with the initial reading for July:

Total production has not exceeded its post-pandemic high in June, but with its increase this month manufacturing production is now the highest since early 2019.

Updating my graph from yesterday, mixing production (gold, left scale) increased 1.1% for the month, but remains below its June peak, while utility production (yellow, narrow, right scale) declined -0.8%:

The overall trend in the past six months remains flat to slightly increasing, after strong increases in 2024 into the beginning of this year.

Nevertheless, my conclusion this month remains the similar as it was last month, when I wrote: “Along with retail sales, this is the second coincident positive for the economy this morning.”  Because after revisions total industrial production declined -0.1% this month vs. July, they are neutral vs. positive, but the net of both is that, unless and until consumers pull back, there is no recession.

Consumers say “hold my beer” to DOOOMing about sales

 

 - by New Deal democrat

Retail sales is the first of two very important indicators we got this morning. Per yesterday, with employment growth “dead in the water” since April, consumer spending - which leads future employment - is the single most crucial element of a turning point. 

 It really is incredible how it takes a major shock for American consumers to cut back on spending. Because in August nominally retail sales rose 0.6%, confirming the very positive weekly data that has recently shown up in Redbook. Additionally, July was revised 0.1% higher, from 0.5% to 0.6%. After taking into account consumer inflation in August, which rose 0.4%, real retail sales rose 0.2% for the month, after a 0.4% increase in July.

This means that real retail sales are now at their highest since January 2023, as shown in the graph below (blue):

The above graph also shows real personal spending on goods (gold, right scale), which is a broader measure and tends to trend similarly to retail spending, but won’t be reported until the end of this month.

Further, with several exceptions, most notably in 2022-23, in the past 75 years whenever real retail sales turned negative YoY, a recession was about to begin or had just begun. If it was positive and not sharply decelerating, a recession was unlikely in the immediate future. At present real retail sales are higher YoY by 2.1%, so there is no sign of any imminent downturn in the economy:

Finally, because consumption leads employment, here is the updated graph of real retail sales YoY, together with real personal consumption of goods compared with nonfarm payrolls (red):

Based on historical experience, after the last two good months, real retail sales now suggest that YoY jobs growth will not roll over, but remain in a similar weakly positive range for the next several months.

The big question continues to be whether the continuing chaos of the imposition of tariffs at the highest rate since Smoot Hawley in 1931 creates enough of a shock to derail consumers. So far, (at least perhaps at the top end)  it most emphatically has not.

Employment growth is dead in the water; tomorrow we will find out about production and sales

 

 - by New Deal democrat

With no news today, let’s take a look at why two releases tomorrow are especially important.

Let me begin with employment, which is “dead in the water.”  I’ve written before about how manufacturing and construction employment, and now the entire leading sector of goods-producing employment, is down. But today let me point out how narrow the poor situation in services as well. [NOTE: all FRED graphs in today’s post are normed to 100 as of April of this year].

Below is the graph of total employment (blue), total employment excluding health care (red), and services sector employment excluding health care (gold):

Not only is total employment down by more than -100,000 since April excluding healthcare, but even in the services sector (which is everything except goods production), employment in every other job except health care is up by a grand total of 2,000. Total employment in *all* sectors is up only 107,000 - and it’s all healthcare.

Domestic goods production looks shaky as well. Below is total industrial production (blue), manufacturing production (red), mining (gold), and utilities (right scale, narrow, orange):

Since March nearly all forms of production are either virtually flat or down. Only utilities (probably due in great part to AI data mining operations) are significantly higher. Tomorrow we will find out if this continues or not.

If employment is flat, and if production is also close to flat, what has really been keeping the economy growing has been consumer spending. 

One measure I keep track of weekly is Redbook’s consumer spending report, which is nominal and is only reported YoY:

In the past few weeks there have been strong grains of 6.5% YoY or more.

