Over the past couple of days I’ve been writing about the fabrication of the unemployment rate derived from the Current Population Survey — also known as the Household Survey — that the Census Bureau conducts on behalf of the Labor Department.
Back in 2010, I wrote a number columns questioning the other survey put out each month by the Labor Department. It’s called the Establishment Survey because it queries businesses about job creation. This is the survey which decided, for instance, that there were 204,000 jobs created last month.
The Labor Department conducts the Establishment Survey itself, which, in my opinion, makes it much more reputable. But the Census Bureau had a trick up its sleeve in 2010 that I believe was affecting the Labor Department’s Establishment Survey. (My gripes with this survey have always been about statistical tricks, like the so-called birth/death model, and not outright deception.)
The decennial census was being conducted from 2008 to 2010, and a lot of part-time workers were needed to knock on doors to get information. What Census appeared to be doing was hiring people for very short periods of time, then letting them go only to rehire them.
This churning seemed odd and not very efficient because these census takers were required to go through a relatively long training period whenever they came back to work.
But each “hiring” showed up as a created job. What’s more, that census taker who was let go, even if she was soon to be rehired, didn’t qualify for unemployment insurance and never showed up as someone who’d lost her job.
The Census Bureau, obviously, denied everything at the time and that story never got much traction because it was a little too complicated to fit into the evening news.
Congress is now looking into the story I broke earlier this week about Census’ manipulation of the unemployment rate survey. One former census worker says he was told by a superior to fake data that went into the survey. The superior denied it.
The House Oversight Committee is now on the case, looking to talk with my anonymous source who claims the ex-census worker, Julius Buckmon, was just the tip of the iceberg and that others were also fabricating data. Moreover, the committee is curious about the hiring and firing of workers during the decennial census.
Here’s something from one of the many field managers for the 2010 census: “As a former manager in the decennial census, I witnessed operations that strongly suggested the Census Bureau manipulated nationwide hiring of tens of thousands of temporary workers to manipulate employment data in 2008 and 2009,” Ron Brochu wrote me in an email.
“Operations that were to include 200 area workers for six weeks suddenly became ones in which 1,200 workers were to complete the work in 10 days. This was just for northern Minnesota,” Brochu wrote. “Multiply that by the entire nation … and it can tip the scales.
“The entire organization was a rat’s nest run by political order-takers who often bent rules,” he told me when I spoke to him Wednesday.
Complaints ranged from veterans not getting the job preference they were promised to training material being substandard and conflicting,
And because workers were subject to very serious fines if they spoke out in public, few did.
-
A lot of news organizations have been trying to follow up on my story about the former census field representative who said he was told by a superior to fake data that went into the calculation of the unemployment rate.
And many media organizations are falling back on the old cop-out — that the former rep was a rogue creating an isolated instance of manipulation. That one person in one office couldn’t possibly move the US unemployment rate.
We’ll see when the full investigation is done, won’t we?
But here’s something to chew on. When asked about the rep’s 2010 charges that a higher-up told him to fake the results, the supervisory survey statistician wrote in a Feb. 2, 2011, statement: “The claims … are completely false. I made calls asking everyone to do whatever they can to get these interviews even if they got partial household information.”
“In the conversation, I never mentioned falsification or re-interview,” the supervisor wrote.
This higher-up for the Philadelphia region was under pressure to meet a strict quota of successful interviews. Was “partial household information” enough to satisfy the Labor Department’s quota standards? And was someone filling out the missing part of those interviews?
This boss still hasn’t returned my call, and the Census Bureau, which is investigating, hasn’t made him or anyone else available to me.
Maybe there really was one rogue in Philadelphia. But was it the field rep, or his supervisor, who controlled a lot more surveys than just the one guy who cheated?
No comments:
Post a Comment