reader bruce hall Think “Apart from travel-related and luxury-related items, monthly adjustments[to the CPI] Often relatively small…”. I beg to differ.
Here is the CPI seasonal component:
figure 1: Log ratio of nsa CPI to sa CPI (blue) for all cities. Dates of peak-to-trough recessions as defined by NBER are shaded in gray. The nsa CPI is the CPIAUCNS of the FRED series, and the sa CPI is the CPIAUCSL of the FRED series. Source: BLS via FRED, NBER and author’s calculations.
In July 2022, the unadjusted CPI looks a third of a percentage point higher than the adjusted one. Is it “small”? I do not think so.
Does converting to growth rate solve the problem? Only if there is a 12-month difference (note: this does not apply to Chinese data, where the Lunar New Moon – tomorrow is 2023 – moves back and forth from January to February). Below are the monthly seasonal components of annualized CPI inflation for all cities.
figure 2: First difference of log ratios of nsa CPI to sa CPI (blue) for all cities. Dates of peak-to-trough recessions as defined by NBER are shaded in gray. The nsa CPI is the CPIAUCNS of the FRED series, and the sa CPI is the CPIAUCSL of the FRED series. Source: BLS via FRED, NBER and author’s calculations.
In April 2022, annualized month-on-month inflation looks like 2.7 Using the unadjusted data looks a few percentage points higher than using the adjusted data. Is it “small”? I do not think so.
What about core CPI?
image 3: Log ratio of nsa core CPI to sa core CPI (blue) for all cities. Dates of peak-to-trough recessions as defined by NBER are shaded in gray. nsa CPI is CPILFENS of FRED series, sa CPI is CPILFESL of FRED series. Source: BLS via FRED, NBER and author’s calculations.
Before the Great Recession, the seasonal factor was about a third of a percentage point. Now they are about one in six.
on off-farm employment.
Figure 4: Ratio of log nsa nonfarm payrolls to sa payrolls (blue). Dates of peak-to-trough recessions as defined by NBER are shaded in gray. nsa nonfarm payroll is FRED series PAYNSA and sa is FRED series PAYEMS. Source: BLS via FRED, NBER and author’s calculations.
Unadjusted nonfarm payrolls were nearly a percentage point higher than adjusted.i think we can all agree no “Minor”.
