About


The Atlanta Fed's macroblog provides commentary and analysis on economic topics including monetary policy, macroeconomic developments, inflation, labor economics, and financial issues.

Authors for macroblog are Dave Altig, John Robertson, and other Atlanta Fed economists and researchers.


« Deep questions from Jackson Hole | Main | Does the GDP deflator lie? »

August 28, 2008


Monitoring the inflation bees, striving not to get stung

As pure blogging fodder, Willem Buiter’s Jackson Hole “advice” to the Fed is a gift that keeps on giving. From Buiter’s paper:

“It has not always been clear whether the Fed actually targets core inflation or whether it targets headline inflation in the medium term and treats core inflation as the best predictor of medium-term inflation.”

As luck would have it, our boss (Federal Reserve Bank of Atlanta President Dennis Lockhart) had a few words to say on exactly that topic yesterday:

“Attempts to measure the aggregate rate of price change—no matter how sophisticated—remain imperfect. As a result, when it comes to measuring inflation, judgment is needed to distinguish persistent price movements that underlie overall inflation from the relative price adjustments. Separating the inflation signal from noise involves much uncertainty—especially when making decisions in real time. Discerning accurately the underlying trend is difficult.”

The difficulty of precisely “separating the inflation signal from the noise” is not a new problem. In fact, this difficulty can be traced at least as far back as the development of index numbers to measure economic aggregates—and aggregate inflation in particular—by the famous economist Irving Fisher. Fisher struggled with the idea of being able to separate out the general movement in prices from the relative price disturbances:

“It would be idle to expect a uniform movement in prices as to expect a uniform movement for bees in a swarm. On the other hand, it would be as idle to deny the existence of a general movement of prices ... as to deny a general movement of a swarm of bees because the individual bees have different movements.”

The distinction between the direction of the swarm and the individual bees is an important one that is the direct path to discussions of measures of “core” inflation. The conceptual issues were nicely articulated in a recent article by Dallas Fed economist Mark Wynne, and they go something like this: Suppose we thought of the percent changes in prices of individual goods and services between two periods as containing a common component (core) and price changes that are unique to the supply and demand conditions of a particular products markets. The object of our desire (the honey if you like) is the level and direction of the common component. The problem is how to measure it.

In his commentary on the “will-o’-the-wisp of ‘core’ inflation,” Professor Buiter decides on a selective concept of core:

“The only measure of core inflation I shall discuss is the one used by the Fed, that is the inflation rate of the standard headline CPI or PCE deflator excluding food and energy prices. Other approaches to measuring core inflation… will not be considered.”

We added the emphasis because we want to contrast that comment with these words from President Lockhart’s speech:

“It is essential for those of us who have responsibility for responding to these trends to use a wide variety of core measures and inflation projections to make the most informed judgment we can.”

The variety of core measures is in fact wide. Some are familiar—the traditional statistics that exclude food and energy prices, the Cleveland Fed median CPI, and the Dallas Fed trimmed-mean PCE are examples. Some important, but less familiar, measures focus on persistence over time in individual price changes and exploit correlation over time in the common and product-specific components. Work by Michael Bryan and Steven Cecchetti and Domenico Giannone and Tyler Matheson  are examples.  An alternative approach is to define core inflation by decomposing headline inflation measures into permanent and transitory components, identifying core inflation as the permanent component. Examples include research by Jim Nason and James Stock and Mark Watson.

Michael Kiley just recently extended the Stock and Watson approach and found the common trend in inflation during the 1970s and early 1980s was attributable to persistent movements in both energy/food and nonenergy/food prices. More recently, that trend has been less influenced by food and energy inflation.

Maybe that isn’t all good news. It is noteworthy that the traditional measures of underlying trend inflation have moved higher over the past year or so, some more than others.

Consumer Price Index Year-Over-Year % Change

As President Lockhart noted at several points in his remarks yesterday:

“No matter how you measure it, the aggregate inflation we are experiencing in the United States at the moment is uncomfortably high…

“Measures of core inflation in the United States suggest that overall price pressures have been on the rise, perhaps because higher commodities costs have begun to affect prices paid by consumers and businesses across a broader range of other goods and services…

“I'm acutely aware that the current FOMC has inherited the inflation policy credibility that was hard won by our predecessors. One thing that has impressed me since taking my position last year is the seriousness with which my colleagues approach the duty to protect that legacy. I am confident that the Federal Reserve's institutional commitment to maintaining low and stable inflation will prevail.”

Professor Buiter raised several theoretical challenges to the core inflation concept that deserve discussion. It really is time, however, to lay to rest the straw-man assertion that central bankers are diverted by a pursuit of single and overly simplistic notions of core inflation.

By John Robertson, vice president in the Atlanta Fed’s research department, and David Altig, senior vice president and research director at the Atlanta Fed

August 28, 2008 in Federal Reserve and Monetary Policy , Inflation | Permalink

TrackBack

TrackBack URL for this entry:
https://www.typepad.com/services/trackback/6a00d8341c834f53ef00e5548faeba8834

Listed below are links to blogs that reference Monitoring the inflation bees, striving not to get stung :

Comments

Thanks for the excellent post gentlemen.

Thankfully, I am not so blind as to believe that the Fed is pursuing single and overly simplistic notions of core inflation.

