Just When Did the Labor Market Begin to Soften?
Data on Job Gains and Losses Suggest It’s Earlier than Previously Thought
While net employment changes are usually tracked as a major indicator of the growth or decline of the economy, these numbers mask the underlying process that begets the net results, a process that includes employment turnover, job creation, and job destruction. To track this underlying process, the Bureau of Labor Statistics launched its Business Employment Dynamics (BED) in 2003. BED is a set of statistics that tracks gross job gains and losses between periods of net employment reports. Gross gains represent the sum of all jobs added at the opening and expanding of establishments, and gross losses are the sum of all jobs lost at the closing and contracting of establishments.
The most recent BED data show gross job gains totaling 7.25 million in the third quarter of 2007 and gross job losses totaling 7.5 million. Before this quarter, the difference in the series, or net employment growth, had not been negative since the second quarter of 2003, in the period following the 2001 recession now referred to as the jobless recovery. The rate of gross job gains fell to 6.4 percent from 6.7 percent in the previous quarter, its lowest since the series began. This decline is mostly explained by existing (or expanding) establishments, where the rate of gains dropped from 5.5 percent to 5.1 percent. At new establishments, the rate of job creation actually increased from 1.2 to 1.3 percent. The rate of gross job losses, in the meantime, increased 0.1 percent to 6.6 percent. Again, existing (or contracting) establishments were solely responsible for the rate change.
In fact, most gross job gain and loss activity occurs at existing (expanding or contracting) establishments rather than at new or closing facilities. In the third quarter of 2007, for example, 80 percent of all job gains occurred at expanding establishments and 82 percent of job losses occurred at contracting establishments. The graph below illustrates this fact by showing that activity at expanding and contracting establishments constitutes a much larger share of total employment than that of new or closing firms.
It is interesting to note that BED data show negative employment change beginning in the third quarter of 2007, while negative activity in the monthly Current Employment Statistics (CES) Employment Report does not begin to surface until January 2008. Since BED data are quarterly and the CES report is monthly, comparing quarterly sums of the CES data to the BED data is informative. It is also necessary to remove government employment from the CES for consistency with the BED. When viewed in this way, net job loss turns up in BED a full two quarters before the Employment Report shows net loss starting in the first quarter of 2008. This evidence from the BED implies that labor markets started to weaken much earlier than the CES data suggest.
There are a number of differences between the two data sets besides their frequency. Business Employment Dynamics is released quarterly with a lag of about nine months. Not until the May 21 release did we see the numbers for the third quarter of 2007, while the CES Employment Report is released monthly with just a one-month lag. Although not as timely as the Employment Report, BED data are based on the Quarterly Census of Employment and Wages (QCEW). The QCEW requires all employers subject to state unemployment insurance laws to submit employment and wage information, so BED data are based on a virtual census covering about 98 percent of all nonfarm employers. The CES Employment Report, on the other hand, is based on much smaller monthly sample surveys and is benchmarked to the QCEW data once a year.