The Bias in Reporting Job Losses

Each day's news brings more stories about layoffs at major companies. The stories get a big play in mainstream media. The leading bloggers also cite the stories and encourage readers to keep a summation of job losses.

This is quite misleading. Job losses occur in highly visible chunks, as we can readily see. New jobs are created a few at a time, both in existing businesses and in new businesses. Even sophisticated observers do not recognize the ongoing job creation from the invisible hand of the market.

Try This Headline

Suppose that the New York Times or the Wall Street Journal had a headline:

100,000 New Jobs Created Today.

Actually, they could run that headline every business day, even during the recession, and it would be accurate.

How do we know? As usual, we start with data. The best illustration available is from the last recession, so let us look back to 2001.

The 2001 Example

The data presented here are drawn from the Business Dynamics series from the BLS. The data are not from surveys. The evidence is from state employment data. Since no one reports employment, and pays taxes, on phantom jobs, these are data that we should believe. Here is the evidence.

Source: Bureau of Labor Statistics seasonally adjusted data. (Unadjusted data show the same result for the year).

As one can readily see, over 35 million jobs were lost during the year. That is what you would get if you added up all of the layoff announcements and also included job cuts that did not make the newspaper. What most people do not realize is that over 32 million jobs were added. This development is not publicized.

By focusing on gross job losses we get a false impression of the problem. The net losses are bad enough; there is no reason to exaggerate.

Conclusion

There are three important conclusions:

1. About 90% of announced job losses were offset during the same month by job gains. We should be taking a 90% haircut to those newspaper articles.
2. There was substantial creation of new jobs in opening establishments, a total of over 7 million for the year. That is something to remember the next time someone scoffs at the idea of business births during a recession.
3. President Obama dropped the tax credit for new jobs, and that is a good thing. There is no way to separate the new jobs from the credit from those that would have occurred anyway. If the credit were paid for gross new jobs, the money would be gone in a couple of months.

Most importantly, this shows that we should remember that net job change is the key economic concept. That should also be our policy target.

Comments

[brokerinbir...:]

I would tend to agree with the writers who have stated that employers do not pay taxes on phantom jobs. If you disagree, when was the last time you asked a chasier o go ahead and add an extra % to the sales tax? When ws the last time you padded your income tax return to purposely pay more taxes? The is no incentive to do that unless you're a left-winger who actually beoieves you get your money's worth out of paying extra income taxes & extra payroll taxes for phantom employees.

Jan 29 11:19 AM

[jetbeagle:]

old limey: "This single example does not tell us anything, either."

My single example wasn't meant to prove anything. it was meant to show that John Gordon's statement:

"new jobs if created are @ lower wages"

is not correct. Some new jobs are low wage. Some command high salaries.

Jan 29 11:43 AM

[Jeff Miller:]

I appreciate the spirited discussion. It is always helpful in sharpening things up for future articles.

For the moment, I would like to add a couple of thoughts. This is not an "opinion" piece or a matter of looking at something as half-empty or half-full. My articles on the monthly employment change have warned of even greater losses than the consensus estimates. A net job loss of over 500K per month is a very serious problem.

I am providing information that you probably have not read anywhere else, because nearly everyone focuses on gross job losses as the story. If you start thinking about the labor market as a place where millions of jobs are both lost and created, in good times and in bad, you will be able to interpret data more objectively.

Finally, for those skeptics who will not believe any data from the government, even when based upon tax receipts, please try this little exercise. Initial jobless claims have been running at about 500K per week for some time. If there was not new job creation ---- and plenty of it -- we would be seeing net job losses of over 2 million per month.

Think about it. This is a dynamic system. Most media and bloggers focus on only one part of it. I am trying to help everyone see the entire picture. It is a difficult topic to explain.

Thanks again for all of the comments!

Jeff

Jan 29 12:31 PM

[jetbeagle:]

OldLimey,

I just realized your point about "burth/death stats" referred to business creation and firms going out of business.

I think you are correct that BLS estimates off-the-radar job loss/creation. Some critics contend those estimates are so inaccurate as to bias completely the total employment figures published by the BLS.

