Day 2 of the Cleveland/Dallas Fed Conference on Price Measurement for Monetary Policy, and the topic on the table is measuring inflation expectations. Regular readers of macroblog know that I have particular affection for expectations derived from inflation-indexed Treasury securities, especially those adjusted for liquidity premia reported by the Cleveland Fed. In the Cleveland series the adjustments are, by design, quick and dirty. If, however, you have been hoping for a more sophisticated approach, now you have it for the U.S. courtesy of Stefania D'Amico (and co-authors from the Federal Reserve Board of Governors) and for the euro area thanks to Peter Hohrdahl (Bank of International Settlements) and Oreste Tristani (European Central Bank).
The methodologies of these two papers differ somewhat -- for the aficionados among you, D'Amico et al use a latent factor approach, while Hohrdahl and Tristani employ a more structural strategy -- but the conclusions are essentially the same: For both the U.S. and the euro area, extracting inflation expectations by comparing yields on inflation-indexed securities with those on non-indexed securities requires some sort of adjustment for liquidity or risk premia. The picture that tells the story for the euro area (based on French inflation-indexed bonds specifically):
The thing to focus on in that picture is the solid thin line, which represents expectations estimated without including adjustments for liquidity/risk premia, and the dark solid line, which represents the premia-adjusted expectations calculations. (The dashed lines delimit the statistical confidence intervals about the expectation estimate, and the dotted line represents consensus survey expectations.) The unadjusted series suggests that inflation expectations in the euro area have fluctuated considerably since 1999, and are at historically high levels today. The adjusted series tells quite a different story: Correcting for risk premia -- Hohrdahl and Tristani prefer the designation "risk premia" to "liquidity premia -- generates estimates of expected inflation that have been remarkably stable since the inception of the eurosystem.
The D'Amico et al analysis suggests a similar conclusion for the U.S. though, as with the simpler Cleveland Fed procedure, it appears that liquidity premia have all but vanished in the last several years:
The balance of the day moved from the extraction of expectations from market asset prices to discussions of expectations derived directly from surveys. More on that to come.