The Corner

Beware the Misuse of Labor Statistics

Last July, North Carolina became the first state to exit the federal government’s extended-benefits program. Six months later, at the start of 2014, the extended-benefits program ceased in the rest of the country. President Obama and some politicians and analysts want to renew the program for at least some of 2014. Others think extended benefits are no longer a wise use of taxpayer money.

Given the stakes, policy analysts across the spectrum have focused intently on the labor-market statistics in North Carolina, which show a dramatic drop in the unemployment rate (from 8.8 percent in June to 6.9 percent in December, more than twice the drop the nation as a whole experienced). I wish these analysts’ intense focus on the issue had been accompanied by an extensive knowledge of how the statistics were produced and what they meant.

For example, if they truly knew what they were talking about, these analysts would not be claiming that North Carolina’s exit from extended benefits caused the state’s civilian labor force to contract. They would know that the apparent decline in North Carolina’s labor force began in February 2013, months before the end of extended benefits.

They would also know that the labor-force count began to decline markedly in other southeastern states around the same time, even though these states retained extended benefits. Indeed, since they would know that Georgia’s decline was about the same as North Carolina’s, and Tennessee’s was actually larger, they would never have been foolish enough to blame the labor-force decline on North Carolina’s unique UI policy.

Here’s something else an informed analyst of North Carolina’s labor-market trends would know: the decline in the labor force from June 2013 to December 2013 was not large enough to be statistically significant. According to BLS, because of the small size of the household-survey sample in North Carolina, “the approximate thresholds for statistically significant changes . . . over a 6-month period (using the change from June-December 2013) are as follows: 68,000 for civilian labor force; 61,000 for employment; 32,000 for unemployment; and 0.7 for unemployment rate. So, our official estimates indicate that unemployment (both in terms of the level and rate) in North Carolina has gone down significantly since June, but neither the labor force nor employment level changes were significant.”

Now, there is a larger, broader set of data from the BLS household survey that bears a closer look here. The standard “unemployment” measure I discussed above, and most people pay most attention to, is called the U-3 rate. It includes working-age people who are not currently employed and say they are actively seeking work. (The actual source of the measure “civilian labor force” is a simple sum of employed persons plus U-3 unemployed persons.) So what happens to people who don’t fit either definition? Some of them are called “discouraged workers.” They have given up looking for work because they couldn’t find what they were looking for. Another group are other “marginally attached workers” who are currently not looking for work but don’t cite discouragement as the reason. They have returned to school, for example, or moved in to take care of a sick parent, or are in the process of relocating. Finally, there are workers who are currently in part-time jobs but would prefer to be in full-time jobs, often called the underemployed.

The BLS household survey includes questions to tease all these alternatives out. Add discouraged workers to U-3 and you get U-4. Add other marginally attached workers to U-4 and you get U-5. Add the underemployed, and you get U-6. But because of small sample sizes, BLS doesn’t publish these broader estimates every month. They using rolling annual averages, updated every quarter.

The 2013 averages just came out. If North Carolina’s exit from extended benefits had caused a large flight from the labor force, the effect should show up in broader measures. So what do they show? Not much. If you compare the U-5 rates for 2012 and 2013, the share of N.C. workers who were discouraged or otherwise marginally attached declined a bit. So did the share of N.C. workers who were underemployed. Overall, North Carolina’s U-6 rate averaged 14.7 percent during 2013, down from 16.3 percent in 2012.

The other BLS survey that generates labor-market statistics, the payroll survey, has a much larger sample size than the household survey. Thus its margin of sampling error is lower. According to the payroll survey, North Carolina employers added 51,400 jobs from June to December, a statistically significant increase. Moreover, that rate of increase (1.3 percent over six months) exceeded the national average for the same period (0.8 percent).

North Carolina’s job-creation trend also exceeds the national average over the longer run. From June 2011 to December 2013, the state added about 184,000 new jobs, a growth rate about 13 percent above average.

There was, in fact, one recent period during which North Carolina lagged the nation. That was the first six months of 2013, when the state added only 13,100 new jobs, a growth rate of just 0.3 percent. It was an outlier. North Carolina’s job creation surpassed the national average during 2012 and during the last six months of 2013, as I’ve already indicated. Unfortunately for the Left’s flimsy narrative, January–June 2013 was the “wrong” period for the state to underperform, since it preceded the end of extended benefits.

To sum up, then, here’s what the available evidence tell us about North Carolina’s July 2013 exit from extended benefits: 1) it was followed by a large, statistically significant drop in both the count and rate of unemployment; 2) it was followed by a statistically significant increase in filled jobs as measured by the payroll survey.

Other claims, about household employment and labor-force participation, lack statistical significance at this time. Does that mean liberal editorialists will stop making these claims?

You already know the answer to that question.

John Hood — John Hood is president of the John William Pope Foundation, a Raleigh-based grantmaker that supports public policy organizations, educational institutions, arts and cultural programs, and humanitarian relief in North Carolina ...

Most Popular


Fire Brenda Snipes

Brenda Snipes, the supervisor of elections in Florida’s Broward County, does not deserve to be within a thousand miles of any election office anywhere in these United States. She should be fired at the earliest possible opportunity. Snipes has held her position since 2003, in which year her predecessor, ... Read More
PC Culture

The Lonely Mob

Just before the election, an Andrew Gillum intern named Shelby Shoup was arrested and charged with battery after assaulting some college Republicans on the campus of Florida State University. It was rather less exciting than that sounds: She went on a rant about “Nazis” and “fascism” — Gillum’s ... Read More

The Georgia Smear

Back in 2016, when Trump refused to say he’d necessarily accept the result if he lost, we were told that this was a terrible violation of democratic norms. Now, refusing to accept that you lost an election is the highest form of patriotism. Not only are the media and the Left not pressuring Stacey Abrams to ... Read More

How Immigration Changes Britain

Almost nothing is discussed as badly in America or Europe as the subject of immigration. And one reason is that it remains almost impossible to have any sensible or rational public discussion of its consequences. Or rather it is eminently possible to have a discussion about the upsides (“diversity,” talent, ... Read More

Sorry, Brian Kemp Still Won

Here was the state of play as of yesterday per the Kemp campaign’s breakdown of publicly available information: As of Saturday, November 10, 2018 (12:00 p.m.) *Information below is public.  Total votes reported: 3,924,658 Kemp: 1,975,162 (50.33%) Abrams: 1,912,383 (48.73%) Metz: ... Read More