Wednesday, March 27, 2024

A detailed look at manufacturing, and an update on frieght

 

 - by New Deal democrat


As I wrote on Monday, the big question for this year is whether the recessionary effects of the Fed rate hikes have just been delayed, or whether, because the rate hikes have stopped, so has the headwind they normally produce. Watching manufacturing and construction, especially housing construction, is what I expect to supply the answer.


On Monday I focused on housing construction and sales. Since there’s no big economic news today, let’s take a more detailed look at manufacturing.

There are three manufacturing metrics that are “official” components of the index of Leading Economic Indicators: the ISM manufacturing new orders subindex, average weekly hours of manufacturing workers, and capital goods new orders. Note that since manufacturing makes up less of the US economy than it did in the 20th century, it takes a steeper downturn in these components to be consistent with a recession than it used to.

Let’s start with the ISM manufacturing index and its new orders component, which were last updated at the beginning of this month:



These are in a definite uptrend, although neither has definitively broken above the dividing line of 50 which separates expansion from contraction.

Next, let’s compare capital goods new orders, which were reported yesterday for February (dark blue), with industrial production (red), a premier coincident indicator, and also manufacturing employment from the payrolls survey (gold), all YoY for easier comparison. First, here’s the historical look:



Note that capital goods orders are very noisy (one reason I typically don’t highlight them), and did not turn negative in advance of the Great Recession. Nevertheless, they generally do turn in advance of industrial production, which typically turns in advance of manufacturing employment.

A similar dynamic has existed since the pandemic:



YoY gains in new capital goods orders decelerated first, followed by industrial production, followed last by manufacturing employment. The two first metrics are generally flat YoY, and manufacturing employment is only slightly positive.

Now let’s compare the average manufacturing workweek (black) with capital goods orders, both again YoY and first historically:



The average manufacturing workweek is even more leading than capital goods orders, turning first, but is even more noisy, and over-sensitive to the downside. That is, sharp declines in manufacturing hours always happen before recessions, but a downturn in hours frequently does not presage a recession at all.

Here’s the post-pandemic look at these two metrics:



Hours turned negative first, and if anything are getting “less bad” in recent months, while capital goods new orders, as already indicated, are essentially flat YoY.

Put the data together, and you get a relatively mild manufacturing recession in 2023, which appears to be recovering this year, as the ISM new orders index and the manufacturing workweek are trending higher (if not positive yet), while capital goods orders and production are flat. Manufacturing employment growth -the least leading metric - appears still to be decelerating. 

Before I conclude, let’s take a brief updated look at transportation. Remember, the theory is that everything that is produced must be shipped to market. So to confirm a trend, both should be moving in the same direction.

Here is the Freight Transportation Index through January (dark blue), the noisier and more negatively biased Cass Transportation Index (light blue), compared with heavy weight truck sales which has been an excellent leading indicator (red):



In February, preliminarily there was a steep drop in excess of -3.5% in the Freight Transportation Index, but it has not been updated at FRED, possibly because at least one component (air freight) was withheld pending further seasonal adjustments.

There was a steep drop off in all of these metrics late last year following the Yellow Freight bankruptcy. The Cass Index and truck sales may be showing the beginning of a recovery from that, although the data is too noisy to say anything definitive.

One final note: we’ll bet a detailed updated look at the spending side of construction and production via the personal spending report this Friday.

Tuesday, March 26, 2024

Repeat home sales price declined slightly in January; expect deceleration in the CPI measures of shelter to continue

 

 - by New Deal democrat


As I noted again yesterday, house prices lag home sales, which in turn lag mortgage rates. Yesterday we got the final February reading on sales. This morning we got the final January read on prices, for repeat sales of existing homes.

Last week’s report on existing home sales showed a sharp increase in February, a repeat of the seasonally adjusted sharp increase last February, which was almost completely taken back over the next two months. YoY sales remained down by over 3%, but the median price of an existing home remained higher by 5.7% - very much *unlike* new homes, where sales have firmed, but price remain almost 20% down from their peak.

This morning the FHFA purchase only price index through January declined -0.1% on a seasonally adjusted monthly basis, while the YoY gain decelerated from 6.6% to 6.3% YoY. Meanwhile the Case Shiller National index declined -0.4% for the month of January on a seasonally adjusted basis, but accelerated from 5.1% to 6.1% YoY. Since the FHFA index (dark blue) has frequently led the Case-Shiller index (light blue) at turning points by a month or two, I put more weight on that Index.

But first, here’s what the monthly numbers look like for the past five years:



Next, here is the long term graph of both of them below, compared with the CPI for shelter (red, *2.5 for scale) shows that the YoY gain particularly in the FHFA index is actually similar to gains during the majority of the past 25 years outside of recessions:



Here’s the close-up view of the last five years, better to show the current trend in both prices and shelter inflation:



For the next seven months the comparisons will be against an average 0.7% increase per month in 2023. Because house price indexes have shown a demonstrated lead over shelter costs as measured in the CPI, if present trends continue, as these YoY comparisons drop out, the YoY deceleration in OER in the CPI index should continue towards its more typical rate of between 2.5% to 4% YoY in the ten years before the pandemic. How soon it gets there will have a lot to do with when the Fed might begin to lower interest rates. 

