By Hugh, who is a long-time commenter at Naked Capitalism. Originally published at Corrente.
The BLS jobs report covering September sparked controversy because of the large three-tenths of a percent drop in the unemployment rate, down to 7.8%. This fall corresponded to a reduction of 456,000 in the number of unemployed. As dramatic as this decrease was, it was dwarfed by an increase of 873,000 in the number of employed. This figure represented not only the 456,000 unemployed who supposedly found jobs but a further 418,000 who entered into the labor force from outside it. These data were all from the Household survey were fundamentally at odds with an anemic increase in jobs of only 114,000 from the larger Establishment survey.
The large increase in employment and decrease in unemployment were internally consistent within the Household survey, but, contradicted as they were by the jobs data from the Establishment survey, weren’t real. This raises the question of why they happened.
In part, the two surveys have very different levels of statistical significance. The larger Establishment survey has a threshold of 100,000 while the Household survey with this month’s problematical data is 400,000. But the size of the discrepancy, especially with regard to employment, in the September report goes beyond a simple mismatch in the data’s significance.
One large source of growth in employment in September was a 582,000 increase in the number of involuntary part time workers. While large, it is in keeping with August to September increases in this category over the last two years:
2010: 579,000
2011: 483,000
The problem with these August-September increases is that they aren’t real, as the graph above shows. The light blue line is the seasonally unadjusted data. The dark blue line is seasonally adjusted with the August-September peaks marked in red. These peaks are phantoms. In 2010 ane 2011, they don’t correspond to or compensate for anything in the unadjusted data. In 2012, there was a 268,000 increase in the unadjusted August-September number (far right of graph), but there is no rationale in the data to assume that the adjusted number should be larger, indeed twice as large.
At least as regards involuntary part time workers, this contradicts the argument that this month’s report was doctored for election purposes. The BLS has had the same recurring glitch in this category for the last three years. The real question is why they haven’t fixed it yet.
The other category I found significant was unemployment.
In this graph, I marked the unadjusted data for 2012 in red. If we look at the unadjusted data, it has an “M” like configuration (increased unemployment in the winter (bad weather) and summer (school’s out) with seasonal decreases in the spring (onset of good weather) and the fall (return to school and Christmas). This is clearest in 2012, but you can see it in 2010 and 2011 as well. The dark blue seasonally adjusted line, on the other hand, shows a series of baselines dropping year by year.
Now if we look at the two together, in 2012, as the winter surge in unemployment tails off into the spring in the unadjusted line, there is only a 173,000 decline in unemployment seasonally adjusted. Yet in the fall with very nearly the same configuration, we see a decline in the seasonally adjusted line (far right of graph) starting in August (133,000) and continuing into the September drop of 456,000. The big question is why. It could be said that intra-year comparisons aren’t valid, but if we look at the corresponding seasonally adjusted declines in unemployment for 2011, it occurs in November and in 2010, in December.
If you are conspiracy-minded, there is some evidence for you here. There very likely will be a further seasonally unadjusted decrease in unemployment going into Christmas, but this usually isn’t reflected in the seasonally adjusted “official” number for another couple of months. This is not so much a question of a phantom number as one temporally out of sync.
If we take the differences in accuracy of the two surveys into account, the phantom number of the involuntary part timers, and the temporally out of place one of the unemployed, we can explain most of the discrepancy between the September Household and Establishment surveys. I have no problem with the unemployment number, but I do have one with its timing, in terms of the data, not the election.
60 comments
Comments are closed.
Let’s see how the flash GDP report for the 3rd quarter comes in on Oct. 26th.
Expectations are for 1.6% growth, but incumbents would prefer a number more like 3.0%.
If flash GDP don’t beat expectations, heads are gonna roll …
So Hugh, now _there’s_ a geek-worthy parsing meriting real applause!
+ 1000.
I am very happy to see that Hugh’s reports are appearing regularly on NC.
