by Calculated Risk on 9/08/2025 11:58:00 AM

Early in February, I expressed my "increasing concern" about the negative economic impact of "executive / fiscal policy errors", however, I concluded that post by noting that I was not currently on recession watch.

In early April, I went on recession watch, but I'm still not yet predicting a recession for several reasons: the U.S. economy is very resilient and was on solid footing at the beginning of the year, and perhaps the tariffs are not enough to topple the economy.

In the short term, it is mostly trade policy that will negatively impact the economy.  However, there other aspects of policy that bear watching - especially immigration.

Here is some of the data I'm watching.  

Housing:  Housing is the basis of one of my favorite models for business cycle forecasting.

This graph shows the YoY change in New Home Sales from the Census Bureau.  Currently new home sales (based on 3-month average of NSA data) are down 3% year-over-year.

Usually when the YoY change in New Home Sales falls about 20%, a recession will follow.  An exception for this data series was the mid '60s when the Vietnam buildup kept the economy out of recession.   Another exception was in late 2021 - we saw a significant YoY decline in new home sales related to the pandemic and the surge in new home sales in the second half of 2020.  I ignored that downturn as a pandemic distortion.  Also note that the sharp decline in 2010 was related to the housing tax credit policy in 2009 - and was just a continuation of the housing bust.

The YoY change in new home sales in late 2022 and early 2023 suggested a possible recession.  I was able to look past the pandemic distortion and was able to predict a pickup in new home sales due to the low level of existing home inventory and because homebuilders could offer mortgage incentives that would somewhat offset the sharp increase in mortgage rates.

There are no special circumstances now, and if this measure falls to off 20% a recession seems likely.

Yield Curve: The yield curve is a commonly used leading indicator.  I dismissed it when the yield curve inverted in 2019 and again in 2022. Both times dismissing the yield curve was correct (the recession in 2020 was obviously due to the pandemic, so we will never know if the yield curve failed to predict a recession in 2019).

Here is a graph of 10-Year Treasury Constant Maturity Minus 2-Year Treasury Constant Maturity from FRED since 1976.

 

The yield curve reverted to normal a year ago and is currently positive at 0.59.  If this inverts, this might suggest a recession is coming.

Click here for interactive graph at FRED.

Heavy Truck (and Vehicle Sales): Another indicator I like to use is heavy truck sales.  This graph shows heavy truck sales since 1967 using data from the BEA. The dashed line is the August 2025 seasonally adjusted annual sales rate (SAAR). Note: "Heavy trucks - trucks more than 14,000 pounds gross vehicle weight."

Heavy truck sales were at 422 thousand SAAR in August, down from 442 thousand in July, and down 15.7% from 501 thousand SAAR in August 2024.

Usually, heavy truck sales decline sharply prior to a recession and sales have been a little soft recently.

Vehicle sales were over 17 million SAAR in March and April as consumers rushed to "beat the tariffs".  Then sales were depressed in May and June.

This graph shows light vehicle sales since the BEA started keeping data in 1967.   This is more of a concurrent indicator than heavy trucks. 

Light vehicle sales in August (16.07 million SAAR) were down 2.9% from the sales rate in July, and up 6.2% from August 2024. (Note: Sales in August were boosted due to the expiring EV credit).

Unemployment: Two other concurrent indicators are the unemployment rate (using the "Sahm Rule") and weekly unemployment claims.

Here is a graph of the Sahm rule from FRED since 1959.  The Sahm rule is a measure of changes in the unemployment rate. It compares the three-month moving average of the unemployment rate (U3) to the minimum of the three-month averages from the previous 12 months.

The Sahm Rule was at 0.23 percentage points in August. 

 If this increases to 0.5 it will suggest a possible recession.  

Note that this increased to 0.56 in August 2024, but Dr. Sahm (and I) argued this was not indicating a recession.

And weekly unemployment claims always rise sharply at the beginning of a recession (other events - like hurricane Katrina - can cause a temporary spike in weekly claims).

