As a rule of thumb, central banks believe it takes 12 to 18 months for policy changes to take effect. This means they rely on forecast models to guide policy. If the consensus forecast suggests inflation is likely to overshoot the target in 18 months this would generally indicate an increase in interest rates should be implemented soon.
The notion that action today is directed toward where we expect the target to be in the future (and not toward where it is now) is common in everyday life. When passing a ball to a running teammate we aim for where we expect them to be when the ball arrives, we don't aim the ball at where they are when we execute the kick or throw.
Consequently, central banks devote a good deal of attention and research to understanding the "transmission mechanism" of monetary policy--how changes in short term policy rates effect changes in the variables they care about, inflation and employment. Once the structure of the relationship between policy rates and the real economy is well understood and parameterized, central banks embed this knowledge into forecast models.
Chairman Powell made an important admission today, 8/27/2020. Current models relating employment and inflation have for a significant period of time been overpredicting inflation. In particular, the models have been predicting higher wage growth and inflation than have actually been associated with the strong labor market the US experienced in the several years before the outbreak of COVID-19. The overprediction of inflation has been consistent and not confined only the United States.
One consequence of relying on forecast models that overpredict inflation is that central banks will tighten policy too soon thus curtailing economic expansion unnecessarily. The FOMC is recognizing this possibility and is indicating that it will keep interest rate increases on hold even when forecast models predict it will overshoot its inflation target some time in the future--let us say 18 months away. If it is right to keep policy settings unchanged, the economic expansion can continue and inflation will remain below or at target. If it is wrong, the FOMC has expressed its willingness to live with above target inflation for a time.
Although one could get very technical with discussing exactly what "flexible average inflation targeting" means, and how to set it up in a mathematical model, in practice this will amount in my opinion to following a version of the US Revolutionary War maxim "don't fire until you see the whites of their eyes". While we certainly should not expect the Fed to wait until inflation is above target to raise rates, they are now likely to put much less weight on forecast models and look for hard evidence that inflation is either at or well on its way to exceeding the target before raising rates. The "18 month" rule of thumb may very well become the "3 month" rule of thumb.
Chairman Powell also made interesting remarks that I would characterize as a rethink of the concept of full employment. In macroeconomic models the work force is generally characterized as homogenous. Though it may sound paradoxical, one of the key elements of macro models is the full employment unemployment rate. In a large economy there are always persons transitioning from one job to another and persons moving into and out of the labor force. That is, almost all persons spend some time of their lives searching for work even though they may never be out of work for a significant period of time. Some economists have suggested this semi-voluntary search time is akin to waiting in airports for flights. It is basically impossible to spend no time waiting in airports and still get to/from your destination. Thus, search time tends to put a lower bound on the feasible unemployment rate for the overall economy. What that lower bound is, expressed as a percentage of the work force, 5 percent, 4 percent, 3 percent, has always been a guestimate and certainly varies from economy to economy, often deemed to depend on the flexibility of the labor force and wages. But that overall number, say 4 percent, is an overall average of the work force. Labor economists have known for a long time that for any overall average unemployment rate, the unemployment rate for different elements of the work force--characterized by age, educational level, race, and gender differ in predictable ways. In the US, in particular, it is well established that unemployment rates for less educated, younger, and non-white labor are consistently higher than white college educated labor. Consequently, while statistical measures of "full employment" are indicative of the situation in certain parts of the work force, they may be a poor indication of the situation among less advantaged groups. Chairman Powell was quite clear that following the Fed "outreach" events, the FOMC is going to rethink full employment as a concept not applying only to the "average" worker but also to various work force segments who traditionally suffer from higher than average unemployment. What these means in practice is that the Fed will be increasingly willing to allow higher-than-target wage growth (inflation) in certain sectors of the economy in order to bring down unemployment and raise wages among lower income categories.
Both of these factors, the lessened reliance on inflation forecasts and the increased concern with employment in demographics with traditionally higher unemployment rates imply that the Fed will be slower to raise rates and quicker to cut rates than has traditionally been the case. Whether this will lead to significantly higher average US inflation rates over the next decades compared with the last remains to be seen. But the FOMC clearly seems convinced that the societal value of attempting to reduce unemployment rates of traditionally disadvantaged demographics is well worth that risk.