IMO, BLS estimates of jobs are probably more accurate than any other available. BLS has responded to criticism and improved their methods. BLS certainly has more eyes watching their estimates and methods than are watching any other projections.

The alternative to using government data is, I suppose, to go with one's gut feel. I have no idea how anyone could have an unbiased gut feel that accurately measures a $14 trillion economy.

Jan 29 01:07 PM

[canonicalman:]

The population of the US grows by about 2.7m/yr which is about 0.9%/yr. This means that the total number of jobs must grow each and every year by roughly that same fracti8on. This is another reason why jobs number are misleading. Even if the net jobs number stays even, this means an effective rise in unemployment because the new jobs for the new workers were not created.

Jan 29 02:07 PM

[Thoreau:]

How do you know it is a $14 Trillion dollar economy? 20% to 30% of the GDP is based on hedonics and imputations, they aint real dude plus the deflators for real GDP are all over the place which makes the number at best an indication of activity. But keep on believing dude the whole economy is one big fantasy, ponzi scheme.

Jan 29 02:31 PM

[phaf:]

Of course, the unemployment figures put out by the government are based on its model, which is constructed from actual data (reflection of reality) with certain assumptions factored in to the structure of the model. Like all models, inaccuracies arise from less than perfect data and less than perfect assumptions. I'm sure the reported figures are decent, because if they were way off, someone would devise a better model. The marketplace of ideas!

It would be nice if the figures that get reported more closely reflect "unemployment" without the qualifiers (i.e. who is counted). To me, the unemployed are people out of work. I would be in favor of the SGS alternate figures being the ones that are reported; while they sound worse, they more closely reflect reality (how many out there are unemployed!) What you do with the guys under the bridge who are happy right where they are, just leave them out.

Jan 29 02:47 PM

[satguru:]

I thought that unemployment figures are based on working people losing their jobs. What has birth rates got to do with employment. Babies are not employed many years after birth. Death rate would affect unemployment figures only for people who are employed and dies. The government are bogus because they do not include people who have fallen off the unemployment roles when it expires. They do include part workers who should be called fractional workers and allotted a fractional number as being employed.
On the point of new jobs being created offseting the ones that are given the pink slips, I beg to differ. If this was the case we would not see so many businesses shuttered. Look at Circuit City alone, over 600 stores being shut down, Linen'N Things, the shuttered Wall Street Investment banks, the banks and brokerages and the list goes on and on. All I can say from personal experience from my own close family members almost 40% have been given notices of layoffs or their hours cut or cuts in pay.
These are unprecedented times and the govrenment stats are bogus - I worked for the FEDs and we were ordered to make numbers reported
make the General for the base look good- and guess what it always did!
Look at all the home foreclosures - it's all the result now of people losing their jobs. We are in a state depression not a recession.
God help us all.
On Jan 29 02:07 PM canonicalman wrote:

> The population of the US grows by about 2.7m/yr which is about 0.9%/yr.
> This means that the total number of jobs must grow each and every
> year by roughly that same fracti8on. This is another reason why jobs
> number are misleading. Even if the net jobs number stays even, this
> means an effective rise in unemployment because the new jobs for
> the new workers were not created.

Jan 29 03:05 PM

[john s. gordon:]

jetbeagle - congratulations on your ability to stem the tsunami of lost jobs. wish more companies could do what you are doing.
> jack

Jan 29 03:14 PM

[jetbeagle:]

canonicalman: "This means that the total number of jobs must grow each and every year by roughly that same fraction. This is another reason why jobs number are misleading."

Well, I think the total number of jobs is a misleading number if one misuses that number. I don't think anyone claimed that the total number of jobs indicates how well the supply of jobs is meeting the demand for jobs. So i don't see how anyone should be misled.

The total number of jobs does indicate whether and how much the workforce is growing or shrinking. Yes, we would hope that it would grow as fast as the number of available workers. But that wasn't the point of Jeff's article, of course. He simply pointed out that focusing on total job losses rather on net job change is misleading.

