Expectations of inflation are a “bitch”. What to make of them?

Gauging the state of inflation expectations is not a trivial enterprise. The hawks, for example, “expect” inflation to rise and oppose more “stimulus”, regardless if unemployment is high. The “doves” think the opposite: more “stimulus” is required and inflation is not a worry. On the other hand, in a depressed economy, a rise in inflation expectations is a positive sign for the future level of activity, drives asset markets and improves the debt situation.

Expectations of inflation are not “observed”, they have to be assessed or estimated. And as always different ways of assessing or estimating expectations will produce different and sometimes “contradictory” results.

When I wrote this post yesterday I put up the following picture:

After posting it late at night, just as I went to sleep I wondered why short run (1 year) inflation expectation – as gauged by the Cleveland Fed – was so much more volatile than their estimated long term (10 years) inflation expectations. Coincidentally today I read this answer by Scott Sumner to a comment of mine in his blog:

Marcus, I am going to offer a challenge to someone. Explain the intuition behind the Cleveland Fed approach. Not the technical stuff, the intuition. Those inflation expectations for 2 and 3 years out seem wildly implausible to me. Where do they get them? Why are they so dramatically different from the TIPS spreads? Why does the Fed itself completely ignore them?

So I´ll give it a try, avoiding any discussion of “technical stuff”. And I´ll do it “my way”, that is through illustrations.

First up is the previous chart covering a longer period accompanied by the chart that depicts headline and Core inflation.

One reason for the greater “volatility” of short term (1 year) expectations might be that they reflect temporary shocks (things like oil and commodity prices) that affect current (or short run) inflation. When you look at the chart that depicts headline and core CPI inflation you see that headline inflation is much more volatile than core inflation. If inflation is “under control” (i.e. the Fed has “credibility”) one would expect that those effects will only be temporary (“noise”), not affecting inflation in the longer run. That´s why many recommend that the Fed “target” Core and not Headline inflation, since the former better describes the inflation trend.

It is natural, then, if the Fed has credibility, to assume that short run inflation expectations will be more volatile than longer term expectations. And they are, so the Cleveland Fed estimate of inflation expectations is not inconsistent with that “empirical regularity”.

I don´t have 1 year TIPS spreads to gauge the behavior of short term inflation expectations from those spreads. But we can examine the 5 and 10 year inflation expectations from both the Cleveland Fed and TIPS spreads and see how they reflect the “underlying story”.

Both measures of inflation expectations trend up after 2003. That´s the moment when “fears of deflation go away” with nominal spending rising so as to bring the economy back to trend, which happens to occur in 2006 (see chart below). By that time TIPS spreads had already flattened but the Cleveland Fed measure of inflation expectations only then stops increasing.

The TIPS spread measure stays “flat” all through to the moment (mid 2008) that nominal spending “drops through the floor”. On the other hand, the Cleveland Fed measure reverses direction in mid-2007, as soon as the fall in house prices (which had peaked in early 2006) begins to cause “trouble for the financial sector” AND nominal spending growth starts to fall below trend.

So far, I believe the measure of inflation expectations gauged by the Cleveland Fed better reflects the “underlying reality”, but let´s plow on.

When “push comes to shove”, inflation expectations drop strongly, more so under the TIPS spread measure. That seems to indicate that “agents” were “caught by surprise” and drastically reviewed their expectations. But expectations of inflation also reverse more rapidly in this case.

Here again I see the Cleveland Fed (CF) measure as being more “realistic”. Note that the CF expectations only reverse when something actually happens to indicate the need for reviewing expectations. That “something” was the implementation of QE1 in March 2009.

From that point, both measures reflect the “switch off” of QE1 and later the “switch on” of QE2 and still later the “switch off” of QE2. After that, inflation expectations fall under both measures, and stop falling when the Fed introduced “changes in communications” and “threats” of a new round of “stimulus” (this is shown in the “ellipse” drawn at the end of the chart).

At the last point of the inflation expectations chart above, 10 year TIPS spreads are indicating that the markets expect inflation that is even a bit above target. That has led Scott Sumner to write in the above mentioned post:

I believe that Bernanke has come to the realization that if the US is going to get a robust recovery, we will need a bit higher inflation. And I think this is starting to occur.

The “low rates until 2014″ policy is pretty meaningless as officially stated, but in my view the markets are able to read between the lines, and see that Bernanke is actually signaling; “we will hold rates near zero at levels of inflation and real growth that would have normally triggered rate increases.

Expectations gauged from the Cleveland Fed seem to be indicating something quite different. The interpretation from the last Fed statement could be:

We expect the economy will remain weak going forward and so that everyone is on the same page we will keep the FF rate at exceptionally low levels for a long time. Or, even more “transparently”: Don´t expect us to act to change that expectation!

So yes, I prefer the CF measure of inflation expectation. My feeling is that the implicit or explicit 2% inflation target is an “attractor” to the TIPS spread (recall that I´m not considering technical details such as inflation risk premium), irrespective of the actual state of the economy and/or the “errors” having been committed by the Fed.

On the other hand the CF measure doesn´t appear to “suffer” from this “bias” and indicates that “optimism is not warranted”.  And so the Fed “ignores” that measure of inflation expectations at its own peril!

4 thoughts on “Expectations of inflation are a “bitch”. What to make of them?

    • Lulu
      I just checked for the 1 year horizon. My view is that expectations of inflation over the next 12 months are pretty much “dependent” on what current and recent 12 month inflation turned out to be.
      As far as implications for CB policy, I´m very skeptical of IT. It´s a “credibility” game that sometimes (when supply shocks hit) gets a bit “messed up”.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.