Chicago Federal Reserve Bank President Austan Goolsbee warned on Wednesday (May 6) that the economy could overheat if large-scale business investment and consumer spending increase significantly before actual productivity improves, amid widespread market expectations of a boom led by artificial intelligence. In that case, the Fed may need to raise interest rates rather than lower them.

“In that scenario, the Fed clearly wouldn’t necessarily need to cut rates; it might actually need to raise them,” Goolsby said in a panel discussion at the Milken Institute conference. He added, “If artificial intelligence is as powerful as it’s touted to be, it will…make us rich. But if those benefits are still primarily realized in the future, I think we need to be more cautious and keep a close eye on things.”

The acceleration of productivity growth and whether it will become a lasting economic dividend are currently hot topics of debate among the Federal Reserve and financial markets.

The relationship between productivity and interest rate policy sparks disagreement

Some have drawn on the experience of the 1990s to conclude that faster productivity growth could lower inflation, thus creating conditions for lower interest rates. Incoming Federal Reserve Chairman Kevin Warsh believes that artificial intelligence will significantly boost productivity, thereby suppressing inflation and providing the Fed with room to cut interest rates.

Cathie Wood, CEO of investment firm Ark Invest, said in the same panel discussion that she expects artificial intelligence to continue to drive U.S. real GDP growth by up to 8% through significant productivity gains, and to bring a “massive undercurrent of deflation.”

Goolsby questions this: Is the current prosperity based on existing technology, or on an over-extension of recent improvement trends? He argues that if it's the latter, then marginal returns will diminish.

He cited self-driving cars as an example, pointing out that at the time it was predicted that self-driving technology would rapidly become widespread and that all professional drivers in the United States would lose their jobs within five years.

"The results didn't happen so quickly," Goolsby said, "because people extrapolated the rate of improvement at the time and assumed it would continue, but that's not the case. Instead, we encountered diminishing marginal returns."

He further stated, "So when I see the massive investments in data centers today, I can't help but wonder how much of the improvement we're seeing is already close to the limits of growth."

Expectation management becomes crucial; overheating must be guarded against.

Goolsby points out that, from an economic perspective, whether interest rates rise or fall depends on whether productivity growth exceeds expectations or has already been anticipated by the market.

In the mid-1990s, then-Federal Reserve Chairman Alan Greenspan argued that productivity gains could explain profits, employment, and inflation data even before improvements were evident. In that context, lowering interest rates was a natural policy response.

But Goolsby emphasized that if consumers and businesses anticipate significant future productivity gains, their behavior today will change, making interest rate policy more complex.

He stated, "We need to pay close attention to how much of the productivity surge is still based on future expectations. The more hype there is, the more interest rates need to rise to prevent the economy from overheating."

Overall, Goolsby's remarks serve as a reminder to the market that amidst the AI boom, one should not be blindly optimistic about future productivity dividends, but rather be wary of the overheating risks arising from premature overspending. The direction of the Federal Reserve's monetary policy will depend heavily on the pace and magnitude of actual productivity improvements.