Q&A: How government measures, reports jobs data

WASHINGTON — Few economic reports carry more weight these days that the monthly report about jobs from the Bureau of Labor Statistics. Here's a look at how these important gauges of employment in the economy actually are compiled.

Q: What does the monthly BLS report measure?

A: When the Employment Situation Summary comes out on the first Friday of every month, it actually contains data from two different surveys — the Current Employment Statistics survey, from which estimates are made about non-farm employment; and a household survey, known as the Current Population Survey, from which the unemployment rate is calculated.

Q: How do they differ?

A: The employment-situation survey samples about 160,000 businesses and government agencies, representing about 410,000 work sites. This survey gauges non-farm payroll employment during the week that includes the 12th day of any given month. When compared with the prior month, estimates of job losses or gains are derived.

The population survey involves about 2,200 Census Bureau specialists fanning out to interview people in the household survey, which involves roughly 60,000 participating households who together make a representative sample of the U.S. population. These participants are asked questions about work activity and job search during the reference week. From this, the jobless rate and a host of specific employment information are compiled.

Q: Which measure is more accurate?

A: Because the employment-situation survey is so large, it has a smaller margin of error when looking at employment changes over the month. Considering that public opinion polls sample 1,000 or 2,000 individuals to get a random sample, the BLS data on each survey is far more exhaustive than the political and media polling that forms public opinions.

Q: Is participation in the survey mandatory?

A: For the population survey, on which the unemployment rate is calculated, no. For the employment-situation survey, it's voluntary in all but three states — North Carolina, Washington and Oregon. In South Carolina, it's mandatory for companies with 20 or more employees. California recently stopped making participation mandatory.

Q: How is information collected from employers?

A: As recently as two decades ago, almost all of it came by mail. Now much of it comes through electronic-data interchange, with businesses automatically sending an electronic file from their payroll system. About 23 percent of information comes via Internet filing and another 17 percent arrives from a computer-assisted telephone interview involving a real person calling the company.

Q: Isn't the unemployment rate derived from unemployment insurance?

A: No. "There is a persistent misconception about that," said Jim Borbely, an economist in BLS's Division of Labor Force Statistics.

Unemployment benefits don't factor into the monthly calculation of the unemployment rate. A separate agency in the Department of Labor reports unemployment insurance claims data, but the claims data don't take into account all of the people who are unemployed. These claims are paid by states, and since benefits eventually run out, those who are still unemployed after benefits run out wouldn't be counted if unemployment checks were the only gauge of joblessness.

Q: Once the data is collected, what happens to it?

A: The BLS offices, located near the U.S. Capitol, go into a lockdown mode during the final week of the month ahead of the release of jobs report the first Friday of every month. Access is restricted as experts crunch the numbers that have come in from establishments and households across the nation.

Q: Who gets to see this data first, and when?

A: There are numerous ways someone could benefit from early release of this information, so there are tough rules on who gets to know what and when.

The Wednesday before each month's release, the Federal Reserve gets information about industrial employment and hours worked in the mining, manufacturing and utilities. The White House Council of Economic Advisers gets the information Thursday afternoon. The secretaries of labor and commerce get the information half an hour ahead of the release Friday. Everyone who gets an early peek at the numbers is prohibited from discussing it until one hour after the release.

Q: Do the surveys get updated?

A: From time to time, what's included in the monthly summary gets updated. One example begins this month, when the population survey will begin collecting more detail about the duration of unemployment, or how long people remain jobless. To date, the duration has been capped at 99 weeks or more. Now it will be capped at five years or more.

Q: Why does that matter?

A: This additional detail should provide greater information about those unemployed for extremely long periods of time. It will provide a better estimate of the average duration of unemployment.

"Our measure of average duration may increase as a result of this change," said Karen Kosanovich, a labor economist in the Division of Labor Force Statistics.