Saturday, October 6, 2012

Last Week's Jobs Posting

It might not seem like it, but today's newly released job numbers are actually really bad news for Obama.

The Bureau of Labor Statistics released today news that the unemployment rate had dropped from 8.1% in August to 7.8% in September. Awesome, right?

Maybe not.

Let's analyze: the report says that according to the Payroll Survey, we gained only 114,000 jobs in September. Also the labor force participation rate has held flat at 63.6%. Then, later in the report we are told that the labor force grew by 873,000 new jobs according to the Household Survey.

Anyone else confused? How did unemployment fall to under 8% for the first time since 2009 if labor participation rate is unchanged? And how do we have such radically different job counts in the same report?

The reasons are why Obama may be in for some bad news if people latch on to these numbers too much.

The Household Survey and the Payroll Survey are two different surveys that attempt to analyze the same data from different perspectives. The Household Survey contacts approximately 60,000 households as to their current employment status and includes the declared self-employed, family workers, agriculture jobs, receiving any money for any work, and those who are absent from their jobs without pay. The Payroll Survey is of 141,000 businesses and government agencies and includes none of the "jobs" I just listed.

Though they often track well together, you can probably see why results from one can dramatically differ from the other. Beyond the sample size differences, the way the data is reported in the Household Survey can skew wildly depending on a number of factors. In fact, the Household Survey is fairly volatile, leaving lots of outliers in the tracing of employment trends.

My intuition is that this month was an outlier. Either the respondents to the survey over-reported jobs or the smaller sample size skewed the results.

If this month's number is relied on by the Obama administration to prove a recovery and it is belied by next month's numbers, It could end up costing Obama the election.