Economic Indicators | Wages, Incomes, and Wealth

Jobs Picture, November 2, 2007

November 2, 2007

Tale of Two Surveys: Strong Payroll Growth, Yet Falling Employment Rates

by Jared Bernstein with research assistance from James Lin

The nation’s payrolls grew much faster than expected last month, adding 166,000 jobs, the best month for job growth since May, according to today’s report from the Bureau of Labor Statistics (BLS). Based on problems in housing and credit markets, along with weakening consumer confidence, economists had been expecting around 90,000 net additions to payrolls.

The unemployment rate held steady at 4.7%, but job growth in the survey from which the jobless figure is derived—the BLS household survey—has been falling in three of the last four months, down 250,000 in October. Due in part to its much smaller sample size, the household survey is a lot more volatile on a monthly basis and widely considered a less reliable short-term indicator of job growth. However, as noted below, rates from the survey (as opposed to levels), such as the share of the population at work are more reliable, and they may be picking up signals the payroll survey is missing right now.

Averaging the data over the past three months helps reveal the underlying trend in job growth. Doing so for the payroll survey shows that job growth has slowed from last year’s pace. The average payroll gain over the past three months was 118,000 per month, compared to 164,000 a year ago. Data from coming months will reveal whether October’s jump in payrolls is the beginning of a new, stronger trend.

The strength in the payroll survey came largely from a few sectors. Professional services—jobs in offices and businesses—grew by 65,000, driven partly by an increase in temporary help, up 20,000, fully reversing last month’s loss of about this magnitude (and temp work had been on a negative trend all year). Some analysts argue that this reversal symbolizes employers’ positive expectations regarding labor demand, but, especially in periods of uncertainty, employers may turn to temp hires to ‘test the waters’ before committing to more expensive permanent hires.

Health care added 34,000 jobs last month and is up 400,000 jobs over the past year alone.  In fact, the pace of growth in this juggernaut sector has accelerated this year, growing 3.1% over the past year, compared to 2.5% one year ago (overall payrolls are up 1.2% over the year). Two other big gainers in October were restaurants and bars, up 37,000, and local government, up 43,000.

Some other important sectors continued to lose jobs. Despite large gains in the exporting of manufactured goods, factory employment fell 21,000 in October and has lost an average of 17,000 jobs each month for the past year. Homebuilding shed 11,000 jobs and is down 222,000 from its peak in March of 2006. Though consumers spent freely in the third quarter according to this week’s GDP report, the nation’s retailers continue to shed jobs, with employment off 22,000 in the sector last month, and down 38,000 since July.

As noted, the household survey is telling a considerably less positive story about the job market than the payroll survey and there is one negative trend that warrants close attention: the employment rate—the share of the working-age population with jobs. This indicator of employer demand has fallen consistently this year, down from 63.4% last December to 62.7% in October, and as such, it is clearly flashing a different signal than the payroll data. The largest declines in employment rates have come from African American men and young persons.

It is difficult to reconcile what these two surveys seem to be telling us. The payroll survey neatly reflects the strong growth in the macro-economy, with two consecutive quarters of real GDP growth close to 4%. Growth of that magnitude should be expected to spin off strong demand for jobs.

However, the household survey, especially the negative trend in employment rates (see Figure) tells a much less sanguine story. Though this trend matches neither the GDP nor payroll reports, it does accurately reflect consumer’s sentiments about the economy and the job market. In October, the Conference Board’s consumer confidence measure fell to its lowest level in two years, and that survey’s measure of job availability is also slumping near two-year lows. 

Employment rates, January 2006-October 2007

The sources of this divergence may be statistical—some analysts claim the payroll survey has yet to pick up losses that are showing up in the household survey, by missing, for example, undocumented workers losing jobs in residential construction. Additionally, some argue that the payroll survey’s birth/death model (adjustments from establishments that start-up or close down) is adding too many jobs right now (the model added 103,000 last month). While these explanations may prove to be right, there is no solid evidence to support them at this point.

Thus the question remains: will October’s strong showing mark the beginning of a new, positive trend or will future employment reports more closely reflect consumer’s negative views about the economy and the job market?

To view archived editions of JOBS PICTURE, click here.

The Economic Policy Institute JOBS PICTURE is published each month upon release of the Bureau of Labor Statistics’ employment report.

EPI offers same-day analysis of income, price, employment, and other economic data released by U.S. government agencies. For more information, contact EPI at 202-775-8810.

See related work on Income and wages | Wages, Incomes, and Wealth

See more work by Jared Bernstein