February's Jobs Report Only Deepens the Debate On How Jobs Data is Calculated

Early each month, the Bureau of Labor and Statistics releases its employment statistics for the prior month. In February, the highlights of that report, as generally reported by the media, were that the economy added 295,000 jobs and the unemployment rate fell to 5.5%.

What most Americans don't know is that those two statistics come from two separate monthly surveys  and that a debate still rages as to whether or not which survey is more accurate at calculating jobs added to the economy.

In February, the 295,000 jobs added comes from something called the "Establishment Data" (Table B)  which is a result of surveying the employment activity of 190,000 supposedly representative businesses in the U.S.   And, from that 190,000 surveyed businesses, the hiring practices of the more than 27 million firms in the country are then estimated.  On the other hand, the "Household Data" (Table A) extrapolates employment data from nearly 50,000 phone calls to households across the country.  According to that survey, the number of unemployed fell by 274,000 but, at the same time, 178,000 left the workforce. So, net-net, the economy only actually added 96,000 jobs; leaving the labor participation rate at a 37-year low. 

For decades, the consensus has been that the Establishment Data is more accurate for counting job additions because it looks at job activity within the businesses themselves.  At the same time, the Household Data is the only real means to determine unemployment rates because it samples real people.  But, the problem with both assumptions is the low sampling rate.  Sampling 190,000 businesses of the 27 million firms in this country is a representative sampling rate of about 7 tenths of one percent.  Sampling 50,000 households (assuming each household is equal to 2.1 persons on average) means that out of a population of 320 million, the sample rate is less than 3 one-hundreds of a percent.  Either way, there is certainly a lot of wiggle room to be wrong on any numbers being collected.

The bottom line is that I, personally, wouldn't bet the bank on the accuracy of any employment numbers that we see from the federal government.  On the other hand, there have been numerous other reports that have not been as "glowing" as this last jobs report.  For that reason, I believe that the Household Data on jobs is closer to the truth with only 96,000 jobs created in February.  But, that is just my opinion. 

References:

February 2015 Employment Situation Report: http://www.bls.gov/news.release/empsit.nr0.htm

Establishment vs. Household Survey—The Debate Continues: http://research.rerc.com/blog/establishment-vs-household-survey-the-debate-continues

Quick Facts From the U.S. Census: http://quickfacts.census.gov/qfd/states/00000.html

Soft U.S. data hints at near-term hiccup in economic growth: http://www.reuters.com/article/2015/03/05/us-usa-economy-unemployment-idUSKBN0M11IA20150305

U.S. Economic Data Not Meeting Expectations: http://www.wallstreetsectorselector.com/investment-articles/editors-desk/2015/03/u-s-economic-data-not-meeting-expectations/ 

US jobless claims rise; Q4 productivity revised down: http://www.cnbc.com/id/102476643

Weekly Economic Calendars: http://www.briefing.com/investor/calendars/