Candidates for Kansas governor last week released statements on recent job figures in Kansas. The releases from Sam Brownback and Paul Davis appear to contain conflicting views of Kansas employment.
Brownback released a statement containing this, in part: “In the past year, we have seen more than 20,000 new jobs in Kansas and a total of 45,600 new jobs created from January 2011 through October 2013.” (Click here for the full statement.)
Davis released a statement containing this, in part: “From January 2011 – Oct 2013: Period during which Brownback cites 46,500 new jobs … Employed: +3,634 (not 46,500, which is what was claimed by Brownback)” (Click here for the full statement.)
So which campaign is correct? The answer is not easy to provide. That’s because there are two series of employment data provided by the Bureau of Labor Statistics. The two series don’t measure exactly the same thing, and each campaign has chosen to use the series that benefits their campaign. Nearby is an example of just how different the two series can appear.
A document from BLS titled Employment from the BLS household and payroll surveys: summary of recent trends explains in brief: “The Bureau of Labor Statistics (BLS) has two monthly surveys that measure employment levels and trends: the Current Population Survey (CPS), also known as the household survey, and the Current Employment Statistics (CES) survey, also known as the payroll or establishment survey. … These estimates differ because the surveys have distinct definitions of employment and distinct survey and estimation methods.”
Another BLS document explains in detail the differences between the CPS and CES data. For example: CES: “Designed to measure employment, hours, and earnings with significant industrial and geographic detail” CPS: “Designed to measure employment and unemployment with significant demographic detail.”
Another difference: CES: “Self-employed persons are excluded.” CPS: “Self-employed persons are included.” (See Understanding the employment measures from the CPS and CES survey.)
I’ve prepared a table showing the claims made primarily by the Davis campaign (since it provided the most detail) and gathered data from both the CES and CPS series. I’ve also showed the seasonally adjusted data compared to the raw data when available. Sometimes the numbers match exactly with the claims made by the campaigns, and sometimes the numbers are a little different. Click here for the full table.
I’ve also created an interactive visualization of the CPS and CES data for Kansas. Click here to open it in a new window.
Each campaign uses the data that best makes its case. Generally speaking, the CES data shows larger employment gains.
So which data set is more reflective of how our economy is actually doing? I’m a bit naive, perhaps ignorant of how some of these figures are calculated – much like the rest of the general public. So – being the curious person that I am – I decided to go research some of this. I was quite shocked at how vastly different these two economic surveys really are. It is even more interesting to see how the political groups who get media attention twist and warp these figures into statistics and pseudo-evidences to support their agenda in ways the survey companies, the government and the actual design of those indicators were never made to do.
I found this interesting overview of these 2 specific surveys from the BLS – published by the BLS on their website to help clarify these surveys.
It’s a wonder why the media does not do more fact checking and hold these party spokespersons to their word by making them understand the true limits and usefulness of a quoted statistic before they use it to launch such debates.
Here is what I found out – the CES survey is a much more comprehensive study consisting of over 160,000 US employers of all sizes with over 400,000 worksites. It’s intended purpose is to more accurately depict a real number of “jobs” with the following disclaimer:
“The survey counts nonfarm payroll jobs only—with no age
restriction on the employed—and does not include the self-
This number is probably the most accurate depiction of institutional and traditional “employment” that I have seen as it is such a huge survey with verifiable data sets tracked to each employer and it covers a much wider variety of the economy – not limited to nearly as much subjectivity or “targeted” data collection from specific industries or regions. The nice thing about the CES data is that it breaks down the numbers by each region – I would trust these numbers – but there are some things things about this institutional report which are a little misleading as well… This website – http://www.dlt.ri.gov/lmi/faq/programs.htm
has a easy to read chart ( also in the previously mentioned pdf ) that sums up some of the differences and limitations. Of note are the fact that self-employed are not included in the CES data – perhaps not showing the true growth in the economy and not giving as high a count for “jobs created” figures to rely on by politicians – but it is possible for an individual to also be counted twice in this survey if she holds two or more institutional jobs that participate in the survey. This might inflate the number a little – but how much? The answer is, quite interestingly, very minimal at less than 6% of the employed labor force. That data comes from the CPS report which breaks down this figure into demographic groups – and is one of the purposes for the CPS survey as discussed.
The CPS report was not designed to give specific number of “jobs” but rather to provide a higher level overview of the demographics of employment – but it only surveys 60,000 people households.
To rely on the CPS report for support in jobs claims the way these politicians are doing is either disingenuous or just reflective on their ignorance just as much as it was of mine in not knowing before I did my own research. I sure wish more media would expose some of these numbers and surveys and point out when a politician is using the data for things it was not designed to be used for.