Two inflationary scenarios
Continued supply shocks and input cost pass-through vs a transitory inflation with little input cost pass-through
In my post yesterday I looked at how reliable the Atlanta Fed’s sticky/flexi price framework is for forecasting and then I did some forecasting.
In doing so I found something interesting: during the recent inflation sticky prices have not been responding as much to rises in flexi prices as they have historically. I think the most reasonable interpretation of this is that manufacturers and retailers are reluctant to raise prices for end consumers in the face of rising input costs. Basically, they are trying as hard as they can to not pass through rising input costs.
This raises the possibility that they might eventually throw in the towel and pass on the input costs. If this were to happen, inflation could accelerate markedly. How much? This is something I left to one side yesterday. But let’s pick it up today.
We are going to run two scenarios.
Flexi prices continue to accelerate. Perhaps because supply chain problems continue due to conflict in Eastern Europe and lockdowns in China. Meanwhile, sticky prices respond to previous flexi price increases as they have historically. This would mean that manufacturers and retailers throw in the towel and pass on rising input costs to consumers.
Flexi prices revert to normal levels by the end of the year. Basically, the supply chain issues evaporate. Meanwhile, sticky prices continue to respond to rising flexi prices in the same muted fashion we have seen until now. So, manufacturers and retailers continue to keep a lid on end consumer prices and are buoyed by the decline in flexi prices.
Here is what the two scenarios look like in the Atlanta series.
These two forecasts — one representing an explosive scenario and the other representing a transitory scenario — can then be used to predict what happens to the CPI index overall. Here are the two forecasts.
Doing two extreme cases allows us to map out a whole field of possibilities. If the assumptions of this modelling holds, then future inflation should fall within those two bounds depending on what happens in the future with respect to supply chain disruptions and the capacity for manufacturers and retailers to keep a lid on end consumer prices.
Until now retailers have succeeded in not passing on most of the rising input prices to consumers. If input costs keep rising and they are forced to pass these on, we could see CPI rise to over 20% year-on-year growth. Scary.
Is that the most extreme scenario we can squeeze out of the data? No. As I keep stressing, sticky prices are not responding to rises in flexi prices as they typically do. This gives rise to a very different regression than would be the case if we used the full time series.
Historically — i.e. using the full time series — a 1% rise in sticky prices has led to a 0.32% rise in CPI while a 1% rise in flexi prices has led to a 0.68% rise in CPI. Since the start of the present inflation, a 1% rise in sticky prices has led to a 0.79% rise in the CPI while a 1% rise in flexi prices has led to only a 0.26% rise in the CPI.
Again, this is because retailers and manufacturers seem reluctant to pass price rises through to end consumers. So, the flexi prices — which are mostly inputs — are not getting as much traction on the CPI as they have historically. This means that if we want a true extreme case we should use the regression coffeicients from the full time series to show what might happens if retailers and manufacturers stop trying to keep a lid on end consumer prices. Here are the results:
Now that is scary. Not my basecase scenario, to be sure. But the very fact that we can generate these forecasts using reasonable assumptions should in itself be cause for concern.