As I noted earlier, I'm not sure how to estimate the economic damage caused by these tariffs and other policies.   The economy is clearly slowing due to policy.

There are also boycotts of U.S. goods and less international tourism based on both the tariffs and the inflammatory rhetoric of the current administration.  

None of these indicators are suggesting a recession.  For now, I'll focus on the leading indicators (especially housing) and I'll update this post monthly while I'm on recession watch.  

by Calculated Risk on 9/08/2025 08:11:00 AM

Altos reports that active single-family inventory was down 1.7% week-over-week.  Inventory usually starts to decline in the fall, and then declines sharply during the holiday season.

Inventory is now up 35.6% from the seasonal bottom in January.   Usually, inventory is up about 20.5% from the seasonal low by this week in the year.   So, 2025 saw a larger than normal increase in inventory.

The first graph shows the seasonal pattern for active single-family inventory since 2015.

Click on graph for larger image.

The red line is for 2025.  The black line is for 2019.  

Inventory was up 20.4% compared to the same week in 2024 (last week it was up 22.4%), and down 10.6% compared to the same week in 2019 (last week it was down 10.3%). 

Inventory started 2025 down 22% compared to 2019.  Inventory has closed more than half of that gap, and it appears inventory will still be below 2019 levels at the end of 2025.

This second inventory graph is courtesy of Altos Research.

As of September 5th, inventory was at 847 thousand (7-day average), compared to 861 thousand the prior week. 

Mike Simonsen discusses this data and much more regularly on YouTube

by Calculated Risk on 9/07/2025 07:10:00 PM

Weekend:
• Schedule for Week of September 7, 2025

Monday:
• No major economic releases scheduled.

From CNBC: Pre-Market Data and Bloomberg futures S&P 500 are up 9 and DOW futures are up 82 (fair value).

Oil prices were up over the last week with WTI futures at $61.92 per barrel and Brent at $65.60 per barrel. A year ago, WTI was at $69, and Brent was at $73 - so WTI oil prices are down about 10% year-over-year.

Here is a graph from Gasbuddy.com for nationwide gasoline prices. Nationally prices are at $3.16 per gallon. A year ago, prices were at $3.23 per gallon, so gasoline prices are down $0.07 year-over-year.

by Calculated Risk on 9/07/2025 08:09:00 AM

From the Association of American Railroads (AAR) AAR Data Center. Graph and excerpts reprinted with permission.

The AAR Freight Rail Index (FRI), which measures seasonally adjusted month to-month rail intermodal shipments plus carloads excluding coal and grain, fell 0.5% in August 2025 from July 2025, its fourth decline in the past five months. Still, the index remains relatively strong: since 2008, it’s been higher than it was in August less than 15% of the time.
emphasis added

Rail intermodal volumes are closely tied to port activity (especially in the west) and the consumer side of the economy. In August, U.S. rail intermodal shipments were up 0.5% over last year. Average weekly intermodal volume in August 2025 was 284,316 containers and trailers, the most for any month since May 2021 and the most for August since 2018.
...
Meanwhile, year-over-year total U.S. rail carloads rose 0.7% in August 2025 over August 2024, marking six consecutive monthly gains. Eleven of the 20 major carload categories tracked by the AAR saw gains in August, the sixth straight month in which at least half the categories saw increases. Total carloads averaged 230,184 per week in August 2025, the most for any month since October 2022. In 2025 through August, total carloads were up 2.5%, or nearly 192,000 carloads, over last year.

by Calculated Risk on 9/06/2025 02:11:00 PM

At the Calculated Risk Real Estate Newsletter this week:

Click on graph for larger image.

• Q2 Update: Delinquencies, Foreclosures and REO

• Freddie Mac House Price Index Declined in July; Up 1.4% Year-over-Year

• What will happen with House Prices?

• Asking Rents Mostly Unchanged Year-over-year

This is usually published 4 to 6 times a week and provides more in-depth analysis of the housing market.

by Calculated Risk on 9/06/2025 08:11:00 AM

The key economic report this week is the August Consumer Price Index (CPI).

