RALEIGH — Dean Baker is an economist who co-directs a left-wing group in Washington called the Center for Economic and Policy Research. Liberal groups in North Carolina and elsewhere often cite Baker as an authority on the effects of public policy on economic growth.
That practice may be unwise, at least if one were to judge by Baker’s fumbling attempt to rebut my recent Wall Street Journal piece on North Carolina’s experience with unemployment-insurance reform. He does not seem to know his way around the relevant kitchen, or how to follow basic recipes for producing sound policy analysis.
For starters, the policy in question was a set of changes the North Carolina legislature made to the amounts and duration of UI benefits. These changes, including an end to the state’s eligibility for federal extended benefits, took effect on July 1, 2013. Evaluating the impact of the changes would require that one compare trends before that date with trends after that date. In my writings for the Journal and CJ, that’s what I did.
Baker didn’t. He started his analysis with the June data, not the July data. That makes a difference because North Carolina experienced no net job growth in June 2013 even as the nation as a whole added 200,000 jobs. Whatever the cause of North Carolina’s poor showing that month, it obviously could not have been a UI policy implemented the following month.
His poor choices didn’t stop there. Rather than compare North Carolina’s job-creation trend to the national average, as I did, Baker limited his analysis to the nine jurisdictions (including D.C.) that make up the Census Bureau’s South Atlantic Region. I certainly don’t mind making regional comparisons, too, but why ignore the national comparison?
Next, Baker passed over the most important time period for evaluating North Carolina’s policy — the six months from July to December 2013 during which the state was the only one in the country where extended benefits disappeared. Instead, he ran his numbers from May 2013 to May 2014, which includes five months in which extended benefits were no longer in effect anywhere in the United States (plus a month, June 2013, in which extended benefits were available in North Carolina and everywhere else, as noted above).
Finally, Baker declared that because some South Carolinians cross the border into Charlotte to work, the proper frame of reference would be to exclude the entire Charlotte metro from the North Carolina total. Of course, South Carolinians also cross the border to work in the Wilmington region, but that region was not excluded. As for the sprawling Charlotte metro — which includes populous counties such as Iredell and Cabarrus that don’t abut South Carolina — the vast majority of its workers reside in North Carolina. To remove one of the state’s two fastest-growing metros from the analysis obviously biased the results.
So, after putting both thumbs and his entire left foot on the scales, Baker was able to conclude that part of North Carolina posted a somewhat slower rate of job creation than some of our neighbors did during a period that partially overlapped the actual period in question. To call such a conclusion meaningless would be generous.
Using the correct starting point (June 2013) and restoring the Charlotte metro, I re-ran the numbers through May 2014. North Carolina’s rate of job creation (1.9 percent) exceeded the average for the region Baker preferred, the South Atlantic States (1.7 percent). Using the proper ending point for analyzing the state’s solitary exit from extended benefits (December 2013), North Carolina’s rate of job creation (1.5 percent) was half again as fast as the regional average (1 percent).
Not content to ruin only one batch of statistical cookies, Baker then put another slab of goo in the oven. “As far as the great news on unemployment that Hood cites,” he wrote, “this is entirely a story of North Carolina workers giving up looking for work and leaving the labor market.” Did Baker cite the relevant data from the Bureau of Labor Statistics to prove his point? Certainly not, because the data show something entirely different. From June 2013 to May 2014, fully 98 percent of the decline in North Carolina unemployment was attributable to employment gains, not to a drop in the labor force. In fact, during this period, North Carolina's labor force dipped then rose, ending up with a net loss of fewer than 2,000 workers, or about four-hundreths of a percent. The number of employed North Carolinians rose by about 85,000.
Baker would have had a somewhat-better argument if he had had sense enough to focus only on the last six months of 2013, when the number of unemployed North Carolinians dropped by 65,000 while the number of employed North Carolinians rose by 26,500. Even so, those figures can’t be construed as showing that the decline in unemployment was “entirely” due to workers leaving the labor market. More importantly, the decline in North Carolina’s labor force that continued during those six months actually began in February 2013, long before extended benefits went away, and also occurred in other nearby states where extended benefits remained in place throughout the year. Since I discussed the issue at some length in the article to which Baker was responding — pointing out that the end of extended benefits in North Carolina couldn’t be the primary cause of the labor-force decline during the latter half of 2013 — he either didn’t read or didn’t understand it.
For good measure, Baker then tossed in another set of numbers — for employment as measured by the household survey — for North Carolina and the South Atlantic Region. He used them to allege that North Carolina’s employment gain lagged behind the regional average. Again, however, he started his series in the wrong month, May 2013. Since June 2013, total employment in North Carolina has grown 2 percent vs. the South Atlantic average of 1.6 percent.
Finally, for his pičce de insistence, Baker alleged that since the end of extended benefits in the nation as a whole, “what is most striking in the data for 2014 is the sharp drop in labor force participation.” Huh? In December 2013, the national labor-force participation rate was 62.8 percent. In June 2014, the national labor-force participation rate was — 62.8 percent. Even if you zoom out to quarterly averages, the labor-force participation rate was 62.9 percent in the 4th quarter of 2013, compared to 63 percent in the 1st quarter of 2014 (the first after the end of extended benefits) and 62.8 percent in the 2nd quarter of 2014. Do you see a “sharp decline” here?
What Baker did, again, was go back further in time to mask the real trend. He compared the average labor-force participation rate in the first six months of 2014 to the average rate for the first six months of 2013. The former was lower than the latter, yes, but all of that decline occurred during 2013, before the end of extended benefits.
In my Wall Street Journal piece, I cited additional BLS measures such as the employment-population ratio and broader unemployment measures that included discouraged workers who dropped out of the labor force. These measures also showed that North Carolina had one of the largest labor-market improvements in the country after exiting extended benefits in July 2013, and the nation as a whole has seen clear labor-market improvement since it exited extended benefits in January 2014.
Baker chose not to attempt to challenge this additional evidence for the proposition that ending UI extended benefits is associated with higher rates of both job acceptance and job creation. That was about the only wise choice he made in preparing what was, at best, a quarter-baked mess.
John Hood is president of the John Locke Foundation.