“One large source of growth in employment in September was a 582,000 increase in the number of involuntary part time workers.”
Perhaps both Obama and Rommney can take credit for this month’s report. It would be useful to look back to 2008 and 2004 to see if similar outsized shifts happened. Could it be that the upcoming November elections have goosed the figures with temporay campaign workers littering the landscape with signs and flyers?
The rise of the involuntary part time worker is also a reason for the great job creation legacy of the Clinton era. As I recall, 22 million jobs were to be had when he was in office. This is no lie, everyone I knew had at least 2 or 3 for themselves, most of their family members had at least 2. Go part time jobs, the secret sauce in an MBA’s labor cost control strategy.
Considering that Reuters is pegging this as a 6 billion dollar election I wouldn’t be surprised if that is right. Both campaigns have been raising money like mad and spending it, mostly on ads to raise more, so in effect this campaign season has been a jobs bill in its own right.
I wonder if the data itself has sufficient detail to let us check that.
My wife has been trying to contact unemployment for two months straight to renew her claim – which was stopped for some unknown reason in the middle. She can’t get a straight answer on the web (“you must contact by phone and speak to a representative…” and yet when she tries to call on the phone the voice mail ALWAYS bounces her off saying “we’re sorry, our representatives are all busy, please call back”. To get to this point takes about ten minutes each time. 8AM to 5PM, dozens and dozens and dozens of attempts. She tries, I try.. hell, even the cat might get a shot at it.
It doesn’t take a genius to figure out where this reduction in new/renewed claims comes from if ya never allow people to contact a rep to file the fucking claim in the first place.
That’s appalling, and an interesting data point in the mystery of the falling UI claims.
On a practical note, is there a bricks + mortar UI office where your wife could go in person (and raise hell, after getting satisfaction)?
@ Shutterbuggery. In dealing with municipal and state bureaucracies I have found it very effective to find the direct phone number or email for the General Counsel’s Office (in your case, whatever State DOL, probably), and to send him/her/it (the office) a concise summary of the facts (with dates if possible), and a closing sentence “looking forward to your prompt attention to this matter.” Finding GC offices at state agencies is not that hard with Google and some creative searching. As well, Martindale Hubbell online may help, or Linkedin. Good luck.
Ms G, good tip, thanks for that. We’re in the FoolsGold State here so we’ll give it a try.
I always appreciate the stats breakdown by experts on this site. But the anecdote can be revealing too. And the street level of increased despair. I know two people that were ‘disappeared’ from state info, and couldn’t get checks. Both got them later, but if they had not had friends that helped cover their rent, they would be homeless. Besides the disgusting to me, lauding of the unemployment rate lowering because of PART TIME jobs…while still losing full time jobs….I am also disgusted by every upper level ‘economist’ that would not in a chance in hell walk our ‘dangerous’ areas of our cities to see the effects of budget cuts. The numbers may be close to ‘true’ , but the life of us down here has nothing to do with the argument or the numbers. And the experts never had to jump through the hoops we have to, just to get some money from a fund we’ve paid into all our working lives. Demeaning, turns to rage. …and there is rage.
And again, what state are you from? Because if I knew, I’d watch out for it…
Shutter:
There are still many people collecting unemployment and also many applying for unemployment. Many do it the old fashion way via phone instead of computer and they have assigned times to call into the office. If she really want to catch them in the Unemployment Office, find out when the office opens up. Using your cell phone’s clock, as soon as the second hand goes beyond 7:00 AM (or whenever the phone lines open up for calls, get on the phone (with a redial button) and call and repeat call.
You will not get them much otherwise and lunchtime is terrible also. The next best thing is too go down there personally and stand in line at the Unemployment Office the old fashion way. The effort has to be on you. Don’t raise hell and forget the GC’s office. Best to treat this as a mistake and how do we fix it.
Hi, what state are you in? I know FL is a disaster for the disemployed. So I wonder how many other states are like FL…
Lost in all this nonsense talk about what real unemployment numbers are is the following fact, rarely mentioned anywhere:
7.8% unemployment is a lousy number.
