In few words, our main result is that the broad inflation regime that households expect matters much more than the exact level of inflation they report. More specifically, we emphasize the difference between households expecting prices to remain stable or to increase.
This extensive margin of inflation expectation is important to understand the way households form their expectations. In particular, the share of households expecting prices to remain stable (0% inflation) is large. Variation of average inflation expectations heavily relies on variation of this extensive margin.
This extensive margin is also key to understand the connection between inflation expectations and consumption decisions. Expecting prices to increase leads to a higher propensity to consume durable goods and we find a significant relationship between inflation expectations and consumption only along this extensive margin. Durable good consumption is important since it accounts for most of the variations in overall consumption.
By contrast, differences in the precise quantitative inflation rate households expect— the intensive margin – matters much less: households with different positive quantitative inflation expectations make similar durable consumption choices.
We also find evidence that the extensive margin has a positive impact on households’ willingness to spend on durable goods in US and German surveys of households. In particular, our conclusions hold true even when considering US data from 1978-1984 – a period known for its high inflation rates.
But why do we care about the importance of such extensive margin?
First, we show that this attenuates a lot the inflation expectation channel compared to a standard New Keynesian model – which is the workhouse model in macroeconomics to think about business cycle fluctuations. A back of the envelope calculation suggests that this channel is attenuated by a factor of 7!
This is important as the New-Keynesian model is known to lead to a very (too) strong inflation expectation channel leading to e.g., the forward guidance puzzle. This puzzle refers to the prediction that interest rate cuts or higher inflation in the future lead to large effects on current aggregate demand, which does not receive support when confronting the model to the data.
Second, policies relying on inflation expectations to stimulate current aggregate demand may run out of ammunition: once households all expect positive inflation, they no longer adjust their consumption decisions – at least if they do not change their views about regimes of inflation.
Overall, our results imply that policies aiming at stabilizing demand through their impact on inflation expectations may be much less effective than what standard macro models predict.
Our results also have implications for our times of higher inflation taken literally, our results suggest that the rise in inflation expectations should have only limited effects on current aggregate demand and, hence, on current inflation. Of course, this holds true to the extent that households do not change the way they connect inflation and consumption decisions.
Authors’ note: The views expressed here are those of the authors and do not necessarily represent those of the institutions with which they are affiliated.