Introductory Remarks to the Price Measurement for Monetary Policy Conference

May 24, 2007 Dallas, Texas

This conference on inflation measurement has been jointly organized by Mark Wynne and Jim Dolmas of the Dallas Fed and David Altig and Michael Bryan of the Cleveland Fed. For their dedicated effort and superb accomplishment in getting participants and presenters from over 21 countries, I am most grateful.

The presentations today are devoted mainly to the question of core inflation measurement, while the presentations tomorrow will look at various approaches to the measurement of expected inflation.

During the next two days, we will hear presentations by leading students of the subject—academics and economists from the Federal Reserve System and participants and presenters from a number of other central banks, including the European Central Bank, Bank for International Settlements, Bank of Italy, National Bank of Belgium, Reserve Bank of Australia and National Bank of Poland. Governor Mishkin will be the keynote speaker at tonight’s dinner.

The discussants, who play an equally important role in determining the success of a conference like this, have been drawn from the Chicago, Kansas City and San Francisco Feds, as well as from the Bank of Canada, IMF, Bank of Mexico, Bank of Sweden, Brandeis University, the University of Illinois and London Business School. We also have people attending from the Reserve Bank of New Zealand, Bank of Iceland, Bank of England, Swiss National Bank, National Bank of Romania, Bank of Korea and the Central Bank of the Netherlands, among others, along with many local area academics.

All in all, we have a very diverse group participating in this conference. For those of you who have come from afar, we welcome you. Texas is a very friendly place. We are honored that you are here, and we hope you will enjoy our hospitality.

The first paper this morning will be presented by Diana Weymark of Vanderbilt University. The paper looks at ways of constructing operational measures of inflation pressure and specifically proposes three new indexes for measuring such pressure, the extent to which such pressure is alleviated and the effectiveness of monetary policy in reducing expected inflation. One of the interesting findings of the paper is that the proposed indexes show the extent to which the Fed under Chairman Greenspan consistently resisted inflationary pressures and capitalized on deflationary pressures to bring inflation down whenever the opportunity arose.

Professor Weymark’s paper will be followed by a paper by Julie Smith of Lafayette College. Professor Smith is one of the few academic experts who continue to think about ways of coming up with better measures of core inflation, and indeed, that is the title of her presentation. In her view, core inflation should be defined as the measure of inflation that best predicts future inflation. The central idea in her paper is that core inflation should be measured by adding up the components of the PCE deflator, but weighting them according to how they behave over time.

One of our main criticisms here at the Dallas Fed of much of the core inflation literature is that it lacks theoretical coherence. It reminds me of the time-honored saying that an economist is someone who sees something work in practice and then wonders if it can work in theory. Today’s third paper, presented by Stefano Siviero of the Bank of Italy, tries to address this issue by proposing a measure of core inflation that weights individual prices on the basis of their usefulness in helping the central bank achieve its goal of price stability. One of the interesting findings of the paper is that a central bank that responds to the traditional “ex-food and energy” measure of core inflation does no better than one that responds to headline inflation. This raises important questions about the utility of the core measure in policy deliberations.

Measuring core inflation can get quite technical, and the fourth paper for today, presented by Richard Anderson of the St. Louis Fed, is an excellent example in this regard. I won’t try to summarize it for you unless you have a strong desire to learn about “wavelets” from someone whose closest encounter with them was while surfing as a teenager. The central idea in the paper goes way back to the great 19th century Irish economist Francis Edgeworth, who proposed that when calculating overall inflation, individual prices should be weighted not by their share in aggregate expenditure but rather by their volatility, with more volatile prices getting a smaller weight in the overall index.

The trimmed mean approach to core measurement, which is the way the Dallas Fed prefers to evaluate inflation, is analyzed at some length in the two papers that follow. One, by a pioneer in this line of research, Mike Bryan of the Cleveland Fed, argues that the trimmed mean approach is particularly good at detecting shifts in trend inflation in low-inflation environments. The other, by Anthony Richards of the Reserve Bank of Australia, provides additional evidence on the superior performance of trimmed mean measures using data for the U.S., euro area, Japan and Australia. One innovation in this paper is to break up the owners’ equivalent rent (OER) component of the U.S. CPI along regional lines to further improve the performance of the trimmed mean.

But lest you think that that settles the debate in favor of the trimmed mean approach, the last paper today, by New York Fed economists Robert Rich and Charles Steindel, casts doubt on the usefulness of any core measure. Their point is that the superior performance of various measures is very sensitive to the sample period being looked at; some measures do well in some periods and some do well in others, leaving little basis for trusting in one measure at all times.

Tomorrow, the conference will look at various approaches to the measurement of inflation expectations. The first two papers, to be presented by Stefania D’Amico of the Board of Governors and Oreste Tristani of the ECB, look at how we can use the prices of financial instruments to make inferences about inflation expectations. Dr. D’Amico and her fellow coauthors at the Board of Governors find that the TIPS break-even rate is a useful proxy for inflation expectations in the U.S., despite the fact that the yield on TIPS is known to contain a significant liquidity premium. Dr. Tristani and his coauthor from the Bank of International Settlements are concerned with the interpretation of term premia in the euro area and find that these premia seem to reflect primarily real rather than inflation risks.

An alternative approach to measuring inflation expectations is to simply go out and ask people what they think inflation is going to be over the next year or two or 10. Our very own Mark Wynne spent a good deal of time working at the European Central Bank on the European Monetary Union. One of his enduring contributions to the EMU was to design the ECB’s Survey of Professional Forecasters, which is now featured every three months in the ECB’s Monthly Bulletin. Two papers on Friday, one presented by economist Juan Angel García of the ECB and another by Aidan Meyler, also of the ECB, will look at the usefulness and track record of this survey and the similar one that is run by the Philadelphia Fed. Well-designed surveys of forecasters can shed light on inflation risks, a topic of perennial interest to central bankers.

Of course, it is not just the expectations of professional forecasters that should be of interest to central bankers. The inflation expectations of firms and households matter as well. The penultimate paper of the conference, by three economists from the Polish central bank and presented tomorrow by author Ryszard Kokoszcynski, will look at the performance of measures of household inflation expectations in a group of central European countries. One of the problems with the European Commission’s survey of household inflation expectations is that it elicits qualitative rather than quantitative information, and these responses then need to be further “massaged” before they can be used in assessing the inflation outlook.

In addition to asking households about inflation expectations, the European Commission survey also asks about inflation perceptions. Normally the perceived rates of inflation in the EU and the rates measured by statisticians track each other quite well. But at the time of the euro cash changeover, there was a huge discrepancy between what the statisticians were measuring and what households said they were experiencing, raising questions about the credibility of the official inflation measures. The final paper of the conference, presented by Marianne Collin of the National Bank of Belgium, looks at what happened and concludes that there was simply a structural break in the relationship between measured and perceived inflation at the time of the cash changeover.

That is the menu for this feast of thought. I hope you enjoy the discussions and that all of us who are privileged to partake in this conference will leave with renewed vigor and determination to better understand inflation and use that understanding to inform those of us who have responsibility for making monetary policy.