The New Inflation Debate

Policymakers are involved in another debate about inflation, but this time its because some think reported inflation is too low, not too high. The new view is not based on any formal review or research. Rather, some policymakers think a consistent failure to hit the 2% price target risks "unanchoring" inflation expectations, resulting in a much lower ceiling on official interest rates than otherwise would be the case, thereby limiting policymakers ability to respond to a weakening economy. As a result, some policymakers have floated the idea that an intentional over-shoot of inflation, or a so-called makeup policy should follow any under-shoot.

One would think a good starting point for this debate would start on price measurement, what’s included and what’s not, as there is no absolute perfection or agreement on how to measure the change in prices. It was not too long ago (mid-1990s) policymakers led by then Fed Chairman Alan Greenspan complained that published price statistics over-stated inflation. That comment started a political outcry for change, the naming of an Congressional Advisory Commission and the issuance of the “Boskin” report---which proposed a number of changes in price measurement, all with the intent to lower the published rate of inflation.

The statistical agencies ended up introducing a number of changes including product substitution, quality adjustments and faster introduction of new goods. There is no consensus on how much these changes reduced the rate of consumer price inflation. Estimates range from a few tenths to as much as two percentages points, and it is unclear if these reductions in the rate of inflation are a permanent feature each and every year, or fluctuate from year to year.

Roughly the same time government statisticians were implementing recommended changes from the "Boskin report" they also decided to abandon the rent series for homeowners and to use the renters rent series for homeowners and the rental market. Research conducted by two Federal Reserve staff economists found this change had the effect of keeping “measured inflation artificially low” during economic expansions. In other words, the findings of the Fed staff concluded that the natural lift to reported inflation from a rise in housing activity during business upswings has been neutralized if not reversed by this change in measurement.

Ordinarily, changes like this would not matter but the combined rent series account for 30% of the overall index. And estimates from the new measurement of housing costs, depending on the status of housing market, range from as a little as a few tenths to as high as two percentage points---or fairly close to the estimates of the proposed changes in the “Boskin” report.

Biases in consumer price measurement of a few tenths of the annual rate of inflation do not matter much when reported inflation is high. Yet, these biases do matter when, as is the case today, reported inflation is low, and policymakers remain puzzled over why cyclical inflation has not been consistently higher.

Controversy over the consumer price index is not new. Back in 1997, then Fed Chairman testified before Congress stating that the aim of any measure of inflation is that it needs to be unbiased. Mr. Greenspan argued "an examination of all available evidence" should be conducted to determine if the current measurement of prices had an equal chance to overstate and understate reported inflation. Mr. Greenspan's concluded based on the evidence at that time there was a 100% probability that inflation was being over-stated. Similarly, one could conclude that based on the research of the Fed's staff on the measurement of housing costs there is a 100% probability that reported consumer price inflation is being understated, and has been for some time.

As policymakers undertake a review of their policy framework it would be prudent to also examine it's key focus, the measurement of consumer prices. The real danger in this debate is that policymakers make a change in policy to address a problem that doesn't actually exist.

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