(For this article, I will assume the conventional view that interest rate policy can be used to control inflation. I have severe concerns about this, but if we do not assume that the central bank can fine tune growth to match arbitrary trajectories, the story is a bit more plausible.)
My view is that the basic inflation target framework followed by some countries is the most practical version. If we assume a 2% central target, just say that the central bank definitely wants to keep inflation between 1% and 3%, and be vague after that. (Yes, the central banks typically have more strict targets, but since nothing happens when they miss, that is no big deal. If the target was supposed to 2%, just forecast that inflation will be 2% at the end of your forecast horizon.)
The alleged problem with the 2% target is that if the 2% level is treated as a ceiling during the expansion, then the average over the cycle will tend to be below 2% -- since inflation drops in a recession, and recovers slowly.
The Fed dealt with this problem by adding a vague averaging clause. However, this was only an issue in the last cycle -- inflation bounced above target during the expansion in various countries in earlier cycles with inflation targeting. (Pre-2008, central bankers in a few countries were crowing at what a great job they did hitting their target on average.)
The only practical difference of this change is that this gives the Fed more room to wait on rate hikes. However, that really is only going to be true if the inflation experience repeats. It is not a law of nature that inflation must stick below 2% for a decade. Meanwhile, if inflation is in fact stable at a low level as in the last cycle, then the rise in rates would be mild, and probably would have no measurable effect on the economy.
Although the policy has been described as an average inflation target, it was unclear to me how symmetric the averaging is. (An earlier document only discussed deviations below a 2% average.) If inflation has averaged 2.5% for an extended period, will the central bank attempt to drive it down to 1.5% to balance this out? Neoclassical models suggest that the central bank can make the inflation rate follow a complex trajectory. I have doubts that the inflation rate can be fine-tuned that easily amid an expansion -- disinflations tend to be driven by recessions. Is the central bank going to risk inducing a recession if the inflation rate is stable at 2% in order to meet the average inflation mandate? If not, then we have to accept this is not really average inflation targeting, this is just a tweak to try to avoid the problems of the last cycle.
A simpler alternative to averaging is to just to tell the central bank to be happy with any inflation rate between 2% and 2.5%. If you want an objective function, it has a dead zone. This means that the central bank will sort-of target 2.25%, but will not care if it drifts from that precise target. I would not view this as being much of a change, as this sort-of described the inflation trajectory in the pre-2008 cycles. (After all, some of the central banks hit their inflation targets on average.)
Allowing for 2.5% instead of 2% inflation might scare some people, but they are being unrealistic. Adding 0.5% to the average realised inflation rate only would make a difference on a long horizon -- a decade or more. The shift of the consumption basket means that price level comparisons over such horizons are sketchy anyway. There is no sign that any entity is sensitive enough to the CPI to care about such a small difference. Firms are worried about their input costs (including labour), and their specific outputs. People fret about inflation, the existence of websites that make money telling people that "inflation is really 8%" tells us that they have no idea what the actual rise in prices is.
(c) Brian Romanchuk 2020