The Federal Reserve recently announced a shift to a new “average inflation targeting” regime, indicating that it will use a longer time horizon to judge its mandate of price stability. Rather than aim for its target inflation rate of 2 percent in each period it sets monetary policy, the Fed will set policy consistent with an average rate of 2 percent over time. This new regime allows for inflation to go above and below the target so long as the trend is 2 percent. While some have applauded the change in policy, it is not likely to work as well as its advocates claim. Unlike price-level targeting, a similar monetary policy framework, average inflation targeting lacks the explicit commitment.
The target set by the central bank is important because it tells us the stance of monetary policy — whether policy is too loose, too tight, or just right. If a central bank targets a particular inflation rate, and the actual inflation rate is above its target, this is an indication that monetary policy is too loose. A realized inflation rate below the target indicates that monetary policy is too tight.
There are two ways of targeting inflation. The first is to target the inflation rate itself. In this case, the central bank sets a target of, say, 2 percent and continuously adjusts its policy to hit that rate. The second way is to target the price level. Because inflation is just the rate of change in some price index, or price level, we can think of inflation as the journey and the price level as the destination. Suppose that the price index today is 100. A price-level target might state that the price level should be equal to 110.4 after 5 years, equivalent to an inflation rate target of 2 percent.
The crucial difference between these approaches pertains to what the central bank does when it misses the target. For example, suppose that under an inflation target, the price index is 100 today. If the inflation rate for the current year is 1 percent, the central bank will conduct policy to raise the inflation rate to 2 percent the following year (if it succeeds, the price level will reach 103). However, under a price-level target, the central bank will conduct policy to try to make up for its mistake. One should expect the price level to be higher than 103 and possibly 104 at the end of the following year. In other words, one should expect the central bank to produce an inflation rate above 2 percent and possibly as high as 3 percent in order to make up for past mistakes and get back to the target growth rate in the price level.
By setting an explicit, long-run goal, a price-level target not only helps to anchor long-run inflation expectations, but also allows periods of expansionary policy to follow periods in which monetary policy was a drag on economic activity. A below-target inflation rate implies that GDP growth is below its potential. Under a simple inflation target, the Fed tolerates lost growth in any given year and tries to do better the following year. Over time, lost growth adds up and leaves GDP smaller than it could otherwise be, just as in the opposite scenario excessive inflation adds up and erodes purchasing power.
Price-level targeting has grown popular in some policy circles in light of the Federal Reserve’s perceived policy errors following the 2008 financial crisis. Although the Fed has had an inflation rate target of 2 percent since 2012, actual inflation has been consistently and persistently below that rate. This suggests monetary policy was too tight, a drag on economic activity and a potential explanation for the anemic recovery.
Advocates of price-level targeting might therefore be inclined to support average inflation targeting, since it seems like it would do the same thing. Sometimes inflation comes in below the Fed’s 2 percent target, and sometimes it comes in above 2 percent. With average inflation targeting, any period in which the inflation rate is below 2 percent will be offset by a future period in which it is above 2 percent. By the same logic, an inflation rate that is too high will be offset in a future period by an inflation rate that is below the 2-percent target. That seems a lot like a price-level target.
While price-level targeting and average inflation targeting seem similar, there is one important difference. Under a price-level target, the Fed has a mandate to meet a specific target for the price level, either explicitly stated or implicitly known from the preferred trend rate of inflation. This price-level target is always known for each point in time and can be evaluated in comparison to the actual price level.
With average inflation targeting, there is no explicit long-run target, but rather a vague promise to correct for past mistakes. This raises a couple of questions. How long can something like this remain a credible commitment? At what point do bygones become bygones?
With a price-level target, consumers and businesses can judge the Federal Reserve’s credibility by observing the actual price level relative to the targeted trend path. If it is clear that Fed policy is not consistent with the price-level target, causing a growing or persistent gap between the actual data and the target, policy ceases to be credible. This explicit target is also a way to hold the Fed accountable for its policy decisions. However, with average inflation targeting, there is no explicit target and no clear timetable for a return to a trend in the price level. Furthermore, there is no clear pronouncement on the length of time over which the average is calculated. The shorter the time horizon, the more likely misses above or below the target will not be corrected.
The thing about average inflation targeting is that it is quite similar to plain old inflation targeting. Typically, there is some margin of error around an inflation target that is deemed acceptable. A central bank targeting an inflation rate of 2 percent might deem anything between 1.8 percent and 2.2 percent acceptable. Assuming the central bank is successful, it is likely to hit 2 percent on average, over time.
The problem with the Federal Reserve over the last several years is that its inflation target has not been symmetric. It has allowed errors, including pretty significant errors, below 2 percent, but has rarely allowed inflation to rise above 2 percent. The pattern has been so persistent that some argue (here and here) that the Federal Reserve’s inflation target was actually an inflation ceiling. What about this recent experience suggests that the Federal Reserve can commit to symmetrical errors around its inflation target?
Average inflation targeting gives the Federal Reserve greater discretion than a price-level target. The Fed always has the option to let bygones be bygones and does not have to answer to Congress about gaps between the actual price level and its target, because no such explicit target exists. With average inflation targeting, the Federal Reserve is promising not to make the same mistakes as it did in the wake of the financial crisis and is promising to correct for any future policy mistakes. The problem is that there is no clear commitment mechanism. A cynic might say that is the intention.