Some people, like Tim Nerenz, noticed a rather disturbing disparity between the two main measures of employment earlier this month. One measure, the Local Area Unemployment Statistics, based on the Current Population Survey, said that 21,570 (rounded up to 21,600) more Wisconsinites were working in March 2012 than in March 2011. The other measure, the Current Establishment Survey, said that there were 30,000 fewer jobs in March 2012 than in March 2011.
Department of Revenue chief economist John Koskinen addressed that disparity last Thursday at a meeting of the Association of Government Accountants….
To wit, Koskinen noted that the CES slide in employment was not supported by the all-establishment Quarterly Census of Employment and Wages (for the first quarter of disparity, the third quarter of 2011), tax revenues received by Wisconsin, per-capita income growth in 2011, or initial unemployment claims. For those of you interested in the PowerPoint portion of the presentation, Christian Schneider posted the slides that are included, in somewhat-pixelated form, on the video.
Before I continue, however, I do have to quote for the benefit of the lefties who might think Koskinen is a Walker stooge his short biography included in the DOR press release:
Prior to joining the Wisconsin Department of Revenue agency in 2007 as Chief Economist, John Koskinen served as a Staff Economist for the Wisconsin Department of Administration from 1979 to 2007. He started his professional career at the Wisconsin Legislative Fiscal Bureau. Koskinen has his B.A. and M.A. in Economics from Marquette University, as well as additional graduate studies in Economics at Northwestern University.
That’s right – Koskinen became DOR’s chief economist in the middle of Democrat Jim Doyle’s administration.
QCEW begins to break the tie
A quick explanation of what is covered by the three measures of employment is in order. The CPS/LAUS survey, covering 60,000 people on a national level and roughly 4% of Wisconsinites of working age, is the smallest of the three, though it covers every conceivable form of legal employment. The CES, covering 440,000 worksites on a national level and approximately 10% of Wisconsinites of working age, misses those who are self-employed and thus not covered by a form of unemployment insurance. The QCEW, covering every one of the approximately 9.7 million employers who pays unemployment taxes (thus missing the self-employed, railroad employees and religious institution employees), is a trailing indicator as it is released 6 months after the quarter that it covers.
For the first 6 months of 2011, the year-over-year changes in all three measures were on essentially the same slope. Starting in July, the year-over-year change in the CES started to separate from the year-over-year changes in the CPS/LAUS and QCEW. As Koskinen somehow used seasonally-adjusted data for the CPS/LAUS data while using unadjusted CES data and actual QCEW data, I redrew the chart to use the same measure for all three sets of data:
The CES really diverged from both the climbing CPS/LAUS and QCEW in August. Against the QCEW, the disparity grew from an average of the QCEW year-over-year change being 5,000 higher than that of the CES for the first half of the year (and 5,700 higher in June) to the QCEW year-over-year change being 32,500 higher in September, the last month QCEW data is available. Against the CPS/LAUS, the disparity went from an average of the CES year-over-year change being 16,600 higher than that of the CPS/LAUS (and 21,200 in June) to the CPS/LAUS year-over-year change being larger than the CES year-over-year change starting in September, growing to a 49,400 disparity in December, and reaching a 51,570 disparity in March.
Koskinen blamed the fact that the second quarter was used as the yearly “benchmark” of the CES rather than the third quarter. I cannot properly evaluate that claim, but the DOR produced a chart supporting this allegation:
Wages and tax collections support the CPS/LAUS numbers
The Bureau of Economic Analysis said that per-capita personal income in Wisconsin grew by 4.8% in between 2010 and 2011. That is not only significantly higher than the national average of 4.3% growth, but was the 11th-highest in the country.
In part because of that, and in part because the Republicans repealed the “millionaires’ tax” and combined reporting instituted by the Democrats when they had total control of state government in 2009, general-purpose revenue increased by an adjusted 4.3% for the first 10 months of FY2012 from FY2011 (that adjustment is downward from 6.0% due to more pay periods this time around). That includes an adjusted 4.5% increase (7.8% unadjusted) in individual income taxes, a 4.8% increase in sales taxes, and 5.4% in corporate taxes. Of note, FY2012 started in July 2011, when the CES measure of employment began to wildly diverge from the other two measures.
Initial unemployment claims for 2011 well below that of 2010, with the last 7 months at pre-recession levels
Perhaps the data that is most damning of the CES “job loss” is initial unemployment claims. The DOR produced a chart showing that those claims are the lowest in 5 years. Once again, I created my own chart, partly to remove the “clutter” of 2009 and 2010 from the DOR chart, partly to align the weeks to the week being reported instead of the week the report was issued, and partly to further demonstrate the point by choosing 2006 instead of 2007 (after all, the Great Recession supposedly started in December 2007).
Throughout 2011, initial unemployment claims were below 2010 levels. Indeed, by the 40th week, it was virtually indistinguishable from 2006 levels, and that trend continued through this year.
That is a measure more of a 1-month change than a 12-month change. So, how do the years compare? Allow me to give you one more chart, this time directly from the BLS:
Something just doesn’t add up, and it’s rather clear it’s the CES numbers everybody has been taking as the last word on jobs.