Nobody Expected Our Current Interest-Rate Configuration

A draft for Project Syndicate...

Nobody—well, very few people—here in the United States expected our current interest-rate configuration.

From early March to mid-May 2022 the long-term real safe interest rate in the United States—the thing that along with “financial conditions” governs incentives to build and also, through its impact on the exchange rate, the balance of net exports—jumped by more than 1%-point as the bond market noted that the Federal Reserve would not long be pedal-to-the-metal in the hope of speeding employment recovery as fast as possible. Then, from late August to early October 2022, this governing real interest rate took another upward jump, this time by an annual 1.5%-points, as bond traders speculated that the Federal Reserve might have to move monetary policy much tighter in the context of a fear of persistent inflation in a U.S. that had reattained full employment.

That essentially gave us our current configuration, in which interest rates are far, far higher by 2%-points or so above what anyone five years ago, back before the plague, would have guessed at the value of the “neutral” interest rate that, in the words of John Maynard Keynes, “bring[s] about… [proper] adjustment between the propensity to consume and the inducement to invest…” And with an interest rate universally seen as above the neutral level, nearly all observers were confident that recession, or at least sufficient weakening that it would take enormous fancy footwork and a good deal of luck to avoid recession, was on the way.

Given the propensity to consume, interest rates above any possible view of their neutral level would sooner or later prove too high to support a sufficient inducement to invest and a sufficiently small balance of net exports to maintain the economy at full employment. It was, the general near-consensus maintained, simply a matter of time. And when the time came it was a question of whether the Federal Reserve could recognize approaching weakness in time to cut soon enough and fast enough for the landing to be smooth and the taxi to the gate to be successful.

This configuration has now lasted for seventeen months.

Things have not been fully static. There was from August to October 2023 a surging wave, an additional 1%-point boost, that the bond market then gave back as speculation shifted to when the Federal Reserve would start and then how fast the Federal Reserve would cut interest rates, and that has left our long-term real Treasury interest rates today much where they were in October 2022. And the long-run expected inflation-premium side of the market has been rock-solid stable since June 2022.

This configuration has now lasted for seventeen months. The landing has been smooth. The engines-in-reverse have done their work. Yet the Federal Reserve in the cockpit has not dared start to move the thrust throttles out of reverse detent and into idle. Why not? Because nonfarm payrolls keep pushing forward: seasonally-adjusted, 275,000 more jobs in February than in January; a hair under 250,000 per month on average over the past six months; 230,000 per month on average over the past year.

My sense is that the Federal Reserve is profoundly uncomfortable with interest rates substantially in excess of what it confidently believes the neutral rate to be at a time when inflation is very near to the Federal Reserve’s 2% PCE-index target. But the Federal Reserve does not dare start to end this post-soft landing touchdown period of reverse thrust until it sees signs of reduced forward employment groundspeed.

As always, there are three things that could be going on to produce this situation: an error in the analysis behind the conclusion that interest rates right now are in excess of the neutral rate; an error in measurement of the state of the economy; and Wile E. Coyote.

Consider Wile E. Coyote first. This is the possibility that weakness is coming, and that in six months the Federal Reserve will wish that it had started cutting rates in January—that decisions to cut back on building and to re-source purchases to overseas were paused in 2022 and delayed until 2023 as people waited to see how much the Federal Reserve would tighten; that those decisions were then made between May and October as the Ten-Year Treasury bond did its temporary run-up from 3.53% to 4.93%, and that the impact of those decisions on employment patterns will hit the economy starting… now. Wile E. Coyote in pursuit of the Roadrunner zagged when he should have zigged, ran off the edge of the cliff, but will not begin to fall until he looks down and sees that he is trying to run on air. I was fairly confident that this was the likely scenario six months ago. But as time passes without it happening my brain is torn: I think that the time it will happen if it does happen is surely coming closer, but also that the chance that it will ever happen is lower.

How about wrong measurement? The consensus has long been that, as far as gauging the employment side of the economy is concerned, the payroll survey is superior to the household survey. However, in the past couple of years the surveys, which were dead-on once one takes account of differences in coverage on the eve of the plague, now differ by perhaps 4,000,000. And while the payroll survey tracks 2.7 million more jobs than a year ago, the household survey tracks only 700,000 more people working in those jobs than a year ago.

But a labor market as weak as the household survey suggests should have already made itself felt in weakness in consumer spending as well. And it has not.

How about wrong analysis? The near-consensus, as of the start of the plague, was that there were powerful fundamental factors that were keeping the neutral interest rate very, very low. The near-consensus now is that there have been no significant changes in those fundamentals over the past 4 1/4 years: nothing big enough changing to significantly move that neutral rate. Thus the neutral interest-rate should still be very low—with interest rate levels today highly restrictive and, therefore, inappropriate to an economy at full employment with inflation near its target.

But if there is one lesson I have learned in 40+ years of trying to understand the business cycle, it is that there is no empirical regularity in the macroeconomy, no matter how well grounded in historical data, in statistical estimation, or in theoretical concepts, that can be trusted not to crumble beneath our feet in a remarkably short time.