But more importantly, below are real personal spending on services (blue) which almost always grow even during many recessions, real personal spending on goods (gold) which tend to turn down shortly before recessions, and real retail sales (red) which also turn down prior to recessions, and while similar to goods spending are more sensitive to the downside:

Real retail sales are down from their tariff front-running March peak. Tomorrow they will be reported for August. Keeping in mind that consumer inflation was 0.4% last month, unless there has been strong nominal growth, real sales are likely to be negative.

Weekly Indicators for September 8 - 12 at Seeking Alpha

 

 - by New Deal democrat

My “Weekly Indicators” post is up at Seeking Alpha.

While job growth has almost completely stalled, and inflation shows signs of picking up, both consumer spending and the stock market continue to plow forward at full speed. It’s an odd situation that may be powered almost exclusively by people at the top end of the income distribution.

In any event, clicking over and reading will bring you up to the virtual moment as to the economic data, and reward me with a penny or two for collecting and organizing it for you.

August real average wages and nonsupervisory payrolls: some signs of flagging but no recession signal yet

 

 - by New Deal democrat

Now that we have the consumer inflation number for August, let’s take a look at real wages and income for ordinary workers.

In the jobs report last Friday, we learned that both average hourly earnings and aggregate payrolls for nonsupervisory workers increased 0.4% in August. Yesterday we learned that consumer inflation also rose 0.4%, so unsurprisingly growth in both real average wages and aggregate payrolls rounded to zero.

First, here is the historical pre-pandemic graph of real average hourly wages, both YoY (red, left scale) and in absolute terms (blue, right scale). As you can see, a decline in YoY real wages has been a decent - though far from perfect - antecedent to recessions:

The metric is badly complicated by gyrations in the work force itself. In particular, from the early 1970s through the mid-1990s, with the entry of the huge Baby Boom generation, as well as the majority of women, into the workforce, real wages underwent a generation of depression. Once entry of the last Boomer and woman was digested, real wages started rising again. Even during that period, when wages declined more than trend, it was a warning signal.

Now, here is the post-pandemic record:

Somewhat with fits and starts, real average hourly wages have been rising since June 2022, the inflection point when gas prices fell from $5 to $3/gallon, and the supply chain un-kinked. 

The increase in real average wages stands at +1% YoY, with no significant sign of decoration at this point. 

The much more reliable indicator is that of real aggregate nonsupervisory payrolls. This tells us how much the vast majority of consumers have to spend. When it rolls over, consumers pull back, and a recession almost always begins.

Here is the historical, pre-pandemic record:

This indicator is almost flawless. If real aggregate payrolls are rising (blue line) the economy is not in recession. With one exception (2002-03), shortly after it peaks, a recession has always begun, typically within two months of when the YoY% change crosses the zero line (red).

Post-pandemic, this indicator has held up as well, with several periods of weakness (late 2022, the beginning of 2024) but never crossing the zero line:

Currently YoY growth is at 2%.

Finally, as the below graph, normed to 100 as of this March shows, we appear to have entered our third period of weakness in real aggregate payrolls (thick, red line):

These have risen only 0.3% in the five subsequent months, for an annual rate of 0.7%. Meanwhile real average hourly wages (thin, orange line) have increased 0.4%.

It would be wrong to project either of these forward, since needless to say, they don’t forecast their own future trajectory. What we can say is that, if weak job growth translates to weaker nominal wage growth, and if tariffs and the weaker US$ result in higher inflation, real aggregate payrolls could cross the zero threshold, signaling recession, by early next year.

As consumer inflation shows more signs of re-acceleration, the Fed is being forced to pick its poison

 - by New Deal democrat

The Fed is really facing a no-win situation. Between the recent employment reports, the QCEW, and even this morning’s jobless claims report, the jobs market has clearly been weakening, and may be on the very cusp of contraction, implicating the Fed’s dual mandate to strive for full employment. But this morning’s CPI report shows that reviewed inflation is beginning to percolate through the economy as well. About the only silver lining is that shelter inflation continues to abate, and is almost at its pre-Covid range.

