The Fed’s ‘Average’ Is More ‘Flexible’ Than ‘Target’

Here is an assignment that, in theory, should be simple: Compute the average of an economic data series over the medium term. Just pick one—retail sales, industrial production, inflation, whatever! The data are pretty easy to find—and freely available—at the St. Louis Fed’s data tool, FRED. But here is the thing: You can’t do it. Why? We didn’t give you enough information. “Medium term” is a fuzzy, unspecified period. Which brings us to the Fed and inflation. The Fed has come under fire for the inflation rate far exceeding its 2% (ish) target. Some suggest it should chuck that target outright, lest it lead the central bank to overreact to spiking oil prices and risk hiking rates aggressively as the economy slows. What they miss: The Fed did that in August 2020. We argued this at the time, and it bears repeating, as knowledge is power for investors.

When Congress established the Fed’s dual mandate with 1978’s Humphrey-Hawkins Full Employment Act, there were no numerical targets—just marching orders to pursue maximum employment while promoting price stability. No one interpreted “price stability” as zero inflation, as most everyone accepts that some inflation is a healthy side effect of a growing economy, which usually means more money in circulation. Conventional wisdom globally eventually zeroed in on a 2% annual inflation rate as reasonable, and many central banks have long had that rate as their target. The Fed followed suit officially in 2012, establishing a target of 2% y/y for the headline Personal Consumption Expenditures Price Index (PCE).

Between then and August 2020, the Fed managed to hit that target all of three times.^[i] For most of that span, inflation undershot 2% y/y, despite all the alleged “stimulus” sloshing around courtesy of quantitative easing (QE). Thus, there developed a view that the target itself was the problem—that the Fed was still acting as if 2% were a hard ceiling and thus keeping policy tighter than it otherwise could be. One school of thought argued central banks should target a certain level of nominal GDP growth instead. Another claimed that if central banks merely targeted an “average” inflation rate, it would give them more latitude as periods of higher inflation would cancel out periods of below-target price increases, thus enabling the Fed to step on the gas.

The Fed adopted this approach in August 2020, saying it would target “inflation that averages 2 percent over time,” a policy framework dubbed, “Flexible Average Inflation Targeting.” But it never defined over what time period, leaving “average” open to interpretation, bringing us back to our introductory conundrum. As the St. Louis Fed explains, the Fed would “seek inflation that averages 2% over a time frame that is not formally defined.”^[ii] This is the problem. How can you be accountable for an average inflation rate if you don’t commit to an iteration? We don’t know if the Fed wants the inflation rate to average 2% over 12 months, 24, 36 or more. It isn’t even clear the Fed knows. This matters, because rolling averages aren’t etched in stone. Data points are regularly coming in and dropping out.

Using an undefined average quickly goes to a strange place. Technically, as one economist we follow pointed out last month, the Fed could argue today’s fast inflation has met the Fed’s target. December’s 5.8% y/y rise in the PCE price index brought it to the same level it would have arrived at if the inflation rate were a steady 2% y/y since January 2005, a totally arbitrary starting point.^[iii] This would be the first time it was back at that mark since 2014. Does anyone really think the Fed should be doing a victory lap, patting itself on the back because a hot inflation rate at the end of 2020 hit a made-up target? Of course not! We will concede that thinking in these terms helps put more recent price increases in perspective, but we are quite certain no one thinks the Fed should hope for hot inflation to balance out below-target price increases over the preceding eight years. That is extremely cold comfort for those trying to manage sharp cost increases today. Steady and predictable is the goal, after all.

Second, how is the Fed calculating this average? Are we talking a simple arithmetic mean (add all the iterations and divide by the number of iterations), or are we using a geometric mean to account for the effects of compounding? 

Newsflash: If you try to target an “average” inflation rate but refuse to commit to strict parameters, then you don’t have a target. Society has nothing specific to hold you accountable for. In the UK, the Bank of England Governor has to write a letter to the Chancellor of the Exchequer whenever the monthly year-over-year inflation rate exceeds 3%. That is a clear boundary. The Fed has no such boundary. Forget the old adage about trying to hit a moving target—the Fed is shooting at something that doesn’t exist.

More to the point, at January’s Fed post-meeting press conference, a reporter asked Fed head Jerome Powell if, with inflation running well above 2%, “Do you want to go below 2%, so that on average you get a 2% inflation rate?” Powell’s response: “So no, there’s nothing in our framework about having inflation run below 2%. That we would do that. That we would try to achieve that outcome. So the answer is no.”^[iv] Hmmm. This … isn’t how averages work.^[v]

In our view, this background knowledge can help you make sense of Fed chatter. Numerous Fed people made headlines over the past two weeks for their seemingly conflicting views on how many times—and by how much—the Fed should hike rates in the coming months. Is this because they disagree on the necessary prescription for today’s elevated inflation? Or because the “average” target means different things to every one of them/ Oh, for the days when competing views of data were the main problem—competing views of the target takes it next-level.

However, we don’t think this changes much for the predictability of Fed moves. Our view of future Fed policy has always been they will do what they do when they do it. If the target really meant much, they wouldn’t have started a tightening cycle in late 2015, when the headline inflation rate was flirting with zero. Moreover, the Fed generally follows market-set rates at a lag. They don’t so much set short-term rates as catch up to them. That was the case before they had a target, while they had a target, and we think it remains the case now that they have an imaginary target-in-name-only. Finally, as we have written, the Fed has a pretty limited ability to rein in supply-driven inflation like today’s, given it cannot drill for oil or mass-produce semiconductors.