Measuring unemployment

Posted: Jan 06, 2004 12:00 AM

 On Friday, the Bureau of Labor Statistics will announce the employment report for December.  It likely will show a further decline in the unemployment rate--the most politically potent of all economic statistics--since initial claims for unemployment compensation have dropped sharply in the last several weeks.  The White House will cheer and claim that the news is confirmation of the wisdom of its economic policies.  To counter this good news, Democrats will have to work even harder to find a dark lining behind the silver cloud.

 For some time, Democrats have been arguing that this has been a jobless recovery despite a fall in the unemployment rate from 6.4 percent in June to 5.9 percent in November.  They have made this argument by using different data than that used to calculate the unemployment rate.  Lately, they have also taken to recalculating the unemployment rate in nonstandard ways in order to make it seem higher than it really is.

 To understand what is going on, one needs to know that the Labor Department collects employment data in two different surveys.  The first, called the household survey, is based on telephone interviews with about 60,000 households per month.  This survey is used to calculate the official unemployment rate, which consists of people not working but looking for work as a share of the labor force (those working plus those looking for work).  Those not looking for work, such as retirees and stay-at-home mothers, therefore, are not counted as unemployed.

 The second survey is called the payroll survey and is based on the actual employment records of domestic businesses.  Economists generally consider this survey to be a more accurate measure of month-to-month changes in national employment.  However, there is evidence that during cyclical upturns, such as we are in now, the payroll survey misses many new business startups, causing it to understate employment growth.  Eventually, the Labor Department finds these businesses and adjusts its data upward, which it probably will do for recent payroll data when revised figures are released on Feb. 6.

 For some time, there has been a growing divergence between the two labor surveys.  The household survey has shown strong employment growth--an increase of more than 2 million jobs between Nov. 2002 and Nov. 2003 (including a statistical adjustment last January).  In the latest month it showed 138,603,000 jobs in the U.S.  But the payroll survey has shown anemic job growth over the same period.  Indeed, between Nov. 2002 and Nov. 2003 it shows a net decline of 235,000 jobs.  According to the payroll survey, there are only 130,174,000 jobs--far fewer than shown in the household survey.

 There are a number of technical reasons why the two surveys will always report different figures.  Among these are people with more than one job, who may be counted twice in the payroll survey, and the self-employed, who are counted only by the household survey.  Although attempts are made to reconcile differences between the two employment surveys, the gap cannot entirely be explained according to a recent BLS study.

 Liberals contend that the payroll survey is the only one worth paying attention to, as a recent Economic Policy Institute study argued, because it confirms their story about a jobless recovery.  But conservatives will continue to point out that the more favorable data from the household survey is and always has been "official" for calculating the unemployment rate.

 To counter this last point, liberals have lately taken to adjusting the household survey numbers so as to get the "true" unemployment rate.  One technique is to add to the number of unemployed those working part time for economic reasons, temporary workers and others marginally attached to the labor force.  In November, these adjustments would have given us an unemployment rate of 9.7 percent instead of the official figure of 5.9 percent.

Princeton economist Paul Krugman claimed in a Dec. 30 New York Times column that the adjusted data "indicate the worst job market in 20 years."  However, the BLS shows unemployment rates as high as 12.8 percent in the mid-1990s using the same methodology.  Therefore, Krugman's claim is simply wrong factually.

Most economists believe that employment is poised to rise significantly in coming months.  It always lags behind increases in output, but eventually catches up.  That now seems to be happening, as indicated by the employment index from the Institute for Supply Management, which rose sharply in November and December.

Over the longer term, some economists are starting to worry about labor shortages for demographic reasons, as the large baby boom generation retires and is replaced by a "baby bust" generation that is much smaller.  Of course, job growth can't happen too fast for those now unemployed.  But it really is happening and they should take heart.