There are approximately 545 people in DC who are responsible for this. None of them are affected by this number. Until DC plays by the same rules as the rest of us (do I hear a Social Security amen?), you can kiss your sweet economy goodbye.
Clint:
Better keep talking to the chair, as it can’t challenge your silly comment on federal government employment which also includes state and local governments funded by the federal govenment which has been has been falling under Obama. When we add in state and local govenment jobs cuts the numbers offset gains in the private sector. http://krugman.blogs.nytimes.com/2012/09/09/government-employment/
In fact, Obama is one of the few President who has overseen cuts in government jobs when compared for Bush senior and junior and Reagan.
run75441,
So, since Obama came into office he’s reduced the unemployment rate by .1% by cutting government workers? And that cost $7Trillion? I think an empty chair could have done it for less.
Clint:
And again you miss the point, presidents (Republican Presidents) before him oversaw a increases in government workers.
$7 trillion? Not sure how this fits into the conversation on the growth in government jobs.
My 2 cents on a way the numbers might have been manipulated.
http://fetchdoggyfetch.wordpress.com/2012/10/09/how-the-obama-campaign-or-any-other-interested-party-could-have-rigged-the-employment-numbers/
You’re the good dog and Rick Santelli’s the bad dog?
http://www.mediaite.com/tv/rick-santelli-and-cnbc-panel-get-into-crazy-shouting-match-over-unemployment-numbers/
Quite interesting. I tend not to be very conspiracy minded but that is a very clever way to manipulate the data. The thing is if the BLS would be able to pick it up in its interviews. Another is this would still be unadjusted data. So it would be necessary to modify it to result in a 0.3% drop in the adjusted data. A final consideration would be to keep the con going through at least next month so that it would blend in with the seasonal Christmas drop in unemployment.
If this is correct and I don’t do statistical sampling theory or methodology caveat caveat caveat:
then campaign hiring could easily affect the numbers, no? (Any campaign, or even the aggregated campaign hiring, could do so, no “conspiracy” or foil involved (what an old Corrente poster named Shystee called “emergent conspiracy”).
Especially because I would bet that campaign call centers, registration efforts, petition drives, and even online trolling would all tend to draw temporary workers from the unemployed still in the labor market. (That is, if somebody’s long term disemployed and have a System D solution worked out, they won’t be very enthusiastic about re-entering the system for a month and filling out a 1099 miscellaneous…)
Well you could also just hack into the BLS computers.
I think people assume that the hard numbers reported by the government are not forged because no one else is collecting the hard numbers. If you get too paranoid, you have to throw out all data. We would have no idea about the state of the labor market.
But I think its safe to assume that the numbers are not crudely manipulated or hacked. The reason is that the BLS has adopted a number of changes or refinements to its methodology and definitions over the past three decades that consistently wind up lowering the number of unemployed. There is no need to do a crude hack of the data if there are more subtle methods available, and likewise why use the fancier methods if you are doing a crude hack?
Nicely done, Hugh.
Thanks again.
Please divide 582,000 by 873,000. The result is .66666666667%, or exactly two-thirds, no matter how many decimal places you take it to. The 582,000 is a plug, or was simply guesstimated from the 873,000. This is not an organic compilation of employment stats.
People don’t believe these numbers of 7.8% we all know real numbers are 14.1% Obama just told labor number to just put numbers down help his get more votes. I hope you all realize it. President failure just can’t take it he gotten his small blk ass kick butt by Romney. I hope and pray president failure Obama will not win election on Nov 6th.
22%
The reason why this conversation seems silly to me is that the “conspiracy” pretends that these numbers were ever real before September 2012.
I have no problem with you saying that the numbers were always made up. But, please give me a break that they were real before September 2012 as the GOP tries to pretend.
My problem ultimately with many of your articles here by the way is that I sense that you have this thing bout Obama where you want to blame him. I get that urge. I think he’s part of the problem too.
