The economy is not falling apart. But important parts of it are barely moving.
The major exception is investment. Business spending on equipment and intellectual property remains strong, with the AI buildout clearly part of that story. And private domestic demand more broadly has not stalled: real final sales to private domestic purchasers rose at a healthy 3.9% annualized rate in the second quarter.
So the weakness isn’t everywhere. It is showing up most clearly in hiring, labor-force growth and household purchasing power. At the same time, households have been saving less: the personal saving rate fell from 4.5% in January to 2.7% in June. That can support spending today, but it leaves less cushion if income growth weakens.
That distinction matters.
Hiring has nearly stopped. Payroll growth has averaged just 34,000 jobs per month over the past year. From 2016 through 2019, payrolls grew by about 186,000 jobs per month on average. We are running at less than one-fifth the pre-pandemic pace.
But the supply of workers is shrinking too. The labor-force participation rate fell to 61.4% in July, its lowest reading since February 2021, and is down 0.7 percentage points just since January.
Inflation looked better in July, but falling energy prices did much of the work just as oil was moving sharply higher. Consumers are still spending, but once you adjust for prices, retail spending is mostly treading water.
Put it together and you get an economy caught in an uncomfortable equilibrium: businesses are hesitant to hire, consumers don’t have much room to accelerate, and supply constraints are keeping inflation too high for the Federal Reserve to come to the rescue.
Nobody has cracked. Nobody is really moving forward either.
The labor market is stuck on both sides
Start with jobs.
Payroll employment fell by 23,000 in July, and job growth has averaged only about 34,000 per month over the past year. The June JOLTS report told essentially the same story from another angle: the hiring rate was just 3.4%, while quits remained subdued at 2.0%. Companies aren’t firing workers in large numbers. They just aren’t hiring many new ones.
Ordinarily, that kind of collapse in hiring would produce considerably more unemployment.
It hasn’t, in part because the supply of workers is weakening at the same time.
The labor-force participation rate was 61.4% in July, down 0.7 percentage points since January. Meanwhile, one of the biggest sources of labor-force growth in recent years – immigration – has slowed dramatically. Census estimates show net international migration peaking at 2.7 million in 2024, falling to 1.3 million in 2025, with its Vintage 2025 estimates projecting roughly 321,000 in 2026 if the current trends continue.
You can already see it in the labor-market numbers.
The foreign-born civilian labor force was about 550,000 smaller in July than a year earlier. But this isn’t only an immigration story. The native-born labor force fell by roughly 861,000 over the same period even though the native-born civilian noninstitutional population age 16 and older grew by nearly 1.9 million. Falling participation is shrinking labor supply from the inside too. These nativity figures are not seasonally adjusted, so the cleanest comparison is July to July.
That helps explain one of the strangest features of this labor market: demand for workers is weak, but workers aren’t exactly plentiful either. And the lack of opportunity can cause some job seekers to give up looking for a job.
Immigration matters for more than the unemployment rate
This is where the immigration debate runs directly into economics.
Immigrants have historically participated in the labor market at higher rates than the native born. The Census Bureau’s 2022 American Community Survey put labor-force participation at 66.9% for the foreign born versus 62.9% for the native born. The difference was especially large among men: 76.8% versus 66.8%. Foreign-born women actually participated at a slightly lower rate than native-born women, so demographics and the composition of the immigrant population matter enormously.
The same pattern remains visible today. In July, foreign-born participation was 65.5% versus 61.1% among the native born. For men, it was 76.1% versus 65.7%.
So when immigration slows sharply, the economic effect isn’t simply fewer people. It means slower growth in a population that has relatively strong attachment to the labor market.
There is also a productivity channel.
The economics literature doesn’t say that every additional immigrant immediately raises productivity, or that there are no distributional effects. Workers with similar skills can compete with one another, and the Congressional Budget Office estimates that the recent immigration surge initially put some downward pressure on wage growth for workers with lower educational attainment.
