Weekly Indicators for May 13 - 17 at Seeking Alpha

 

 - by New Deal democrat

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

Several of the monthly reports for April - retail sales, residential construction, manufacturing production - all demonstrated a significant “soft patch.” But that has not shown up in the high frequency data updated weekly.

As usual, clicking over and reading will bring you up to the virtual moment as to the status of all the data, and reward me with a little pocket change for organizing it for you.

Real wages, payrolls, and consumption vs. employment, and their forecast implications: April update

 

 - by New Deal democrat

With this week’s inflation report for April, we can update several measures of the real economic status of average American workers, as well as their forecast for further job and economic gains.

First, here is real average hourly wages for nonsupervisory workers. In April, nominal average wages increased 0.2%. Since consumer inflation increased 0.3%, real nonsupervisory wages declined -0.1%, the third monthly decline in a row:

Real nonsupervisory wages are up 2.8% since just before the pandemic, and while the sharp increase in 2020 can be discounted due to compositional effects (many more low wage service workers were laid off during the pandemic closures than more highly paid office workers), still real average hourly wages have made no progress at all since July 2020.

On the other hand, on a YoY basis, real average hourly wages are up 0.6%, which historically is not bad:

For the real amount of wage income available to the American working and middle class as a whole, we turn to real aggregate payrolls. Because the number of hours worked actually declined during April, these only increased 0.1%, so in real terms they declined -0.2%:

With the exception of the pandemic, these have always peaked between 4 and 10 months before the onset of recessions. As the longer historical graph below shows, one month decline is not unusual, but obviously if this continues for several more months it would be of increasing concern:

On a YoY basis, these are up 2.1%. Typically they have sharply declined to negative territory coincidently or very close thereto the start of recessions:

There is no sign of any such sharp decline at present. So these remain a very positive sign for the economy now.

Finally, retail sales for April were unchanged, meaning real retail sales declined -0.2%. As I have written many times over the past 15 years, although the relationship is somewhat noisy, consumption leads employment. Here is the historical graph of the 25+ years before the pandemic:

And here is the post-pandemic record, with real retail sales down -0.3% YoY as of April:

Typically employment changes at about 1/2 the rate of consumption, so the above graphs divide YoY consumption by 2. One year ago job gains were averaging just under 300,000 per month, so real retail consumption suggests that employment gains will be significantly below those levels in the next few months, although still positive. 

April housing: Uh-oh, housing units under construction has stopped levitating

 

 - by New Deal democrat

This morning I pointed out that manufacturing production is -1.8% below its 2022 high, and may be in a slightly declining trend. Which means that added attention has to be paid to whether the other leading production sector, construction, is holding up. 

Instead, this morning brought the first sustained evidence that housing units under construction, the “real” economic measure of the residential building sector, which had been levitating at or near its  post-pandemic high for over a year, has finally turned down.

To cut to the chase, housing units under construction declined -1.4% in April, after a -1.1% decline in March. It is now -5.6% below its October 2022 peak (single family and multi-unit numbers also shown):

In the past, it has typically taken a -10% decline or more before the onset of a recession. After a minor decline in 2023, in the past four months the increased pace of decline has taken us over half the way to that mark:

Earlier this year, while noting that I expected to see more of a decline in the actual hard-data metric of housing units under construction, I wrote that “With permits having increased off their bottom, I am not expecting such a 10% decline in construction to materialize.” Here’s the updated look.

Below I show total permits (dark blue) compared with the much more volatile and slightly less leading total starts (light blue) and also compared with single family permits (red, right scale) which are the most leading and least volatile of all:

Starts rose a sharp 5.7% for the month, while total permits declined -1.0% and single family permits declined -0.8%. The dip in total and single family permits in the past two months has taken us back to January 2023 levels, between -20%-25% below that peak. Historically it has taken more than a -30% decline, or even a -40% decline, before a recession has begun:

April’s decline in permits in turn was most likely caused by the increase in mortgage rates back above 7%. As per usual, mortgage rates lead permits:

Importantly, it’s worth noting that even so, we have not seen mortgage rates anywhere near last October’s 7.79% peak.