For the record, let’s start with the month over month numbers for headline inflation (blue), core inflation (red), and inflation ex-shelter (gold) for the past two years:

Note that I am no longer including the big inflationary spike of 2021-22. Note that the last three months of both headline and core inflation show no deceleration at all, and are actually closer to their highest monthly readings of the past 24 months. In other words, the deceleration in consumer inflation has stopped.

Here is the YoY% look at the same data:

This now clearly shows an uptrend in non-shelter inflation and a smaller but notable increase in headline inflation, with no deceleration in the past 12 months flat YoY core inflation.

Now let’s look at the silver lining: shelter, as usual comparing the YoY% changes in the repeat home sales indexes, which lead by about 12-18 months (/2.5 for scale), to CPI for shelter (red). YoY home price increases are near or at multi-year lows, and shelter inflation has followed. While shelter CPI increased 0.4% in August, on a YoY basis it is up 3.6%, its lowest level since October of 2021. The below graph includes several years before Covid to show that this is actually at the very top end of its 3.2%-3.6% range during the latter part of the last expansion:

On a monthly basis, actual rent increased 0.3%, while fictitious owners’ rent increased 0.4%. On an YoY they advanced 3.4% and 4.0% respectively, the lowest YoY% increases since the end of 2021:

Let’s take a look at a few other areas of interest.

First, new car prices continue to be largely unchanged, down -0.1% for the month and up only 0.7% YoY, while the story for used car prices is completely different, as they  increased 1.0% monthly and are up 6.0% YoY. Still, on a long term basis the two are within their historic relative ranges, as shown in the below long term graph which is normed to their early 1980s prices:

I suspect the rebound in used car prices is because car loan interest rates may be causing a bigger percent of purchasers to go to lower cost used vehicles.

Next, transportation services (mainly car repairs and insurance) lag the prices of new and used cars. Inflation here has returned to below 4.0% YoY this year. But note that inflation in maintenance and repairs has increased from 5.0% to 8.5% YoY in the past three months:

I suspect this is a direct result of the impact of tariffs.

Next, recently price increases in medical care services have also re-accelerated, and again this month increased 0.3% for a 4.2% YoY increase:

Finally, gas for utilities and electricity costs have also turned up sharply this year. In August the former increased 2.5% and the latter 1.0%. On a YoY basis, they are up 13.8% and 6.2%, respectively:

At least some of this is probably due to a sharp increase in demand caused by the enormous use of electricity in data-mining plants used for AI. Ordinary residential customers are not going to be thrilled, to say the least.

In sum, August’s consumer inflation report continued the trend of the two previous months, in which I wrote that consumer inflation was in a transitionary period. In August, the transition is further along, with shelter having disinflated to the cusp of its pre-COVID range, while inflation elsewhere has re-accelerated. The Fed is in the unenviable position of having to pick its poison, while there is massive political pressure to print free money for T—-p.

 

Initial claims have a Texas-sized increase

 

 - by New Deal democrat

I’ll post about the CPI later this morning. But unusually, the biggest news of the morning was initial jobless claims, which spiked to 263,000, an increase of 27,000 from the previous week. The four week moving average increased 9,750 to 240,500. Meanwhile, with the typical one week delay, continuing claims were unchanged at 1.939 million:

For comparison, the weekly number was the highest since October of 2021, but note that the four week average was higher just in June, and also in the summer of 2023:

This may well just be a one week outlier. A scan of the State inputs indicates that initial claims in Texas nearly doubled, from 16,600 to 31,900, last week. It’s unclear which employer(s) are the culprits, as there have been a number of articles in the Texas media about increased layoffs in a number of industries in the past month.

As usual, the YoY% changes are more important for forecasting purposes. While the one week number was higher by 14.3%, the four week and monthly averages are much more important. And the four week moving average was only higher by 4.3%. Continuing claims are up by 5.1%:

So take this week’s number with a liberal dose of salt. Still, it is consistent with my view that there were some unresolved seasonal distortions this summer, which would resolve this month. The bottom line is that jobless claims continue to score “neutral,” suggesting a weak but still expanding economy. 

