Karl Smith is thinking about inflation forecasts and directs us to the figure below. Here the daily nominal interest rate on 10 year treasury securities is graphed along with the daily real interest rate on 10-year Treasury inflation-protected securities (TIPS) :
Karl notes the difference between the two series is the bond market's 10-year forecast of inflation and currently it is about 2.5%. Interestingly, outside the financial crisis the forecast value is on average close to this 2.5% value. This can be seen more easily by plotting the difference between the two series as is done in this figure:
This caught my attention because if one goes and looks at the quarterly forecast for the average annual CPI inflation rate over the next 10 years in the Philadelphia Fed's Survey of Professional Forecasters the value is also around 2.5%. Here is the forecast data:
The similarity of inflation forecasts from the bond market and the survey is remarkable.* Even more remarkable is that this long-run value of 2.5% has persisted since the late 1990s and has not been seriously affected by the crisis. It suggests that this value is the Fed's long-run inflation target. As I have said before, the stability of this value attests to the Fed's inflation-fighting credibility. Bond markets, apparently, see no unanchoring of inflation in the future. Of course, this stability of the long-run inflation target also lends support to the view that the Fed was been too tight with monetary policy over the last year and half or so. And on the flip side such a rigid long-run inflation target may also present problems when there are sustained productivity gains.
Update: Peter N. Ireland sheds some light on the Fed's target inflation rate.
*Yes, there is a big difference between the two series during the financial crisis, but much of that may be due to the sharp rise in the liquidity premium on TIPS during this time. Also, the bond market data is daily while the survey is quarterly so some of the day-to-day variation is lost in averaging for the survey.
Karl notes the difference between the two series is the bond market's 10-year forecast of inflation and currently it is about 2.5%. Interestingly, outside the financial crisis the forecast value is on average close to this 2.5% value. This can be seen more easily by plotting the difference between the two series as is done in this figure:
This caught my attention because if one goes and looks at the quarterly forecast for the average annual CPI inflation rate over the next 10 years in the Philadelphia Fed's Survey of Professional Forecasters the value is also around 2.5%. Here is the forecast data:
The similarity of inflation forecasts from the bond market and the survey is remarkable.* Even more remarkable is that this long-run value of 2.5% has persisted since the late 1990s and has not been seriously affected by the crisis. It suggests that this value is the Fed's long-run inflation target. As I have said before, the stability of this value attests to the Fed's inflation-fighting credibility. Bond markets, apparently, see no unanchoring of inflation in the future. Of course, this stability of the long-run inflation target also lends support to the view that the Fed was been too tight with monetary policy over the last year and half or so. And on the flip side such a rigid long-run inflation target may also present problems when there are sustained productivity gains.Update: Peter N. Ireland sheds some light on the Fed's target inflation rate.
*Yes, there is a big difference between the two series during the financial crisis, but much of that may be due to the sharp rise in the liquidity premium on TIPS during this time. Also, the bond market data is daily while the survey is quarterly so some of the day-to-day variation is lost in averaging for the survey.
If a market participant expects inflation to be 2.5%, then he would buy bonds whenever the spread is higher than that or buy TIPS when the spread is lower, and he would be indifferent when the spread is exactly 2.5% - or would he? If you expect 2.5% inflation, and find that the market spread is exactly 2.5%, do you split your investment evenly? Does the ability of the TIPS to preserve purchasing power in a high inflation period factor in, eg meriting a price premium - or even a discount in a deflationary situation? If TIPS merit a price premium most of the time, then a market spread of 2.5% would indicate a real inflation expectation higher than 2.5%, say 3% with a 0.5% yield premium paid to the seller. I'm curious to hear the blogger's view on this. I personally observed a market premium on deflation protected TIPS (new issues with almost no built-in principal growth) during the crisis-period deflation scare. I "feel" like there is a priced-in inflation premium right now. What is the academic perspective?
ReplyDeleteAnonymous:
ReplyDeleteYes, there should be a liquidity premium on the TIPs--a point I mentioned in the footnote--which makes it all the more remarkable that the two measures of long-run expected inflation are so similar.
I don't have any estimates of how big the liquidity premium is now, but obviously it is less than it was during the peak of the financial crisis. The Cleveland Fed has a great resource that provides sophisticated estimates of the liquidity premium. Unfortunately, for the time being they have quit making these estimates: http://www.clevelandfed.org/research/data/tips/
Is there a website where I can get up-to-date plots of inflation expectations based on TIPS prices? In other words, where did you get the plots in your posting?
ReplyDeleteBrad:
ReplyDeleteProbably the easiest way to get the data for free is to go the Fred database:http://research.stlouisfed.org/fred2/
Look under the interest rate link for the data.