My sense is that, while housing units under construction will decline further, unless interest rates increase further, permits and starts will stabilize, and after a period so while units under construction, without crossing into recession warning territory.

At the same time, this is the most negative housing construction report since the end of 2022, when there were a lot of other signals that suggested a recession might be very close. 

To return to my opening comment, this year I am paying extra close attention to the manufacturng and construction sectors, because a significant turndown in both of those simultaneously would be a danger signal for oncoming recession. This morning we did get reports showing declines in both. Only one month, but these are the two most negative economic reports of this entire year so far.

Jobless claims still positive, still suggest at least a slight improvement in unemployment in coming months

 

 - by New Deal democrat

[Programming note: I’ll discuss housing permtis, starts, and units under construction later today. Preliminarily, this morning’s housing report may have been the most significant negative data of the entire year so far. Stay tuned.]

Initial jobless claims declined -10,000 last week back into its recent range, to 222,000. The four week moving average rose 2,500 to 217,500. With the usual one week delay, continuing claims rose 23,000 to 1.794 million:

This takes back some - but not all - of the caution raised by last week’s big jump, with the important caveat that there may be some unresolved seasonality at work, given the similar increase last summer that began in May.

The more important for forecasting purposes YoY figures, initial claims were down -1.3%. The four week average was down -0.2%, and continuing claims were up 4.9%:

Although the recent trend is tending towards the negative, as of now both initial claims comparisons remain positive for the economy. Continuing claims, while negative, with the exception of three weeks at the turn of the year have been in the range of 1.750-1.830 for the past ten months. Unless there is a significant change, continuing claims are boing to be much less of an issue in about eight weeks.

Two weeks into May, let’s update our Sahm Rule comparison as well. This year has averaged out to 227,000, which is exactly equal to last May’s average. This tells us that the YoY comparisons with the unemployment rate one year ago should improve in the next few months:

Since the unemployment rate last May was 3.7%, and the May-July average was 3.6%, I continue to expect the unemployment rate to improve slightly. Even using the less leading continuing claims metric, since the YoY comparisons are getting less negative, I would not expect any further incrrease in the unemployment rate at very least over the next few months.

Industrial production still flat, manufacturing slides

 

 - by New Deal democrat

Industrial production, one of the premier series the NBER has historically used to declare recessions vs. expansions, has faded in importance since China was admitted to regular trading status in 1999. As you can see in the first graph below, both total and manufacturing production peaked in 2007. Further, manufacturing has continued to fade, as its post-pandemic peak has not equaled its 2010’s peak either:

In March, total production was unchanged from net upward revisions to February and March. Without those revisions, production would have been up 0.1%. Nevertheless it is still down -0.7% from its September 2022 post-pandemic peak. 

The news was significantly negative for manufacturing production, which declined -0.3% from net downward revisions of -0.2% for February and March, and is down by -1.8% from its post-pandemic peak as well:

Note that in 2023, like 2015-16, and 2019,  production was again down YoY with no recession. As of April, manufacturing production is down -0.1% YoY, while total production is down -0.4%.

Through March, production had essentially been flat. With all the revisions, as of April the trend may still be neutral, or it may have turned slightly negative. This is what we have been seeing for a year or more in both the ISM manufacturing index and in the Fed regional indexes. 

My overall theme for this year has been to watch both manufacturing and construction for any signs of strength or weakness. If we have - at least as of April - a confirmed sign of weakness in manufacturing, then construction becomes even more important. In my post later today, I’ll update that sector.

Real retail sales back to negative YoY

 

 - by New Deal democrat

Here is today’s update on one of my favorite indicators: retail sales. In April they were unchanged on a nominal basis. Adjusted for inflation they declined -0.3% for the month. They are also down -6.2% from their 2021 peak and -2.9% since January 2023:

On a YoY basis, they have also returned to the negative side, down -0.3% (note two graphs below adds 0.3% to show at the zero line):

Here is the historical comparison going back 30 years:

With the notable exception of last year, such a YoY decline has only happened during recessions (and with few exceptions the same is true going back 75 years with the predecessor series).

