How $4/gallon gas could take the economy from a nearly complete stall into outright recession

 

 - by New Deal democrat

So, first some bad news: my tech issue has resurfaced, so only links to graphs rather than graphs themselves, hopefully just for a day or two. Basically, unless I keep a bar up open to the blog page, Google and Apple sever their “handshake,” and I have to start from scratch to drag them back into it. Think of it as the tech version of herding cats.

And it’s a particular shame because, well, it’s always a bad day for the economy when the most exciting drama on TV is the financial channel. So today let me take a look at the state of the economy, ex-gas prices; and then what gas prices of $4/gallon or more might do to it. In that context I’ll also update an important graph on real retail sales, which were reported for January on Friday.

First, a couple of months ago I mentioned that gas prices under $3/gallon were a new, real tailwind for the economy. I showed this by dividing that cost by average hourly wages for non-supervisory workers. The resulting graph showed how much labor it required for an ordinary worker to be able to buy a gallon of gas. In January it was close to the lowest since the beginning of the new Millennium.

Here’s the link to an updated graph. Since as of last week’s report gas was just over $3/gallon, I’ve normed the result so that gas at $4/gallon divided by the average hourly wages shows at the 0 line:

 https://fred.stlouisfed.org/graph/fredgraph.png?g=1TeXc&height=490 

The simple summary is that gas prices at $4/gallon would no longer be a tailwind, but they wouldn’t be much of a headwind either. Rather, they would be about average (compared with wages) for the past 25 years.



All things being equal, gas prices deteriorating from a significant positive for the economy to merely neutral wouldn’t be that big a deal. But all things are never really equal. 

Because the economy as of the end of 2025 was balancing just at the edge of recessionary readings. The below link goes to a graph of the four main monthly datapoints used by the NBER to determine whether or not a recession is underway - jobs, real personal income minus government transfer payments, real manufacturing and trade sales, and industrial production. Because business sales have only been updated through last November, I also include real retail sales, which as I noted above were just updated through January last Friday (declining -0.3% for the month, and -0.8% below their most recent interim peak last August). Additionally, I wanted to show the impact of AI related data center construction by removing utilities from the industrial production measure: 

 https://fred.stlouisfed.org/graph/fredgraph.png?g=1TeZS&height=490 

All of the above metrics went basically sideways in 2025. *All* of them are below their respective peaks in various months from April through September. It’s already been an open question whether the government shutdown last autumn formed the peak of last expansion. Either the expansion just barely scraped by, or we were already in a very shallow recession.

In other words, gas prices turning from a tailwind to simply neutral, even if they don’t go much above $4/gallon, may well be enough to tip over the above metrics into outright recessionary readings.

In that regard, my final link is to one of my usual real retail sales graphs, showing how it (and similarly real personal spending on goods) typically leads employment by a number of months: 

 https://fred.stlouisfed.org/graph/fredgraph.png?g=1TeU4&height=490 

Both Real retail sales and real spending on goods were negative YoY in December, before the former rebounded to +0.7% YoY in January (because January 2025 was even worse). Jobs are only up 0.1% YoY as of last Friday’s release for February. Except for the near “double-dip” of 2002-03, going back 85 years job gains have *never* been only higher by 0.1% YoY without a recession being either imminent or already in progress.

Weekly Indicators for March 2 - 6 at Seeking Alpha

 

 - by New Deal democrat

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

Unsurprisingly, the big news of the week was the skyrocketing of oil and gas prices. A little surprisingly, the US$ gained as a “safe haven” or perhaps “least dirty shirt” trade.

As usual, clicking over an reading will bring you thoroughly up to date on the economy, and reward me with a penny or two for my efforts.

February jobs report: Main Street lays an egg

 

 - by New Deal democrat

I described last month as “the month the birds came home to roost…. In particular, the *entire* gains over the past year were reduced from 584,000 to 181,000 - an average of only 15,000 jobs gained per month.”

Well, this month the nesting birds, to butcher Edgar Allen Poe, started screeching “recession.” 

Below is my in depth synopsis.

