Production rebounds sharply in December, but hold the celebration

 

 - by New Deal democrat

Industrial production, the now-deposed King of Coincident Indicators, rebounded strongly in December. 

I say “deposed” because before 2001 and especially before the Great Recession, any YoY decline in production *always* coincided with or at least immediately heralded a recession. But since the accession of China to regular trading status in 1999, downturns of even -5% or more, as in 2015-16 and 2018-19 have not necessarily meant recession.

In December total industrial production rose 0.9%, while manufacturing production rose 0.6%. The November number for each was also revised higher by 0.3%. While this is obviously a big increase, mainly it just reverses the likely hurricane induced declines of the past few months. Total production made a 6 month high, while manufacturing production only made a 4 month high. Further, both remain down from their late 2022 peaks by -0.3% and -1.1% respectively:

With the big improvement this month, on a YoY basis total production is now up by 0.6%, and manufacturing production by 0.1%:

Despite the lackluster performance of manufacturing and production for the past two years, the economy has continued to be powered forward by consumption of services, and also by construction sector, and in particular construction employment, which as I pointed out earlier this morning has continued to increase.

Residential housing construction gives *very* mixed signals in December; recessionary red flag continues

 

 - by New Deal democrat

The most lading components of the long leading housing sector rose in this morning’s report for December, while the most important “hard” economic datapoint continued its decline into recession territory.

First let’s compare the most leading datapoint, permits, and the less noisy single family permits, with starts. To reiterate, starts are much noisier and generally lag permits by one or two months.  And that is what we saw this morning.

Total permits (gold) declined -10,000 on an annualized basis to 1.483 million. Single family permits (red) increased 16,000 to 992,000. Starts (blue) rebounded sharply, by 205,000 to 1.499 million annualized:

This is consistent with historical patterns, as permits bottomed late last spring and have generally slowly risen since. This has now been confirmed by starts particularly as averaged over three months. Also a reminder that some of the past few months’ volatility in starts was the effect of hurricanes in the Southeast, which delayed activity that has now been made up.

The above ought to result in a near term increase in actual construction. But the previous declines earlier in 2024 are still feeding through into this metric, which declined another -6,000 to 1.431 million units, down -16.4% down from its peak, and the lowest number since August 2021:



In the past units under construction have declined on average -15.1% and by a median of 13.4% before the onset of recessions. Indeed, with the exception of one month in the 1980s, when construction has been down more than -10% YoY, that has always meant recession. Currently units under construction are down -14.8% YoY:

Four months ago I hoisted a yellow flag “recession watch” for housing construction, and last month I hoisted the red flag indicating that housing is forecasting recession. Nevertheless, for the past few months I have cautioned that because permits (and now starts) had bottomed, units under construction would probably not fall much further, which this month’s minor -6,000 decline is consistent with. When recessions have occurred, units under construction continue to decline into the first few months of the economic downturn.

Additionally, employment in residential construction, which typically follows units under construction with a lag, has - surprisingly - continued to increase:

In the past, residential construction employment has been the last shoe to drop before a recession begins. Note that in the laste 1980s this did not occur until nearly two years after construction peaked!:

The very big fly in this ointment is the item that leads permits as well: mortgage rates (red in the graph below). These have recently increased back over 7%. The below graph shows them inverted to better illustrate how they lead permits (/1.5 for scale with single family permits):

The increase in mortgage rates over the past 3.5 months has already led to fresh declined in mortgage applications, and is sufficient to forecast a reversal of the recent increase in permits and starts. If - and most likely when - this occurs recession risks increase.

While the housing sector is forecasting recession in the near future, the jury is very much out on other important forecasting components such as corporate profits in the long leading range, and consumption and employment indicators in the short leading range. As I pointed out above, construction employment is still increasing. And the broader goods employment measure has declined only slightly. Real sales and consumption have also continued to increase. All of these would need to turn down to signal a recession in the short term future.

Real retail sales remain positive for the economy, but suggest further slowing in employment gains

 

 - by New Deal democrat

Since I posted earlier about why I follow jobless claims so closely, let me briefly restate why I pay a lot of attention to real retail sales.