Because the two series tend to trend similarly, below I show the historical record for the past 15+ years of both YoY real retail sales (dark blue) and YoY real personal spending on goods (light blue), a similar but more comprehensive measure. The two metrics tend to trend together over time, although the latter has tended to increase more (hence I adjust to bring the trends more in line):

Here is the close-up post-pandemic view:

With the addition of this month’s data, the bigger picture is that real retail sales, which recently had trended neutral, now appear to be trending slightly downward, while real spending on goods has continued to trend higher.

Unless and until there is a confirming downshift in real personal spending on goods, which when positive have almost always meant continued expansion, there does not appear to be any reason for alarm.

April consumer prices: still an interplay of gas and house prices, with a side helping of motor vehicle insuance

 

 - by New Deal democrat

First, a programming note: I’ll post about retail sales later today.

Consumer inflation in April continued essentially to be an interplay between shelter and gas prices, with a side helping of auto insurance and repairs. During late 2022 and early 2023, shelter was still accelerating or steady at a high rate of inflation, while gas prices were falling. Beginning in late 2023, the dynamic reversed, as shelter inflation was slowly decelerating, while gas prices had bottomed. That remained the case in April.

So first, let’s look at the month over month change in headline inflation (blue) vs. inflation less energy (red) and inflation less shelter (gold) for the past two years:

All three rounded to +0.3% increases in April, about par for the first two for the past twelve months, and above average for CPI less shelter.

On a YoY basis, the trends become clearer, with the increase in gas prices leading to an increase in all items less shelter, steady headline inflation, but a continued deceleration in CPI less energy - which is another way of saying that energy prices have increased, while shelter price gains have continued to abate:

In particular, shelter has continued to behave as I expected. Here is an update to the 12-18 month leading relationship between house prices (as measured by the FHFA) and Owners’ Equivalent Rent in the CPI:

House prices are currently increasing at about their average pre-pandemic rate, which has translated to OER and the other measures of shelter inflation to continue to decelerate YoY, but at a slower pace than their initial rapid decline. On a YoY basis, OER has increased 5.8%, a -0.1% decline from its YoY rate in March. Rent of primary residence (not shown) has followed a similar trajectory, currently up 5.4% YoY vs. 5.7% YoY in March. I expect this trend to continue in the coming months.

Although I won’t bother with a graph, the former problem children of new and used vehicle prices have reached a new equilibrium. Used car prices have actually declined -6.9%YoY, including -0.4% in April. New car prices are also down -0.4% YoY.

The remaining problem areas of inflation are:

 (1) food away from home (fading), which peaked at 8.8% YoY just over one year ago, and is now down to a 4.1% increase, close to its pre-pandemic average of 2.5%-3.0%;

 (2) electricity, which has followed gas prices higher, rising from 2.2% YoY last August to 5.1% in April, although it declined -0.1% for the month; and 

 (3) transportation services - mainly car repairs (unchanged for the month, but up 7.6% YoY) and insurance (up 1.8% for the month and up 22.6% YoY!) - which has rocketed from its pre-pandemic range of 2.5%-5.0% to as high as 15.2% in October 2022, and is now still up 11.2%.

Although I won’t repeat the graph this month, based on the past inflationary period of 1966-82, it is clear that transportation services lags increases in vehicle prices by 1-2 years and even more, sometimes increasing right through recessions

To summarize: if we exclude the well-documented historically lagging sectors of shelter prices and motor vehicle insurance, consumer inflation continues to be well behaved. To repeat: ex shelter, consumer prices are only up 2.2% YoY. Any surprises in the month ahead will likely be due to changes in gas prices. If gas prices become well-behaved again, headline inflation should go below 3%

April producer prices reflect some building pressure from a strong economy with full employment

 

 - by New Deal democrat

Tomorrow and Thursday a plethora of data will be released, on consumer inflation and spending, production, housing, and jobless claims. In the meantime today we got a chance to look at upstream pressures on inflation.

