The 4 shocks jolting the US economy towards recession

 

 - by New Deal democrat

This week the normal empty period for new data after the monthly jobs report is of course compounded by the government shutdown. There is no noteworthy new private data coming out until Thursday, so don’t be surprised if I play hooky tomorrow.

But today let me follow up on some macro analysis. 

A few months ago I pointed out that, going back 60 years, every US recession has been associated with a shock to the system. In other words, the normal progression of interest rates, building, sales, income, and jobs have caused waxing and waning in sectors of the economy, but haven’t by themselves been enough to cause it to contract. For that, you needed an economy that was vulnerable, and a shock administered to that vulnerable economy.

Sometimes it has been an oil shock (1974, 1979, 1991, and partly 2007). Sometimes it has been an interest rate shock (1969, 1974, 1979, 1981). Sometimes it has been a financial shock (2001, 2008). And of course in 2020, it was the Giant Flaming Meteor of Death a/k/a COVID. I don’t mean to suggest that these recessions have been monocausal, just that a shock to the system has always been a part of the equation.

Going into 2025, the US economy was certainly vulnerable. High interest rates had taken a toll on the housing market. Employment, especially in manufacturing, was waning. The post-COVID spike in vehicle prices, repairs, and insurance, was still making its way through the system.

But this year, at least 4 shocks have been administered to the system, some (likely) transitory, but some more chronic:

1. Tariffs

2. Medical cost increases.

3. The suspension of food stamp benefits.

4. AI-related layoffs.

Let’s briefly discuss each of these in turn.

1. Since “Liberation Day” at the beginning of April, almost $200 Billion in tariffs have been imposed, an increase of over $100 Billion since 2024. In Q3, this was roughly 3x the amount collected before this year, and 6x the amount collected just before the 1st T—-p Administration:

The evidence so far is that in the aggregate companies are bearing about half of the increased burden, and are passing on about half of the burden to customers. In October alone, $60 Billion in tariffs were collected. If customers paid half of that by way of price increases, that amounts to $50 for every man, woman, and child in the US. That may not be a significant burden for affluent households, but for lower income households it is going to have a real impact.

2. Medical cost increases. As part of the Big Bad Billionaire Bust-out Bill, Obamacare subsidies were ended. Since about 90% of all Obamacare enrollees make use of these subsidies, that means that 22 million people are directly affected:

These households are either going to have to pay (in some cases astronomical) increases for the cost of coverage, or drop medical coverage completely. If they pay, that is another shock to the household budget; if they don’t, it affects the coverage, and the profits of insurers and health care providers. The proposed ending of the government shutdown will make this effective immediately.

3. The suspension of food stamp benefits. As has been widely reported this month, about 44 million people in the US make use of food stamps:

To the extent meager household income has to be diverted to buy that food, needless to say it is not available for any other kind of consumption (including payment or rent or utilities). While the ending of the government shutdown should mean the resumption in payments, keep in mind that the proposed continuing resolution only lasts through January. In other words, in 75 days we are right back here, and the successful use of SNAP benefits as a fiscal weapon this time means it will surely be implemented again then.

4. Finally, as I noted on Friday, layoffs as counted by Challenger Gray were at their second highest monthly level since the COVID lockdowns, and the highest October level in a quarter of a century. Many of these were concentrated in tech companies, as live human beings were replaced (or, “replaced”) by AI algorithms:

While this hasn’t shown up in any appreciable increase in new jobless claims, as I noted yesterday the number of continuing claims is close to its highest level in over three years. If this continues, it seems very likely we’ll see it show up in an increase in initial claims as well.

So there you have it: four somewhat independent shocks to the US economic system that individually or collectively might be enough to push it over the edge into a recession. In that regard, let me re-post this graph of the likely combined impacts of tariffs and the Big Bad Billionaire Bust-out Bill on household incomes by percentile:

This only includes the first two of the 4 shocks described above.

Needless to say, the shutdown of official government economic data could hardly have come at a worse time. While we have reasonable proxies for the employment situation, alternative data for income, sales, and spending is spotty. In particular, how has spending on durable goods by companies and consumers held up during this period? Have they cut back, or have stock market gains and the wealth effect so generated more than overbalanced the above described shocks. We simply don’t know. Assuming government data releases resume shortly, I will be particularly focused on that information.