I sincerely don't know how any person could read ANY of the analysis coming from the Kansas City, Cleveland, Atlanta or even Dallas Feds could come to such a conclusion. I'm simply baffled.

At any rate, I'm a little intriuged by your metaphors of bees in this post.

First, I'd like to ask if you gentlemen feel that it is appropriate to use some Wiener's work, the foundations of statistical mechanics, improving Einstein's modeling of Brownian motion.

Some have found statistical mechanics to be extraordinarily useful in modeling group behaviors containing stochastic processes.

Although I clearly am not a mathematician, my judgment is that statistical mechanics might be a very helpful supplement to vector autoregression analyses, especially if you think, as I do, that the swarming behavior of bees is a useful metaphor for modeling markets.

Any such models would then of course be supplemented by the examining the role of less-informed hornets (such as Buiter and Cramer) upon the direction of those swarms.

Secondly, can you point me to any papers that would expand the work of Nason and Stock and Watson to include migration of variables from transitory to more permanent expectations?

Although that is obviously a slippery concept, such math would attempt to model rational consumers like myself who KNEW after Volcker's Belgrade moment on October 6, 1979 that the NEXT long term trend in inflationary expectations was CLEARLY downwards and indeed that was the case for at least 29 years now.

So I would like to see the "stickiness" funtion measured of inflation expectations as well.

Please forgive my mathematical shortcomings obvious from the post. I am working sincerely and diligently to correct them.

Matt Dubuque

Posted by: Matt Dubuque | August 28, 2008 at 03:29 PM

Another way to frame what I seek is some links to quantitative analyses describing how the anchoring of inflation expectations changes over time, and why.

Thanks for the good work,

Matt Dubuque

Posted by: Matt Dubuque | August 28, 2008 at 05:35 PM

I'm not sure I understand why measuring the movement of the swarm of bees is of much practical use. In keeping with the agricultural analogies, to help illustrate my point; isn't measuring the swarm's movement akin to measuring the cows that have run out of the barn after the barn door was left open? It doesn't get us to the answer of how to keep the barn door shut, or how to keep the swarm from moving in one direction or another. Rather, measuring movements of the swarm simply tells us a problem is already occurring. There is nothing in this proposed excercise that allows us to stop the swarm from moving once it is in motion. In fact, the tools we have to stop the movement of the swarm are widely believed to work with a significant time lag. It seems to me that the debate should be over what can be done to keep the swarm from moving in the first place, rather than simply measuring it once it is already in motion. Said another way, should we not be focused on keeping the barn door shut rather than simply counting cows? Professor Friedman taught us how to do this, and Professor Taylor refined the approach into something quantifiable. Still, we avoid their lessons and somehow are focused on counting cows. I just don't understand the practical implications of this debate.

Posted by: Ken | August 29, 2008 at 11:56 AM

For me, the practical implications of this debate are how do we learn about and model differential inflationary expectations among consumers that "swarm" and aggregate around certain points and means, and co-vary in different ways.

Increased inflationary expectations feed through directly into elevated "real" interest rates that are so harmful to us all. Nobody wants this and our goal to faithfully fulfill the mandate of the Fed is not at issue.

In terms of modeling swarms of bees (and the math behind this can be quite sophisticated and is used by Google in server allocation planning in cloud computing scenarios, for example) the DIRECTIONALITY of the swarm (up or down) can be modeled with these tools.

Surely inflationary expectations can COLLAPSE as well as rise.

If the Fed DRAMATICALLY (and inexplicably) raised the target for the Fed funds rate to 36% tomorrow morning while simultaneously taking several other highly restrictive steps, surely inflationary expectations would collapse.

Inflationary expectations can BOTH rise and fall.

Bee swarms can also RISE and FALL.

Cows only move in two dimensions, within a planar topological space as it were.

ALL OTHER THINGS BEING EQUAL, adding another dimension to these analyses can make our tools more powerful.

Mathematics did not stop progressing in the 1960s, when Friedman published his best work. Fast Fourier transforms are probably the most obvious example of this to the lay person.

Why should we unduly halt new analytical frameworks with higher granularity from better informing our decisions and judgments?

Medical science has certainly adopted some of these new tools and so has the National Security Agency.

Let's not condemn the Fed to analytical tools that are becoming increasingly blunt instruments in this noisy and deregulated banking environment.

Matt Dubuque

Posted by: Matt Dubuque | August 29, 2008 at 06:03 PM

"It really is time, however, to lay to rest the straw-man assertion that central bankers are diverted by a pursuit of single and overly simplistic notions of core inflation."
As one of the 14 or so "strawmen" savers remaining, let me respond: NONSENSE, UTTER NONSENSE!
Welcome back, Dave A.

Posted by: Bailey | August 30, 2008 at 11:55 AM

Hello,

Just wanted to let you know, US month to month CPI/ Core CPI as well as annual CPI can be found under Prices and Indices tab in FXEconoStats (http://www.fxeconostats.com)
In case you want to compare CPI across multiple countries check the comparative charts section.

Posted by: Sogol | September 19, 2008 at 10:20 AM

I have to agree with Bailey, I don't think there are significant advantages to waiting around for your savings to mature.

Posted by: Mike J. Riley | November 07, 2008 at 05:09 AM

Post a comment

Comments are moderated and will not appear until the moderator has approved them.

If you have a TypeKey or TypePad account, please Sign in

Google Search



Recent Posts


Archives


Categories


Powered by TypePad