Coibion and Gorodnichenko on the missing disinflation

Here is their abstract (pdf):

We evaluate possible explanations for the absence of a persistent decline in inflation during the Great Recession and find commonly suggested explanations to be insufficient. We propose a new explanation for this puzzle within the context of a standard Phillips curve. If firms’ inflation expectations track those of households, then the missing disinflation can be explained by the rise in their inflation expectations between 2009 and 2011. We present new econometric and survey evidence consistent with firms having similar expectations as households. The rise in household inflation expectations from 2009 to 2011 can be explained by the increase in oil prices over this time period.

Writing on the paper, here is Jim Hamilton’s bottom line:

The phenomenon identified by Coibion and Gorodnichenko would undermine the Fed’s ability to stimulate the economy in a number of important respects. First, it makes it much more difficult for the Fed to try to justify its actions to the public on the grounds that inflation is currently too low. Second, if makes it harder for the Fed to stimulate the economy without raising inflation, particularly if one byproduct of stimulus efforts is an increase in the relative price of oil. Third, it implies that ex ante real interest rates, if we base that concept on the perceptions of large numbers of economically important decision makers, are extremely negative at the moment, casting doubt on the claim that a primary policy objective should be to make them even more negative.


More evidence in favor of modest supply-side factors in the recovery.

Yay! Economics has finally caught up with what normal people have known for years!

Normal people have known that they have been over-weighting prices at the pump in their perceptions of inflation?

Well, my take on it was more along these lines: here are a couple of economists who are actually acknowledging the "absence of a persistent decline in inflation." And normal people have been forced to acknowledge this absence during the vast majority of the recession.

(And even if I take into account the declining prices of real estate, I see that a great fraction of the normal people in our economy have been unable to benefit much because of stricter loan standards and competition from all cash buyers.)

Sigh...again: expectations of inflation are not inflation.


(1) at the right data (i.e. not just consumer price inflation but at domestic demand inflation which also captures the prices of investments and government consumption instead of just household consumption (government consumption being the price of education and the like paid for by the government)
(2) and taking account of long and variable lags in the economy (the ECB estimates that one of these lags, i.e. the rather imperfect relation between money and inflation) is about five to eight years)

it seems that there has been persistent, although limited, disinflation. By the way - domestic demand inflation in for instance the UK is way lower than consumer price inflation. A graph on Eurozone inflation:

Consumer price inflation in Spain is at the moment -0,1%. That's 3,6% disinflation in one year. And the Troika is pushing for more.

See also their summary on Vox: which encouraged me to write this comment.

First this is a paper worth thinking about ... the formation and adjustment of inflation expectations is central to conventional money/macro theory and policy ... and yet we know relatively little about empirical reality, particularly at the individual level. So I am happy to see this paper but am less happy to see policy implications yet, as in Hamilton's post.

I also use the Michigan survey data in my research and have even had the opportunity to listen some of the interviews. I am convinced from my research, the work of others, and some macro forecasting that this type of household-expectations data has signal, but it also has noise -- and it's probably not just random wackiness that averages out in a representative sample. One needs to exercise a lot of care with the inference and this paper makes some good first steps.

Some comments/critiques:

1) There should be more exploration of the various inflation measures on the survey. For example, it is more common to use the median not the average of the reported inflation expectations in macro models. There are some extreme outliers and imputations in the data (these are not easy questions) see Curtin: So does using the median change the story any? And why should you take the average response in the household sample, shouldn't it be weighted by consumption shares? If the lower income respondents drive the covariance with oil prices, then it would be muted when consumption-weighted aggregate measure. And why not use the gas price expectation questions? Or the qualitative questions on info used in price responses? There's more information in the survey to support or shade the current micro results.

2) More conceptually, why use the year-ahead inflation expectations and not the five-to-ten-year ahead ones? The latter show much less co-variation with oil/gas prices and has moved in a pretty narrow range. I would have thought theory should favor using the longer horizon measure (as is the basis for the models in:, but maybe there's a case to be made for the short-term measures.

3) Also it would be more convincing if you found this mechanism in other episodes or countries. It seems like the repeated cross sections and the short-panels going back in the US or episodes in other countries would provide better identification ... with the one case of this business cycle and commodity price movements you have to worry that it's just coincidence. Can you explain the anchored survey expectations in Japan during deflation?

4) I liked the point about the professional forecasters (economists with standard models) having different inflation forecasts than households and maybe firms (who may well be using different 'models' than economists). Would be interesting to see how that holds up and what's really behind it.

Very cool paper, very important questions ... I would call the conclusions still speculative but worth exploring.

Comments for this post are closed