BLS documents the extent of revisions going from 1st to 2nd, 1st to 3rd, 2nd to 3rd releases of NFP, here. Using the 2019 mean absolute revisions, the range of plausible values for employment growth (not percent growth) is shown below.
Figure 1: Change in nonfarm payroll employment (black), and 2 times Mean Absolute Revision (MAR) band (gray), using 1st to 3rd for October, and 2nd to 3rd for September. MARs are for 2019. Source: BLS Employment Situation October release, BLS, and author’s calculations.
So, even with this uncertainty regarding revisions (what Manski terms transitory statistical uncertainty), we know that employment is growing or shrinking (mean revision is 5, and 8). Now, 2019 was before the pandemic. What if we use the 2020 revisions? This yields Figure 2.
Figure 2: Change in nonfarm payroll employment (black), and 2 times Mean Absolute Revision (MAR) band (gray), using 1st to 3rd for October, and 2nd to 3rd for September. MARs are for 2020. Source: BLS Employment Situation October release, BLS, and author’s calculations.
So, if one uses some measure of spread regarding data revisions (BLS is using MAR, instead of a standard deviation because changes look non-Normal), we can guess that it’s likely that the September decline in growth will persist to the third revision, and the October increase in growth will persist to the third revision (this kind of addresses reader rsm‘s query, insofar as it can be made coherent).
I, like many of my countryfolk, have been inspired by Covid to become an entrepreneur. Why should low-wage workers and lucky duckies have all the fun?
I want to open a chain of drinking establishments, places where a working stiff can relax and squander a Covid-bloated paycheck. Places where you can get sloppy drunk without apology. Places where nobody knows your name.
I’m gonna call ’em “Error Bars”.
I hear the new drink rsm is quite popular. What goes into this new drink?
I’m in!
Likely, there will be trillions more rolling around after they shove through [given BOOMING jobs now unnecessary] two MORE, massive, Federal spending boondoggles.
The menu would consist of beer and whiskey/whisky; and pickled eggs.
In other news – Thank God! – my liquor super store shelves are full.
Only thing, Musk and Brandon’s billionaire buddies won’t spend their $555 billion government money in any beer and ball saloon in the Bronx.
Point of information: in fact, the little guy got screwed. The billionaires made out – as intended – on the massive lockdowns; huge deficits; near-zero rates; vaccine billions; cars/crypto/equities/real estate bubbles, and other COVID insanities.
[ Waves arms around like Model B-9 ] “Danger Will Robinson!!!!! Danger Will Robinson!!!!!”
https://www.nytimes.com/2021/10/25/dining/drinks/craft-beer-bars.html
This should have been a reply directly to macroduck’s comment, my mistake obviously.
Why not report the standard error, though? Don’t they bury it deep in the footnotes? Why not show confidence intervals instead of revisions?
In other words, do the revisions have a standard error too?
Why ignore confidence intervals, for both initial and revised reports?
Laziness? Intellectual dishonesty? Ignorance of basic statistics?
rsm: Did you read/understand the post you are commenting on? I suspect either you do not, and/or you don’t understand statistics. At least not Classical statistics. I’m pretty sure you don’t understand Bayesian statistics either.
One could do essentially what I did using the MAR — but I do want to ask you, which MAR should one use? One based on 12 observations — either 2020 or 2019 — or one with hundreds but encompassing (or not) the 2020-21 period? (And I don’t think you understand why a MAR is reported instead of a Root Mean Squared Revision).
My heartfelt suggestion, if you wish to cease further embarrassing yourself — read a statistics textbook, perhaps starting at high school level.
Perhaps rsm was a student in same the Statistics class Kopits claimed he was a TA of?? This actually makes sense, as I wager you’ll remember Kopits also had a problem with confidence intervals.
rsm,
Again with the inappropriate standard errors? You accuse Menzie of “intellectual dishonesty” when you are massively guilty of intellectual stupidity and ignorance.
I have a suggestion. Go get drunk in that new business macroduck has opened, his Error Bar. I am sure he will have some good drinks there that will make you forget all about your embarrassingly silly comments here.