Remember the attack on core inflation? Right about this time a year ago there was a wave of criticism aimed at the idea that core inflation—headline inflation less food and energy prices—is a useful predictor of overall pricing pressures. But a funny thing happened on the way to the lynching of core: the much-maligned concept for looking ahead turned out to be reliable… again.
Consider the rolling one-year percentage changes for headline and core CPI through this past April. Each is higher by 2.3% over the previous 12 months. But the path to equality has been a volatile trek, as it always is. That’s not surprising. But while core and headline are once again running neck and neck, it won’t last. But when the next round of divergence comes, and you’re looking for some perspective on the outlook for inflation, history suggests we should start by considering core inflation.
That’s a radical notion in some circles. A year ago, the idea that core could tell us anything useful about the inflation outlook generally had fallen on hard times. For example, St. Louis Fed President James Bullard publicly rejected a fair amount of research when he asserted in May 2011:
One popular argument for focusing on core inflation is that core inflation is a good predictor of future headline inflation. I think this is wrongheaded, as well as wrong.
A year later, core doesn’t look quite so rotten. In fact, it looks pretty good. A year ago, if you had looked to core as a guide for where headline would be, you’d have done quite well in the forecasting game. By contrast, some pundits we’re telling us that the then-rising 12-month pace of headline inflation was a harbinger of things to come and so the increasing inflation rate would roll on. Low inflation was history, they said—we were on the cusp of surge in headline inflation. But core inflation’s lower rate of increase suggested otherwise, and a year on we now know that core was the better predictor.
What changed? Energy prices are no longer skyrocketing. In fact, crude oil has fallen rather dramatically in recent months. After reaching nearly $110 a barrel earlier this year, the benchmark West Texas Intermediate has sunk to the mid-$80s.
It’s no surprise that most of the folks who were projecting higher headline inflation a year ago were also warning that $200-a-barrel-oil was just around the corner. Yes, those risks can’t be dismissed, even today. But if there’s an imminent danger of those outcomes, we’ll likely see an early warning in a steadily rising pace of core inflation. For the moment, those potential risks look rather minimal.
Looking to inflation less energy and food prices is, in fact, an old idea, and one that starts with Irving Fisher’s debt-deflation theory. As I wrote a year ago, “the focus on core has evolved via years of research that suggests that this narrower gauge inflation is a better predictor of headline inflation for the medium- to long-term future.” You can find a few examples of the research from recent years in that post.
Why, then, do so many analysts effectively ignore this body of analysis, and the empirical record? Part of it is probably related to the pressures of the moment. When you’re watching oil and gasoline prices rise over, say, the previous six months, it’s hard to dismiss the trend and argue that a lower core inflation rate may be a better measure of where prices generally are headed. The basic concept of core inflation, in other words, is counterintuitive in real time. So many people fall into the trap of deciding that core can’t be right tomorrow because they see energy and food prices rising today. But looking in the rearview mirror isn’t necessarily the best way to project trends beyond a few months. That may be obvious now, but the insight will again go out the window the next time oil prices surge.
This post originally appeared at The Capital Spectator and is posted with permission.
2 Responses to “A Year Later, Core Inflation Doesn’t Look So Rotten”
Nice piece. In addition, I always like looking at the Cleveland Fed's 16% trimmed mean inflation and median inflation as alternative measures of underlying price trends. They throw out the most volatile prices each month whether they are food, energy, or something else. Recently, not just the headline inflation and the core, but also those other measures have converged.Nice piece. In addition, I always like looking at the Cleveland Fed's 16% trimmed mean inflation and median inflation as alternative measures of underlying price trends. They throw out the most volatile prices each month whether they are food, energy, or something else. Recently, not just the headline inflation and the core, but also those other measures have converged.
I would like to see the energy and food prices included as a 12-month (maybe up to 36-month) running average used in the calculations. That would permit those items to reflect the real inflation rate but would eliminate the inherent volatility in the prices of the commodities.