Tabulated state level jobless claims continue neutral trend

 

 - by New Deal democrat

As I have done since the beginning of the government shutdown, the unadjusted number of initial and continuing claims can be calculated based on reporting by the States, plus DC, and Puerto Rico. Then, by applying the same adjustment as was used for the same week last year, the seasonally adjusted number can also be estimated closely as well.

Indeed, since my forecasting method relies on the YoY% changes, it is almost never an affected by that seasonality. 

So tabulated, for the week ending November 2, unadjusted initial claims totaled 216,238 vs. 212,743 in 2024, an increase of 1.6%.  

Last year this week the seasonal multiplier was *1.0388. Applying it gives us an estimated seasonally adjusted number of 225,000.

We can similarly calculate the four week moving average, since the last four weeks of claims were 224,000, 230,000, and 220,000, as well as this week’s 225,000. That gives us an average of 224,750, which is -2,000, or -0.9% lower than the number of 226,7500 one year ago. Of note, this is the last week in which the 2024 comparison will be affected by the hurricanes that temporarily depressed claims for several weeks in October. Excluding that week, the average of the three last weeks this year is slightly higher.

Using the same methodology, unadjusted continuing claims for the week ending October 25 totaled 1,711,947 vs. 1,646,920 last year, an increase of 3.9%.

The seasonal adjustment for the applicable week last year was *1.14152. Applying it gives us an estimate of 1.954 million continuing claims, or -7,000 lower than one week ago. Still, continuing claims during the government shutdown have all been close to their highest levels since 2021, which was 1.968 million this past July. 

For graphic comparison, here are initial claims (blue), the four week average (red), and continuing claims (gold) all normed to 0 as of this week’s tabulation, compared with their readings in the past two years before the shutdown:

As with the past several weeks, absent hurricane distortions this continues the general neutral trend of initial and continuing claims, higher than one year ago but much less than 10% higher, forecasting a weakly expanding economy for the next several months.

Weekly Indicators for October 3 - 7 at Seeking Alpha

 

 - by New Deal democrat

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

This was the week that the 800 pound gorillas of Wall Street reported Q3 earnings, and - to use a World Series type metaphor - hit it out of the park. 

Meanwhile down in gruntland, the amount of goods being moved from ports to markets hit a major air pocket.

More signs of a “K shaped”, or bifurcated economy where Wall Street titans are doing well, fueling upper income spending, while down below there are signs of exhaustion.

As usual, clicking over and reading will bring you up to the virtual moment as to the state of the economy - with data almost entirely unaffected by the government shutdown - and reward me with a penny or two of lunch money for my efforts in collecting and organizing it for you.

October employment situation: stagnant hiring, increased firing, continued wage growth

 

 - by New Deal democrat

On a normal first Friday of the month, I would be crunching the official jobs report data to try to provide not just a coincident report on the jobs market, but also to focus on the leading indicators within that report, such as the manufacturing and construction sectors and also aggregate real payrolls. For the second month in a row, I can’t do that. But what I can do is aggregate some of the most reliable alternate sources for the state of the labor market; and fortunately, there is no big divergence among those: they all point to a stagnant but not meaningfully contracting jobs sector.

Let me start with the number of jobs, and then follow with alternatives to the unemployment rate. And further, let me start with the three best sources: the ISM reports, the regional Fed reports, and ADP.

To begin with, the ISM reports are diffusion indexes. They don’t report the number of gains or losses, but whether more firms than not are hiring vs. firing. Any number below 50 means more firms are letting people go than hiring them, and ss I reported Wednesday, ISM services showed slight contraction, at 48.2. The ISM manufacturing report had previously also shown contraction for October, at 46.0:

Since services account for about 75% of all jobs, the economically weighted number is 47.6 for the month. This would translate to a continued downturn in jobs in the goods producing sector, to a virtual standstill or even small losses in the service-providing sector as well.

