August’s Poor Employment Report
The Bureau of Labor Statistics (BLS) released its initial estimate of total nonfarm employment today, coming in up 96 K from the prior month. The 96 K gain in employment coincided with a drop in the unemployment rate from 8.3% to 8.1%, as 400 K fewer individuals sought work. The government’s accounting was strikingly different than ADP’s August report, which gave an initial estimate of 201 K jobs gained, with 185 K in the service sector and 16 K in the good producing sector.
Household and Payroll Employment
When gauging employment conditions, the BLS uses two separate surveys – one of households and one of businesses – with the employer survey usually being the source data for the monthly employment growth numbers reported in the press, and the household survey used for the unemployment rate and labor force figures. Although some professional analysts always prefer the employer survey, because of the size and scope, the household survey is not irrelevant and is generally a more accurate leading indicator heading into an employment recession, as is shown in the following chart from 2007 to 2009 (the employer survey is considered a more accurate survey coming out of a recession). The household survey has given two consecutive months of negative employment growth, with August’s figure indicating a decrease of 119 K. The economy may not be headed for a recession, but anyone that feels confident solely based on the employer survey is a poor analyst.
The Labor Force and the Unemployment Rate
August’s employment report gave an unemployment rate of 8.1%, a decline of 0.2% from the 8.3% of July. Why such a decent decline in the unemployment rate when employment only increased 96 K? The answer lies in the BLS’ estimate of the group of individuals referred to as the “labor force,” which dropped another 368 K this past month, on top of the 150 K of July. This is represented by the blue dot and the right axis in the above chart.
Depending upon the family structure, the 518 K drop in the labor force may not be a good sign, if it represents individuals giving up looking for work who really want to work. On the other hand, it could be a good sign if, for instance, it represents women leaving the workforce to spend more time with their kids (presuming there’s no aggregate decrease in productivity). In looking over the longer horizon of the recovery, it still remains unclear whether there is a long term structural change in the percentage of individuals that want to work.
What does all this mean? Overall, the recent employment report increases the probability of at least one month of negative employment growth within the next year, based upon the employer survey, to about 64%, a terrible sign if you’re an American worker. The American labor market is dynamic and BLS figures are subject to large revisions, so hopefully the initial numbers turn out to be simply false alarms.