The policies of the current Administration have received a lot of attention, and that’s no surprise considering how much they’ve differed from historic norms. We’ve experienced:
• The biggest increase in tariffs in our lifetimes
• A sharp drop in immigration and a surge in deportations
• A war with Iran that has caused energy prices to soar
These are all what economists call negative supply shocks, which raise prices and reduce output. Despite them, the U.S. economy and financial markets have performed remarkably well. That’s because we are in the midst of an historic technological spending boom. Global spending on data centers alone–heavily concentrated in the U.S. – is expected to amount to as much as $7 trillion (more than 20% of US GDP) over the next few years. The net effect of all these influences has been solid economic growth, higher inflation, strong growth in corporate profits, a robust stock market and higher bond yields. This suggests that the current stance of monetary policy is stimulative at a time when inflation has been consistently above the Fed’s 2% target. New Fed Chair Kevin Warsh has been clear about his commitment to that inflation target, but the Federal Open Market Committee nevertheless decided to maintain its policy rate at its late July meeting. The financial market reaction reflected the need for higher rates by bidding up bond yields to their highest levels since 2007.
The near-term outlook for the economy is bright because the burst of spending on AI shows no sign of slowing down. Both the users and producers of AI as well as their suppliers believe that it is critical to their future success and are competing for supremacy. Investors don’t want to be left behind either, so they continue to bid up stock prices.
Higher energy prices usually bring demand destruction, but the income created by massive spending on AI is boosting overall demand. While second-quarter GDP growth slowed to 1.5% (from 2.1% in Q1), domestic demand (equal to consumer plus business spending) accelerated to 3.9%, up from 1.7% in Q1. Retail sales, car sales and travel are all strong, despite much higher prices for gasoline, electricity and airfares. Strong demand usually means a healthy labor market, and clear signs of improvement have appeared there as well (despite the weaker-than-expected jobs report for July). Job growth averaged 75,000 per month in the first half of this year compared with an average decline of 8,000 in the second half of last year. The unemployment rate has dropped from a peak of 4.5% in November to 4.1% in July. Jobless claims for unemployment insurance have moved down, consistent with a declining unemployment rate.
The notion that AI will reduce overall employment is misguided. Hiring for some entry-level positions has slowed, but that is being more than offset by new jobs generated by the huge AI infrastructure build-out. Over time, new business lines will develop that were not available before the widespread adoption of AI. Recall that during the dotcom boom, online shopping was expected to kill retail stores, and it has displaced a significant number of jobs at many outlets. But that loss of employment has been more than offset by staffing at warehouses and an explosion in the delivery business.
The Fed’s preferred inflation measure–the core PCE index–rose by 3.4% in the second quarter, down from 4.4% in Q1 but still well above its 2% target. Inflation is likely to remain elevated even as oil prices come back down. The pass-through from higher energy prices to related items takes several quarters to play out, and the costs of production inputs, transportation, and travel are all still rising. AI is presumed to be deflationary because it will automate work and increase productivity, and that will probably be the case eventually. But the building of AI is inflationary because it increases overall demand before the productivity gains fully materialize. Memory chip prices have surged after falling for most of their history, while the building of AI data centers is putting upward pressure on electricity prices.
The implications for Fed policy are clear: It needs to begin raising rates now if it wants to lower inflation back to its 2% target. The Fed reduced short-term interest rates from well over 5% in August 2024 to just over 3.5% at the end of last year, where they are now. The driving force was a significant drop in job growth, from over 200,000 per month to modest net job losses. The labor market is now improving while the war with Iran kicked off a significant rise in inflation. Even before that, inflation has been above the Fed’s 2% target for five years. The unemployment rate has moved down and nominal GDP growth has accelerated. Financial conditions remain significantly supportive: stock and bond issuance has surged, credit spreads are tight and the stock market is up solidly this year after a nearly 80% rise in the previous three years.
While the outlook for both the economy and financial markets is positive for this year, there are signs of a bubble emerging. Technology revolutions that transform the economy have a long history of attracting more investment than near-term returns justify. The internet turned out to be the gamechanger everyone thought it would be, but that didn’t prevent stock prices from falling some 50% and dotcom stocks by even more. The AI buildout is by far the largest capital spending surge in U.S. history, swamping what was spent during the internet boom: U.S. investment during the dot com boom doubled over 5 years; AI spending has skyrocketed by a factor of 4.5 in under 3 years. Investors have bid up the valuations of companies expected to dominate, lenders have funded an unprecedented infrastructure buildout, and suppliers have expanded to meet the demand. The economy has become reliant on AI investment at a time when near-term profit gains from AI are relatively small because it is not yet ready for full-scale deployment.
Signs of excess in financial markets have already emerged. Private credit funds–which exploded after the financial crisis as regulations discouraged banks from making high interest loans–are now a $3 trillion industry, attracting both institutional and retail investors. They perform banking functions by taking in investor money and lending to companies that tend to have a high risk profile but are much less regulated. Banks have made an estimated $1.4 trillion in loans to private credit firms, which are now facing heavy withdrawal demands. Meanwhile, margin debt has soared to record highs, implying considerable leverage in the stock market. Brokerages are tightening margin requirements, heightening vulnerability to a market correction.
While we seem to be in a period of excess that will end badly at some point, it is very difficult to predict when bubbles burst, and they usually last longer than most expect. Former Fed Chair Alan Greenspan coined the term “irrational exuberance” during a sharp rise in the stock market in late 1996, and the internet bubble didn’t burst until March 2000. There were also signs of a bubble before the global financial crisis emerged: House prices in the U.S. peaked in early 2006, the Bear Stearns credit funds failed in June 2007 and the stock market didn’t crash until September 2008.
The bottom line is that AI will probably deliver the super-charged productivity boost that markets expect, but getting there may involve an historic investment cycle, financial vulnerabilities and even a recession.
