The Fed has a dependency problem that needs fixing
Summary
The U.S. central bank needs to kill its policy of “data dependency”—and get investors thinking about the bigger picture.Investors have once again been head-faked by the Fed. After seven false alarms where markets priced in a dovish pivot by policymakers, the Federal Reserve finally turned dovish in a big way—only for bond yields to rise sharply and investors to price in fewer rate cuts than they had before.
What’s going on?
The answer can be summed up in two words: Data dependency.
The Fed says it sets policy based on incoming data, especially on inflation and jobs. And those data have been both unreliable and far more volatile than usual, confusing investors into a series of rapid reversals. The data point first to economic weakness and then, sometimes after revisions, strength.
Since the Fed cut last month, economic data have come in much stronger than expected. The economy also appears strong. The weak jobs figures that spooked the Fed into a double cut of half a percentage point last month reversed in this month’s report, which was the third-strongest of the year. Live estimates of economic growth by the New York and Atlanta Feds are both above 3% for the third quarter, up from 2% in late August.
The Fed should, of course, look at the data. But “data dependency" has come to mean looking only at recent data, ignoring projections for the effects of interest rates on the economy in future. Data dependency has made bond markets unnecessarily volatile.
In the summer, data suggested high rates (along with a global slowdown, troubles in China and wars in the Middle East and Eastern Europe) were hurting the economy as the jobs market slowed. Now it seems as if the economy is fine, China’s stimulus might mean it stops being such a drag, and inflation may be stickier than was thought.
Use only the most recent data, and both short-term gloom and mini-booms get extrapolated, generating big swings in expectations for Fed interest rates, and so for Treasurys.
Investors are chasing groupthink, rather than expressing the wisdom of the crowd. As hope built of a half-point cut coming in September, the probability priced into Federal-Funds futures of further cuts of 1.75 percentage points or more by June soared to 77%, according to CME Fedwatch.
Could there really be cuts equivalent to more than a quarter point at the next six meetings? No. Traders now say the probability is back to zero, where it stood before weak jobs figures for July sparked a serious, but brief, selloff.
Data dependency at the Fed has exacerbated what psychologists call recency bias in the market, with shifts in a couple of months of jobs or inflation data showing up in major moves.
Being data dependent made some sense when the Fed was extremely worried about inflation. It feared that high inflation would feed through into consumer and business expectations of future inflation, in a self-fulfilling upward spiral. To break the spiral, it yanked rates higher, predicting it would cause a recession and get price expectations under control. If reported price rises lead, via expectations, to future price rises, then it is reasonable to focus on reported price rises.
There is a case to be made for relying on past data when rates were stuck at zero, too. For many years after the global financial crisis of 2007-09, the Fed wanted to convince investors that it would be slow to raise rates, waiting until the economy got going before it moved. The aim was to lower long-term yields and prevent economic green shoots being crushed by rapidly rising rate expectations.
Both worked, with 10-year yields reaching their lowest ever and better economic data leading investors to price rate rises not at the next meeting after the data, but further ahead.
That changed in the postpandemic inflation, according to a study by the Cleveland Fed, which found futures traders were supersensitive to the inflation figures just before each Fed meeting—true data dependence.
Data dependency in this short-term way makes much less sense as the Fed tries to navigate a soft landing. It can afford to look through noisy month-to-month data and focus on the big picture. The jobs market may still be hot, but it is no longer red-hot. Inflation is still above target, but it is no longer scary.
Step back from the data, and the big question isn’t how to knock a few more tenths of a percentage point off inflation, but how fast and how far to bring down interest rates. That requires predictions about what level of rates the economy can put up with in the long run.
Cynics quite rightly point out that past Fed predictions have been terrible. Skeptics, including me, point to the wide disagreement among Fed policymakers about where interest rates will eventually come to rest.
But navigating a soft landing using the rearview mirror isn’t an option when it takes half a year or more for rate changes to have an effect on jobs and inflation. It is time for the Fed to kill its “data dependency," and try to get investors to think about the longer run.
Write to James Mackintosh at james.mackintosh@wsj.com