August producer prices largely a reflection of volatile food and energy prices

 

 - by New Deal democrat

Consumer price inflation will be reported tomorrow. In the meantime, this morning producer prices for August were reported. Normally I don’t pay too much attention to producer prices - and I won’t this month, either. But let me put that in some context.

In the past, when producer prices have outstripped consumer prices, that has meant that producers aren’t able to pass on the full amount of price increases to consumers.

Since the summer of 2024, final demand producer price gains have been approximately equal to consumer price gains. If producer prices were to spike even higher, we should expect that to show up in corporate profits within another quarter or two, and possibly even this quarter. And when corporate profits turn down, they think about scaling back hiring, and even layoff off workers.

July’s report suggested that such a spike in producer costs, probably engendered mainly by tariffs, but also by the weakened US$, has begun. This month there was a reversal, but it was largely driven by volatile food and energy prices, as well as an anomalous slight decline in services PPI.

Total final demand producer prices for finished goods increased 0.1% in August. Once food and energy are taken out, they increased 0.3%. Meanwhile producer prices for services declined -0.2%:

On a YoY basis, total PPI for goods are up 1.9%, and core goods ex-food and energy are up 2.8%, while PPI for services is up 2.9%, a deceleration from last winter when they peaked at 4.5%:

Raw commodity prices were unchanged for the month, while headline final demand PPI declined -0.1%. Consumer prices, which will be reported tomorrow, are also shown in red:

On a YoY basis, commodity prices are up 2.7%, the highest since January 2023, while headline final demand is up 2.6%, about par for the course for most of this year, vs. CPI, which was up 2.7% one month ago:

With the continuing upward pressure on commodity prices, in great part due to the relative depreciation of the US$, as well as to tariffs, producer prices later in the process are being squeezed. And their equivalence to CPI suggests profit margins are stagnating as well. 

Tomorrow we will get the more important CPI, and see if producers are continuing to “eat” most of the pressure, or whether it is being passed on to consumers. In the meantime, I expect food and energy prices to remain volatile, so I would pay more attention to the core PPI readings.

The “gold standard” QCEW suggests there may have been *no growth at all* in jobs so far this year

 

 - by New Deal democrat

The Quarterly Census of Employment and Wages (QCEW) for Q1 of this year was released this morning. Perhaps more importantly, the numbers for last year were finalized. This forms the “preliminary benchmark” for the actual reported changes to payrolls over that period which will show up in next February’s report for January.

To reiterate, the QCEW is an actual census of 95%+ of all employers, who must report new employees for purposes like unemployment and disability benefits. It is the gold standard, and is used for the final revisions, a/k/a benchmarks, for monthly jobs numbers, which are estimates based on surveys.

Per the release, there were -911,000 fewer jobs created in the period than are currently reflected in the monthly payrolls totals. The report is not seasonally adjusted, but here are the YoY% changes as currently reported by the payrolls survey vs. the new preliminary QCEW-based benchmark:

[YoY% change; NY Times via Ben Casselman]

On a YoY basis, for all of 2024, about 500,000 fewer jobs were created than we thought based on the monthly payroll series. But even at the end of 2024, on a year over year basis employment grew by about 1.4 million, or 0.9%. These are final numbers.

Then in the first quarter of this year, comparisons fell off a cliff again. On a *preliminary* basis, only about 675,000 jobs were added YoY, or an increase of only 0.4%. 

These are not seasonally adjusted numbers, so although we can only estimate what the seasonally adjusted monthly change would be in the first three months of this year, preliminarily the 333,000 gain in payrolls turns into a -12,000 *decline.” This is based on a 0.9% seasonally adjusted YoY gain through December 2024, adjusting down the March 2024 number based on the final benchmark, and then multiplying that by 1.004. More sophisticated methods will arrive at somewhat different estimates, but suffice it to say that as of now the QCEW is suggesting there might not have been any job growth at all this year.