I just don’t buy that he’s anything unique. The jobs numbers just reinforces a narrative that’s been there forever.
If you want to make a real point, please all of this into that wider problem. That we don’t have a choice going into Nov 2012. That regardless, those numbers would be fake.
Instead, by trying to make Sept 2012 seem special, the implication is that the rest were just fine before now. This is likely now your intent because I have seen other articles here on the subject, BUT, the GOP conspiracy is not that these numbers are always fake. The GOP conspiracy is that only Obama’s numbers are fake.
So, you are playing into the game that they want to use to control the debate. Not a debate about both parties. But just the Democrats. This is like one of several articles I have seen here like that. When reading it, one gets a sense of disconnect from reality. Not because you aren’t right to criticize Obama or the Democrats, but the sense that you don’t make it clear each time that the problem is both parties and that this needs to be repeated until its hammered home that, for example manipulating the numbers, is a part of how the powers that be stay in power.
Instead, the article feels more like “there’s something unique here”
Jane Doe, thanks for offering a sincere concern about the stuff you’re talking about. I don’t think Hugh had an
“intent.” And I don’t think you really thought so either, except to make a point. On the bright side, we can learn from each other, right?
Then why mention conspiracy, which is really the GOP talking point, if not to say there is something unique here. There’s no conspiracy.
So placing “intent” in quotes once again seems off to me.
Its like the strange articles from the other day in which someone mentioned how Romney should challenge Obama from the Left.
I remembered thinking “sure, if that were the nature of things, then we wouldn’t really be having this debate at all because the country would be moving in the right direction”
The problem is that we aren’t, and this all seems like foil to bring up what you want to discuss.
So why not just discuss it rather than pretending there is something unique going on here.
Why mention Romney the other day. Why not do what Glen Greenwald did- say both parties don’t want to discuss these issues.
Why not here say this isn’t a conspiracy to screw with Sept 2012, but a problem with the way both parties have manipulated the numbers for decades.
If you want to have that discussion, lets have that discussion.
Great! Hugh?
The conspiracy context was supplied by Jack Welch, of blessed memory. Perhaps I should have punched up the headline to make that clear, but Hugh is a pretty sober, technical kind of guy, so I left the headline as he wrote it.
The best indicator with respect to the jobs market and economy is to count the number of people you know personally who are pissed off because of the economy and who erupt into tirades at the slightest mention of the names Obama, Romney, Congress, Banks, etc. Based on this data, I can tell you that economy is still screwed up.
I am not sure what you are reacting to. Many of us here don’t see any significant difference between Democrats and Republicans, between Romney and Obama.
If Obama gets mentioned more, it is because he is the President. When Bush was in office, I wrote about how poor job creation was in his Administration. But if you notice, I do not mention Obama in this post, and even in the last phrase, I try to steer clear of any electoral implications.
Yes, the “problem” has always been there. I am one of the people who has been pointing this out. The definitions and models the BLS uses have limitations and weaknesses. This has led me to find a way to recover those the BLS defines out of the labor force. And this in turn resulted in my introduction of measures for real unemployment and disemployment, It has also led me to looking much more closely at the unadjusted data, both in its own right and how the adjusted data relates to it. However, although there have always been problems with the BLS data, these became especially acute after December 2007 and the start of this particular recession. Most of what has happened since has been on Obama’s watch.
What this post was about is not about general problems in the BLS data but two specific anomalies (employed, and unemployed) in one jobs report. It was not meant to address larger issues.
For what it’s worth, I think Jane Doe was reacting to the context-free style of your piece. She should speak for herself.
(1) Because this isn’t the first article I have seen at this site that lacks context in the last few months. I like the content here, but context does matter. That’s why I mentioned the criticism Obama from the left by bringing Romney up as if he’s ever going to move left.
(2) Because of the mention of conspiracy, which indicates again a lack of context. If you want to argue this is inline with manipulations that have been happening for a long time. Say that. Don’t jump on the bandwagon of conspiracies that the GOP does.
(3) I don’t care if you did this with Bush as much as Obama. Reacting to the moment rather than providing context is exactly the problem with the analysis. There was nothing unique when Bush did it. And there’s nothing unique with Obama does it. That’s the context to me.
I am don’t know how much more I can say to make my point clear.
You are unclear. There is a legitimate discrepancy in the September data. So if the conspiracy-minded want to see a conspiracy in the jobs report, it isn’t like they are totally inventing something. They could be misreading it, but there is something there. That’s all I’m saying.
As for context, the context of the controversy in the September report is well known. The wider context of the deficiencies in how jobs and employment are reported are dealt with in my monthly analysis of the jobs reports. If the idea is that we are supposed to just state that there are problems in the jobs reports, stop there, and write off everything about them, then that pretty much kills any discussion or analysis of jobs and employment. You will have all the context that you want, and no content.
This is a blog. I don’t know what you mean by “lack of context”. If you mean every post is supposed to go and cover all the background on every issue, you should stop reading blogs entirely. We don’t spoon feed readers. I suggest you read either the MSM, which pretends to do that but actually just choses a frame, or PhD dissertations, which by convention give surveys of the relevant literature.
Hugh began his 2nd to last paragraph in his blog — “If you are conspiracy-minded, there is some evidence for you here.” I’m guessing that’s what triggered Jane Doe’s response. He also begins his response to her — “You are unclear.” A put down.
I’m sure you appreciated A Damasio’s INET keynote address in Berlin. There’s too much unnecessary emotional argument going on here. Just my opinion. I don’t see this one as black and white.
I don’t understand the problem with criticizing Obama from the left. Given the conservatism in his policies over the last 3 3/4 years, there is a large part of the political spectrum, including the center, to the left of Obama. Personally, I am bemused that seeking solutions to problems that actually work puts me on the left. I see less and less of what goes on in this country in left-right terms.
That said, Jane Doe was unclear. Not all problems are the same either qualitatively or quantitatively. Undercounting of the unemployed has always gone on in the BLS reports. However, in 2008, this problem began to balloon. Before, it only occasionally hit above and about 2.5 million, by my calculations, and never stayed there for long before falling back. But in 2008, it began a build that now puts it at over 8 million. That’s a quantitative change, and almost all of it has taken place during the Obama Presidency.
The problems I was addressing in the September report are a different kind of problem. They are not based on the BLS’ definition of unemployment. They have to do with problems in the seasonal adjustment model the BLS is using. This is a qualitative difference.
Talking about general problems with the BLS in a post that is focused on two specific problems in one month’s report is not relevant. The BLS does have problems, even many problems with its definitions, data, and models, but this doesn’t mean that nothing can be learned from studying their data.
This post was about pointing out glitches in the September data relevant to the manipulation charge and explaining why they were problematic. On the employment issue, the data suggests a recurrent problem, not a one off manipulation. On the unemployment issue, a drop is defensible, but its timing remains unexplained. In the absence of an explanation, conspiracy theorists can, and will, construct their own.
Thank you. Your statement is actually closest to what was bothering me about the article. Rather than seeking to understand that, what I got instead, and this now bothers me more, are wild statements such as if you want to read MSM… I mean- what’s MSM about saying that they have been manipulating the employment numbers for decades? The real problem seems to be that I made the mistake of making a criticism about the way an article was written. It bothers me that this is not allowed here without such a reaction as ‘go away” being the response.
For the most part, I like your site.
I am making a specific criticism about a tendency to frame issues in term of Obama rather than placing Obama in a wider context.
How this context is a bad thing I do not know.
I think is interesting that apparently that criticism is not allowed without the over the top statements about MSM and other comments coming up.
I have seen this happen at other sites (Firedoglake or Daily Kos) where everyone who has an interested in reading the site is supposed to think the same to the point that any difference of opinion is met with a high level of unnecessary hostility.
Its sad to see that happening here. Its your site. Do what you want. Its your choice. Its just saddens me to see a response about a criticism. As someone else pointed out, part of the article mentions “conspiracies” so its not exactly like I am bringing up the issue with out justification.
I think you are catching a certain level of frustration because your criticism seems to be principally about Obama and is meta to both the post and the site. That is what threw me and I expect others as well.
If I understand your position correctly, I don’t think it is going to get a lot of traction around here. From my perspective, it looks like you are asking us to pull our punches because Obama is just this guy who is in, and maybe part of, a bad system. Well, the system is bad but that doesn’t lessen Obama’s responsibility for his actions a single iota. Indeed in that sense some of us have been contextualizing Obama for more than 4 years now. And we have had this conversation many, many times. We have in this country a Chief Executive with near dictatorial powers. If Obama had wanted to and if he disagreed with its product, he could have shaken up the BLS just as he could have prosecuted fraud and looting on Wall Street and torture in the Bush Administration. He could do this without any new act of Congress, indeed with legitimate powers that predate the construction of the imperial Presidency and unilateral Executive. This is a man who has made it standard practice to wage undeclared wars without reference to Congress or even the figleaf of the War Powers Act and has given himself the power to execute US citizens with zero due process. So please do not try to tell me that he just a captive of the system. He isn’t. He is a proud and powerful enabler and abettor of it.
Jane,
With all due respect, the controversy over Jack Welch’s attacks on the latest employment releases have been on the front page of every business section for days. His op ed defending himself after he quit Fortune and Reuters because they criticized him was the most widely read WSJ article for well over a day.
If you wander into a top business story and aren’t up to speed, as I said, it is NOT the job of this site to spoon feed you. You are the one who is not well informed and you seek to make that our problem. Sorry, it isn’t. This is a finance and economics blog, not the MSM. And your comments re context ARE a demand that we act like an MSM site.
I do not take well to unjustified attacks on guest bloggers and that’s what you have engaged in, persistently to boot. Nor do I take well to your concern trolling either. Hugh frankly has been way too patient with you.
Jane,
They don’t want criticism here. They think every thinking person who disagrees with anything they say is a troll. They want approval, and for the most part they get it. What they really want is adulation, and they get lots of that too. They just never seem to get enough.
@jake… the tea party is at the Koch shit house, your late.
Skippy… don’t make me drag all your old posts up, begging, Yves to let you co chair… you (Yves) need my (Jake’s) help to make things happen.
Speaking of context, long-time and/or careful readers would know that this statement addressing Hugh, “Don’t jump on the bandwagon of conspiracies that the GOP does” — in the imperative mood, no less! — is ludicrous in the extreme. Hugh has a track record, and he’s not about to jump on any kind of bandwagon, let alone a GOP one. You gotta know the territory, and you gotta know the bloggers. And speaking of the Doe family…
UPDATE Adding, there’s a famous joke, which I cannot remember, about a reviewer who eviscerates a book because it is not the book that they themselves would have written.
Are you referring to Kate Schutt? I’m outta my league.
Excene Cervenka — Mrs. John Doe.
I wish the press would discuss labor force participation.
I agree. It is the great untold story. It also can not be easily fudged because it is based on the non-institutional population over 16. Not only is the NIP a very big number and so quibbles about its exact size become less important but it inherently takes into account population growth.
It decreased from the 67% heights of the Clinton era to just over 66% by about August 2003. It then plateaued with some fluctuation at that level until August of 2008 and since then it has declined steadily to its current level of 63.6. When you consider that every percent equals 2.3-2.4 million people, that’s a big decrease, especially when you consider that these are people taken completely off the table and treated as if they don’t exist in terms of jobs, unemployment, etc..
I hope this is helpful. “Pepe” got caught between comments. Time stamps on “Hugh” and “Jane Doe” should be noted.
Hugh:
It is not much of an untold story. Bill McBride at CR has talked about it, Spencer England at Angry Bear has talked about it, Laurent Guerby addressed it with Thoma, and I have addressed Participation Rate at Slate as well as a few other places. U3 without addressing Particpation Rate is only 1/2 a loaf since the end of the 2001 recession at which point it was ~66.7% and has been on an acclerated downward trend. Few understood the politics of U3 and how little it meant once Participation Rate began to recede.
Male Participation Rate has been declining with the decrease in Manufacturing jobs while Female Participation Rate has been increasing as Serviced Related jobs. In any case the two earner income is still not enough as income has not kept up with increases. Really, is all of this Obama’s fault with the polices Bush and the Republican Party of Norquist has pursued? I think not. Obama may have his faults; but, isthere the alternative of R&R better?
Hugh:
I would look to the Bith/Death Rate as it causes wide fluctuations.
Re the lesser evilism argument,
If I offered you two shit sandwiches, and I told you one contained a little less shit in it than the other, would you eat it? I doubt it. So why are you asking the rest of us to?
I am not sure to what you are referring when you mention the birth/death rate. The non-institutional population over 16 means the population over 16 years of age with some exceptions. So birth is not involved. Growth in this population has a summer peak at around 200,000 and lows in winter of something over 140,000. It does not fluctuate wildly.
Or are you referring to the birth/death business model? This is used in the Establishment survey (jobs), but what I was addressing in this post comes from the Household survey. As I pointed out in the post, the Household survey has a much higher threshold of significance (400,000) and this does allow for more fluctuation in the data than in the much larger Establishment survey.
Hugh:
I am not the one who needs convincing on the issues of Unemployment and Employment. Much more could have been done and wasn’t by a conflicted Congress and a President who could have said more to force the issue. Much of the public believes there is a conspiracy afloat and also buys into the proverbial “shit” sandwich being offered to them concerning Obama. This has been going on far longer than our current President.
I am more interested in how the 873,000 came to be besides an uptick in Participation Rate.
You acknowledge an inconsistency here:
“These data were all from the Household survey were fundamentally at odds with an anemic increase in jobs of only 114,000 from the larger Establishment survey.”
and then allude it may be from 582,000 involuntary and part time workers. Is there more to this?
If you look at the graph of part timers, the uptick in part timers represents a recurrent glitch in the seasonal adjustment model. The unadjusted numbers represent actual people working on the ground. The adjusted numbers don’t although they are always reported as if they do. They are supposed to expose the underlying trend by smoothing the data, basically by suppressing the seasonal hills and valleys in the data. But in the graph, there is a smaller hill (268,000)in the unadjusted numbers which is paradoxically being amplified, rather than suppressed, in the adjusted data (582,000). Now because the adjusted August number was already higher than the unadjusted number, when the unadjusted number increased in September we should expect to see some or all of that rise erased in the September adjusted number. That would take out a major chunk of the 873,000 increase in the employed.
Something similar but with opposite sign happens with the unemployed numbers. The adjusted number chases the unadjusted one into a valley rather than smoothing the valley over.
Mix into this the greater variability in the Household numbers and we end up in the neighborhood of the 114,000 increase, seasonally adjusted, in jobs in the Establishment survey.
Parenthetically, the increase in jobs in September, seasonally unadjusted, was 574,000 which was again in the ballpark with the 775,000 increase in the employed, seasonally unadjusted. (The employed in the Household surve includes more classes of workers than jobs in the Establishment survey) Again add in greater less precision in the Household number and these two numbers aren’t necessarily that far off.
What is far off is what we saw in the adjusted numbers: 114,000 jobs and 873,000 employed. This requires explanation and that was what this post was about.
Sorry for the typos, especially “greater less precision” > less precision
Hugh:
Thanks for the return. Appreciated the remarks. Not too concerned with the typos as my are littering the enire column :).
Pepe:
Read Calculated Risk and Bill McBrid as the press couldn’t figure it out and are not interested.