The Economy

To get a clearer look, I'd probably start with the stuff that John Williams is doing at Shadowstats.

Shadow Government Statistics - Home Page

 

Of course there would be a reason to do contextual seasonal adjustments, but their models have gotten convoluted way beyond simple measures.

On top of it, you are giving absolutely no context, when you are continually creating new adjustments (that people only find out about on the day of release, and only those with the wherewithal to actually read through their stuff and decode their methodology) that just send the numbers jerking around so that what you thought you knew just a month before suddenly is being presented as the exact opposite.

And none of that even addresses their basic definitions. ... about what is and isn't a job (and how they double count in some instances) , about who is looking and who is discouraged, etc.

 

Just to give an example of how arbitrary what they are doing has become in practice. ...

Jan Hatzius at Goldman Sachs used to be the most consistent forecaster on Wall Street. Year after year. He predicted a loss of 250K jobs last month. The number came in at a gain of 467K. That number was 4 times the median consensus estimate! These are people with the money and resources to survey on their own and mirror the work as well or better than the BLS. There is something seriously wrong isn't there?

There may or may not be anything nefarious going on with the adjustments they did (or the times in the past this has come up), but if there isn't, then the question is, what is the purpose of what you are doing, when you took December's number from 199K to 510K, and November from 249K to 647K? And why should anyone have thought the numbers were meaningful in November and December or that the new numbers are meaningful today, because who knows what you are going to come at us with in March?

And if this is the pile of droppings that the Fed supposedly bases it's own alchemy on, wouldn't we be just as well served by central-planning done via dart and dart board?

 

Um, there are countries within the EU that are printing inflation number north of 10 percent year over year right now.

 

You can't throw away a 3-sigma beat and just say, "the generation of the actual report" is all I care about. Well, maybe you can, because you don't care about this. But anyone trying to give these reports significance can't do that.

The numbers aren't the problem. But for the purposes of this conversation, an ever-changing methodology thrown over the numbers is demonstrating just how drastically their adjustments change the employment picture (in both directions) from the non-adjusted data. And curiously, in ways that benefit the people who are creating those numbers.

You said there is nothing nefarious going on (at least nothing necessarily nefarious), and I actually would guess that as it relates to this particular report the insane revisions weren't some nefarious thing entirely, at least. ... I think they started out trying to do nefarious things a long time ago that they thought would give them "better" numbers, and over time they turned what they were doing into a soup they can't control. And now they are just flailing. They got F'd by Covid, and were adjusting their numbers based on patterns that were no longer all that predictable. That is obvious. So this month they are laying a "Covid adjustment" over the count, and voila, 2.8 million jobs lost now = 467,000 jobs gained. There has never been a seasonal adjusment anywhere near this magnitude on a January number. We are zipcodes away. Apply the typical adjustment applied in January over the last 10 years and the number would have been about 165K. ... i.e. in the ballpark of the consensus that was being forecast.

Now someone can come back and argue that their new seasonal adjustment is warranted, but then why didn't they do this immediately after those huge numbers they reported over the summer (i.e. -- Biden was talking about all the jobs he is creating), which are now being adjusted down by more than a million jobs (but nary a peep now), while they are focusing on the more recent numbers (August to December was revised up by 817K jobs), the last two months turned on their head?

Yeah, I am being cynical. But when you see this shit over and over again. ... At worst, the cynicism is warranted, but at best they are chasing their asses.