But immigration also expands the labor force, encourages investment, allows workers to specialize in different tasks and can raise innovation. One influential estimate from economist Giovanni Peri, using U.S. state data, found that a 1% increase in employment attributable to immigration was associated with roughly a 0.5% increase in income per worker, with productivity rather than displacement doing much of the work. Immigrants and native-born workers tend to specialize in different tasks, allowing native workers to shift toward jobs that rely more heavily on communication and language skills.
At the higher end of the skill distribution, researchers Jennifer Hunt and Marjolaine Gauthier-Loiselle found an especially strong innovation channel. Immigrants patent at about twice the native rate, largely because they are disproportionately represented in science and engineering, and states that experienced larger increases in skilled immigration subsequently generated more patents per capita.
None of that means less immigration automatically causes productivity to fall next quarter.
It means something more basic: reducing the growth of the workforce reduces one source of potential economic growth, while also removing some of the specialization, investment and innovation channels through which immigration can raise productivity over time.
And that brings us to a much bigger question.
What productivity boom?
As economist Neil Dutta has pointed out recently: what productivity boom?
We hear constantly about the AI productivity boom. The investment boom is obvious. The productivity boom is much harder to find in the aggregate data.
Look at the nearby figure from the Federal Reserve Bank of San Francisco.
Total factor productivity (TFP) is the part of output growth left after accounting for measured labor and capital inputs. We often interpret that residual as technological progress. But over short periods, ordinary TFP is noisy and highly cyclical. Factories run harder when demand is strong. Workers put in more effort. Equipment sits idle when the economy slows. All of that can move measured productivity without anyone inventing anything.
That’s why the San Francisco Fed also publishes a utilization-adjusted series. It attempts to remove changes in labor effort and capital utilization and describes the result as an improvement over more “naive” TFP measures for identifying technological change at high frequency.
And here’s the striking part.
Over the four quarters ending in the second quarter of 2026, ordinary TFP increased 1.1% but utilization-adjusted TFP fell 0.42%.
In the second quarter alone, ordinary TFP fell at a 0.34% annualized rate while utilization-adjusted TFP fell 2.19%. Meanwhile, capital input grew at a 3.19% annualized rate.
In other words, companies are unquestionably spending enormous amounts of money on technology and capital. But we do not yet have evidence in this measure of an economy-wide AI-driven productivity boom.
That doesn’t mean it won’t come. Technological revolutions often require large investments in equipment, software, organizational change and worker training before their productivity effects become visible. It does mean we should distinguish an AI investment boom from an AI productivity boom. They are not the same thing.
And if those productivity gains fail to materialize, it would eventually challenge valuations that depend on today’s extraordinary AI investment producing extraordinary earnings growth.
Inflation looked better. But that could be short-lived.
Productivity is one of the few ways the economy can grow faster without creating more inflation. But so far, inflation remains a major problem.
July CPI was genuinely encouraging. Consumer prices rose just 0.1% during the month, headline inflation eased to 3.4% from 3.5%, and core inflation slowed to 2.5%. Services excluding energy are now running at only about a 2% annualized pace over the past three months, a meaningful slowdown in momentum.
But the timing was unusually favorable. Energy prices fell 1.5% in July and gasoline fell 2.9%, even though crude oil surged late in the month. Pump prices lag crude prices, meaning much of the latest increase in oil prices simply arrived too late to show up fully in the July CPI.
So July’s inflation report told us what happened to consumer prices largely before the latest energy shock worked its way through.
Producer prices added another complication.
Headline PPI was flat in July, and its annual rate fell to 4.7%. Again, energy did a lot of the work: producer energy prices fell 3.1% and gasoline fell 5.7%.
Strip that out and the picture gets less comfortable. Producer prices excluding food, energy and trade services rose 0.4%, while final-demand services excluding trade, transportation and warehousing rose 0.6%. More importantly for the Fed, several producer-price categories that feed directly into PCE – including financial services, health care and airfares – firmed. Portfolio-management prices alone jumped 6.5%.
Our backtested nowcast, which combines core CPI and PPI core services, currently points to July core PCE of roughly 0.24% for the month, or about 3% annualized.
That’s not moving in the direction the Fed needs.