The BLS will release the preliminary employment benchmark revision on Tuesday.

----- Monday, September 8th -----
No major economic releases scheduled.

----- Tuesday, September 9th -----
6:00 AM: NFIB Small Business Optimism Index for August.

10:00 AM: the Bureau of Labor Statistics (BLS) will release the Current Employment Statistics Preliminary Benchmark (National) for March 2025.

----- Wednesday, September 10th -----
7:00 AM ET: The Mortgage Bankers Association (MBA) will release the results for the mortgage purchase applications index.

8:30 AM: The Producer Price Index for August from the BLS. The consensus is for a 0.3% increase in PPI, and a 0.3% increase in core PPI.

----- Thursday, September 11th -----
8:30 AM: The initial weekly unemployment claims report will be released. The consensus is for initial claims to increase to 240 thousand from 237 thousand last week.

8:30 AM: The Consumer Price Index for August from the BLS. The consensus is for a 0.3% increase in CPI, and a 0.3% increase in core CPI.  The consensus is for CPI to be up 2.9% year-over-year (up from 2.7% in July) and core CPI to be up 3.1% YoY (unchanged from 3.1% in July).

12:00 PM: Q2 Flow of Funds Accounts of the United States from the Federal Reserve.

----- Friday, September 12th -----
10:00 AM: University of Michigan's Consumer sentiment index (Preliminary for September).

by Calculated Risk on 9/05/2025 03:58:00 PM

Hotel occupancy was weak over the summer months, likely due to less international tourism.  The fall months are mostly domestic travel.

From STR: U.S. hotel results for week ending 30 August

The U.S. hotel industry reported mostly positive year-over-year comparisons, according to CoStar’s latest data through 30 August. ...

24-30 August 2025 (percentage change from comparable week in 2024):

• Occupancy: 63.4% (-0.8%)
• Average daily rate (ADR): US$155.87 (+1.0%)
• Revenue per available room (RevPAR): US$98.88 (+0.2%)
emphasis added

The following graph shows the seasonal pattern for the hotel occupancy rate using the four-week average.

Click on graph for larger image.

The red line is for 2025, blue is the median, and dashed light blue is for 2024.  Dashed purple is for 2018, the record year for hotel occupancy. 

The 4-week average of the occupancy rate is tracking behind last year and close to the median rate for the period 2000 through 2024 (Blue).

Note: Y-axis doesn't start at zero to better show the seasonal change.

The 4-week average will decrease seasonally until the Fall travel period.

On a year-to-date basis, the only worse years for occupancy over the last 25 years were pandemic or recession years.

by Calculated Risk on 9/05/2025 03:11:00 PM

From BofA:

[O]ur 3Q GDP tracking has moved up a tenth to 1.6%. Also, our 2Q GDP tracking is down a tenth to 3.2% from the second estiamte of 2Q GDP by the BEA. [September 5th comment]
emphasis added

From Goldman:

We lowered our Q3 GDP tracking estimate by 0.1pp to +1.6% (quarter-over-quarter annualized). We left our Q3 domestic final sales estimate unchanged at +0.7%. [September 4th estimate]

And from the Atlanta Fed:

The GDPNow model estimate for real GDP growth (seasonally adjusted annual rate) in the third quarter of 2025 is 3.0 percent on September 4, unchanged from September 2 after rounding. After recent releases from the US Bureau of Economic Analysis, the US Census Bureau, and the Institute for Supply Management, increases in the nowcasts of real personal consumption expenditures growth and real gross private domestic investment growth from 1.7 percent and 5.3 percent, respectively, to 2.1 percent and 6.0 percent, were more than offset by a decline in the nowcast in the contribution on net exports to GDP growth from 0.69 percentage points to 0.28 percentage points. [September 4th estimate]

by Calculated Risk on 9/05/2025 12:05:00 PM

Today, in the Real Estate Newsletter: What will happen with House Prices?

