A new Brookings analysis by Jens H. E. Christensen and Glenn D. Rudebusch examines why estimates of r*—the natural or steady-state short-term real interest rate—have risen by roughly 1 percentage point in the United States since 2020, after decades of decline. Economists have increasingly pointed to two possible explanations: expectations of higher government debt and stronger productivity growth driven by artificial intelligence. But the authors' event-study evidence suggests that neither explanation adequately accounts for the recent increase.
The finding on AI is particularly interesting. If investors believed that generative AI would substantially raise future productivity and economic growth, major AI announcements should push up estimates of the natural rate. Instead, the researchers find that news surrounding major generative-AI model releases was associated with an overall decline in measures of r*. In other words, the financial-market evidence does not support the simple story that the AI productivity boom is responsible for today's higher equilibrium interest rates.
Fiscal developments appear to have had some upward effect, but it was relatively modest. The authors find that U.S. fiscal-policy news during the first half of the 2020s provided only a limited lift to the natural rate. They also find that monetary-policy news does not explain the recent increase, despite earlier research linking persistent movements in longer-term yields to Federal Reserve policy announcements.
That leaves an important economic puzzle. Something appears to be pushing the natural rate higher strongly enough to overwhelm downward influences associated with AI, monetary developments, demographics and other factors. The implication is that policymakers should be cautious about attributing today's higher interest-rate environment to the AI boom or fiscal expansion alone. The study instead suggests that a still-unidentified structural force is driving much of the increase in r*—and identifying that force matters because r* is central to judging whether Federal Reserve policy is restrictive or stimulative and to assessing the long-term sustainability of U.S. government debt.