Retail sales have been tracked for over 75 years. When they are lower YoY, that has historically been a good (not perfect) indicator that a recession is near. That’s because that same 75 year history empirically demonstrates that consumption leads jobs. In other words, it is the change in sales that causes employers to add or lay off employees (not the other way around, as I have sometimes seen claimed).

And the news for December was mildly positive. Nominally retail sales rose 0.4%. Inflation also clocked in at 0.4%, but after rounding real retail sales rose 0.1%, continuing their general uptrend for the past six months:

One of the rare false signals of YoY sales took place in the last several years, as consumers binged on goods purchases with their pandemic stimulus money in 2021 and early 2022, resulting in a downdraft for the next several years:

As you can see, this has abated in the last few months, with YoY real retail sales up 1.0% in December. Much of this pattern has to do with what we saw in motor vehicle prices in the inflation report yesterday. After the big jump in vehicle prices in 2021-2022, consumers balked in 2023, and new vehicle prices remained flat. In the past few months, demand has increased and prices have begun to rise again as well. 

In any event, real retail sales are a positive for the economy in the next few months.

Finally, per the above paradigm, here is the update on real retail sales vs. employment. Because the distortion in shelter prices has had so much effect on consumer inflation in the past few years, last month I also added the comparison with inflation ex-shelter (light blue):

Real sales suggest that employment gains should continue to slow. Indeed, we might very well find out with next month’s annual re-benchmarking of the employment numbers that they already have.

A refresher on why I pay so much attention to jobless claims; and why they are neutral now

 

 - by New Deal democrat 

This might be a good time to reiterate why I post each week on jobless claims, and what my system is.

Initial jobless claims in particular are a recognized leading indicator. In fact, they are one of the 10 official components of the Index of Leading Economic Indicators. Additionally, together with the YoY% change in stock prices they form my “quick and dirty” forecasting tool. 

Based on the nearly 60 year history of initial jobless claims: when initial claims are lower YoY, that is positive for good economic growth in the next few months. When they are higher by less than 10%, they are neutral — still indicating growth, but more anemic. When they (especially the 4 week moving average) are higher by over 10%, that is a yellow flag indicating the risks of recession are significant. Finally, when they are higher by 12.5% or higher for a period that persists for at least two months, that constitutes a red flag recession warning, because under those circumstances a recession has almost always been close at hand. We almost triggered that red flag 18 months ago, but the high YoY change in claims backed off just short of the two month trigger. It has turned out that there is some residual seasonality that hasn’t been massaged out of the numbers that first really appeared in 2023, and that is the type of reason why I need the signal to persist for at least two months.

With that out of the way, let’s look at this week’s numbers.

Initial claims rose 14,000 to 217,000, while the four week average declined -750 to 212,750. Continuing claims also decreased -18,000 to 1.859 million:

On the YoY% basis I use for forecasting as described above, initial claims were up 11.9%, while the more important four week moving average was up 6.0%. Continuing claims were also up 7.6%:

Per the above, although the weekly number came in higher by more than 10%, because the four week average remains under that threshold, this was a neutral reading. It suggests slow improvement in the economy in the months ahead.

Finally, since I mentioned it above, here is the current state of the “quick and dirty” model (note the four week average of initial claims is inverted so that a negative reading shows as below the zero line):

Note there was a brief 2 month period in 2023 when both were negative, but jobless claims were not higher by more than 10% during that period, so there was no true recession signal.

December CPI: rebounds in gas and car prices outpace deceleration in shelter and insurance laggards

 

 - by New Deal democrat

December consumer prices indicate that we are leaving the immediate post-covid era and seeing a rebalancing of sectors, as sectors that declined sharply in the past several years rebound.

As in November, the only two categories of “hot” numbers showing price increases of 4.0% a year or more are two laggards: shelter and transportation services. All of the former problem children are now more or less in line.

As per usual, the biggest dichotomy in the numbers is shelter vs. ex-shelter, but let’s start with the headline and core CPI readings. For the record the former increased 0.4% for the month, while the latter increased only 0.2%. On a YoY basis, headline prices are up 2.9%, an increase of 0.5% from their 2.4% low three months ago. Core prices excluding food and energy are up 3.2% YoY:

Now let’s look at CPI for shelter vs. ex-shelter:

Shelter prices increased 0.3% for the month. Everything else all together rose a sharp 0.5%. On a YoY basis, shelter increased a little under 4.6%, its lowest such reading since January 2022. Despite the “hot” monthly reading, all other prices increased 1.9% YoY, the 20th month in a row they have risen less than 2.5%.