Vehicle sales in August looked pre-recessionary

 

 - by New Deal democrat

The QCEW for Q1 of this year will be released at 10 AM Eastern time this morning. It should also finalize the numbers for last year. Why is that important? Because it will also set the preliminary benchmark revisions for the monthly jobs numbers last year and into this year. I expect to report on that later, but in the meantime here is something else of interest . . . 

Last week the monthly sales numbers of light vehicles (cars, SUVs, pick-up trucks) and heavy trucks were reported for August, and they suggested that this important industrial and consumer durable good sector is rolling over.

First, here is the historical look:

The important thing to notice here is that the heavy trucks component typically rolls over first, and more decisively, while light vehicle sales are noisier, although when smoothed are also a leading indicator going in to recessions (heavy truck sales also pick up later, making them a lagging indicator coming out of recessions).

Here is the post-pandemic picture:

Heavy truck sales in August made a 3 year low, down almost -25% from their peak. 

This is a typical decline right on the cusp of past recessions. 

Additionally, after some tariff front-running and backlash, August car sales came in significantly off their previous peak at the end of last year, suggesting that these sales too may be trending down. 

Only one indicator, of course, and no indicator is perfect. But recall that the typical paradigm is that after housing, durable industrial goods and then durable consumer goods turn down in advance of recessions. This report is evidence that the latter is happening now.

Scenes from the no good, rotten, terrible, and abysmal August jobs report

 

 - by New Deal democrat

Let’s take a more in-depth look at the leading indicators for the economy in Friday’s abysmal employment report.

As shown in the graph below, employment in goods has historically turned down first; indeed, in some recessions employment in services doesn’t turn negative YoY at all:

Which is a shorthand way of saying that the leading jobs in the employment report are all in the goods-producing sector. To wit, below is a graph of employment in manufacturing (gold), total construction (red), residential building construction (orange) and goods-producing as a whole (blue), all normed to 100 as of April with the exception of residential construction, which peaked in March:

With Friday’s report for August, all 4 series have now turned down. Manufacturing employment is down -0.3%, residential construction employment is down -0.4%, total construction employment is down -0.1%, and goods-producing employment as a whole is down -0.3%.

One of the actual “official” 10 leading indicators is average weekly hours for manufacturing employees. While the “official” number simply relies on the absolute number, since the 1980s the typical number of hours worked in manufacturing has included significant overtime. Thus while a decline is negative, in the past 40+ years the economy has typically not been in the “danger zone” until this declines to 40.5 hours or less:

While we did decline -0.2 hours in August, the actual number is 40.9 hours, so we aren’t in recessionary territory yet by this metric.

Another leading indicator in the jobs report is the number of short-term unemployed. These are people who have been unemployed less than 5 weeks. This metric is somewhat noisy, but generally accords with initial jobless claims. 

Here is the post-pandemic record updated through this month:

August was the highest number for such short duration unemployment since the end of 2020.

Next let’s take an updated look at real aggregate nonsupervisory payrolls. Recall that this is an excellent “fundamental” indicator, tellling us how much average American working families in total have to spend in real terms. When that turns down, so does spending, and a recession almost always quickly follows. This had been stagnating this year before improving to a new record in July. In August nominal aggregate payrolls increased 0.4% (orange, left scale), so depending upon revisions we might set another record, although there are clear signs of deceleration (blue, right scale):



Next, here is a look at the YoY% change in total employment going all the way back to World War 2:

Currently employment is up about 0.85% YoY. In the past 80 years, only once - in 1952 - has such anemic growth not occurred either in or just prior to a recession.

Finally, let’s take a look at the main monthly coincident indicators of recession monitored by the NBER, including employment, look like over the past 12 months:

There was pretty strong growth in the latter part of 2024 into early 2025, but since the beginning of spring, there is evidence of sharp deceleration or even stagnation. Industrial production and real personal spending on consumer goods look like the crucial reports later this month.