Jan 29 03:15 PM

[Lilguy:]

The question is not so much "bias"--implying an intent to mislead by either omission or commission--as it is error in federal unemployment and employment reporting.

The BLS monthly report on employment/unemploymen... includes a jobs "birth/death" model, for example, which has been tremendously inaccurate over the last year (which I think your table tries to capture). In particular, it has been recording increases in employment in construction and entertainment over the last year, two of the hardest hit sectors of the economy. Who knows, it may be understating growth elsewhere, but I doubt that.

The problem is that the statistical models in use DO NOT account for sharp turns in our economy until well--as in quarters--after the event. That is a key reason that the consensus of economists has been understating the downturn in about every economic variable it has tried to project, from weekly unemployment numbers to quarterly GDP, much less next year's GDP value.

They'll catch up sometime, but probably not until the market and economy have turned upward.

Jan 29 03:21 PM

[jetbeagle:]

phaf: "It would be nice if the figures that get reported more closely reflect "unemployment" without the qualifiers (i.e. who is counted). "

I think what you seek is the employment to population ratio. This statistic, published monthly by the BLS, compares the number of employed adult civilians with the total number of adult civilians. The denominator would thus include not only "the guys under the bridge who are happy right where they are", but also the infirm, the retirees, the independently wealthy heirs, etc. as well as those working and seeking work.

The employment to population ratio:

dec-2007......62.7
mar-2008......62.7
jun-2008.......62.4
sep-2008......61.9
dec-2008......61.0

So, yeah, it's down, just as it goes down in every recession. But it's nowhere near what it would be in a depression.

www.bls.gov/news.relea...

Jan 29 03:40 PM

[gordon:]

The birth/death model the BLS uses tries to capture the newly created jobs, but it is NOT ADJUSTED FOR RECESSION/CONTRACTION! They have
added retail jobs before a dismal holiday season, when retailers
said they did NOT hire, they add construction jobs, finance, at a time when the mortgage industry is going BK. It is a SHAM, and they are supposed
to recapture these SHAM PHANTOM jobs in Feb, the report coming up
is suppose to be adjusted for these BOGUS added jobs. Musings around
today in the media hint they may have changed the methodology again to COOK THE NUMBER. Bush appointed the commissioner in 2007. He
installed a BUSH LOYALIST in every dept of government, per Karl Rove. I want this guy investigated. They deliberately lowball the inflation imput
costs to RAISE THE GDP. They add DEBT CREATION as a positive imput to GDP. They count PRISON INMATES AS MANUFACTURING, as
is fast food. They use the same 60K list for the hgousehold survey repeatedly, only calling about half. The created the BOGUS birthing of jobs by going to each Federal Reserve Bank District and combing the new filings for corporations, many of which are "S" or just sole proprietors looking for a tax-dodge, and the BLS artificially decides how many employees these new corporations will hire. This is for starters!

Jan 29 08:52 PM

[gordon:]

One more thing. Try to find information on the commissioners @ the BLS on the internet! You can't find any history on their party affiliation, family in politics, loyalties and cronyisms. I say they are there. Schooling, yes, formidable, if you can even find that. Secrecy abounds.

Jan 29 08:57 PM

[Mr. Ed, Jr.:]

At many businesses, work days, hours, pay and/or benefits are being cut to avoid having to cut employees loose. These reductions have an overall cumulative effect that is equal to x number of job losses. What statistical adjustment is being made for these cuts ?

Jan 29 10:14 PM

[Glen L.:]

Lies, damn lies, and statistics. It's a depression if you're unemployed, a recession if those other guys are unemployed.

My wife and I went out to eat tonight at a local tapas restaurant that was jammed-packed the last time we went there on a Thursday night. Tonight it was less than 1/4 full. I talked to the owner, who admitted things are really slow, even considering that it's January and the weather has been pretty bad lately. But this is in the supposedly recession-proof D.C. area with all of its government workers (oh, did anyone mention that government employment is UP in spite of the recession?), and the owner says this is pretty scary.

