On Thursday the Wisconsin Department of Workforce Development (DWD) released revised jobs data alongside the January figures, data that will be released by the Federal Bureau of Labor Statistics on Monday.
DWD press releases of this monthly data typically pick and choose the brightest spots to highlight - if unemployment should tick down, or jobs in a particular sector of the economy should tick up, they draw attention to it. Fair enough: it's a normal part of trying to sell the state positively as an economically active area of the country even if the changes aren't usually statistically significant.
But not this time.
This time the Current Employment Statistics (CES) revisions to jobs data in prior months was significantly upwards and especially the figures since the second half of 2011. This was expected, so the DWD took the time to crow about they were right when they told us CES was doing it wrong in 2011 and QCEW (the Quarterly Census of Employment and Wages) was right.
One critical way they phrased things was:
The Federal government’s initial estimate for total non-farm jobs December 2012 was off by 67,100 when benchmarked
Then we come to the
Journal-Sentinel's reporting of this highlight:
Those revisions found that the preliminary estimates for total non-farm jobs last year undercounted non-farm job creation by 67,100, the agency said.
See the difference? This is what the press release and the data very notably did
not say, but wanted the media to take for granted:
The Federal government’s initial estimate for total non-farm jobs December 2011 was off by 0 when benchmarked
In fact it was revised up by nearly as much as the December 2012 figure.
2012 total job creation was originally put at 9,400 (going from 2,724,200 in December 2011 to 2,733,600 in December 2012) and revised to 14,100 (going from 2,785,600 in December 2011 to 2,799,700 in December 2012).
So the Journal-Sentinel fell for the lie of omission and reported a revision to the 2012 job creation that was much more than 10x what it actually was.
Details and an assessment of how Walker's 250,000 jobs promise is shaping up after the jump...
Here I have plotted not the revisions to the data, but the what they do to the 12-month trends:
The revision for 2011 (the 12-month trend to December 2011) was big, but the revision for 2012 was very, very small. The Milwaukee Journal-Sentinel don't realize this.
What happened was that CES mistakenly stated job losses in the second half of 2011 that have taken until the end of 2012 to work their way out of 12-month trends.
Here I've subtracted the 12-month CES trends - before and after revision - from the 12-month QCEW trend. This chart uses the DWD pre-released QCEW data for its last 3 months (it'll be released by the BLS at the end of this month):
CES trends now slightly over-estimate QCEW trends
The thing to bear in mind about this second chart is that only the changes in the second half of 2011 were the subject of this revision: the smoothness of the revised data since then is all CES, all by itself. Conclusion: CES has been near-perfect in 2012.
So where does Walker's 250,000 private sector jobs promise stand?
There are two ways of figuring this, both of which give approximately the same numbers. The first way is the Politifact way: they use seasonally-adjusted CES changes and substitute in whole-year QCEW changes once they become available. Using this:
Before revisions, the QCEW change for 2011 was +29,800. The CES change for 2012 was +9,700 for a total of +38,500. After revisions, the QCEW change for 2011 was (still) +29,800. The CES change for 2012 was revised to +14,200 for a total of +44,000.
The other way of doing it is to use as much QCEW as you can get your hands on, make up a complete year with seasonally-unadjusted CES (QCEW does not get seasonally-adjusted so we have to stay consistent), then add seasonally-adjusted CES for any more months you might have available.
Before revisions, the QCEW change for December 2010 - September 2012 was +71,973, the CES unadjusted change for September 2012 - December 2012 was -30,100 for a total of +41,873. After revisions, , the QCEW change for December 2010 - September 2012 was (still) +71,973, the CES unadjusted change for September 2012 - December 2012 was -27,700 for a total of +44,273.
+44,000 is an awful lot like the +44,273 that uses 9 more months of QCEW data: that's a testament to how closely CES has been anticipating QCEW in 2012.
But now we have the January 2013 CES changes too, which have Wisconsin private sector jobs up 12,400 (seasonally-adjusted). So 25 months into Walker's term, the best estimate we have for total private sector job gains is +56,673. This is how that compares:
+ 56,673 Best estimate of private sector job gains.
+101,600 If Wisconsin was on the national pace.
+112,500 If Wisconsin was beating the national pace by as much as we were in 2010.
+119,000 Image Journal-Sentinel is creating in uncritical readers' minds (actual + overstatement of revision).
+126,900 If Wisconsin was on the pace needed to meet Walker's promise.