Economic Indicators | Inequality and Poverty

Jobs Picture—October 8, 2004

October 8, 2004

September’s job growth weaker than expected

The nation’s payrolls expanded by a less-than-expected 96,000 in September and unemployment was unchanged, according to today’s report from the Bureau of Labor Statistics.  Over the past four months, payrolls have grown by an average of 101,000 per month, well below the growth rate needed to significantly lower unemployment and absorb the considerable degree of slack remaining in the labor market (see for a discussion of benchmarks for employment growth).  Importantly, this recent trend of diminished growth is well below the average payroll gains of 225,000 per month which prevailed in January through May of this year.  For the private sector (excluding government), the deceleration is even stronger: only a monthly average of 77,000 jobs gained in June through September compared to 225,000 from January through May.

Since the recession began in March of 2001, payrolls are down by 940,000.  Since 1939, when the BLS began tracking such data, by this point in every business cycle (42 months past the prior peak) job growth had surpassed—in some cases, far surpassed—the prior employment peak.

The Bureau explicitly addressed the impact of the recent hurricanes on today’s report, noting that such events cause both negative and positive effects on employment (the latter due to rebuilding and accommodating evacuees).  They conclude that, while they cannot “precisely quantify the weather effects,” they “do not believe that the net result of these factors materially changes the national employment situation for September.”

The BLS report revealed considerable weaknesses throughout the economy.  The household survey, from which unemployment is measured, reported a decline of over 200,000 in both employment and the labor force.  While this survey is more volatile on a month-to-month basis, it clearly reveals a labor force that is growing half as fast as has usually been the case in recent years (0.5% per year so far this year compared to 1% in prior years).  The deceleration is a signal that many potential job seekers are avoiding job searches in such a weak labor environment.  Most importantly, since these persons are not counted among the unemployed, it suggests that the 5.4% unemployment rate paints too rosy a picture of the tautness of the job market.

It is also notable in this regard that the share of the population employed—62.3%—remains two percentage points below its pre-recession level, another sign of lagging employment opportunities.  Over the past year, the employment rate has crept up only 0.2 percentage points.  At this point in previous recoveries, employment rates were increasing more quickly.   For example, at this stage in the 1990s recovery, which also began with a jobless phase, employment rates were up 0.8 points over the prior year; the comparable figure for the 1980s recovery was 1.1 percentage points.
The lack of job creation is also evident in the data on long-term unemployment, which show that 21.8% of the unemployed have been jobless for at least half a year, up from 20.7% last month.  Since October 2002, more than one-fifth of the unemployed have fallen into this long-term jobless category, a range typically seen during periods of much higher unemployment rates. This represents the longest stretch on record in which at least one-fifth of the unemployed are long-termers.
Employment changes by industry reveal pervasive weakness in hiring.  Manufacturing employment, which had recently reversed trend and started growing this year, contracted by 18,000, the sector’s worse month since last December.  Private-sector service jobs were up by 72,000, but 33,000 of these net new jobs were in temporary help, a continuing sign of employers’ cautious hiring practices.  Retail trade is now in a three-month slide, down 15,000 last month and 34,000 since June, reflecting tepid growth in retail sales, particularly among large, lower-end retail outlets.  Information services has also lost ground over the past three months, shedding 12,000 jobs in September, led by a 9,000 decline in telecom. 

Of those sectors adding jobs, the largest was government, adding 37,000, exclusively among state and local governments.  Other than a strong month for real estate and rentals—up 15,000—there was little evidence of strong private-sector labor demand.

Average hourly earnings were up 2.4% over the past year, ahead of the annual growth rate in recent months, but still below the most recently published annual rate of inflation (2.7% for August 2003-04).  Average hours per week have been flat for the past three months, stuck at 33.8 hours per week, which is also the average thus far this year.  This is a half-hour less work per week than the 2000 average of 34.3 hours.  For someone working 50 weeks per year, this amounts to 25 fewer hours per year, which at the average wage equals a loss of almost $400.

As is their practice, the BLS announced in today’s report the preliminary estimate of the so-called benchmark revision for the payroll survey.  According to this estimate, the employment level for March 2004 will be 236,000 higher than the current published levels, meaning that this many more jobs were added between March 2003 and March 2004 (the Bureau “wedges” the revision in at a rate of just below 20,000 per month—236,000 jobs divided by 12 months—from April 2003 through March 2004).  Thus, if this preliminary estimate holds up (and such estimates can change significantly by the time the final value is derived), payroll comparisons that cross the affected period will reflect this higher level of job growth.  For example, the number of jobs lost since the recession began in March 2001 will equal about 700,000, not 940,000.  The latter number, noted above, is still the official BLS estimate until the revision is officially incorporated in the payroll series early next year.

Today’s report confirms that although the jobless recovery is over, the nation has yet to reach a level of employment growth that will absorb the slack remaining in the job market.  In fact, September’s payroll gains of 96,000 appear to be part of a new, diminished rate of job growth that began a few months ago and stands in contrast to the stronger growth rate earlier in the year.  At this point, the main factor keeping unemployment from rising is the number of people withdrawing from or not entering the labor force.

Today’s report was written by EPI senior economist Jared Bernstein with research assistance from Yulia Fungard.

For more information on the most recent job and wage data, go to EPI’s web feature

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 Wages, Incomes, and Wealth | Inequality and Poverty

See more work by Jared Bernstein