Weekly Indicators for February 24 - 28 at Seeking Alpha

 

 - by New Deal democrat

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

Last week I wrote that exogenous factors - like political decisions - could have nearly simultaneous effects across all timeframes of indicators. In other words, the long and short leading indicators as well as the coincident indicators, could all react at the same time.

This week there was evidence of exactly that.

As usual, clicking over and reading will bring you up to the virtual moment as to the state of the economy, and reward me a little bit for organizing the data for you.

January income and spending show unresolved seasonality in a typical late cycle configuration

 

 - by New Deal democrat

There were two important points in this morning’s personal income and spending report for January. The first is that there appears to be some unresolved seasonality at work. The second is that nevertheless both were weaker than one year ago.

Nominally personal income rose 0.9%, while spending declined -0.2%. Since the deflator increased 0.3%, real income rose 0.6%, and real spending declined -0.4%. Income was a new record high, while spending obviously pulled back:

But a look at the monthly comparisons for the last several years shows that the sharpest increase in real income all last year was also in January, up 0.9%, and the sharpest decrease in real spending was also in last January, down -0.3%:

This strongly hints at unresolved seasonality.

On a YoY basis, the chain type price deflator (blue) was up 2.5%, average for the past six months. Indeed (not shown) since last April prices are up 1.5% for a 2.1% annualized rate over the past nine months. Decomposing the price deflator shows that for goods the index rose a sharp 0.5% for the month, while for services it rose 0.3%. On a YoY basis, the index for services (gold) rose 3.4%, the lowest such increase since March 2021, while for goods (red) it rose 0.6%, the biggest increase since September 2023:

This points to the end of the deflationary period in commodities as being the underlying reason why the PCE inflation index has not declined further since early 2024.  While I won’t bother with the graph this month, despite the decrease in YoY inflation in services consumption, the number is high compared with the 30 years before the pandemic. As I wrote last month, it is noteworthy that goods inflation increasing, and services inflation remaining elevated, is a typical late cycle configuration.

More evidence for residual seasonality was that the personal saving rate jumped from 3.5% to 4.6% on a month over month basis. As the graph below shows, the very same jump happened in January of last year:

Ordinarily I would consider such a jump as a potential precursor to recession, showing an increase in consumer caution, but compared with last January’s rate of 5.5%, the suggestion is that the real trend is a continued slow decline in saving - again, a typical late cycle phenomenon. In other words, as expansions continue, consumers get further out over their skis, and so more vulnerable to an adverse shock.

Some good news was that real income less government transfers, one of the metrics that the NBER looks at to determine economic expansions vs. recessions, increased by 0.3% in January to another new high:

Finally, the PCE price index is used to calculate real manufacturing and trade sales (with a one month lag), another metric used by the NBER to determine if the economy is in recession or not. These jumped 0.7% in December, continuing their post-pandemic uptrend:

Because of the apparent unresolved seasonality in the January numbers, I think we have to turn to the YoY trends for more insight this month. These show continued but decelerating growth in both income and spending, as well as a likely continuing ebbing of the saving rate. Prices of goods are firming and picking up, while services remain elevated. This continues to be a typical late cycle type of configuration.

The end of “steady as she goes” in jobless claims?

 

 - by New Deal democrat

This week’s report on initial jobless claims was of particular interest, because of the issue of whether Federal employees laid off by the new Administration would cause an increase. It appears they did.

Initial jobless claims rose 22,000 for the week to 242,000, and the four week moving average rose 8,500 to 224,000. With the typical one week delay, continuing claims declined -5,000 to 1.862 million:

As you can see from the above graph, initial claims tied for the second highest weekly number in the past six months. A preliminary check of the state by state data indicates that there was a sharp increase in claims in Washington DC, but that would only marginally increase the weekly number. There was also a significant increase YoY in Virginia, but not in Maryland. So as an initial matter it appears that the Federal layoffs were only a part of this increase.

