No, Minnesota is not ‘dead last’ in the Midwest in job growth

[raw shortcodes=1]

Quick quiz for those who’ve been paying attention: Was Minnesota’s private sector job growth slower or faster than Wisconsin’s from March 2013 to March 2014?  

The correct answer is, of course: yes.

If you read the news, or listened to Jeff Johson in the first gubernatorial debate, you know that Minnesota’s private sector job growth was — at 0.8 percent — not only slower than Wisconsin’s from March 2013 to March 2014 but was also “dead last” in the Midwest and 41^st in the nation.

But those numbers were drawn the from the Quarterly Census of Employment and Wages, which is only one of three sources from which to draw job growth data. For example, a different metric — the Current Employment Survey — shows Minnesota’s private sector jobs growing by 1.7 percent and Wisconsin’s growing 1.3 percent. Now it’s Wisconsin that is near the bottom in job growth, not Minnesota.

Like so many things having to do with the economy, the picture looks different depending on the data you consult. So let’s take a look at what’s available on state job growth and then dig into the numbers. 

State and federal government statisticians compile a variety of databases on state-level job growth. The three most commonly cited are Local Area Unemployment Statistics (LAUS); State and Metro Area Employment, Hours, & Earnings; and the Quarterly Census of Employment and Wages (QCEW).

To compile the numbers, the Bureau of Labor Statistics conducts two surveys each month, one of households and one of business establishments. The household survey, formally known as the Current Population Survey (CPS), is the source for the LAUS data. The business establishment survey — the Current Employment Statistics (CES) survey — provides data for the State and Metro Area Employment, Hours, and Earnings reports.

But the QCEW is different: Rather than taking a sample from either households or businesses, the data are drawn from “a comprehensive tabulation of employment and wage information for workers covered by State unemployment insurance (UI) laws.”  In other words, the QCEW attempts to actually count all employees in establishments covered by state unemployment insurance law rather than make an estimate based on a sample of these establishments. The Bureau of Labor Statistics refers to this as a “near census” because they aren’t able to count every single worker in the United States.

Theoretically, figures from a census should be better than estimates computed from samples because the former is a true count of employment while the other is a statistical inference that carries all of the statistical baggage carried by estimates. Further, the QCEW forms the benchmarks for the two surveys; that is, the survey data are measured against the QCEW and adjusted over time so that they better match the census data.

So, what do we see when we look at these data?

Here are the QCEW data for March 2004 to March 2014:

Now, take a look at the CES data for the same period:

Both pictures are drawn on ratio scales so that the slope of the lines indicate the growth rate of employment — the steeper the line, the faster the growth rate.

The QCEW data show Minnesota slowing down from March 2013 to March 2014. But the CES data do not. The truth is probably somewhere in the middle — and that’s why we collect data on employment from multiple perspectives.

The long term trends, however, are quite clear. First, private sector job growth in Minnesota was faster after the Great Recession than Wisconsin’s. Second, Wisconsin is falling far short of the 250,000 net new private sector jobs that Gov. Scott Walker promised would be created in his first term.

The latter point is relevant given Johnson’s promise to “go all Scott Walker on Minnesota.” Gov. Walker’s economic policies, as I’ve talked about before, haven’t helped Wisconsin catch up to Minnesota in per capita income, unemployment rates, or other indicators of economic health.

Of course, candidates often cherry-pick statistics that fit their particular policy positions (or that make their opponents look bad), but this episode also brings up an important point. Whomever wins the governor’s race, let’s hope they take a more careful look at all the available data when it comes to making decisions — not just latch on to the set of numbers that favor their political point of view.

MP.highcharts.makeChart(‘.qcew-chart’, $.extend(true, {}, MP.highcharts.lineOptions, { xAxis: { categories: [‘2004′,’2005′,’2006′,’2007′,’2008′,’2009′,’2010′,’2011′,’2012′,’2013′,’2014’] }, yAxis: { title: { enabled: false }, min: 2000000 }, legend: { enabled: true }, series: [{ name: ‘MN’, data: [2146734,2181780,2248274,2251767,2260613,2158576,2112390,2156377,2208379,2253729,2272747] }, { name: ‘WI’, data: [2258717,2298785,2323020,2338434,2342358,2224513,2172883,2214233,2253989,2280050,2308762] }

] }));

MP.highcharts.makeChart(‘.ces-chart’, $.extend(true, {}, MP.highcharts.lineOptions, { xAxis: { categories: [‘2004′,’2005′,’2006′,’2007′,’2008′,’2009′,’2010′,’2011′,’2012′,’2013′,’2014’] }, yAxis: { title: { enabled: false }, min: 2000 }, legend: { enabled: true }, series: [{ name: ‘MN’, data: [2249.2,2281.0,2349.1,2359.8,2362.3,2261.6,2210.4,2257.7,2310.8,2356.3,2396.1] }, { name: ‘WI’, data: [2378.4,2415.2,2442.2,2462.1,2463.2,2351.4,2296.9,2331.3,2375.4,2402.4,2433.2] }

] })); }(jQuery));[/raw]