And those upstream pressures do seem to be building slightly, reflecting a strong economy with full employment.

Commodity prices increased 0.9%. These are very volatile, so this was not particularly out of the ordinary, as shown in the below graph of monthly changes for the past 10 years:

YoY commodity prices are just 0.1% above unchanged (red, left scale in the graph below). They are very well behaved compared with just after the pandemic (blue, right scale):

By the time we get to finished products, we can see some pressures building up particularly as to goods (blue), which increased 0.7% in April and so far this year are up 4.1%. Service inflation (red) for producers rose 0.6%, the second highest monthly increase in two years:

By contrast, as shown in the below graph, in the six years before the pandemic, final demand goods and services typically rose on the order of 0.2% or 0.3% monthly:

Nevertheless, on a YoY basis both producer goods and services inflation is well within its normal range prior to the pandemic:

There is reason to believe, at least when it comes to goods, that the recent bigger monthly increases may not last. Below is the most recent update of the FRBNY’s Global Supply Chain Pressure Index:

Each horizontal line represents one standard deviation from the long term average (including the pandemic).  Since late last year there had been some tightness compared with the pre-pandemic average, but in April this entirely dissipated.

So to reiterate, it looks likely that there is a little producer pressure due to a strong economy and full employment, but nothing out of the range of normal on any significant time frame at this point.

Tomorrow we will see how this plays out with consumer prices. Because gas prices increased 5.4% in April, I am expecting the headline number to come in a little hot as well. Meanwhile I expect shelter inflation to continue to abate but more slowly than in the past 12 months. We’ll see.

Weekly Indicators for May 6 - 10 at Seeking Alpha

 

 - by New Deal democrat

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

The majority of short leading and coincident indicators continue to show strength rather than weakness. This week it was commodity prices’ turn to show that the global economy is getting stronger.

As usual, clicking over and reading will bring you up to the virtual moment as to the economic data, and reward me with a little pocket change for my efforts.

The Household Survey isn’t the only data series sending up caution flares

 

 - by New Deal democrat

I’ve written two posts earlier this week delving into the big divergence between the Establishment Survey portion of the Employment Report, which shows moderate growth, and the Household Survey, which is most consistent with a recession already having started.

At any given time, some data will be positive and some will be negative. That’s why I follow a whole series of reports with longer term proven reliability. Most of those at present are positive.

But the Household Survey isn’t the only negative data point. 

Here is a graph from six months ago showing the historical record over the past 25 years of both the ISM manufacturing index and the ISM non-manufacturing index. The former has a 75 year history, but the latter was only started 25 years ago and somewhat revised 10 years later:

Since the China shock in particular following its being accorded normal trade relations in 1999, there have been a number of “false positives” in the manufacturing index. But when it has been paired with the non-manufacturing index, measuring services, and the latter has *also* dipped below the 50 mark dividing expansion from contraction, the economy *has* been in recession - with the sole exception of one month in 2022.

Since then, the manufacturing index has been generally improving, although in April it dipped back below 50 to 49.2:

Also in April, the non-manufacturing index dipped below 50 for the second time post-pandemic, to 49.4:

Again, only one month. But worth paying extra attention to. If the non-manufacturing index gives us several more readings below 50 without further improvement in the manufacturing index, that would spell trouble.

