Will AI Lower Interest Rates?

During the 1990s, Federal Reserve chairman Alan Greenspanearned a reputation for recognizing the productivity benefits of the IT revolution before they appeared in official data. While many policymakers feared low unemployment would trigger inflation, Greenspan concluded that technological progress was expanding the economy’s productive capacity. Confident that productivity gains would offset inflationary pressure, he kept interest rates relatively steady in 1996 and 1997—a decision that initially proved correct.

Kevin Warsh, the Fed’s new chairman, believes artificial intelligence could produce a similar transformation. Like Greenspan, he argues that central bankers should pay close attention to business anecdotes rather than rely solely on backward-looking statistics. In his view, AI will raise productivity, expand supply, reduce production costs and allow the economy to grow faster with less inflation, justifying lower interest rates.

The short-term picture, however, is more complicated. AI is currently fuelling enormous investment in data centres, semiconductor factories and electricity infrastructure. Combined with rising stock prices that encourage consumer spending, this surge in demand may outweigh any productivity gains for now. Some policymakers therefore argue that AI enthusiasm could actually require higher interest rates to prevent the economy from overheating.

Even if AI substantially improves productivity, lower interest rates are not guaranteed. Greater efficiency increases incomes, which can generate additional consumer spending and business investment. Stronger growth may therefore push up the economy’s neutral interest rate (r-star)—the rate that neither stimulates nor restrains economic activity. Research suggests AI could raise this equilibrium rate by roughly one percentage point, making a return to the ultra-low interest rates of the pre-pandemic era unlikely.

Looking further ahead, the outcome becomes even more uncertain. If AI dramatically boosts future incomes, households and firms may borrow and invest more aggressively, pushing interest rates higher. Conversely, if AI displaces large numbers of workers, increases inequality and weakens investment opportunities, precautionary saving could rise while overall demand falls, driving equilibrium interest rates lower. Whether AI lowers or raises rates depends largely on how its economic benefits are distributed across society.

The article argues that these contrasting possibilities should make policymakers cautious. AI is unlike previous technological revolutions, making historical comparisons imperfect.

The experience of Alan Greenspan offers a final lesson. Although he correctly identified the productivity boom of the 1990s, he later recognised that the same technological gains also fuelled excessive demand and speculative asset bubbles. The Federal Reserve eventually had to raise rates aggressively, but inflation and financial excesses had already built up before the dot-com bubble burst.

The implication for Kevin Warsh is clear: learning from Greenspan means understanding both sides of the story. AI may indeed increase productivity, but whether it ultimately pushes interest rates higher or lower will depend on how productivity, demand, investment and income distribution evolve together. History suggests that recognizing a technological revolution is only the beginning; managing its macroeconomic consequences is the far harder task.

Source:

https://economist.com/finance-and-economics/2026/06/25/will-ai-lower-interest-rates

関連投稿: