The "Muzzle" on Inflation

The "muzzle" on reported inflation has policymakers and analysts perplexed. Numerous economic explanations and theories have been offered, and policymakers are considering making changes to their operating price-targeting framework. Yet, before any decisions are made policymakers should consider all of the factors that could be keeping a "muzzle" on published inflation. I am going to discuss two.

First, a little more than 20 years ago the Bureau of Labor Statistics (BLS) introduced a number of new measurement techniques in the estimation of consumer inflation. So the current business cycle, which started in 2009, is the second consecutive cycle in which these new procedures have been employed. Statistical changes have been made to account for product substitution, a greater degree of quality changes in products and services and faster introduction of new outlets or ways in which people shop. The introduction of new variables in the estimation of inflation alters the pattern and at various times the rate of change as well.

Prior to their implementation, analysts and government statisticians estimated that the potential reduction in core inflation from all of these statistical changes would range from one-half to a full percentage point. Yet, all of those estimates were looking backwards and there is no guidance from the statistical agencies of the scale of the reduction in reported inflation after implementation.

Odds are high that the impact on reported inflation varies year to year, with some years at the upper end of range of estimate and others at the lower end. Nonetheless, to overlook the impact from changes in measurement would be shortsighted, especially since changes in consumer price of a few tenths of a percent or more do matter a lot when inflation is low, and readings below the 2% target could be misconstrued as a failure of monetary policy.

Second, research conducted by the Federal Reserve staff has found that the shift in the measurement of shelter costs two decades ago to only use only prices from rental market and exclude those from the owners housing market systematically removed the largest single "driver" of cyclical inflation, while it also simultaneously reduced the volatility in reported inflation.

The significance of these findings has not received as much attention as they should. Removing the largest single driver of cyclical changes in inflation means that reported inflation nowadays does not exhibit the same sensitivity to economic growth and interest rates, as was the case in previous cycles. Accordingly, one of the reasons why the trade-off between changes in unemployment and reported inflation has been so benign in the last 20 years is due to changes in price measurement.

The missed signal from housing inflation was readily apparent in the 2000s when core inflation peaked at a relatively modest 2.5% even though house price increases were recording double-digits increases. In previous business cycles in which house prices recorded gains north of 10% core inflation readings were two or three times higher. In the current cycle, house price increases have run ahead of rent increases, but not to same extent as was the case in the 2000s.

These findings strongly suggest that price measurement issues are important to consider when looking at trends in the reported inflation data. For all of the conceptual changes and measurement issues the key question policymakers should be asking is whether the "muzzle" on reported inflation still makes it a useful benchmark for the price-targeting framework. The fact that currently constructed published price measure miss modern day inflation in the asset markets strongly suggests policy may need a new working definition of inflation before they contemplate any changes to the price-target framework.

Viewpoint commentaries are the opinions of the author and do not reflect the views of Haver Analytics.