Is Inflation Changing Its Ways?
Policymakers and academics have noticed that the inflation process in the United States and other countries has changed markedly since 2000. Two formerly characteristic features of the process have been deviating from their historical norms. First, inflation persistence - the degree to which current inflation depends on past inflations - has declined dramatically. Second, an equally dramatic decline has occurred in the degree to which the output gap affects inflation. The output gap is the percent by which actual output deviates from its potential.
Historically, for every one percentage point increase in output above its potential, inflation increased 0.15 percent. Since 2000, the point estimate is actually negative. Although one should not necessarily conclude that the output gap and inflation are negatively related, the decline is puzzling and has important policy ramifications. It implies that a larger output gap has to be opened up to lower inflation.
The policy implications of the decline in inflation persistence are mixed. When the value of inflation persistence is 1.0, as it nearly was at the beginning of 2000, it implies that all shocks to inflation are permanent. If inflation shocks are permanent, it suggests that the only way to offset them is to widen the output gap. But now that inflation persistence has fallen to 0.4, it would appear that shocks to inflation are temporary and policymakers can potentially wait for inflation to return to normal.
Less inflation persistence, however, has another side. If one wants to permanently lower inflation from its current level, less persistence implies that an output gap has to be opened for an even longer period of time than if persistence were higher. Consider the extreme case, where changes in inflation are nearly permanent (persistence is close to 1.0). In this case, an output gap need only be opened for a short period of time because a gap today implies less inflation both today and in the future. Less inflation persistence works the other way.
Some argue that we should not read too much into these declines because they are not statistically significant, which implies that they could be the result of pure chance. But policymakers may not have the luxury of waiting to see if a change in the inflation process is statistically significant before reacting. A researcher's trigger point for starting to seriously contemplate a possible change in the inflation process is not necessarily the same as a policymaker's.
Others argue that the output gap is poorly measured, so we can't be certain about how much output is deviating from potential. If we can't accurately measure the output gap, we can't ascertain the impact of the gap on inflation or know whether it's declining. While the recent declines in inflation persistence and the impact of the output gap on inflation are not necessarily statistically significant, it is worth noting that the effects of lagged inflation and the output gap on current inflation are being estimated with more precision than before the declines. This suggests that the decline in the output gap's effect on inflation is not because potential output is being poorly measured.
To determine why inflation persistence is declining, it might help to examine the relationship between inflation and past inflations during the period over which persistence changed most dramatically--from the first quarter of 2000 to the third quarter of 2002. (For this exercise, we ignore the impact of the gap on inflation.) At the beginning of 2000, every percentage point in the previous quarter's inflation was associated with an 0.8 percentage point increase in current inflation. Six quarters later, that number had fallen to 0.4.
Just 10 years earlier, beginning in the first quarter of 1990, some economists believe that the long-term inflation target began to decrease. Such a decrease could artificially add in inflation persistence. After 2000, long-term inflation was probably fairly constant. If this explanation is correct, then inflation persistence will probably continue to stay low.