There is solid evidence that the Philadelphia office of the Census Bureau didn’t cheat on unemployment data only in 2010. The scam continued until I exposed it this past November.
On Nov. 19, I wrote a column that broke the news about Julius Buckmon, a Census enumerator — or data collector — who was conducting surveys in the Washington, DC, area for the Labor Department’s Current Population Survey (CPS).
The CPS — also known as the Household Survey — is what determines the nation’s monthly unemployment rate, a number recently used by the Federal Reserve to make monetary policy and by many others in government and private industry.
Buckmon’s cheating has been documented and is now under intense investigation by the House Oversight Committee, the Joint Economic Committee of Congress, the Commerce Department’s Office of Inspector General and — I believe — the US attorney’s office in Philadelphia.
And, of course, I’m still conducting my own investigation.
Buckmon was never fired, but he was eased out of the Census Bureau in 2010. Nobody else in the Philadelphia office of Census was either fired or prosecuted, even though Buckmon alleged in a discrimination case he brought against the bureau in 2010 that he was ordered to falsify surveys by higher-ups.
Because the CPS is a scientific survey, each response counts for about 5,000 households. Buckmon, I’m told, created more than 100 false surveys a month, which would have produced inaccurate data on the equivalent of 500,000 American households.
The big question today is whether Buckmon — on orders or not — was the only one cheating. If there were others, in Philly or in any of the other five Census regions, then the jobless rate data in its entirety could be tainted and useless.
I’ve written before that a source with knowledge of the situation has alleged that the practice of falsifying CPS surveys continued after Buckmon left and that it was particularly intense right before the 2012 presidential election. There was a big drop in the unemployment rate two months before President Obama’s re-election that was questioned by many people.
Here’s some of what I think proves that the practice of falsifying data continued to occur until recently.
According to its contract with the Labor Department, the Census Bureau needs to achieve a 90 percent success rate in the CPS interviews. Surveyors don’t cold-call households. People agree in advance to participate in the CPS survey, and from that group the Census Bureau needs to hit the 90 percent rate.
The survey is long and includes such questions as: “Does anyone in this household have a business or a farm? “Did you do ANY work for pay, or profit (the week before last)?” And “do you currently want a job, either full or part time?”
These are nuanced questions, since the Labor Department records different types of joblessness, including those who aren’t considered unemployed because they have given up looking for work.
Answers to this questionnaire, which goes on for three dozen pages, are recorded by a Census agent like Buckman on an HP computer.
There’s an incentive for supervisors to hit the 90 percent: Their bonuses depend on it. But people like Buckmon get paid by the hour.
The average number of monthly interviews, I’m told, is around 35.
According to a source, the Philadelphia Census office regularly achieved a better than 90 percent success rate on its surveys before my November column ran. Since then, it has always been under 90 percent.
Philly achieved just a 88.77 percent success rate in the February employment survey that was released last Friday, according to information I got through the grapevine.
The Census Bureau didn’t get back to me with official data before my deadline, so I have to rely on information leaked to me.
Only the New York office of Census, with just a 84.68 percent interview success rate, got fewer people to answer the CPS survey. (The Brooklyn office of Census was also investigated for data falsification in 2010, but that involved that year’s population count and not employment figures. In the 2010 instance, people were fired.)
By contrast, Atlanta had a success rate in last month’s CPS survey of nearly 93 percent.
And Philly, I’m told, was only able to achieve its 88.77 percent level by leaving its survey open a day or two longer than normal. If it had stuck to the stricter deadline, the rate would have been even lower.
What can be deduced from this? I suppose it could simply mean that Philly is getting less efficient in its surveying. But I’d take a more cynical view: Philly simply can’t get to the 90 percent rate because people there are too afraid to turn in fraudulent surveys with everyone watching.
And remember, Buckmon left the bureau in 2010. So he couldn’t have been the one getting Philly to 90 percent last fall.
Could all of this affect the monthly unemployment rate being released by the Labor Department? Sure. By how much? Someone with access to a lot more information would have to determine that.
But here is something else you need to know: The surveys handed in by enumerators like Buckmon through those HP computers can also be changed by higher-ups. Remember, Buckmon has alleged that he was ordered to do what he did by people above him.
The problem, I’m told, is that changes aren’t necessarily verified.
So if someone in one of the regional offices wanted to be dishonest, conceivably vacant houses and empty lots could suddenly turn into thriving households with plenty of employed people.
And in that case, you, I and the Federal Reserve wouldn’t really get a true picture of the unemployment rate.