Jobless Expansion Puzzling the Fed
November 25, 2025
James Savin
Sign up for our newly launched weekly newsletter here.
November 25, 2025
James Savin
Economic expansion is usually coupled with job growth. Still, a robust GDP and a faltering employment market are leaving Federal Reserve (Fed) officials unsure how to address the problem in future economic policy. Typically, the two numbers would coincide—the Fed would respond to a dwindling GDP and weak job market by cutting rates, and raising them in the inverse—but the two mismatched numbers present a conundrum. Economists attribute the issue to increased investment in manufacturing and AI, as well as a jump in productivity driven by growing automation and workplace restructuring. The Fed has been left unsure of its next steps and is treading carefully around the possibility of a recession.
In the second quarter of 2025, US GDP rose by a staggering 3.8% from the previous quarter. This is, on its surface, a good sign. It’s an extremely high number for a developed economy and suggests the Fed has managed to avoid a severe economic downturn. The Fed’s decision to only conservatively lower interest rates this past year appears to be paying off.
Usually, the job market grows in tandem with GDP. GDP is a measure of goods and services bought by consumers. Thus, when GDP is high, businesses are prompted to invest in more staff to produce more goods for sale. However, this quarter we’ve seen the opposite of this rule. As GDP growth bounds ahead, the job market has stagnated, with the US labor market in Q2 gaining only 187,000 jobs, and in August, job growth was in the negatives as 4,000 Americans lost their jobs.
187,000 jobs gained over three months is concerningly little, as the job market has historically seen growth far exceeding that during comparable GDP growth. During the Clinton administration, GDP growth averaged 4% annually. In that time, 22.5 million jobs were created. This averages out to more than 234,000 jobs created per month—far more than the 62,000 average from this quarter.
This mismatch appears to have two causes: first, a surge in productivity due to automation, and second, growth driven by capital spending rather than consumer spending.
When job growth and loss rates are stratified by sector, there’s a clear divide between those industries least and most impacted by job loss. The majority of industries that have seen growth are those in which workers are difficult to replace by machines or generative AI, with the largest gains being seen in private education, leisure and hospitality and retail trade. Based on high GDP growth, jobs requiring human interaction are increasing, while jobs in business services, manufacturing and transportation are declining. This is coupled with a 3.3% increase in productivity in Q2, indicating businesses are making more money with fewer workers.
Capital spending amounts for a significant portion of the GDP increase we’ve seen this quarter. The Economist reports that AI’s contribution to GDP growth is 40%. One Harvard economist, Jason Furman, even estimated that 92% of US GDP growth in the first half of the year was in AI investments. This overextension is worrying—the US economy is one bubble pop away from catastrophe.
“The divergence between solid economic growth and weak job creation created a particularly challenging environment for policy decisions,” Fed officials said in their October meeting.
Fed Governor Christopher Waller said last month, “Something’s gotta give — either economic growth softens to match a soft labor market, or the labor market rebounds to match stronger economic growth.” It made Fed officials wary of further rate cuts, suggesting that interest rates will remain high. This, too, presents its own problems—the high rates trickle down to already thin–stretched American consumers, making daily life more expensive. It’s on the Fed to ensure a soft landing when the AI bubble inevitably pops, and pressure is building from the American people to cut rates, putting the Fed between a political rock and a hard place.
Read more here: