[kenrexford]

Maybe someone can help me with this.

I read an AP story today discussing the amazing drop (before the election) of the unemployment rate to 7.8%!!! Wow, that sounds good.

Except, I am having truoble with the numbers. The BLS says we added 114,000 jobs in September. The unemployment rate is based upon surveys, and this month's survey indicated that the number of people claiming to be employed soared by 873,000. (The adjustments to prior month's jobs numbers would not matter, because the surveys back then would have accounted for the real situation, not the BLS being supposedly off. Plus, that was only up about 67,000.)

So, how could we add 114,000 jobs but have 873,000 more people employed?

An easy answer could be that (873000-114000=) 759,000 newly "employed" people are simply unemployed people who support Mr. Obama claiming that they are really employed, to help out the cause, but that would be quite cynical.

[Flem72]

kenrexford, on 2012-October-05, 07:48, said:

Maybe someone can help me with this.

I read an AP story today discussing the amazing drop (before the election) of the unemployment rate to 7.8%!!! Wow, that sounds good.

Except, I am having truoble with the numbers. The BLS says we added 114,000 jobs in September. The unemployment rate is based upon surveys, and this month's survey indicated that the number of people claiming to be employed soared by 873,000. (The adjustments to prior month's jobs numbers would not matter, because the surveys back then would have accounted for the real situation, not the BLS being supposedly off. Plus, that was only up about 67,000.)

So, how could we add 114,000 jobs but have 873,000 more people employed?

An easy answer could be that (873000-114000=) 759,000 newly "employed" people are simply unemployed people who support Mr. Obama claiming that they are really employed, to help out the cause, but that would be quite cynical.

Almost as cynical as leaving out the 3M or so non-statistics who are outside the reporting system....

[kenrexford]

Flem72, on 2012-October-05, 08:19, said:

Almost as cynical as leaving out the 3M or so non-statistics who are outside the reporting system....

Of course. I mean, we all know that these numbers are misleading. Ignoring the shift from full-time to part-time is one thing, as is not counting people who are "not looking," as is forgetting population growth.

But, there is an amount of BS that is expected, while this one seems astonishing, unless I am missing something. You just cannot get 873,000 more people employed with 113,000 jobs.

[Cyberyeti]

The BLS is nonfarm jobs. A small part could be people employed seasonally in agriculture. Also it excludes the self employed where the household survey doesn't, have a lot of people started up one man businesses ? It is also a sample which is extrapolated, if the sample isn't representative on the small business front, it's possible the two numbers might be wildly different.

[jdeegan]

I am, actually, a real economist with some knowledge of this subject. The 7.8% number may well be a pre-election scam, but the way the numbers are put together makes it hard to tell just now. To begin with, the 7.8% number comes from the Household Survey (Bureau of Labor Statistics) which is a poll of a random sample of U.S. households, so there is some sampling error of plus or minus (if memory serves) several tenths of one percent.

The jobs number comes from running the ES202 Social Security tapes (from Treasury, I think), which is more or less a census of employers (done by the BLS). They don't do this every month (I think it's now done every three months). The other two months they project based on a sample survey drawn from the most recent ES202 tapes. When the next ES202 data come available they revise the preceding two months.

One must also consider people who drop out of the labor force. Some of them are, in fact, hardship cases. Also, students have been staying in school longer since there are few decent jobs available. However, there has been a big increase in age-related retirement since the front edge of the baby boomers are reaching 65-70 years of age.

Bottom line is that reconciling data from three sources - the Household Survey, the ES202 data and the labor force data (also from the Household Survey, I think)- is pretty near impossible except maybe months or years after the fact. If it is a scam, the perps know it can't be uncovered until after the election, if ever.

Is the U.S. economy getting better? Yeah, but not very fast. There are three indicators:

1.the unemployment rate from the Household Survey-fairly positive

2.the jobs numbers from the ES202-positive, but weak

3.GDP growth from the Commerce Dept.-positive, but very weak

Obama has been a weak and 'do nothing' president. Romney would probably do better.

[WellSpyder]

It is quite normal to get two types of employment statistics - one set based on information from employers about how many people they are employing ("jobs"), the other set based on information from individuals on how many of them are working ("workers"). And it is quite normal for them to tell significantly different stories (my main experience is with the UK data). To some extent this can true - if, for example, the number of job-shares (1 job, 2 workers) or the number of part-timers with a "portfolio" of jobs (2 jobs, 1 worker)changes. But it can also reflect differences in timing, sampling variation, or just plain mistakes in responding to surveys.

[hrothgar]

jdeegan, on 2012-October-05, 08:49, said:

I am, actually, a real economist with some knowledge of this subject.

Tell me, oh economist, who is responsible for calculating these statistics of which you speak?

My recollection is that the bureaucrats responsible for calculating these figures are quite deliberately firewalled off from the political arms of the government.

It is certainly possible for the executive wing to try to "goose the economy" using fiscal or monetary policy, but its kinda hard to reconcile this kind of fine grained policy control with the "do nothing" president that you describe.

[blackshoe]

"The government are fond of collecting statistics, and of basing decisions on those statistics. However, they should remember that the statistics are written down by the village watchman, who writes what he damn well pleases." -- Sir John Stamp, ca. 1895

[kenrexford]

Let me see if I get this correctly. We have 873,000 fewer people unemployed in one month, and that might be because of a sampling error. We also have polls out showing a specific view of the likely voters, but that might be off a bit because of a sampling error.

I think I am going to start using this theory in bridge.

For example, suppose that I want to psych a 1♠ opening with ♠ K 10 9 8 2 ♥ 10 2 ♦ 9 4 3 ♣ 8 5 4. I sample my first two cards (King of spades, and the 10 of spades). I extrapolate out that this means an average of 19.5 HCP, plus something for the 6.5 10's in my hand. Thus, my 1♠ opening, if anything, was an underbid.

[PassedOut]

Here is a description of how the BLS counts jobs and how those numbers have changed for each president since FDR: Jobs created during U.S. presidential terms. Since Clinton left office, the picture has not been good despite the repeated tax cuts. Just to keep up with population increase the US needs to add over 130,000 jobs per month, so only the amount above that actually chops down unemployment.

Where I live, in Upper Michigan, I do see a lot of self-employment. Folks cut up fallen trees and sell firewood. Folks pick wild berries and sell them at roadside stands. Folks put plows on pickups and arrange to plow driveways, or just drive around looking for jobs whenever it snows. Folks put up signs in yards advertising handyman services. And on and on. It would not surprise me to learn that some of the income is not reported to the IRS. And some guys drive to North Dakota to pick up work for a few months each year, sending money to their families back home.

[kenrexford]

PassedOut, on 2012-October-05, 09:27, said:

Here is a description of how the BLS counts jobs and how those numbers have changed for each president since FDR: Jobs created during U.S. presidential terms. Since Clinton left office, the picture has not been good despite the repeated tax cuts. Just to keep up with population increase the US needs to add over 130,000 jobs per month, so only the amount above that actually chops down unemployment.

Where I live, in Upper Michigan, I do see a lot of self-employment. Folks cut up fallen trees and sell firewood. Folks pick wild berries and sell them at roadside stands. Folks put plows on pickups and arrange to plow driveways, or just drive around looking for jobs whenever it snows. Folks put up signs in yards advertising handyman services. And on and on. It would not surprise me to learn that some of the income is not reported to the IRS. And some guys drive to North Dakota to pick up work for a few months each year, sending money to their families back home.

As a criminal defense lawyer, I can also offer several examples of self-employed people who do not report this to the IRS. Maybe 759,000 new drug dealers have entered the market?

[hrothgar]

kenrexford, on 2012-October-05, 09:22, said:

For example, suppose that I want to psych a 1♠ opening with ♠ K 10 9 8 2 ♥ 10 2 ♦ 9 4 3 ♣ 8 5 4. I sample my first two cards (King of spades, and the 10 of spades). I extrapolate out that this means an average of 19.5 HCP, plus something for the 6.5 10's in my hand. Thus, my 1♠ opening, if anything, was an underbid.

You might want to familiarize yourself with something called a "confidence interval" before you go much further down this road

[WellSpyder]

kenrexford, on 2012-October-05, 09:22, said:

Let me see if I get this correctly. We have 873,000 fewer people unemployed in one month,...

No, sorry, I don't think you do get it correctly. The number of extra people with jobs is not the same as the number of fewer people unemployed!

There are lots of reasons for this. New people are joining the labour force all the time, eg school leavers and immigrants. Of course, people are leaving the labour force for opposite reasons, but not as many. And there are lots of people who are neither employed nor unemployed (eg students, carers, housewives/husbands or whatever) and the numbers of these vary from month to month - including variations dependent on the state of the labour market, even though these people may say they are not actively seeking employment. Again, I'm more familiar with UK data than US data, but I think it would be more normal to expect the change in unemployment to be half the change in employment than to expect them to be equal, even before allowing for differences caused by growth in the population.

[jdeegan]

hrothgar, on 2012-October-05, 09:03, said:

Tell me, oh economist, who is responsible for calculating these statistics of which you speak?

My recollection is that the bureaucrats responsible for calculating these figures are quite deliberately firewalled off from the political arms of the government.

It is certainly possible for the executive wing to try to "goose the economy" using fiscal or monetary policy, but its kinda hard to reconcile this kind of fine grained policy control with the "do nothing" president that you describe.

The drudges at BLS are career civil service. Their top level bosses are politically appointed super grades - almost all Democrats for Obama. I rather doubt any foolishness took place. It would be quite a trick. The super grade guys would have to slip one by the rank and file workers. This would be easier done with the Household Survey's 7.8%. The data come in in bits and pieces, so it is conceivable a relatively small cabal of mostly super grades with a few drudges could do it.

'Weak' and 'do nothing' have to do with Obama's inability to implement any kind of coherent fiscal policy. The fool allowed the creation of the 'fiscal cliff'. He seems to be tone deaf when it comes to negotiation. Ask (off the record) almost anyone at the Fed what they think of Obama's economic stewardship. QEIII is a desperate attempt to at least do something.

[kenrexford]

hrothgar, on 2012-October-05, 09:31, said:

You might want to familiarize yourself with something called a "confidence interval" before you go much further down this road

Many of my partners experience a confidence interval problem every time I bid.

[kenrexford]

WellSpyder, on 2012-October-05, 09:40, said:

No, sorry, I don't think you do get it correctly. The number of extra people with jobs is not the same as the number of fewer people unemployed!

There are lots of reasons for this. New people are joining the labour force all the time, eg school leavers and immigrants. Of course, people are leaving the labour force for opposite reasons, but not as many. And there are lots of people who are neither employed nor unemployed (eg students, carers, housewives/husbands or whatever) and the numbers of these vary from month to month - including variations dependent on the state of the labour market, even though these people may say they are not actively seeking employment. Again, I'm more familiar with UK data than US data, but I think it would be more normal to expect the change in unemployment to be half the change in employment than to expect them to be equal, even before allowing for differences caused by growth in the population.

So, 873,000 people got new jobs this past month! That has to be about the best figure of all time. At that rate, we should have about 10.5 million new jobs on one year, or about 42 million more jobs in four years. I now know who to vote for.

[PassedOut]

kenrexford, on 2012-October-05, 09:54, said:

So, 873,000 people got new jobs this past month! That has to be about the best figure of all time. At that rate, we should have about 10.5 million new jobs on one year, or about 42 million more jobs in four years. I now know who to vote for.

Nate Silver puts the number in perspective in his blog today: Jobs News Makes Obama’s Case Easier

Quote

But unlike last month, when a decline in the unemployment rate was caused by the exit of workers from the labor force, the household survey also reflected genuinely good news in September. According to that survey, 413,000 workers joined the labor force in September. But 873,000 more people became employed, causing the unemployment rate to fall to 7.8 percent.

If the September numbers resulted in part from statistical variance, it is certainly possible that there will be some payback in the October report, which will be released the Friday before the Nov. 6 election.

But it is also possible that the strength shown in the government’s report on Friday reflects it playing catch up. The firm ADP, which tracks private-sector payrolls, had reported that an average of 170,000 private-sector jobs had been created each month so far this year. The ADP reports are much maligned because they do not always match the government’s payroll figures over the short run. But in the long run, the numbers tend to converge.

Furthermore, there has been a fairly consistent pattern of upward revisions to the government’s jobs reports recently.

I suppose we'll have to wait until after the elections to see if ADP has been right all along.

[kenberg]

kenrexford, on 2012-October-05, 07:48, said:

An easy answer could be that (873000-114000=) 759,000 newly "employed" people are simply unemployed people who support Mr. Obama claiming that they are really employed, to help out the cause, but that would be quite cynical.

A bizarre explanation more than a cynical one. Think of three quarter of a million unemployed people sitting around, wondering how they can help Obama keep his job. Suddenly, spontaneously across the nation, they realize they should lie to the pollsters. No, I don't think so.


But I completely agree with what I take to be your larger point, the numbers we see in the media are often incomprehensible junk. The reporters just take whatever numbers they are given, hand them out, and ask no questions. I suppose the numbers you cite come from somewhere and are accurate as far as they go. Some number really is 7.8% of some other number, the difference of two numbers really is 873,000, and the difference of two other numbers really is 114,000. But without further guidance, the numbers appear to be irreconcilable. Most any reporter surely has the intelligence to see there is a problem, s/he apparently doesn't care enough to act on this observation.

[awm]

The unemployment rate includes people joining/leaving the work force (i.e. if I give up looking for a job, I'm no longer unemployed; if I go back to school, I'm no longer unemployed) so it can easily go up or down independently of the "jobs created."

The jobs created marks are seasonally adjusted; typically the retail industry starts ramping up for the holiday season around now so this is probably a downward adjustment from "actual jobs" to "seasonally adjusted jobs" in September/October (and the reverse in January/February).

Various different methods are used to compute these jobs, and results may differ, and further they try to adjust for inaccuracies in previous months, and that has some impact.

The overall point is that you can't read much into top-line numbers. The general view is that the economy is headed in a mildly positive direction (with jobs and growth comparable to what we saw in most of 2001-2007), we are not headed into a new recession (unlike much of Europe), but that we have a lot of ground to make up after the disaster of late 2008 and early 2009 and we're not making much progress on that.

[jdeegan]

kenrexford, on 2012-October-05, 09:54, said:

So, 873,000 people got new jobs this past month! That has to be about the best figure of all time. At that rate, we should have about 10.5 million new jobs on one year, or about 42 million more jobs in four years. I now know who to vote for.

Turns out the Household Survey is based on a monthly telephone survey of 50,000 persons. At first blush this looks like a reasonably adequate sample, esp. if it can be stratified. What is spooky to me is what this means in our new world of cell phones. Disposable cell phones. Personal cell phones. People who have a land line and a cell phone. I'm sure they must address this, but it could be a problem inherently too difficult for the bureaucrats doing the job, or maybe anybody for that matter. I was once a consultant to the BLS back in the 1980's working in the ES202 area. Nice people but .................

The jobs data are based on a very frequent census using the ES202's. No real sampling frame problems here. These numbers have to be much more reliable. Their only shortcoming is missing off the grid employment.

After a little thought, I am beginning to believe that any hanky panky can probably be uncovered by analyzing in detail the most recent six months of Household Survey data. The 873,000 number supposedly includes revisions from past months. I don't want to do it, but I'm sure there are plenty of Republican analysts who are working on it right now. We shall see what we shall see.