That “bias,” the distortion cooked into the process, has doubled since 1994, Krueger and his colleagues argue.
Now, into the weeds:
To calculate the unemployment rate, the U.S. Census Bureau and U.S. Bureau of Labor Statistics takes samples of nearly 60,000 households. They are surveyed monthly for four consecutive months, left alone for eight months and then surveyed again monthly for the next four months.
That means that in each month, there are eight “rotation groups,” each of them intended to be representative of the population. The government then weights the groups to come up with an official unemployment rate.
So that weighting has a large effect on the result.
Krueger and his collaborators found that during the first half of 2014, the unemployment rate among people in the first month of being interviewed was 7.5 percent. However, for those in the final month of being interviewed, it was only 6.1 percent.
But the Bureau of Labor Statistics weighted the first interview more heavily, so the official unemployment rate for this period was 6.5 percent.
And here’s the thing, Krueger said: Maybe they are right. It’s hard to say.
“It is unclear which rotation group provides the most accurate measure of the unemployment rate,” he said.
The Krueger paper proposes changes, including revisions in how the surveys are collected, the questions asked and the interviewing methods.