HEADLINES:

• -92,000 jobs lost. Private sector jobs declined -86,000. Government jobs declined -6,000. The three month average declined to a puny +6,000.
• The pattern of downward revisions to previous months continued. December was revised downward by -65,000, and January was revised downward by -4,000, for a net decline of -69,000. 
• The alternate, and more volatile measure in the household report, declined by -185,000 jobs. On a YoY basis, this series *DECLINED* -426,000 jobs, or an average of -35,000 monthly.
• The U3 unemployment rate rose 0.1% to 4.4%, which is where it was in December. 
• The U6 underemployment rate declined -0.1% to 7.9%.
• Further out on the spectrum, those who are not in the labor force but want a job now rose by 166,000.

Leading employment indicators of a slowdown or recession

These are leading sectors for the economy overall, and help us gauge how much the post-pandemic employment boom is shading towards a downturn. These were mainly negative:

• The average manufacturing workweek, one of the 10 components of the Index of Leading Indicators, rose 0.1 hours to 41.5 hours, and is now down only -0.1 hour from its 2021 peak of 41.6 hours.

• Manufacturing jobs decreased by -12,000, the 11th decline in the last 12 months. It is now at a 3+ year low.
• Truck driving, which had briefly rebounded early in 2025, declined another -500.
• Construction jobs declined -11,000.
• Residential construction jobs, which are even more leading, rose 2,400, continuing the trend of stabilizing since last April.
• Goods producing jobs as a whole declined -25,000.. 

• Temporary jobs, which have declined by over -650,000 since late 2022, declined again this month, by -6,500, but remained above their post-pandemic low set last October.
• The number of people unemployed for 5 weeks or fewer rose 153,000.

Wages of non-managerial workers 

• Average Hourly Earnings for Production and Nonsupervisory Personnel increased $.09, or +0.3%, to $32.03, for a YoY gain of +3.7%, its lowest YoY% gain since the pandemic. Nevertheless, this continues to be significantly above the YoY inflation rate.

Aggregate hours and wages: 

• The index of aggregate hours worked for non-managerial workers declined -0.2%, and is up 1.2% YoY, about average for the past two years.

• The index of aggregate payrolls for non-managerial workers rose 0.1%, and is up 4.7% YoY, also about average for the past two years.

Other significant data:

• Professional and business employment declined another -5,000. These tend to be well-paying jobs. While this remains above its October low, it remains lower YoY by -0.4%, which in the past 80+ years - until now - has almost *always* meant recession.
• The employment population ratio declined -0.1% to 59.3%, vs. 61.1% in February 2020.
• The Labor Force Participation Rate declined -0.1% to 62.0% , vs. 63.4% in February 2020.

SUMMARY

As I wrote at the outset, this was a recessionary report, partly because of the monthly change, and partly because of the sideways to downward trend in employment it represents since last spring.

There were some bright spots, including another increase in the average manufacturing work week, and residential construction jobs, which like much other data in the housing sector, shows signs of “green shoots,” i.e., that the bottom is being or has been formed. The U-6 underemployment rate* also declined slightly. And hourly wages continued to increase at a good clip.

But the vast majority of leading and coincident indicators in the report were negative: manufacturing, construction, trucking, and temporary help employment all declined, as did the goods-producing sector as a whole. Total hours work declined, and it is almost certain that once we get the inflation report, we will see that real aggregate payrolls also declined. The employment/population ratio* and labor force participation rate* declined. Revisions continued to be negative, and the unemployment rate*, against expectations, increased slightly.

[*These figures come from the Household Survey and may have been affected by the annual revisions, which were delayed a month and were reported this morning. But they were applied to the January numbers, so the month over month changes should not have been affected.]

Finally, please note that these figures are from *before* the war with Iran and its affect on gas prices started - so be prepared for worse in the next several months.

To conclude by returning to my opening comments about birds coming home to roost: this month Main Street, in the form of jobs, laid an egg.

“New regime” of lower jobless claims continues - a good sign (but for geopolitical idiocy)

 

 - by New Deal democrat

Let’s take our weekly look at jobless claims. As a reminder, I pay attention to these because they are a good short leading barometer of the economy in general, and the jobs market in particular.