The consumer hasn’t cracked. That’s different from being strong.
Then came Friday’s retail report.
Retail and food-service sales fell 0.58% in July, the first monthly decline in months. But the three-month trend is still growing at about a 2.4% annualized pace, so calling the consumer broken would go too far.
The more important number is what happens after adjusting for prices.
Retail sales are up 5.0% from a year ago in dollars. Using headline CPI as a rough benchmark for inflation, that leaves real growth of only around 1½% to 2%. The retail control group is also growing only modestly after adjusting for prices.
Put differently, roughly two-thirds of the headline increase in retail spending over the past year reflects higher prices rather than more stuff being purchased. That’s a back-of-the-envelope comparison rather than an official real-retail-sales measure, but the basic point holds.
That is resilience, but it isn’t acceleration.
And underneath the aggregate number, the split keeps widening. Lower-income households have been pulling back while higher-income households continue to spend. Even the high end is looking harder for value – trading down on retailers and increasingly embracing private labels rather than simply spending without regard to price.
Wages tell a similar story. Average hourly earnings are up about 3.2% over the past year while consumer prices are up 3.4%.
Wage gains, at least by that measure, are no longer keeping pace with prices.
Businesses are dealing with the other side of the same squeeze. Their input costs are elevated, their customers are increasingly price sensitive, and hiring has slowed to a crawl. Some costs can be passed along. Others have to be absorbed.
That is what an adverse supply shock does. It acts like a negative real-income shock: households lose purchasing power, while businesses that cannot fully pass higher costs through see margins squeezed.
So how do we square that with blockbuster Wall Street earnings?
Part of the answer is composition. The S&P 500 is not the typical American business, and two companies have distorted the headline this quarter. FactSet says S&P 500 earnings are on pace to rise 50.4% from a year ago in the second quarter. Exclude Alphabet and Amazon – whose results included unusually large investment-related gains – and that falls to 32.0%.
But 32% is still exceptionally strong, so this isn’t a story of corporate America broadly losing money. Ten of the 11 S&P 500 sectors are reporting earnings growth and eight are reporting double-digit gains.
The squeeze is uneven.
Large public companies with scale, pricing power and access to capital can protect margins even while households and more vulnerable smaller businesses are under pressure. That matters because small businesses punch far above their weight in job creation – creating about 60% of net new jobs over the past few decades.
And strong profits do not require strong hiring. In fact, controlling labor costs is one way businesses protect those profits. That helps explain how Wall Street can report excellent earnings while payroll growth nearly disappears.
So everybody waits.
Consumers haven’t stopped spending, but they don’t have much room to accelerate. Businesses aren’t firing aggressively, but they aren’t eager to hire. Wage growth is cooling despite a decline in labor supply.
We are all holding on and hoping something gets cheaper.
Which leaves the Fed stuck too
This is why I don’t think the latest data make a compelling case for either direction on interest rates.
The Fed held its policy rate at 3.50%-3.75% in July, but the vote was 9-3. Beth Hammack, Neel Kashkari and Lorie Logan all wanted a quarter-point hike.
The hawks have a point. Inflation remains above target. Energy presents another upside risk. Tariff pass-through isn’t necessarily finished. Household inflation expectations have moved up. The PPI categories feeding PCE aren’t behaving particularly well.
But raising rates now would mean tightening policy into an economy in which payroll growth has essentially stopped, hiring is weak, wage gains are struggling to keep pace with inflation and retail spending is barely advancing in real terms.
The doves have the opposite problem. Cutting rates would provide relief to an economy that could use it, but it would also mean easing while inflation remains above target and several important supply-side risks are still pointing upward.
So my read is simple: No cuts. No hikes.
Not because everything is fine, but because the economy has managed to put the Fed between a rock and a hard place.
Supply constraints are keeping inflation elevated. Weak hiring is keeping businesses and households cautious. The consumer is still swimming, but barely moving forward.
The Fed can push harder and risk pulling the consumer underwater. Or it can cut rates and risk giving inflation another breath.
For now, the least-bad option is to stay where it is.
