How To Estimate “Rational” Market Expectations Of Future Inflation

by
Barkley Rosser
My articles About meMy booksMy videos
Follow on:

 I am not a fan of rational expectations, hence the quotation marks around "rational" in the subject head here.  Nevertheless I have become aware thanks to some posts at Econbrowser by the intrepid Menzie Chinn that the usual way this has been measured and reported by most people needs to be modified, with the understanding of this only developing quite recently.  This came from a paper in 2018 by some Fed Board of Governors economists: S. D'Amico, D.H. Kim, and M. Wei, (although Menzie refers to it as the "DKKW model"), "Tips from TIPS: The Informational Content of Treasury Inflation-Protected Inflation Securities," Journal of Financial and Quantitative Analysis, 2018, 53(1), 395-436. 

For a given time-horizon, it has been conventional for those estimating such a "rational" market forecast of expected inflation to take the appropriate Treasury security nominal yield of that time horizon (say 5 years) and simply subtract from it the yield on the same time horizon TIPS, which covers security holders for inflation.  So it has long looked like this difference is a pretty good estimate of this market expectation of inflation, given that TIPS covers for it while the same time horizon Treasury security does not.

Well, it turns out that there are some other things involved here that need to be taken account, one for each of these securities.  On the Treasury side, it turns out that the proper measure of the expected real yield must take into account the expected time path of shorter term yields up to that time horizon.  This time path has associated with it a risk regarding the path of interest rates throughout the time period.  This is called the Treasury risk premium, or trp.  It can be either positive or negative, with it apparently having been quite high during the inflationary 1979s.

The element that needs to be taken into account with respect to the TIPS is that these securities are apparently not as liquid in general as regular Treasury securities, and the measure of this gap is the Liquidity premium, or lp.  This was apparently quite high when these were first issues and also saw a surge during the 2008-09 financial crisis.  In principle this can also be of either sign, although has apparently been positive.

Anyway, the difference between the nominal T security yield and the appropriate TIPS yield is called the "inflation breakeven," the number that used to be focused on as the measure of market inflation expectations.  But the new view is that this must be adjusted by adding (tpr - lp).

In a post just put up on Econbrowser by Menzie the current inflation breakeven for five years out is 2.52%. But according to Menzie the current (tpr - lp) adjustment factor is -0.64%.  So adding these two together gives as the market expected inflation rate five years from now of 1.88%, although Menzie rounded it out to 1.9%.

If indeed this is what we should be looking at it says the market is not expecting all that much of an increase in the rate of inflation from its current 1.7% five years from now. The Fed and others are looking at a short term spike in prices this year, but the market seems to agree with their apparent nonchalance (shared by Janet Yellen) that this will wain later on, with that expected 5 year rate of inflation still below the Fed's target of 2%.

Certainly this contrasts with the scary talk coming from Larry Summers and Olivier Blanchard, not to mention most GOP commentators, regarding what the impact of current fiscal policies passed and proposed by Biden will do to the future rate of inflation.  Not a whole lot, although, of course, rational expectations is not something that always forecasts all that well, so the pessimists might still prove to be right.

Barkley Rosser