Should the Fed Focus on Core Inflation or Headline Inflation?

Charles Bean says central banks are mistaken to focus on core inflation rather than headline inflation in their policy deliberations:

BoE hits at US inflation measure, by Krishna Guha in Jackson Hole, Financial Times: The US Federal Reserve is wrong to focus on core measures of inflation that exclude energy prices, Charles Bean, chief economist at the Bank of England, has suggested.

It should focus instead on headline inflation, which is much higher... Including energy and food costs, US consumer price inflation is running at an annual rate of 4.1 per cent, against 2.7 per cent for core inflation.

Mr Bean told the Fed’s annual Jackson Hole symposium ... that energy prices were rising for the same reason the price of many manufactured goods were falling: the rise of China and other emerging market economies. Since both price trends had a common cause, he said it makes little sense to focus “on measures of core inflation that strip out energy prices while not stripping out falling goods prices as well.”

Mr Bean did not mention the Fed by name but his implication was clear. ... Central bankers in Europe take a sharply different approach. Both the Bank of England and the European Central Bank put greater emphasis on talk of headline inflation, which includes the immediate “first round” effect of rising energy prices.

But the Bank of England and the ECB increasingly take the view that energy prices may be on a long-term upward trend, driven by industrialisation and urbanisation in China and India.

David Altig said recently:

Why We Focus On Core Inflation II, by David Altig: From this month's edition of Economic Trends, published by the Federal Reserve Bank of Cleveland (page 3):

What information households base their inflation expectations on is the topic of frequent academic debate. Rather crude correlations, which examine the relationship between realized inflation rates and households’ expectations, indicate that their year-ahead expectations are most closely correlated with the headline CPI inflation rate, and are especially sensitive to this measure over longer time horizons. Interestingly, expectations for the inflation rate over the next five to 10 years are more closely correlated with the core CPI inflation rate than with headline CPI. The correlation also grows stronger as the underlying core CPI inflation trend becomes more persistent. Indeed, ... this may indicate that households see through the same transitory fluctuations in prices that the core inflation measure is designed to isolate.

At a minimum, this means that the FOMC has been pretty successful at getting the public to focus on core inflation as a reasonable measure of the longer-term inflation trend. But it is unlikely that this situation could persist without a central bank that generally delivers the goods, and as I suggested once before, it does appear that, over the medium term, core inflation is a better predictor of headline inflation than is headline inflation itself.

To the extent that stabilizing inflation expectations is the really important part of getting monetary policy right, these sorts of results begin a persuasive case for the use of core inflation as a guide.

I agree with David - I want to see evidence that we forecast future headline inflation better by dropping core measures, and as noted below there are theoretical arguments for using core inflation as well. For now, as David notes, core inflation has the most predictive power for future headline inflation. Notice, however, a difference in the underlying assumptions that drives the use of different measures of inflation. The Economic Trends article sees the difference between core and headline measures as "transitory fluctuations" whereas in Europe energy prices are seen as following a "long-run upward trend." Finally, core inflation measures can be defended on a more theoretical basis, e.g. see here for a summary of Woodford's argument on this point.

Comments

[spencer]

If a central banks ignores the impact of higher oil prices on inflation does it mean that it will monetize the increase in oil prices?

If wages are partially a function of inflation expectations and these are based on actual inflation
don't higher oil prices work through the wage setting mechanism to generate higher core inflation?

It looks to me like total inflation leads core inflation.

[Cyrille]

It seems to me that, if core inflation is to have a point at all, it should at least be looked at after adding the average extra inflation due to the other components. I agree that removing statistical noise helps seeing an underlying trend, but when was the last time that core inflation was greater than headline inflation?
So, if the difference is always above zero (or even averages above zero statistically significantly), to simply wash it away is to look at inflationless inflation.

All the more so since those increases in energy and alimentation are likely to have a deflationary effect on the rest, because there is less money to spend. Imagine an economy where there are just two goods, energy and equipment, and wages are stable. If energy goes up, people have less money to buy equipment, driving the price down, but of course not enough to buy as much as before, because oil producing countries will be able to buy quite a lot of equipment.
So, in that case you would argue that core inflation was negative. But people get less than before for the same amount...

[spencer]

The last time core inflation was higher than headline inflation was 2002. Since 1960 core inflation has been higher than headline inflation 52% of the time.
Over this period headline inflation averaged 4.26% while core inflation averaged 4.22%.

[bakho]

Why should the Fed be responsible for trying to fix policy failures? High energy prices should be attacked with conservation measures, fuel standards and research into alternatives. Raising interest rates won't fix inaction by Congress and this administration.

[Cyrille]

Well, then it does seem as though in the past 40 years, the difference between headline and core inflation was really statistical noise. Sorry for not knowing that -well here in France it's not exactly common knowledge.
However, I would argue that it may not be so today. Energy is not going back to its low price, and with the increasing climatic problems I would expect that food should become more expensive too. So maybe what once was statistical noise now has a genuine pattern.

[lonesome moderate]

"Why should the Fed be responsible for trying to fix policy failures? High energy prices should be attacked with conservation measures, fuel standards and research into alternatives. "

These kinds of policy failures will have their impact on the economy sooner or later, regardless of what the fed does. It's not the fed's job to fix what the president and congress have broken, but they do have responsibility for maintaining broader economic stability by making the impact as predictable and transparent as possible--the now famous "soft landing".

[lonesome moderate]

If I understand the article correctly, what economists are mostly concerned about is which of the two short term numbers should be perceived by the general public as "the inflation rate". The Fed's own models are of course much more sophisticated than that--they have their equations, which no doubt have both short term numbers as inputs, along with a bunch of other stuff.

Rather than trying to choose one of these two numbers as the inflation sound bite, perhaps a better alternative would be to publish one of their equations. Then the news reports would say "the Fed's long term inflation trends index now stands at X%, based on the latest figures for core and broader inflation." This would hopefully get across the idea is that the most important inflation number is what the Fed thinks it is, rather than one of the CPI rates or any other single measure.