Why friends shouldn’t let friends mix and match seasonally adjusted and seasonally adjusted data when calculating changes

Reader Bruce Corridor feedback in his protection of calculating an 18-month change utilizing non-seasonally adjusted CPI knowledge after which 6 months of seasonally adjusted knowledge:

…pgl insists on utilizing 13.3%, which is what I wrote in my first touch upon this “value stage change” based mostly on non-seasonally adjusted knowledge. Following his “correcting my pondering,” I used the Fed’s seasonally adjusted knowledge, which got here in at 12.6% from January 2021 to June 2022. In fact, everyone knows that the endpoints of seasonally adjusted knowledge are topic to revision because the seasonal elements are recalculated over time. Possibly after just a few months that quantity shall be 12.4% or 12.8%. Nonetheless, the problem is not the precise proportion change, however slightly the general magnitude of that change and the way this pertains to later (June-December) adjustments. June appears to be the purpose of discontinuity, not generally seen within the 12 months/12 months numbers however price studying up on.

Properly, I identified that he dangers making 18 month edits to nsa knowledge right here; however what about including nsa and sa knowledge?

Properly, this is an instance (from the Scott Walker marketing campaign, in a 2016 publish) so as to add not solely completely different sequence but in addition completely different seasonally adjusted (or non-adjusted) knowledge.

Particularly Wisconsin Governor Walker’s marketing campaign bumped into hassle when they’re marketed Job creation figures ensuing from the addition of seasonally adjusted job numbers (from the so-called quarterly employment and wage census) with seasonally adjusted employment figures (from the enterprise survey) with a purpose to acquire a cumulative change in employment. (They did this as a result of the QCEW figures are many months behind whereas the institution survey knowledge are newer). That is proven in Determine 3.

Figures 3: Wisconsin Nonfarm personal employment from Quarterly Employment and Wage Census, not seasonally adjusted (blue), nonfarm personal employment from Institution Survey, seasonally adjusted (purple). Black arrows point out adjustments from the QCEW and institution survey numbers; blue-green arrows over firm survey. Supply: BLS.

Observe that the adjustments are from December 2010 (simply earlier than Walker takes workplace) to March 2012 (newest obtainable QCEW figures of 12 figures obtainable on December 12, 2012). That’s, add 89.1 to six.4 to get 95.5,000, which is near the determine of 100,000 cited in Governor Walker’s marketing campaign. You’ll be able to see why Governor Walker’s marketing campaign officers did this – the proper calculation utilizing the change within the institution census from December 2010 to October 2012 was solely 61.1 thousand.

So, if not in “good” firm, Bruce Corridor is within the firm of the likes of the Scott Walker marketing campaign.