Economists debunk Jack Welch's jobs tweet

• Economists and government agency rebut dubious allegation
• Wharton economist on Welch claim: 'Bollocks'
• Manipulating data would need collaboration of thousands of people

What is this: A conspiracy theory so far-fetched it only commanded the seventh-most prominent headline on the news aggregating website Drudge Report?

Social media, especially the Twitterverse of those who tweet their thoughts and feelings instantaneously, exploded Friday with claims that September's unemployment report, which showed the unemployment rate falling to 7.8% from 8.1% a month earlier, must have been cooked.

Retired General Electric CEO Jack Welch claimed that President Obama's "Chicago guys will do anything. . . . Can't debate, so changed the numbers.''

Whether he was joking or not, there was no shortage of others making the same claims.

A conservative columnist for the Washington Examiner named Conn Carroll argued that Democrats lied en masse to the Census Bureau about finding jobs so the president would look good.

And a conservative blog called the Washington Free Beacon reported that two economists from the Bureau of Labor Statistics have donated money to Obama's campaigns.

"Bollocks,"' Wharton School economist Justin Wolfers said. "Once you understand how the numbers are collected and processed, you understand that it's literally impossible to fool with the numbers.''

Manipulating the unemployment rate would require the cooperation of thousands of people -- not to mention violating federal laws. But it wouldn't be the first time that the allegation would have made in a presidential election year.

The difference is that in the past, the charges might have been whispered behind closed doors or shown up hours, if not days later on talk radio or a newspaper article. Not so in the warp-speed era of the digital age.

The unemployment rate is based on a survey of 60,000 households, conducted at the middle of each month because holidays, which can result in short-term hiring or layoffs, are usually at the beginning or end of a month, said Karen Kosanovich, an economist at the U.S. Bureau of Labor Statistics who works on the survey.

About 1,500 Census workers gather the data for the survey, which dates back to 1940, Kosanovich said.

"It's designed to be representative,'' she said. "People have thought about this for decades.''

Census sends raw data over to BLS, where civil-servant economists compile it and add seasonal adjustments, she said. The work is typically not finished until the day before the report is released, she said.

The White House Council of Economic Advisers isn't told the number until the evening before release. Even the president is barred from speaking publicly about the data until an hour after it is released, according to a 1985 regulation issued by the Office of Management and Budget.

The BLS uses the household survey to estimate the unemployment rate because household heads know whether anyone in the home has a job, or is looking for one, Kosanovich said.

The agency estimates the number of jobs through a larger survey of 140,000 business establishments. The household survey has a larger margin of error than the establishment survey, meaning that its estimates of job growth are often subject to big revisions later on, she said.

The household survey, which said the economy added 873,000 jobs last month while the establishment survey said only 114,000 jobs were added, has a margin of error of plus or minus 280,000 jobs.

But human-resources people who fill out the establishment survey have no way to know how many people are looking for work, making it impossible to calculate the jobless rate through that poll, she said.

Economists never question the integrity of U.S. economic data, said Diane Swonk, chief economist at asset management firm Mesirow Financial.

``You can't fool Americans,'' Swonk said. ``At the end of the day we're a very transparent society.''