Abstract
1- The shrinking upper tail
1-1- Better anchored or better feeling?
References
Abstract
Consumer inflation expectations are highly disperse, with some households reporting very high inflation forecasts. In recent years, disagreement in longer-run inflation expectations has fallen, reflecting compression in the upper part of the distribution. The 75th percentile of the distribution of longer-run inflation forecast has fallen 0.21 percentage points per year since 2012 and is at an all-time low. I show that the decline in long-run inflation expectations at the upper end of the distribution seems to reflect improvement in consumers’ general economic sentiment, rather than stronger anchoring of inflation expectations.
The shrinking upper tail
Panel A of Fig. 1 plots the 25th, 50th, and 75th percentiles of the distribution of longer-run inflation expectations over time. The 75th percentile shows the largest decline in the past few years, having decreased by an average of 0.21 percentage points per year since the inflation targeting announcement in 2012. The median and 25th percentile have fallen by an average of 0.07 and 0.05 percentage points per year, respectively, since 2012.3 Higher portions of the distribution have also fallen—e.g. the 90th percentile by an average of 6 percentage points per year since 2012, from above 60% to around 15%. Another way to see the movement away from extreme high forecasts is in Panel B, which shows the falling shares of forecasts above various thresholds. Panel C shows that the interquartile range (a proxy for disagreement) has declined to an all-time low of 2 percentage points.4 This decline is attributable to the shrinking difference between the 50th and 75th percentiles (π(75, 50)), as the difference between the 25th and 50th percentiles (π(50, 25)) has remained constant. Now π(75, 50) is smaller than π(50, 25), and the difference between mean and median expectations has fallen. As the upper tail of the distribution of inflation expectations is shrinking, so are the disparities in forecasts of various groups of consumers. Most notably, the forecasts of consumers with a college degree were nearly a percentage point lower than those of consumers with only a high school education in 2014, and only 0.15 percentage points lower in 2019.