Okay, this is only anecdotal, but enough anecdotes will eventually prove the point that we are headed into depression. Give it a year to play out before you call me a fool, please.

Jan 29 10:30 PM

[bbzz24:]

the folly in this B**LS*** is that jobs are not created and destroyed, simply people are replaced. it is very hard to believe 1/3 of all employed change the nature of what they are doing for a living.
and then you have the small business birth/death model that adds more B**LS*** to the picture.

Jan 30 02:52 AM

[jetbeagle:]

sittingoutthisrecessio... "These reductions have an overall cumulative effect that is equal to x number of job losses. What statistical adjustment is being made for these cuts ? "

It seems that both you and phaf want the government to report one single number that tells everything happening to the nation's economy. Such a number would be dangerous in the hands of those who couldn't understand it, though.

Here's a couple of other statistics that should probably be reported by the media along with the unemployment rate and the labor force participation rate:


Weekly hours worked (private production and non-supervisory workers)

dec-2007......34.1
mar-2008......33.8
jun-2008.......34.1
sep-2008......33.6
dec-2008......33.3

Again, this number is down slightly from a year ago, but nothing to indicate depression level pain.

Average weekly earnings (private production and non-supervisory workers)

dec-2007......$605.28
mar-2008......$605.70
jun-2008.......$612.44
sep-2008......$611.86
dec-2008......$611.72

In inflation-adjusted dollars, weekly earnings have probably declined slightly the past year. Giiven that energy prices have plunged, it seems likely that those who retained their jobs are no worse off than they were a year ago.

There's no question that a few households suffered layoffs and are having trouble coping. But this recession is still a recession - not a depression.

Jan 30 11:03 AM

[Mr. Ed, Jr.:]

jetbeagle, thanks for the stats.

Jan 31 02:36 AM

[Thoreau up:]

Thoreau, (and Whistlewhat, Angry Citi Investor, Angry Citi Shareholder–I have to admit it is difficult to keep track of all the aliases you are using to post identical comments while appearing to be different bloggers. Do you talk to yourself also?)

I agree. Your comments about Hamersely just don’t make any sense at all. I too am highly skeptical that your bold statements about Hamersley could be based on any reliable evidence at all. I also read Travails in Tax and personally observed Hamersley’s testimony before the Senate Finance Committee. He does clearly seem to be an exceedingly honest guy to me. Yeah, how do you explain that KPMG stated Hamersely had absolutely no involvement or knowledge of tax shelters in its press release to the Senate Finace Committee after Hamersley testified in October 2003? www.pbs.org/wgbh/pages... See also Hamersley Senate Finance Committee Testimony 003 TNT 204-35 online at finance.senate.gov/hea...

Are you suggesting Hamersley and KPMG are in cahoots? Wow, that would be a bold strategy seeing as Hamersley sued the crap out of them. Case No. BC 297209, Los Angeles Superior Court (June 23, 2003.), also reported in Tax Notes Today full copy of complaint 2003 TNT 124-5.



On Jan 29 08:07 AM Thoreau wrote:

> The BLS are Government numbers, who evens knows where they come from.
> Of course U-6 unemployment is about 12.5% and likely understated
> by a few hundred basis points (and that is not even including all
> the recently unemployed illegals). How about that for a headline.
> All government numbers are a fantasy kind of like in the state of
> California where the government is busy confiscating citizens’ income
> by using fraudster tax shelter crusader lawyers like Mike Hamersley.
> Hamersley is a high level tax lawyer fraudster (by his own definition)
> working for the California FTB. Hamersley lied to the Senate, lied
> to his clients and lied to the courts in his lawsuit against KPMG.
> Hamersley destroyed many families' lives with his lies as he helps
> California steal income from ordinary citizens under his hypocritical
> convoluted interpretation of the tax laws. Of course this should
> come as no surprise since the fiscally and morally bankrupt state
> of California is stealing ordinary citizens' income by not paying
> refunds of taxes owed and considering illegally issuing what could
> be described as its own currency, IOUs.
>
>

Apr 08 10:06 AM