Eric Jaffa: Fake Job Numbers Paint a Rosy Picture
Fake Job Numbers from the Bush Administration Paint a Rosy Picture
by Eric Jaffa, June 7, 2004
How should the government count the number of jobs in America?
If it were up to me, the Labor Dept. would use the number of people who pay FICA taxes as the main basis.
However, it's not up to me, and the Labor Dept. doesn't use FICA in their calculation of the number of jobs.
Instead, they send out surveys to a sample of businesses asking them if they increased or decreased the number of jobs.
One of the reason not every business in the sample responds is that some of the companies have gone out of business.
There are also new businesses which the Labor Dept. doesn't write because it doesn't have their addresses yet.
New businesses and businesses which recently closed approximately cancel each other out.
But starting with its report for April 2003, the Bush Administration's Labor Dept. has been making assumptions about those businesses.
It calls those assumptions "The CES (Current Employment Statistics) Net Birth/Death Model."
The Labor Dept. explains this method as follows:
During the net birth/death modeling process, simulated monthly probability estimates containing continuous and imputed employment over a 5-year period are created and compared with population employment levels that contain actual business births and deaths along with the continuous units. Moving from a simulated benchmark, the differences between the series across time represent a cumulative error component. Those residuals are converted to month-to-month differences and are used as input series to the modeling process.
Models are fit using X-12 ARIMA. Outliers, level shifts, and temporary ramps are automatically identified. Seven models are tested, and the model exhibiting the lowest average forecast error is selected for each series.
While this may be a sincere summary from the Labor Dept., it ISN'T detailed enough to allow other economists to replicate their calculation.
It's a principle of science of that one's results should be replicable by other scientists.
Since the Labor Dept. isn't being forthcoming enough to let outside economists repeat their calculations, one must wonder if they run the calculation exactly same way each month, or if there is room for fudging the numbers.
Also, note the part of their summary which says that actual business births and deaths are part of the calculation. Are they saying they can obtain real numbers of businesses which opened and closed from long ago, but not for the past month? From where? What are they doing to get real numbers more promptly?
If they want us to believe these figures, telling us exactly where the raw data is from and step-by-step how it was processed would be the way to build trust.
The lack of transparency gives us no reason to believe the numbers.
But let's assume for a minute the Labor Dept. economists are sincere, and they do run the numbers for the CES Net Birth/Death Model exactly the same way each month.
There is still the question of whether the calculation makes sense.
Maverick by Conscience has issues with it, expressed in his May 11, 2004 blog-entry (emphasis added):
In a piece in the NY Post, John Crudele ferreted out one suspicious lever in the Department of Labor's calculations. It's called the "CES Net Birth/Death Model." Here's the key portion:
"Earlier research indicated that while both the business birth and death portions of total employment are generally significant, the net contribution is relatively small and stable. To account for this net birth/death portion of total employment, BLS is implementing an estimation procedure with two components: the first component uses business deaths to impute employment for business births. This is incorporated into the sample-based link relative estimate procedure by simply not reflecting sample units going out of business, but imputing to them the same trend as the other firms in the sample."
What this says is we don't count the job losses for businesses which fail, which no longer conduct business, instead we count jobs in these by looking at the average firm that is still doing business. If companies still in business are adding jobs then we plug numbers that indicate that the firms going out of business are adding jobs. We do this to offset the problem that job figures aren't available for newly created businesses.
To demonstrate the absurdity of this idea assume for a moment that there are only two companies in the economy and that both employ 10,000 people. Company A goes out of business. There are now only 10,000 jobs in the economy. Some of their customers, however, must now get their products from Company B, so Company B hires 2,500 people to produce products to meet this additional demand. There are now 12,500 jobs in the economy, a loss of 7,500 jobs.
No, that's not right says the Labor Department because Company C may have started up and we don't know how many jobs that might have created. We will adjust for that. Instead of going out of business we'll assume Company A stayed in business and like the average company still in business (Company B) added 25% more jobs. So they now have 12,500 jobs as does Company B and the total jobs in the economy is 25,000. The number of jobs in the economy grew 25%. Wow!