The next source is the 5 regional Fed reports from NY, Philly, Richmond VA, Kansas City, and Dallas. These are also diffusion indexes, with the balance point at zero, so a positive number is expansion, a negative one contraction. As I reported last week, here are the numbers from the five manufacturing reports (1st line) and services reports (2nd line):

NY 6.2; Philly 4.6; Richmond -10; Kansas City 1; Dallas 2.0; AVERAGE 0.8

NY -5.2; Philly -0.5; Richmond 0; Kansas City -4; Dallas -5.8; AVERAGE -3.2

Unlike the ISM manufacturing report, the regional Fed reports show slight gains in the manuacturing employment sector; but like the ISM services report, significant contraction in the services employment sector.

Next, as has been widely reported, ADP indicated that 42,000 private sector jobs were added in October. As the below graph by Prof. Jason Furman shows, the average of the last three months is approximately 0:

Bank of America’s internal data also showed further deceleration over the past two months from the last officially reported numbers:

A gain of only 0.5% for the entire last 12 month period suggests no monthly growth at all in either September or October.

Finally, although I have no idea whether this new source is reliable or not, Revelio Labs reported that in October -9,100 jobs were lost:

Weakness was also shown in new job postings by Indeed, which declined further in October. Note that this source has closely matched the job openings data from the JOLTS series:

Another important metric in the monthly jobs report is wages. These are also covered by some of the regional Fed reports as well as ADP.

Here are the manufacturing (1st line) and services (2nd line) diffusion indexes from the regional Fed reports for October:

Manufacturing: NY n/a; Philly n/a; Richmond 15; Kansas City n/a; Dallas 14.2; AVERAGE 14.6

Services: NY 25.9; Philly 38.3; Richmond 17; Kansas City 21; Dallas 10.7; AVERAGE 22.6

The economically weighted average from both indexes is 16.5, indicating that wage increases continue at a strong pace.

ADP similarly reported continued strong wage growth, at 4.5% YoY for job stayers, and 6.7% for job switchers:

Finally, let’s take a look at unemployment measures.

As I often say, jobless claims lead the unemployment rate. Initial jobless claims are noisier but more leading; when combined with continued claims they carry more signal but lead only slightly. Here is what both measures looked like compared with the unemployment rate as of the last official reports:

Since the shutdown, initial claims have varied between 220,000 and 230,000, and on a four week moving average basis have been slightly lower than one year ago, while continuing claims have been in the 1.930 to 1.960 million range, close to the top of their readings in the past three years. This suggests that the unemployment rate would be no lower than, and likely slightly above its range from one year ago, which was 4.1%-4.2%:

This would put this month’s likely unemployment rate at between 4.2% to 4.4%.

Further, as was widely reported yesterday, Challenger Gray indicated that there were 153,000 job cuts in October, the highest for this month in several decades, and with one exception the highest since the pandemic lockdowns:

And Bank of America indicated that unemployment insurance checks directly deposited into its account, while slightly down from September, were 10% higher YoY - an even bigger YoY increase than official continuing claims:

To summarize, the best alternative jobs data we have for October are almost all in accord. There were roughly no job gains at all, +/- about 40,000. Meanwhile wages continued to grow at a relatively fast pace, in accord with the official data from earlier this year. And the unemployment rate was steady to slightly higher compared with earlier.

Long leading indicator Senior Loan Officer Survey for Q3 was neutral to slightly positive

 

 - 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. Fortunately, since it is reported by the Federal Reserve, it is unaffected by the government shutdown.

The one downside is that the information is only reported Quarterly, and with a one a one month lag. So the Q3 update was only reported on Tuesday.

There are two series that have a long enough record to give us a lot of information. The first is whether banks are tightening or loosening standards. Since tightening is shown as an increase, this is one of those series where higher means worse. 

Historically banks are still tightening, but less so as an expansion begins. As it progresses, they ease lending standards; but  become more cautious well in advance of an ensuing recession. In early 2024, the sharp decline in the percentage of banks that were tightening standards indicated an expansion that was well established.

Interestingly, although at no point during this expansion have banks on net eased, this year the relative progression of banks that moving in that direction has stalled:

This is a neutral reading. It does not look like what has historically happened closer in time to a recession; but on the other hand it does not indicate a strengthening expansion.

In the second series, confusingly, higher does mean better. And in Q3, each measure was slightly positive. Further, while the series are very noisy, there was an unmistakable improving trend in 2023-24, which may be plateauing this year:

In any event, the Q3 result indicates business confidence in expansionary plans.

The upshot is that this long leading credit indicator is presently neutral to slightly positive.