Brief excerpt:

Almost every day a journalist or an analyst asks me what will happen with house prices.

Every cycle is different, and usually I focus on inventory, sales, and months-of-supply to answer this question.

However, there have been significant policy changes this year, especially with trade and immigration. This has led to a period of rising inflation, and a weakening employment situation (rising unemployment rate). A period of stagflation.

These are powerful forces for the economy and housing.
...
But what is the impact of rising unemployment?

The following graph shows the year-over-year in the Case-Shiller National Index versus the Sahm rule (from economist Claudia Sahm). The Sahm rule is a measure of changes in the unemployment rate. It compares the three-month moving average of the unemployment rate (U3) to the minimum of the three-month averages from the previous 12 months.

In general, a rising unemployment rate corresponds to weaker house prices. Of course, correlation does not imply causation. And there are exceptions - like at the onset of the pandemic when the unemployment increased sharply, but house prices took off (mortgage rates fell sharply and most potential homebuyers stayed employed).

This is much more in the article.

by Calculated Risk on 9/05/2025 09:11:00 AM

The headline jobs number in the August employment report was below expectations and June and July payrolls were revised down by 21,000 combined.  A weak report. The participation rate increased, the employment population ratio was unchanged, and the unemployment rate was increased to 4.3%.

Earlier: August Employment Report: 22 thousand Jobs, 4.3% Unemployment Rate

Prime (25 to 54 Years Old) Participation

Since the overall participation rate is impacted by both cyclical (recession) and demographic (aging population, younger people staying in school) reasons, here is the employment-population ratio for the key working age group: 25 to 54 years old.

The 25 to 54 years old participation rate increased in August to 83.7% from 83.4% in July.

The 25 to 54 employment population ratio increased to 80.7% from 80.4% the previous month.

Both are down from the recent peaks, but still near the highest level this millennium.

Average Hourly Wages

The graph shows the nominal year-over-year change in "Average Hourly Earnings" for all private employees from the Current Employment Statistics (CES).  

There was a huge increase at the beginning of the pandemic as lower paid employees were let go, and then the pandemic related spike reversed a year later.

Wage growth has trended down after peaking at 5.9% YoY in March 2022 and was at 3.7% YoY in August, down from 3.9% YoY in July.

Part Time for Economic Reasons

From the BLS report:

"The number of people employed part time for economic reasons, at 4.7 million, changed little in August. These individuals would have preferred full-time employment but were working part time because their hours had been reduced or they were unable to find full-time jobs."

The number of persons working part time for economic reasons increased in August to 4.75 million from 4.68 million in July.  This is above the pre-pandemic levels.

These workers are included in the alternate measure of labor underutilization (U-6) that increased to 8.1% from 7.9% in the previous month. This is down from the record high in April 2020 of 22.9% and up from the lowest level on record (seasonally adjusted) in December 2022 (6.6%). (This series started in 1994). This measure is above the 7.0% level in February 2020 (pre-pandemic).

Unemployed over 26 Weeks

This graph shows the number of workers unemployed for 27 weeks or more.

According to the BLS, there are 1.93 million workers who have been unemployed for more than 26 weeks and still want a job, up from 1.83 million the previous month.

This is down from post-pandemic high of 4.171 million, and up from the recent low of 1.056 million.

This is above pre-pandemic levels.

Job Streak

The recent streak of positive job reports ended in May (Payrolls were negative in June).  The recent streak finished in 2nd place of the longest job streaks in US history (since 1939).

Headline Jobs, Top 10 Streaks
	Year Ended	Streak, Months
1	2020	113
2	2025	531
3	1990	48
4	2007	46
5	1979	45
6 tie	1943	33
6 tie	1986	33
6 tie	2000	33
9	1967	29
10	1995	25
1Recent Streak Ended in May

Summary:

The headline jobs number in the August employment report was below expectations and June and July payrolls were revised down by 21,000 combined.  The unemployment rate increased to 4.3%.

This was a weak employment report.