The sharp increase also showed up in the YoY% changes, which are more important for forecasting purposes. There, initial claims were up 13.6% YoY, the four week average up 7.0%, and continuing claims up 3.2%:

Of note, the average of the last two weeks in initial claims is +11.3% YoY.

These numbers are still neutral for forecasting purposes, since the four week moving average is more important than the noisy weekly numbers. But as per the above, that could change as early as next week.

Finally, let’s update our look at what this suggests about the unemployment rate in the months ahead, since initial claims in particular have a very long record of leading the unemployment rate:

On a monthly basis initial claims are up 7.0% YoY, and initial + continuing claims together are up 10.4%. Since one year ago the unemployment rate was 3.8%, for the first time in many months this suggests upward pressure on the unemployment rate, since 3.8%* 1.07 and *1.10 indicates an unemployment rate of 4.1% or 4.2%, vs. last month’s 4.0%.

More unwelcome news in (new) house prices, while sales continue sideways trend

 

 - by New Deal democrat

I almost always start out my post on new home sales by indicating that, while they are the most leading of all housing metrics, they are very noisy and heavily revised.

That was true in spades for this morning’s report for January. As shown in the graph below, sales of new single family homes declined -10.5%, or by 77,000 annualized, to 657,000, from December’s level of 734,000, which was upwardly revised by 36,000, or about 5%, from its originally reported level of 698,000. 

That’s why I usually compare them with single family permits (red, right scale), which lag slightly but are much less noisy or revised:

Just like existing home sales, which I discussed earlier this week, new home sales have been rangebound in the past two years, varying between a low of 611,000 and a high of 741,000.

So despite the sturm and drang of the monthly decline, really this just shows a steady and flat market.

What is perhaps more important is what is happening with prices. To reiterate my theme from the past few months, Ive been looking at new and existing home sales more in tandem, with a rebalancing of the market in mind. For that to happen we need price increases to abate in existing homes, and prices to remain flat or still declining in new homes.

Sales lead prices, which are best viewed in a YoY% comparison. The below graph shows sales (/1.5 for scale) and median prices of new homes (red) in that format, together with the YoY% change in the FHFA repeat sales index reported yesterday (gold):

You can see that prices followed sales higher with about a 12 month lag, and settled in to a slightly declining trend with a similar delay. 

The unwelcome news here is, just as with the repeat sales indexes yesterday, after about two years of generally declining prices, on a YoY basis the median price of a new home was higher by 3.7%. This isn’t rebalancing, but a renewed push in inflation in both new and existing house prices

As the below graph of actual non-seasonally adjusted prices shows, in January we broke that two year trend of slightly declining prices, with the highest price since summer 2022:

This is not good.

Finally, the inventory of new houses made yet another 15+ year high in January, if fact an all time high except for 2006-07. This is actually “good” news - for the moment - because as the below long term historical graph shows, recessions have in the past happened after not just sales decline, but the inventory of new homes for sale - which also consistently lag - also decline (as builders pull back):

So in summary January continued the “steady as she goes” pattern for sales, but broke that trend for prices (subject to noise and revisions next month!). With mortgage rates still close to 7%, I do not expect any upward breakout in sales soon. Which, while it isn’t *bad* news for the economy, is definitely not good news either.

Unwelcome news for homebuyers and the CPI, as repeat home sales prices continue re-acceleration in December

 

 - by New Deal democrat

There was unwelcome news in this morning’s repeat home sales reports from the FHFA and Case-Shiller. On a seasonally adjusted basis, in the three month average through December, according to the Case-Shiller national index (light blue in the graphs below) prices rose 0.5%, and the somewhat more leading FHFA purchase only index (dark blue) rose 0.4%. Both of these continue the trend of re-acceleration we have seen in house prices in the second half of 2024 [Note: FRED hasn’t updated the FHFA data yet]:

Both indexes also accelerated on a YoY basis, the Case Shiller index by 0.1% to a 3.9% gain, and the FHFA index by +0.5% to a 4.7% YoY increase:

Because house prices lead the measure of shelter inflation in the CPI, specifically Owners Equivalent Rent by 12-18 months, here is the updated calculation of its trend. Despite the increases in the house price indexes in the past several months, there is still every reason to believe that OER should continue to trend gradually towards roughly a 3.5% YoY increase in the months ahead:

The most leading rental index, the Fed’s experimental all new rental index, has not been updated since November, but the similar Apartment List National Rent Report as of the end of January continued to indicate that YoY rent increases should decline further. So the bulk of the evidence continues to point further deceleration in that huge component of consumer price inflation:

Because prices generally follow sales, as a refresher here is the graph of existing home sales. As I wrote earlier this week, for the past 2 years these have remained in a relatively tight range (following mortgage rates):

The takeaway from the increased price pressure in the existing home market as measured by the FHFA and Case Shiller Indexes is that the trend of slowly abating shelter inflation in the CPI may slow even further, although it seems likely to slow down.

Q3 2024 QCEW suggests employment was considerably weaker than we thought last year

 

 - by New Deal democrat

This will be income, spending, and housing week, but that won’t start until tomorrow. While there’s no news today, there was an important update to employment data last week; namely, the QCEW for Q3 of last year.

As a refresher, the Quarterly Census of Employment and Wages is just that, a *census* rather than a survey. It includes something like 95% of all businesses, and is taken from their tax reporting. This means it is not an estimate or guess, but very close to a full itemized count. The drawback being, of course, that it doesn’t get reported until almost 6 months later.

Since at least June of 2023, the QCEW has been telling us that the monthly jobs report has been over counting employment. The recent benchmark revisions downwardly revised 2023 and 2024 employment by about -600,000.

And the Q3 2024 QCEW tells us it will probably have to be revised down further.

Since the QCEW is not seasonally adjusted, the only way to measure is YoY. Also, unfortunately FRED does not pick up the data for easy putting in graphic form. But here is the crucial chart:

To cut to the chase, in the first five months of last year, employment grew either 1.3% or 1.4% YoY. That suddenly downshifted in June, and remained anemic through September, with YoY growth of 0.8% to 1.0%.

Now leet’s compare that with a graph of the YoY% growth in nonfarm payrolls:

With the recent benchmark revisions, the YoY growth rate in jobs for the first five months of last year is now estimated at 1.4%-1.6%, only slightly higher than the QCEW. But in the JUne through September period, it is 1.2% or 1.3%, significantly higher than the QCEW.

According to the crrrent nonfarm payrolls numbers, the economy added almost 2 million jobs during the 12 months from September 2023 through September 2024. If we bring that down to a 0.9% gain (the average of the June through September YoY gains as measured by the QCEW, that brings us down to a 1.4 million job gain, a difference of -600,000, primarily centered on the last four months of that period.

Now here is a graph of the monthly gains in jobs during that same 12 month period:

The current estimate of job gains in the June through September 2024 period is 486,000, including three very anemic months. If we apply the YoY downshifting of the QCEW during those months, all of those gains disappear.

At the same time, it’s important to note that all of the QCEW numbers for 2024 are preliminary at this point. In 2023 there was a similar cratering of the QCEW, strongly suggesting actual job losses - that subsequently disappeared when the QCEW for that year was finalized.

Finally, let me emphasize that the above does not mean there was a recession last year. Almost all of the other important indicators showed continued growth through the period. And the QCEW also reports wage growth, and there it showed continued aggregate wage growth of 4.4% and 4.5% in Q2 and Q3 of last year, very much in line with aggregate nonsupervisory payrolls from the monthly reports:

Still, it’s another cautionary signal that the economy last year (which also means the economy going into the November elections) probably was not as strong as was thought at the time.