Do inflation expectations matter for inflation?

Many people (NYT) are talking about the new paper by Jeremy Rudd on exactly this topic — Rudd is skeptical that they matter very much.  So I went to read the paper, and I have to say I am baffled.  It didn’t change my priors at all.  I didn’t see new empirical estimates, or new theoretical arguments, and furthermore I didn’t see the most relevant factors discussed much.  I did see a lot of pokes at Friedman, Phelps, and Lucas (and there is also an introductory assertion that, even given enough time, markets with flexible prices do not clear.  Then he goes on to deny that the theory of household choice is sufficient to derive downward-sloping demand…why do that???).

I would start by looking at the clearest cases where inflationary expectations do matter.  If inflation rates become quite high (say above forty percent?), many people switch to alternate currencies.  Or in hyperinflations the velocity of money, after some point, accelerates very dramatically, thereby fueling the inflation further.  So inflation expectations really do matter, as supported by both theory and evidence.  Contrary to Rudd, the theoretical case is there, though the question of magnitude at lower inflation rates is largely an open one.  But let’s not slip from “open” to “we are justified in thinking they don’t matter very much at all.”

You can even put aside velocity and the demand for money and the theory for inflation expectations mattering still is there.  For instance, if I look at the simple and fairly general menu costs models, suppliers know they won’t be changing their prices very often.  So they form an expectation of inflation when deciding where to put the price right now, knowing that level will have to be “close enough” for some time to come.  That’s a very simple theoretical mechanism!  No, that is not a priori but it carries a lot of force and Rudd does not consider it.

If I ask “which papers do I, as outsider, already know in this area?” I hit upon Michael F. Bryan, “The inflation expectations of firms: What do they look like, are they accurate, and do they matter?”, 2015, not so ancient and from the Atlanta Fed.  There is lots of careful estimation in this paper, and here is part of the abstract:

Next we show that, during our three-year sample, firm inflation expectations appear to be unbiased predictors of their year-ahead observed (perceived) inflation. We also show that firms know what they don’t know—that the accuracy of firm inflation expectations is significantly and negatively related to their uncertainty about future inflation. And lastly, we demonstrate, by way of a cross-sectional Phillips curve, that firm inflation expectations are a useful addition to a policymaker’s information set. We show that firms’ inflation perceptions depend (importantly) on their expectations for inflation and their perception of firm-level slack.

That doesn’t prove that inflation expectations matter a great deal, but it is certainly consistent with them mattering, as theory would lead us to expect.  It seems to go much further than Judd, and it is not cited by Judd.

Or how about the NBER survey by Olivier Coibion, “Empirical evidence suggests that inflation expectations of households and firms affect their actions but the underlying mechanisms remain unclear, especially for firms.”  There is a whole section of the paper “Do inflation expectations affect economic decisions/”, with the answer mostly being “yes,” in part with the caveat that inflationary expectations are hard to measure precisely.

That all is consistent with my rather basic understanding of the matter.  Again, this paper is not cited by Rudd.  It’s not that I think a contribution has to cite every paper out there, but when you play the “no estimation of my own, just going to poke holes in various claims, and focusing from people decades ago…”…a reader such as I is going to want to see you cite and rebut the main recent attempts to establish relevance for inflationary expectations.

And here is commentary from Joseph Gagnon.  Here is commentary from Ricardo Reis, some good points though I think he overstates the indeterminacy issue.

Most of all, I don’t feel I have a horse in this race.  I am very comfortable with the idea that we, as economists, have a poor understanding of regime shifts, including shifts in inflation regimes.  But when I am told that, as a result of agnosticism, I should infer inflationary expectations are not very relevant…then I feel someone has jumped the gun.


Comments for this post are closed