Foreword
Friend and client, Kevin Murtaugh, writes some very interesting material under the umbrella, "View from Silicon Valley." In his article appearing here, Kevin examines whether what you see is what you actually get when it comes to California employment data. (NOTE: Kevin's views do not necessarily reflect those of Gillespie Research Associates.)
_____
October 21, 2004
View from Silicon Valley - "California EDD: Let's Just Make Up the Answers"
By Kevin Murtaugh
© View from Silicon Valley, 2004. All rights reserved.
_________
California's Employment Development Department's (EDD's) September payroll survey released October 18 claims "California's Unemployment Rate Unchanged at 5.9%" followed by "Payroll Employment up 4,900 jobs." This is trumpeted locally as wonderful news, especially on the heels of August's +4,300 new jobs.
California's EDD reports numbers which seem similar to the Bureau of Labor Statistics. Both the EDD and BLS publish their underlying hard numbers, implying you can determine if their results are just "spin." The EDD and BLS both changed their measuring methodology this year. However, you can still easily find pre-revision EDD data(*).
Reviewing EDD's data and comparing it to their latest press release brings to light a number of concerns:
Problem #1: In a universe of 14,539,100 California payroll jobs, +5,000 new jobs is chump change.
To be fair, the EDD reports unadjusted payrolls last month were +46,000. Is it possible that last month the EDD and BLS were both hoist by their own petard? Did the bureaucracies' endless machinations actually damage the re-election efforts of the incumbent politicians?
The EDD's unadjusted payrolls YTD are +245,000 (+1.69%; +87,100 and +0.6% adjusted). Spread over nine months, in an economy touted as strong, with generation-low interest rates on top of state and federal governments both pumping out record amounts of money over and above collections, this is awful.
Problem #2: September's 14.5M total jobs is down 1.76M from September 2003's 16,349,900. (-10.8% y-o-y???)
California's EDD applied a benchmark revision to the payroll survey based on 2003 numbers. They revised current and history numbers "based on tax records." However, this does not explain why, or how, such a revision could reasonably affect UN-adjusted payroll numbers. (Unless even the unadjusted numbers are adjusted?) With December, 2003 reporting 16.36M payroll jobs dropping to only 14.29M by January, 2004, major changes were clearly made.
So far, 2000 -2004 payroll employment figures are available using this new "2003 benchmark:"
Original after 2003
Unadjusted Benchmark
2000 16.06M 14.49M
2001 16.25M 14.60M
2002 16.21M 14.58M
2003 16.28M --?
2004 14.46M* 14.49M*
*-YTD average
Doesn't this seem a tiny bit convenient? Suddenly, through the magic of a "benchmark revision," 1.6M lost jobs disappear. The EDD would now have us believe California has been steady and even gaining jobs since 2000. Why not just tell the EDD the answer you want and let them work backwards to produce supporting data?
Problem #3: Unemployment insurance claims are not adjusted but unemployment rates are.
Headlines proclaim a 5.7% unemployment or 5.9% seasonally adjusted. How or why a ratio based on a mix of adjusted and unadjusted data is statistically valid is not clear. (Except when you assume adjusted numbers leave bureaucrats more latitude to make the numbers look better.)
Problem #4: California admits to 1,000,400 total unemployed.
Unfortunately for the EDD, "adjusting" employment figures down sharply while sticking with unadjusted unemployment figures will lead a spike up in the unemployment rate. It's simple arithmetic.
Using 1,000,400 unemployed and the new, improved payroll count at 14.49M, our headline unemployment rate should be 6.5%, not 5.9% or 5.7%. So how did they get to 5.9%?
It turns out the EDD conveniently changed horses in mid-stream. The headline jobs gain was based on the payroll report while the same headline's unemployment rate was based on the household survey. (The household survey showed a loss of jobs.) Therefore EDD used "total civilian labor force" at 17.69M to get down to a 5.9% unemployment figure. As a wonderful side effect, according to the EDD's "Household Survey," the BLS version of which is sometimes called "notoriously unreliable," California employment is now at an all-time high!
Maybe the next research project will be to figure out how long the EDD has been twisting their numbers like this...
Problem #5: The unemployment rate is not the whole story.
Even if it was reasonable to switch back and forth between the payroll and household surveys in order to put results in the best possible light, a 5.9% unemployment rate does not tell the whole story.
The household survey shows total unemployed in California fell -171,000 from July to September. However, similar to Santa Clara County for the last three years, the civilian labor force declined -153,000 over the same period. Hence the steadiness in the unemployment rate.
How can the EDD simultaneously report a falling labor force and record employment? Don't think, just do.
Problem #6: There are 1.7M lost payroll jobs but 2M+ more people employed in the space of just two years.
Household's 16.65M workers minus payroll's 14.53M employees, leaves 2M+ workers "missing." Unadjusted 2002 payroll jobs at 16.2M minus the adjusted 14.53M jobs shows 1.7M jobs "missing." If EDD's benchmark was truly effective at distinguishing between people claiming to work vs. people paying taxes, this implies a large fraction of the "missing" 2M workers are working are not paying taxes.
If nearly 12% of the workforce dropped off the tax rolls, what does that say for tax receipts and service levels in California going forward? What does that say about the $15B bond debt on top of nearly $8B revenue deficit this year in California? Never mind, it won't fit as a sound byte on the evening news.
Conclusion: We have all learned that we need to study and question BLS reports. It turns out the EDD's reports, even with extensive study, are even less decipherable...
* * * * *
All data are taken from what are presumed to be reliable sources. The author actively denies any responsibility for any errors.
* * * * *
*- CA Employment Development Department press releases www.edd.ca.gov/urate200410.pdf
Also visit: www.labormarketinfo.edd.ca.gov/
|
