In his January Economic Report of the President, George W. Bush promised that his economic plan would create 2.6 million jobs this year.
At this time, he’s about 1.3 million jobs short (plus another 6 million I’ll get to later).
So his supporters have jumped in trying to make the jobs picture look better, but they’re flat out wrong. (Links open in new windows)
It’s partisan, disingenuous obfuscation – they’re just trying to throw dust in our eyes.
[This is quite long to so if you’re interested, you may want to print it out.]
They do it by playing with numbers from the Bureau of Labor Statistics (BLS). The BLS provides two monthly jobs stats:
- Current Employment Statistics (CES), also known as the Payroll Survey or the Establishment Survey. This audits actual jobs and has proven its accuracy.
- Current Population Survey (CPS), also know as the Household Survey. Never designed to measure jobs – it’s for gathering unemployment estimates – this is a poll.
The good data on jobs is the Payroll Survey. This is based on hard data from a large sample.
The Payroll Survey is based on real employment numbers from 300,000 businesses, representing about 40,000,000 jobs. To put that in perspective, that’s about 30% of all non-farm jobs.
The poor data on jobs comes from the Household Survey. This is based on a small sample and is essentially a household poll.
For the Household Survey, the BLS calls and talks to people in 60,000 households representing about 70,000 jobs every month.
The sample in the Household Survey is thus about 0.05% of the jobs universe, compared to the Payroll Survey’s 30% of the jobs universe. A relatively few professionals too embarrassed to admit to unemployment and saying they’re "consultants" can have a disproportionate effect on the results. Sample size makes a difference in survey accuracy, and the Payroll Survey sample is nearly 600 times larger.
But don’t just take my word for it.
The Payroll survey is more accurate and provides much better numbers .
Payroll Survey and Household Survey accuracy.
The Payroll Survey is adjusted once a year on a lagged basis, with the last benchmark revision done through March. This revision roughly represents the overall survey error (sampling plus non-sampling errors) and has been very small:
In contrast, the Household Survey has a huge margin for error. The BLS says:
This means that last month’s job "increase" of 96,000 in the Household Survey could be off by 365%. Pretty sloppy, particularly when you contrast this to the average benchmark correction for the Payroll Survey.
In an attempt to deny the value of the Payroll Survey, they’ll tell you the BLS Payroll survey is very, very wrong because it misses new business start-ups for a few months.
But missed new start-ups are a non-starter, because the BLS compensates for this.
They use a formula based on known recent business birth and death rates to adjust the numbers before they are released. As a result, the worst-case residual error in the Payroll Survey because of unmeasured start-up businesses is nominal at worst.
The bottom line is that the BLS can say with confidence:
- When Bush took office, there were about 132,388,000 non-farm jobs in the country.
- Job losses started almost immediately and total employment bottomed out at 129,789,000 mid-2003.
- Some jobs have been coming back and last month there were 131,567,000 jobs, but there’s still a net loss. (Note that this is not a measure of the increase in unemployed – I’ll get to this later)
Additionally, the labor force grew by 1.8 million per year during the Bush administration, so another 6,750,000 more need jobs that do not exist.
The situation is thus much worse than the number of jobs available would indicate:
This administration is 7,577,000 jobs short of breaking even on jobs .
So the next time you hear the neoconservative Heritage Foundation, the American Enterprise Institute, Steve Forbes and other rabid Bush partisans making the wild-eyed claim that millions of new jobs have been created during Bush’s stewardship of the economy, tell them to get a real job.
The unemployment situation is grim, too.
January, 2001 started with 5,641,000 unemployed. Last month (09/2004) there were 8,003,000 unemployed.
The base net increase of unemployed in America during the Bush administration is 2,362,000.
Think about that – if you gathered them all together you would have a city nearly the size of Chicago.
But it’s worse than that.
The BLS also tells us that there are 1,500,000 workers counted as "marginally attached" and they need jobs, too. And yet another 4,720,000 who want jobs but are not counted in other government measures.
That adds up to 14,233,000 out of work across this country (about the population of New York, Los Angeles, and Chicago combined).
And the economy is supposed to be getting better?
That’s just voodoo economics totally disconnected from reality.
Powered by Sidelines