Reader Bruce Hall argues that “with the exception of travel and luxury items, the monthly adjustments[to the CPI] It tends to be relatively young…”. Please change.
Here is the seasonal component of the CPI:
Figure 1: Ratio of log nsa CPI all urban to sa CPI (blue). The National Bureau of Economic Research has peak-to-trough recession dates highlighted in grey. nsa CPI is the FRED CPIAUCNS series, sa CPI is the FRED CPIAUCSL series. Source: BLS via accounts by FRED, NBER, and the author.
In July 2022, the unadjusted CPI appeared to be a third of a percentage point higher than the adjusted one. Is this “minor”? I do not think so.
Does switching to growth rates fix things? Only if there is a 12-month difference (note: this will not work for Chinese data, as the new moon – tomorrow of 2023 – goes from January to February and back). Here is the seasonal component of the annual CPI calculated on a monthly basis for all urban inflation.
Figure 2: First difference in the ratio of the logarithms of the nsa CPI per metropolitan area to the sa CPI (blue). The National Bureau of Economic Research has peak-to-trough recession dates highlighted in grey. nsa CPI is the FRED CPIAUCNS series, sa CPI is the FRED CPIAUCSL series. Source: BLS via accounts by FRED, NBER, and the author.
In April 2022, annual inflation appears on a monthly basis 2.7 percentage points higher using unadjusted data than it would appear using adjusted data. Is this “minor”? I do not think so.
What about Core CPI?
Figure 3: The ratio of the nsa log core CPI all metropolitan areas to the core CPI (blue). The National Bureau of Economic Research has peak-to-trough recession dates highlighted in grey. nsa CPI is the FRED CPILFENS series, sa CPI is the FRED CPILFESL series. Source: BLS via accounts by FRED, NBER, and the author.
Until the Great Recession, seasons were about a third of a percentage point. They are now about a sixth.
And about working on the non-farm payroll.
Figure 4: NSA non-farm employment-to-employment ratio (blue). The National Bureau of Economic Research has peak-to-trough recession dates highlighted in grey. The NSA non-farm payroll is a FRED PAYNSA series, which is a FRED PAYEMS series. Source: BLS via accounts by FRED, NBER, and the author.
Unadjusted nonfarm payrolls are about a percentage point higher than adjusted. I think we all agree on this Not “Underage”.