Monday, March 25, 2024

As mortgage rates remain rangebound, so do new home sales

 

 - by New Deal democrat


Let’s begin this post by putting why I am watching new home sales in context.

The economy was kept out of recession last year, despite aggressive Fed rate hikes, in large part by commodity price deflation, much or most of which was triggered by the un-kinking of supply chains after the pandemic. That gale force economic tailwind is gone, but the Fed rate hikes remain. So the big question for this year is whether the effects of the Fed rate hikes have just been delayed, or whether, because the rate hikes have stopped, so has the headwind they normally produce. Watching manufacturing and construction, especially housing construction, is what I expect to supply the answer.

So, to the data, starting with my usual caveat: while new home sales (blue in the graphs below) are the most leading of the housing metrics, they are noisy and heavily revised. There was little this month, as January was only revised higher by 3,000 to 664,000. February gave back -2,000 of that, coming in at 662,000 annualized. In the below graph I also show th slightly less leading but much less noisy single family permits (red, right scale):



Before I discuss this graph a little further, let’s compare sales with the even more leading metric of mortgage rates. Both are shown YoY (rates inverted, and *100 for scale):



Except for the distortion created by the pandemic shutdowns in spring 2020 and the YoY comparisons in spring 2021, we see that new home sales have almost simultaneously followed the trajectory of mortgage rates: the higher the mortgage rate YoY, the lower new home sales YoY. Because mortgage rates remain slightly elevated compared with one year ago, the progress YoY in new home sales has almost completely stopped. As a result, I expect single family permits to follow the more noisy downward trend in month over month comparisons in sales in the immediate future.

In other words, so long as mortgage rates remain in the 6%-7% range, I expect new housing sales and construction to stall out as well, but not to decline significantly either.

Finally, as I always reiterate, prices lag sales. So here’s the YoY update on median prices (red), which are not seasonally adjusted (red) compared with YoY sales:



Again, aside from the spring 2020 and 2021 YoY distortions due to the pandemic lockdowns, we see that prices followed sales higher, and then in 2023 followed sales lower. We will probably continue see negative YoY comparisons in prices for a few months more before they follow sales back higher YoY probably by late this year.

But the big takeaway remains that, generally speaking, I am not expecting much in the way of big moves in new home sales or prices until there is a significant change in mortgage rates.

Saturday, March 23, 2024

Weekly Indicators for March 18 - 22 at Seeking Alpha

 

 - by New Deal democrat


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

With interest rates having come down from their highs of five months ago, I anticipated that the shorter leading indicators might follow suit and improve. This week, there was some more evidence that they have.

As usual, clicking over and reading should bring you up to the virtual moment as to the important indicators for the economy, and will reward me with a little pocket change.

Friday, March 22, 2024

Signs of a thaw in the frozen existing homes market, but a very long way to go

 

 - by New Deal democrat


There’s no big economic news today, but yesterday existing home sales were released. While they have historically constituted up to 90% of the entire market, they have much less economic impact than new home sales, which involve all sorts of construction activity, followed by landscaping, furnishings, and other sales.


Since the Fed started raising rates two years ago, the two markets have gone in entirely different directions, since existing homeowners are largely trapped by new or refinanced mortgages in the 3% range, while builders of new homes can make all sorts of accommodations to entice buyers even with mortgage rates near 7%.

As a result, the existing home market for all intents and purposes froze. Yesterday’s report on existing home sales showed that there is a little thawing going on - but hold the pom poms for now.

The seasonally adjusted annual pace of existing home sales rose 38,000 to 438,000 in February, vs, 4,530,000 one year ago. But as the comparison with non-seasonally adjusted data shows, February is typically one of the slowest sales months of the year. Thus a relatively small change - caused by, say, unusually accommodating winter weather - can make a big difference in the seasonally adjusted rate:



When we step back and look at the data for the past 5 years, we can see that there was a similar jump in February of last year as well, that made for the highest seasonally adjusted total for the entire year:



As usual, take one month’s seasonally adjusted data, especially during winter months, with an extra grain of salt.

That being said, there are signs of some improvement in the market, as both total active listings (blue) and new listings (red) housing inventory has been higher YoY for the past four months:



For new listings, this is the first sustained uptick in three years.

But if this is a thaw, it is only the beginning of the thaw, as we can see when we look at the absolute levels rather than the YoY changes (note separate scales):



The first thing to point out is that this data is not seasonally adjusted. Typically listings bottom in December or January, and improve until mid year. So we need to compare this February with February in previous years.

Before the pandemic, total active listings in February averaged just over 1 million units. New listings averaged about 425,000 for the month. After the pandemic struck, both really plunged, with active listings in February bottoming at 347,000 in 2022, while new listings in February bottomed at 305,000 last year. In February this year, active listings totaled 665,000 and new listings totaled 339,000.

In other words, while there are definitely signs of improvement, the existing home market has a long ways - as in, about 400,000 more monthly active listings and 100,000 new monthly listings - to go.