As it has been for several years, in the broadest terms inflation in excess of the Fed target remains almost all about shelter.

Within shelter, both actual rents and “owners equivalent rent,” the fictitious measure of house prices, increased 0.3%. On a YoY basis, rent increased 4.3% while OER increased 4.8%. Both of these are the lowest since early 2022:

While the deceleration in shelter inflation has been slow, it is continuing as forecast from both the leading house price and new apartment rent indexes. To reiterate what I wrote last month, the Philadelphia Fed’s experimental new and all rent indexes, which are designed to lead the CPI for rents, for the last two quarters have forecast a decline below 4% YoY, and at the current pace of deceleration, that forecast could come to fruition within the next 2 to 3 months.

Turning to the other remaining problem child, transportation services (mainly insurance and repair costs); recall that these lag vehicle prices. In December, new vehicle prices increased 0.5%, and used car prices increased 1.2%. These have been strong monthly numbers in th past few months. But on a YoY basis, new car prices are down -0.4%, and used car prices down -3.3%. In other words, this is a rebound from the pullback since 2023 [Note: instead of YoY, graph is normed to 100 as of just before the pandemic, better to show the surge in prices and the stagnation or decline afterward]:

In December transportation services costs increased 0.5%. On a YoY basis, they rose from 7.1% to 7.3%, which is still nearly the “best” reading in 3 years:

Within transportation services, motor vehicle repairs are up 6.2% while insurance is up 11.3%.This is the real problem child is motor (for which unfortunately FRED does not provide a graph). 

What the above all means is that if we were to take out the two areas that we know lag, shelter and transportation services, consumer inflation would probably be up only something like 1% YoY.

Another former problem child of food away from home increased 0.3% this month, and is up 3.6% YoY, the lowest increase in over 4 years:

Last month I wrote that another emerging sector of concern is medical care services. This month they increased 0.2%, while the YoY measure decelerated to 3.4%, which is good news, as in the context of the past 10 years, this increase is only a little above average:

 

Before I finish, let me also update one important labor sector graph.

Nominally aggregate payrolls increased 0.4% in December, which means that after adjusting for today’s inflation number, real aggregate payrolls were unchanged, but at their record high:

This is consistent with continued economic expansion over the near term.

So let’s conclude. The lagging sectors - shelter, and motor vehicle repairs and insurance - are the primary sources of remaining high inflation. But both are decelerating as expected. At these same time, new and used car prices are rebounding from their stagnation (new cars) and deep declines (used cars) of the past several years, as are energy prices. Because the former are decelerating more slowly than the latter are rebounding, headline inflation has started to increase again.

Producer prices join the parade of yellow flags

 

 - by New Deal democrat

I generally don’t pay too much attention to producer prices, but there are a couple of exceptions. One exception is that sometimes producer prices lead consumer prices by a number of months. That hasn’t been the case recently. But the other exception is that producer prices can tell us whether profit margins are being squeezed or not. *That* is relevant at the moment.

Let me start with the monthly change in raw commodity prices (gold, /2 for scale), final demand producer prices (blue), and consumer prices (red) for the past two years:

The most important component of the above graph is that the downdraft in commodity prices - which are the most upstream prices of all - appears to have ended, as the big declines of 2023 abated except for three months in 2024. And there have been no declines in the past three months. So there is no tailwind likely to help consumers downstream (okay, I’m mixing metaphors but you get it).

Now here’s the YoY look, going back three years:

Again, we see big declines in the year after June 2022, the peak for oil prices after the Ukraine invasion. For the past nine months, commodity prices have been stable on a YoY basis. 

Even more importantly, note that final demand producer prices - the ones just before products and services are finished for consumption - have risen YoY for the past several months, from 2.1% in September to 3.3% last month. That compares with 2.7% for the last month reported in the CPI. 

