Clear84° WeatherClear84° Weather
Working | What the numbers say

Surveys provide data for reports on jobs, joblessness

Each month, when we report on the Labor Department's employment and unemployment numbers, I hear from readers who think the statistics are wrong or manipulated. So, here's some background information and myth-busting about the household survey and the establishment survey that, respectively, produce the unemployment and employment numbers. Diane Stafford, McClatchy-Tribune Newspapers

The unemployment rate undercounts all of us who have run out of unemployment benefits.

Not so. Whether you're drawing unemployment checks or not makes no difference. The monthly unemployment figure comes from a random, statistically viable survey in which about 60,000 households around the country are called by the U.S. Bureau of Labor Statistics.

Adults who answer the phone are asked if they were working or looking for work in the survey week. If the individuals say they were not working, more questions determine whether they were not actively searching for work because of retirement or other reasons.

(Note: A drop in the civilian labor force can cause the unemployment rate to rise; the exact same number of unemployed constitutes a larger percentage of a smaller work force.)

The official unemployment rate is politically manipulated.

Given that top Labor Department officials are political appointees, the charge gains traction in election years. But it would be virtually impossible for the bureau's monthly number crunchers to swing data according to political whims.

A more valid criticism might lie in what the department has designated as the "official" unemployment rate, which was 6.5 percent in October.

Some labor market analysts argue that the official rate, dubbed U-3, doesn't truly reflect job market pain.

U-3 is the "total unemployed, as a percentage of the civilian labor force." That's the percentage of would-be workers who told the household survey that they were jobless and actively searching for work.

But the bureau doesn't hide other measures of job loss. Each month's unemployment report includes "alternative measures of labor underutilization."

In October, the U-4 rate, including "discouraged" workers who had not actively searched for a job because of market conditions, was 6.8 percent.

The U-5 rate, adding in "marginally attached" workers who hadn't looked for work recently but were available for work, was 7.5 percent. And the highest rate, U-6, including workers who wanted full-time work but could only find part-time jobs, was 11.8 percent.

It's easy to see why a professional who has been job hunting fruitlessly in his field for six months and has taken a part-time retail job to help pay the bills would identify more with U-6 than with the "official" rate.

The reports ignore illegal immigration.

Not really. There is no effort to find out in these statistical gathering processes whether workers or the unemployed have legal work status. All are counted the same as citizens.

The bureau knows from other statistical tools that 15.7 percent of the U.S. labor force last year was composed of persons born outside the United States, but that's the extent of the bureau's statistical concern.

The reports magnify job cuts at big businesses. What about the small-business and self-employment gains that are the bedrock of the American work force?

They're included. The household survey reaches the self-employed, and the establishment payroll survey, which produces the widely broadcast number of net job gains or losses, reaches beyond large companies.

The payroll survey each month covers about 160,000 businesses, large and small, and government agencies. The sample represents about 400,000 separate work sites and about one-third of all nonfarm payroll workers.

You can't believe the numbers at all. Just look at the big revisions they sometimes make a month later.

Sometimes, you can make that case. This is, after all, statistical sampling that's only as good as the data fed into it each month. There are occasionally inexplicable swings in some sectors' monthly reports.

Take the October report's announcement that payroll employment in September had fallen by 284,000 jobs — 125,000 lost jobs more than originally estimated.

The revision was largely attributable to what can generously be described as anomalies in three job sectors: educational services, local government and state government.

For reasons that the statistics bureau can't fully explain — they're not attributable to any seasonal or economic force — wild variations were reported.

For example, educational services reported a jobs gain of 4,200 in September. In October, it was revised to show a 47,100-job loss for September. Statisticians were left scratching their heads.

Statistics can be confusing. Sometimes they're "garbage in, garbage out." But they're what we've got to keep tabs on the job market.

The best advice is to understand how the data are gathered, read them with at least a small grain of salt, and wait for time to ferret out "the truth."

Surveys provide data for reports on jobs, joblessness 12/13/08 [Last modified: Saturday, December 13, 2008 3:30am]

Copyright: For copyright information, please check with the distributor of this item, McClatchy-Tribune Newspapers.
    

Join the discussion: Click to view comments, add yours

Loading...