Initial claims jolted awake from snooze-fest by highest number in almost nine months

 

 - by New Deal democrat

After several months of snoozing at almost identical weekly levels, initial jobless claims awoke with a bit of a jolt this week, increasing by 22,000 to 231,000, the highest weekly number since last August. The four week average unsurprisingly also rose, by 4,750, to 215,000. With the usual one week delay, continuing claims rose 17,000 to 1.785 million, still one of the lowest readings since last August:

As usual, the YoY% figures are more important for forecasting purposes. The weekly number was higher for the first time in six weeks, by 2.7%. The four week average is still lower by -1.4%. Continuing claims remain higher, by 4.6%, but are still close to their lowest YoY reading in over a year:

Now that we have all of the jobless claims data for April, here’s what the monthly numbers (right scale) look like compared with the unemployment rate (left scale):

To reiterate, we have 60 years of evidence that initial jobless claims in particular lead the unemployment rate. Continuing claims do also with a much shorter lead time, and sometimes they are coincident. With initial and continuing claims close to unchanged YoY, the unemployment rate should move in that direction as well. Now that the last 3.4% reading of the unemployment rate is out of the picture,  and there is only one 3.5% reading left (for July), I continue to expect that the unemployment rate is more likely to decline towards 3.7% than any other direction. As to last week’s renewed 3.9% rate, Paul Krugman has helpfully noted that it was primarily a rounding issue, as one digit further out it rose from 3.83% to 3.87%. In any event, initial claims continue to indicate that the “Sahm Rule” is not going to be triggered in the near future.

A quick scan of this week’s release does not indicate any special issues in any State. Because last year seasonally adjusted claims rose throughout May and remained high during the summer, it is possible there is a residual post-pandemic seasonal adjustment issue. We’ll have to watch and see if this is just a one-off anomaly or the beginning of a longer change of trend.

The Establishment survey portion of the jobs report continued to be positive

 

 - by New Deal democrat

On Monday I wrote that the Household survey portion of the jobs report was recessionary for the second time in three months. But I pointed out that there was a very large divergence in jobs growth in the past 24 months, amounting to 1.7% of the prime age workforce, between that survey and the Establishment survey, one of the largest such divergences on record.

Today let’s take a look at the Establishment survey, which is much more positive.

Every month as part of my look at the jobs report, I look at the leading employment sectors. These are the ones that usually turn down first before the overall jobs market does.

So let’s look at five of those sectors: manufacturing, the sub-sector of motor vehicle manufacturing, construction, the sub-sector of residential building construction, and goods production as a whole.

To begin with, with the exception of the manufacturing data in several months, almost all of these have been positive every month for the entire last year:

That’s pretty positive, especially when we see how this compares with the historical record.

Here’s the YoY% change in each post-pandemic:

Unsurprisingly, per the above, all are positive.

Here is the historical record:

With the exception of 1960 and the two oil shocks of the 1970s, all or at least most of these had turned negative before or just as the recession was beginning. That’s not the case now.

Even before these average hours in manufacturing have turned down, typically by more than -.5 hours YoY. At present, average manufacturing hours are down -0.1:

Here’s the historical record for comparison:

Especially in the past 30 years, there have been about half a dozen times when manufacturing hours have been down more than they are now YoY without a recession occurring.

In short, when we look at the jobs sectors that we would expect to already be suffering before a recession were to start, there are no such signs of distress at present. And since the Establishment Survey is bigger and less noisy than the Household Survey, we should expect the divergence between the two to resolve in the Establishment Survey’s direction.

Q1 credit conditions showed no significant change

 

 - by New Deal democrat

The Senior Loan Officer Survey is a long leading indicator, telling us about credit conditions that typically turn worse a year or more before the economy turns down, and improve just at the economy is ready to turn up.

The big drawback of this series is that the information is only reported Quarterly, and with a one a one month lag. As I indicated in my introductory note yesterday, data for Q1 was released yesterday.

There are two series that have a long enough record to give us a lot of information: whether banks are tightening or loosening standards; and the demand for commercial and industrial loans. 

Let’s look at each in turn.

The first series is the percentage of banks tightening lending standards, meaning that a positive number means more tightening than loosening standards, i.e., positive is worse for the economy. In Q1, there was a slight increase in the percentage of banks tightening credit conditions:

Versus Q4 of last year, in Q1 there was a 1.1% increase to 15.6% in the number of banks tightening standards for large firms, and a 1.2% increase to 19.7% as to small firms. This is similar to what happened in 2002, is not a significant change q/q, and is much lower than the roughly 50% for both metrics back in Q3 of last year. As with Q4, this is consistent with coming out of past recessions.