Producer prices increasing faster than consumer prices means producers cannot pass on their price increases to consumers, which has the effect of squeezing their profits. And that appears to be happening, as suggested by the latest actual + estimated earnings for Q3 and Q4 of last year as reported by publicly traded corporations (via FactSet):

Earnings barely increased in Q3, and so far are estimated to barely increase again in Q4. And when corporate earnings stall, CEO’s start looking for ways to cut costs.

We’ll see what happens with consumer prices tomorrow, but this is yet another yellow flag suggesting recession risks are rising (albeit still relatively low) for later in 2025.

Scenes from the last employment report of Joe Biden’s Presidency

 

 - by New Deal democrat

Friday’s jobs report was an excellent one for Joe Biden’s Presidential term to end on. In the past 4 years, 17 million jobs have been created. Even if we take out 2021 as being a COVID rebound year, in the past 3 years there were 9.8 million new jobs, an average of 272,000 jobs per month. Meanwhile the unemployment rate declined from 6.7% at the end of 2020 to 4.1%. It rose 0.2% from 3.9% three years ago, ranging from a low of 3.4% to a high of 4.2%, which is still an excellent record. Real average nonsupervisory wages are up 0.6% from December 2020, and 1.3% from December 2021 - not so great, but still positive. And since average wages in 2020 were distorted by layoffs mainly affecting lower income workers, if we measure from December 2019, real wages have risen 4.8%.

Not too shabby.

But let’s focus on a couple of items from last Friday’s report beyond the headlines.

One of the important revelations in the past year has been how the nearly 6 million new immigrants since 2020 have distorted the unemployment rate, even as job growth has been robust. So this month let’s break out this rate for native born vs. foreign born workers. Remember that new workers who can’t find jobs don’t file for jobless claims, and Los Illegales generally do not either.

The below graph breaks out the YoY% change in initial jobless claims (light blue) vs. the total unemployment rate (light red); and compares both with the total of initial + continuing jobless claims (dark blue) and the unemployment rate for the native born only (dark red):

Initial claims always lead, but those unemployed also include people with continuing claims. So while it is less leading, the unemployment rate tracks closer to the total. And in this expansion, the unemployment rate for the native born tracks even closer than that. The inclusion of continued claims helps explain why the unemployment rate never ticked lower YoY in 2024, while the limitation of the unemployment rate shows a much less noisy leading/lagging relationship. As of December, while initial claims were higher YoY by 7.4% vs. the total unemployment rate being up 7.9%*, the combination of initial plus continuing claims were up 3.9% vs. native born unemployment being up 5.7%* (*remember these are percent changes of a percentage).

Now let’s turn our attention to employment in the goods producing sector. Recall that historically, services employment almost never actually goes down more than a fraction except in severe recessions. The decline in employment during recessions is almost all about declines in goods-producing jobs. And goods producing jobs are a leading indicator, always peaking some months before a recession begins:

Goods producing jobs in turn fall almost totally into two categories: manufacturing and construction. In the following graphs I break out the two categories separately, and focus on the recession since the 1980s.

Note that in the mid-1980s manufacturing already had a substantial downturn, but construction jobs grew strongly. It was only when both turned down in 1989-90 that a recession was forecast:

The 2001 recession by contrast was producer-led.  Construction jobs continued to grow very slowly into that recession, but manufacturing turned down sharply in advance:

Manufacturing jobs continued to decline during the “China shock” of the 2000’s, but the downturn accelerated in advance of 2007, and was joined by a downturn in construction jobs in 2007:

At present construction jobs are still increasing, while manufacturing jobs have turned down slightly:

It’s helpful to also compare this information YoY, which the next two graphs do in 20 year increments:

The closest comparison to the present is with the 2001 recession, but by the onset of that recession manufacturing jobs were down -2.1% YoY, while construction jobs were only up 1.3% and in a sharply deteriorating trend. At present, manufacturing jobs are down only -0.7%, while the trend in construction jobs has been one of slow deceleration, currently at 1.6% YoY. 

While the downturn in goods producing jobs as a whole from a peak three months ago is definitely cautionary, such minor downturns have happened many times before without a recession beginning. In other words, we have a necessary but not sufficient leading indicator at present. Most importantly, at present I am focusing on construction jobs, which seem to be levitating in the face of a downturn in the housing market and a small downturn in total construction spending. If the construction sector joins the manufacturing sector in deterioration, it will be much more concerning. But it isn’t now.