Yes, I have simplified the starting economy for clarity, but the point is clearly made: this model is a rosy predictor in bad times. (Hint: you only need one for bad times.) When many businesses are failing it assumes that they remain in business and increase or decrease jobs at the average rate of those firms which are successful at remaining in business. It is designed to show things better than they are.
For April 2004, 94% of the supposed new jobs were estimated based on the "CES Net Birth/Death Model, as this blog entry by air23cal discusses (emphasis added):
In case you have been paying attention, we added 288,000 jobs last month. What the media and the Bureau of Labor Statistics (BLS) fail to publicize is that 270,000 of those jobs are "estimated".
How could this be you ask? Well, the Current Employment Statistics (CES) employs a method of estimating how many jobs are created out of new business ventures. Since there is a lag time between a new business opening (or closing) and showing up on the payroll survey, the CES estimates how many jobs are created/lost based on historic benchmarks.
This has come to be known as the "birth/death model".
Their justification for these estimates is that the birth/death model's "...net contribution is relatively small and stable ". According to the CES, last month 270,000 jobs were added from the birth/death model.
I find it funny that out of 288,000 jobs added in April, the "small and stable" part accounted for 94% of the results.
This is either further proof that even at the current employment rebound, we aren't creating enough jobs or that our government is manufacturing numbers during an election year.
The other scary part about the birth/death model is that the "...models do not attempt to correct for any other potential error sources in the CES estimates such as sampling error or design limitations". Also from the BLS-CES site: "The most significant potential drawback to this or any model-based approach is that time series modeling assumes a predictable continuation of historical patterns and relationships and therefore is likely to have some difficulty producing reliable estimates at economic turning points or during periods when there are sudden changes in trend."
History might be able to make us aware of certain trends but when you start making predictions into facts, I have a major problem with that. All in all folks, these employment numbers are manipulated and can't be trusted.
I am not making this stuff up either, I got this information straight from the official BLS site. I encourage you to check out the following links so you can judge for yourself how we make up these monthly reports:
New York Post reporter John Crudele wrote a May 11, 2004 article on the"CES Net Birth/Death Model" which first alerted me to it.
John Crudele's article notes (emphasis added):
As staggering as the assumption about new companies was in March, the Labor Department got even more brazen in April.
Last Friday, it was disclosed that these imaginary jobs had been increased by 117,000 to 270,000 for the latest month - because, I guess, the stat jockeys got a vision from the gods of spring.
Without those extra 117,000 make-believe jobs, the total growth for April would have been just 171,000 - sub-par for an economy that's supposed to be growing at more than 4 percent a year, but right on the pros' targets.
Take away all 270,000 make-believe jobs and, well, you have the sort of pessimism that the political pollsters are seeing.
With regard to May 2004, I'd like to see an explanation of this statement from the Bureau of Labor Statitics website:
Nonfarm payroll employment rose by 248,000 in May, and the unemployment rate was unchanged at 5.6 percent.
How can the net number of jobs increase by 248,000 in a month without decreasing the unemployment rate?
This seems like an internal inconsistency for the May 2004 numbers, just as the Labor Dept.'s statement that the "Birth/Death Model" is a "small and stable" part of job growth, and yet accounts for 94% of new April 2004 jobs, seems like an internal inconsistency.
Democrats shouldn't shy away from discussing jobs because of these fake, rosy numbers.
We should speak out, saying that starting for April 2003 the Bush Administration used a new way of estimating the number of jobs. A new method which the Labor Dept. seems to be exploiting recently.
There is no reason to believe these numbers when the Labor Dept. isn't forthcoming about exactly how they were produced, to allow the figures to be independently verified.
There are too few jobs, and we need new leadership.
Eric Jaffa is the editor of MoveLeft Media at moveleft.com.