WaPo Report on DC-MD-VA January Job Market Betrays Seasonal, Other Ignorance

March 13th, 2010 11:25 AM
JobSearchNewspaper

Yesterday, Jim Taranto at the Wall Street Journal's indispensable Best of the Web took note of a report by the Washington Post's V. Dion Hayes about the state of the employment market in DC, Maryland, and Virginia, and summarized its findings thusly:

So what looks to the Post like good news that looks like bad news is actually bad news that looks like good news.

Even that assessment turns out to have been overly charitable. Hayes, like most of the press, betrayed that he doesn't understand the crucial difference between raw and seasonally adjusted data by mixing the concepts (perhaps without even realizing it), failed to look at data from previous years, and ended up producing an incoherent report with no supportable conclusions.

The following table identifies all relevant changes in the three employment markets between December and January (sources -- Bureau of Labor Statistics state and selected areas tables 3, 4, 5, and 6 for January):

DCMDVAjobRelatedChangesJan2010

With the above data in mind, a detailed breakdown of Hayes's broken-down report follows:

Rise in Washington area unemployment seen as good sign for economy's recovery
The District's jobless rate increased to 12 percent in January from 11.9 percent the previous month, according to data released Wednesday by the federal Bureau of Labor Statistics. Maryland's rate climbed to 7.5 percent from 7.4 percent, while in Virginia, the jobless rate rose to 6.9 percent from 6.8 percent. That same month, the U.S. unemployment rate dropped to 9.7 percent from 10 percent in December. (1)
... "You do have people coming back thinking things are loosening up and they might get something," said Ann D. Lang, senior economist at the Virginia Employment Commission. She said the state's labor force grew by 30,000 in January, even though the number of unemployed people rose and the number of employed declined. (2)
... The District experienced the nation's largest monthly increase -- 1 percent -- in the number of jobs, according to the federal data. That represents the addition of 6,700 jobs, mainly in the federal government, according to the city's Department of Employment Services. (3)
Officials at the agency say the city's labor force grew by 3,300 from January 2009 to January 2010, and they are seeing more job postings and job fairs held by large employers, including Safeway and Pepco. (4)
... Maryland's labor force grew by 1,400 in January, the first increase since May 2008, state officials said. The number of unemployed people rose by 5,100, but the number of people who lost their jobs totaled 3,700. State officials speculate that the balance of 1,400 were workers reentering the labor force. (5)

Here are the footnote explanations:

(1) -- Though Hayes never identified them as such, all percentages presented in this paragraph are seasonally adjusted. The presentation clearly indicates that what Hayes reported is what he believes happened in reality. It's not; MD's and VA's unemployment rates went up significantly. To be clear, such January jumps are not unusual. January is typically a month of steep job losses and unemployment rate increases because those who were hired for Christmas and other seasonal employment are usually released.

(2) -- As you can see from the table, Virginia's Lang is clearly referring to the NOT seasonally adjusted 29,700 above. Hayes seems to have no clue.

(3) -- That 6,700-job increase is after seasonal adjustment. The government's best estimate of the reality on the ground is that 5,700 fewer people were working. I defy Mr. Hayes or anyone in DC's Department of Employment Services to find those 6,700 net new employees -- in the federal government or anywhere else.

(4) -- Hayes inexplicably uses a seasonally adjusted 12-month change after reporting on single-month changes all along. That the change is the same as the NSA number above is pure coincidence.

(5) -- Again, Hayes gives readers the impression that this is what really happened, when it isn't. The reported "speculation" by the state in relation to seasonally adjusted figures makes no sense.

The fact is that Hayes did not look at how January's real, not seasonally adjusted changes compare to previous years, especially those earlier than 2009, when the entire economy was in a virtual free-fall. That's a problem, even moreso than usual, as I noted last week (at NewsBusters; at BizzyBlog), when I wrote that:

2009 was so bad, that it’s distorting the SA calculation — to the point where it’s really not as reliable an indicator of what is going on as it normally would be in a more predictable situation."

That concern could be even more relevant when looking at state and local data.

Because of that, at least the following contentions by Hayes and those he consulted for quotes have no demonstrated support:

  • Via Hayes -- "(the report) suggests that discouraged job-seekers are feeling more optimistic about their prospects and have resumed looking for work."
  • Via Anirban Basu, chairman and chief executive of Sage Policy Group -- "Maryland, Virginia and the District are where the labor market is advanced in terms of recovery ..."
  • Via Hayes -- "The increase suggests that long-term unemployed people in the D.C. area who had given up looking for work have restarted their job hunt, perhaps because they see evidence that the region's economy is improving and that employers are beginning to hire again." The absolute best that might be said is that fewer employers might be letting fewer people go than occurs in a typical January -- but we don't even know that.
  • Via Ann D. Lang, senior economist at the Virginia Employment Commission -- "You do have people coming back thinking things are loosening up and they might get something."
  • Via Tim Bibo Jr., research analyst at the Maryland Governor's Workforce Investment Board -- "If we have any optimism we've got to keep it tempered, because it is a modest increase -- but it's a step in the right direction."

I would modify Jim Taranto's conclusion noted at the beginning of this post as follows:

So what looks to the Post like good news that looks like bad news may actually bad news that looks like good news -- but based on errors in V. Dion Hayes's presentation and work that the reporter left undone, how can we possibly know?
Cross-posted at BizzyBlog.com.