And the news this week continued to reflect the “new regime” of lower YoY claims that we have seen for the past 8+ months, as well as the post-pandemic unresolved seasonality in which claims generally rise from the beginning of the year until mid-year. 

Initial claims were unchanged at 213,000, and the four week moving average declilned -4,750 to 215,750. With the typical one week delay, continuing claims rose 46,000 to 1.868 million:

As per usual, the YoY% change is more important for forecating purposes. Here, the news was all positive. Initial claims were down -4.9%, the four week moving average down -4.7%, and continuing claims down -1.3%:

All of which is very inconsistent with any near term onset of a recession.

Finally, as per usual let’s take a look at what this might mean for the unemployment rate in the next several months:

Jobless claims continue to forecast downward pressure on the unemployment rate towards 4.2% or even 4.1%. We’ll find out tomorrow. 

The bottom line is that initial claims, like some of the other short leading data like the improvement in manufacturing, suggest that although it is touch and go, the US economy would be increasingly likely to avoid a recession this year. 

I said “would be” rather than “will” because of the ongoing geopolitical idiocy that is the war with Iran — and I believe “war with Iran” is the appropriate term. This is not a “touch and go” like Venezuela. Yesterday our navy sank an Iranian frigate near Sri Lanka, 1000 miles away from the Persian Gulf. And since we just killed all of the immediate family of Iran’s new Supreme Leader, I do not think he is going to be interested in a cessation of hostilities anytime soon.

Strongly positive ISM services report for February gives the best economically weighted reading for the economy in a year, (but also with a big dose of inflation)

 

 - by New Deal democrat

If Monday’s ISM manufacturing report was good (but with a dose of inflation), today’s ISM services report for February was even better (but also with a dose of inflation). Together they negative the likelihood of an economic downturn in the next several months (geopolitical idiocy aside). 

Let’s take a look. Recall that services represent about 75% or all economic activity, with the goods producing sector the other 25%. Also, typically I average the last three months of each reading to reduce noise. As we will see today, I really don’t need to do that, because the message is pretty clear. In all the graphs below, the services reading is in blue, the equivalent manufacturing one in grey.

Let’s start with what has been the most moribund reading — that on employment. On Monday, we saw that manufacturing employment got “less bad,” improving to 48.8. This morning’s services employment diffusion reading was 51.8:

There is strong evidence that there was a bottom in employment last July, and an improving trend since. The three month average in services has been 51.3, i.e., weakly positive. The economically weighted average vs. manufacturing employment’s 47.2 three month average is 50.3 - just barely positive, but nevertheless the best reading in a year.

As similar pattern shows up in the headline number, which for services came in at 56.1 for February. The three month average is 54.6:

The bottom in the headline number was a little before employment, in the May through July period. The economically weighted three month average including manufacturing is 53.7, solidly if not sharply positive.

The more leading new orders component improved to a strong 58.6. The three month average also increased to 56.1:

New orders bottomed in the March through May period of last year. Their economically weighted three month average including manufacturing increased to 55.4, a very positive number — and the most positive number in over 3 years. Needless to say, this is an excellent development for the economy.

But nothing is perfect, and the problem child in both ISM indexes is inflation, in the form of prices paid. The services component did decline to 63.0, bringing the three month average down to 64.9 (which is still a very concerning number for prices), the lowest in over a year:

The economically weighted three month average of prices paid is 63.2, suggesting inflation remains entrenched in the broad economy.

Two final points: first, the ISM services reports have been wildly divergent from the regional Fed services indexes, which have been very negative for months, most recently averaging -10. One of these is giving a false signal.

The second is that the ISM services reports *are* confirming what I have been seeing for months in the Redbook weekly retail spending report, which has not just been positive, but increasingly so in the latter part of 2025 into this year:

Again, I suspect the very strong retail spending data, as well as the services diffusion indexes reported above, has almost everything to do with a wealth effect generated by stock price increases in the past year, which in turn were the result of AI data center related spending. 

In any event, mark down this morning’s services report in the solidly positive column, and hope it does not get derailed by idiocy in the Middle East.