Bad inflation forecasts are what got us here.
Inflation is at 7.9 percent while short-term interest rates remain historically low. Bad inflation forecasts are what got us here.
O n Monday, Federal Reserve chairman Jerome Powell admitted the Fed’s inflation forecasting has not been great: “The rise in inflation has been much greater and more persistent than forecasters generally expected.” This admission gets at the heart of the Fed’s challenges: over-relying on macroeconomic forecasts in its monetary-policy decisions.
According to February’s data, headline CPI inflation is at 7.9 percent (core inflation is at 6.4 percent), and the Fed until this past week has kept interest rates near zero. The producer price index has been at 10 percent year-over-year.
None of these data reflects the economic and inflationary impact of the Ukraine–Russia war and its associated surge in oil prices. Moreover, the poor also tend to experience the highest inflation rates when inflation rises, as economists such as Xavier Jaravel have shown (in part because gasoline makes up a disproportionate amount of poor people’s consumption).
The Fed over the past decade experienced inflation frustratingly stuck below its 2 percent target despite quantitative easing and other measures taken in the wake of the Great Recession. Getting back to 2 percent seemed like a very desirable goal. So much so that the Fed wanted to make up for past misses to ensure the credibility of its inflation targeting objective, so it established “flexible average inflation targeting” in its 2020 framework review. But it did not state the time period that such an average would be measured. Will the Fed try to hit a 2 percent average core inflation target over three years, five years, ten years, or 20 years? The difference in this window would lead to different approaches toward monetary policy and when to tighten it.
Nonetheless, the framework, despite its many critics, is not one of the primary issues for the Fed today. Nor is it the prospect of a weakening economy, despite “stagflationary” warnings from supply shocks associated with the Ukraine–Russia conflict and associated sanctions. The largest threat to the U.S. labor market is arguably from tightening monetary policy, which is why falling behind the curve on inflation is so dangerous. Inflation dynamics empirically act in a rather nonlinear manner — it’s a lot easier to go from 8 percent to 9 percent inflation than it is to move from 2 percent to 3 percent.
Basing monetary-policy decisions on forecasts, as the Powell Fed has done, is something that shouldn’t be taken lightly. What if the inflation forecasts are wrong? For much of the past year, we’ve heard narratives from all over about supply chains (e.g., auto prices rising due to chip shortages) with suggested forecasts that the associated inflation will simply be “transitory” (of course with no indication of how long “transitory” meant).
To be fair, some of these examples did seem to be pandemic reopening stories, such as car-rental companies buying back cars they had sold earlier in the pandemic and driving up used-car prices. We turned to measures such as median CPI or trimmed mean CPI (kindly provided by the Cleveland Fed) to remove such strange product category outliers in the inflation numbers. But by October, even these measures were rising. In November of 2021, the Fed ditched the word “transitory” as it became clear that inflation was not sector-specific as it spread to other areas such as housing (which I previously wrote could happen given that surging housing prices appear in CPI data only through rent increases, which typically adjust at a lag).
These loose narratives surrounding the rise in inflation have continued to influence the Fed’s forecasts along with many professional forecasters and have played a key role in keeping rates near zero for longer than they otherwise should have.
This forecasting mistake has happened in recent memory in the reverse direction. Back in 2018, the Fed signaled that it planned to begin a cycle of preemptive interest-rate hikes given inflation they had forecast using the Phillips curve. The Fed still raised rates while inflation remained below 2 percent, leading to fears that the hikes were unnecessary and slowing growth (echoed by many including myself and Andy Puzder in the WSJ in May 2019). After realizing its forecasts were wrong, the Fed then backtracked and began cutting rates again in July 2019. As Heather Long at the Washington Post put it in a year-end article, “2019: The year the Federal Reserve admitted it was wrong.” Perhaps 2021 deserves that title as well.
The challenge in 2018–2019 was being overly dependent on models such as the Phillips curve. The challenge in 2021 is instead being overly dependent on ill-defined narratives about supply chains and pandemic reopening that now influence our inflation forecasts, which at best can be described as loose conjecture which has turned into groupthink. To be fair, understanding causation in macroeconomics is incredibly difficult, as macroeconomists Emi Nakamura and Jon Steinnson have written in a now-famous article “Identification in Macroeconomics.” It’s difficult to precisely understand how various macroeconomic variables such as monetary policy, inflation, employment, and GDP interact (especially in what magnitudes).
When originally Powell came into the Fed in 2017, he described his approach to monetary policy as “risk management.” While that may sound nice, there’s something deeper to it. It assumes you actually know the risks and their probabilities. Who would have guessed we would be hit with a 100-year pandemic? Who would have guessed that inflation wouldn’t appear until May 2021, quite some time after a year of tame inflation amid significant amounts of fiscal and monetary stimulus? No one can be faulted for not predicting either of these events. What we can do is think about how we want to respond once these things have happened with the acknowledgment that the future will still be difficult to predict.
One insight we do have, however — with lots of empirical evidence — is that inflation does fall conditional on monetary policy tightening. Policy-makers can rely on today’s data and use them in their monetary-policy reaction function to appropriately weigh the risks of inducing unemployment from monetary-policy tightening and the risks of inflation rising further if we don’t tighten. Had it simply been following today’s data as the best forecast of data in the next period, the Fed might not be so behind the curve on inflation right now.