The story was similar as to the second series, demand for commercial and industrial loans (confusingly, in this one higher does mean better). There was a slight downshift q/q of -1.6% as to large firms and -0.6% as to small firms:

Again, this has been more typical of an economy coming out of a recession than going in to one. 

The Chicago Fed also looks at credit conditions, but their data is weekly and thus much more timely. Here the data has been much more consistently good (these series, again, are ones in which a positive number means “tightening” and so is worse for the economy):

The Index adjusted for “normal” credit conditions has never shown any tightness since the end of the pandemic lockdowns in 2020. The more leading leverage subindex did show tigh contentions after the Silicon Bank failure last year, but has been negative or neutral for the past six months. Note that the leverage index has been more prone to false positives than the adjusted financial conditions index.

Unless for some reason the Fed decides to tighten again, or there is another spate of bank failures, credit conditions are either relatively loose (the Chicago Fed indexes) or “less restrictive” (the Senior Loan Officer Survey), which do not indicate any particular credit stress in the system.

For the second time in three months, the Household jobs Survey was recessionary

 

 - by New Deal democrat

First, a brief programming note. This week is particularly sparse in the new economic data department. The Senior Loan Officer Survey will be reported this afternoon, and on Thursday as usual we get jobless claims. Aside from that, nada. So I might take a day or two off.

But I want to spend some time looking more closely at last Friday’s jobs report(s). I use the plural, because last Friday there really were two very divergent reports. The Establishment report was decent, but as I say in the title to this post, for the second time in three months, the Household Report was what I would expect to see in a recession.

Let’s start by comparing the employment level (blue) with the unemployment level (red). The former did increase by a paltry 25,000, while the latter increased by 64,000. On a YoY basis, the employment level is up 0.3% (blue in the graph below), while the unemployment level is up 13.6% (red, /15 for scale). The graph adds or subtracts the current change so that both show at the zero line):

The next two graphs give the historical view, with the same adjustment:

At no point in the past 75 years have both metrics been at their respective current levels except during recessions. Only twice - in the 1950s - did they come even close.

A similar story is told by the U-3 unemployment rate (blue) and the U-6 underemployment rate (red)(this latter statistic has only been reported since 1994). Currently the unemployment rate is 0.5% higher than 12 months ago, and the latter 0.8% higher:

Again, here is the historical view:

Neither one has ever been this much higher YoY without a recession having already started.

Note that the above is different from the “Sahm Rule,” which is a three month average increase of 0.5% over the 3 month average low in the past 12 months. That metric currently stands at .37%:

With only 4 exceptions (and one near miss) in the past 75 years, even at this level a recession has already been occurring:

Turning to the employment side of the coin, the YoY change in the employment level is slightly below the YoY change in the prime age population, i.e., the number of people who became employed is less than the number of people who on net entered this prime employment demographic:

Historically only 4x in the past 50+ years has this been the case without a recession already occurring or at least imminent, and one of those times was only for one month:

At root the source of this divergence dates back to March 2022. Since then, while the Establishment Survey has indicated that jobs have grown by 4.6%, the Household Survey has indicated only a 2.1% gain:

Indeed this divergence between the two measures, on a YoY population-adjusted basis presently at about 1.7%, has only been matched, and usually only for a month or two, 8x in the past 50+ years:

As the above graph shows, while there is lots of noise, there has always been a reversion to the mean. Normally this is because the noisier Household series converges towards the more stable Establishment survey data. 

I suspect what is going on has to do with the formation and closure of new businesses. There is plenty of evidence that in the immediate aftermath of the pandemic, a record number of new businesses were started. If some of the self-employed at those businesses have attritted back into employment by others, that may explain the disparity. If so, I would expect to see a couple of outsized gains in the Household survey in the months ahead. We’ll see.