Debating continued two days after Barack Obama and Mitt Romney met in Denver, not between the candidates but over the monthly ""Employment Situation report"":http://www.bls.gov/news.release/pdf/empsit.pdf showing the unemployment rate dropped to 7.8 percent, the first sub-8.0 percent unemployment rate since January 2009.
The 0.3 percentage point drop was the largest month-month decline since January 2011. What started the ""debate"" was that the decline in the unemployment rate was accompanied by a modest 114,000 increase in payroll employment.
The ""debate"" revealed more about those mounting the arguments â€" and how little they understand about the report â€" than it did about the report itself.
And, what's worse is the answers were right there in the report itself â€" albeit tucked away in boilerplate. That's perhaps more indicting: the information has been there for journalists, economists, pundits to read month after month after month.
The surprise drop in the complicated data point immediately gave birth to ""truthers"" who alleged the books were cooked and the number manufactured to benefit the President. Instead of adding to the body of knowledge, the ""truthers"" revealed they either are stupid or believe the rest of us are.
The two pieces of data â€" the unemployment rate and change in payroll employment â€" are unrelated.
So, why did the unemployment rate drop when payroll employment barely budged?
The answer could be found not only in the footnotes but in two other pieces of data in the monthly BLS report: the number of multiple jobholders and self-employment.
The unemployment rate is computed by dividing the number of people unemployed by the some of those employed and unemployed â€" information derived from a survey of 60,000 households each month with the results extrapolated to the entire population.
Unemployment is a defined term; an individual has to be out of work, available for work and looking for work to be considered ""unemployed"" by the Bureau of Labor Statistics. The number of payroll jobs comes from a companion survey of businesses who report on the number of people on payrolls, paychecks issued, during a specific week in the month. It doesn't matter if an individual has worked one hour or 40 hours â€" or more â€" during the week; if it resulted in a paycheck, it counts as a job.
In September, payroll employment increased by 114,000. It represented the weakest increase in the last four months. In the preceding three, the unemployment rate barely budged, going from 8.2 percent in June to 8.3 percent in July when payrolls grew 181,000 and to 8.1 percent in August when payrolls expanded by 142,000.
The confusing factor is the weak growth in payroll jobs. That may be explained by the increase in self-employment since a self-employed individual would be under the radar of the CES which surveys only a sampling of businesses which have filed unemployment insurance tax returns.
The number of self-employed individuals increased 144,000 in September. To the extent those individuals were in the CPS, they could have correctly described themselves as ""employed"" even though they wouldn't have been included in the count of jobs. Another often-overlooked statistic in the Employment Situation report is the number of multiple jobholders. In September, the number of multiple jobholders declined. If an individual held two jobs and lost one of them, the number of jobs would go down, but employment/unemployment would be unchanged since that individual would still be employed. In September the number of multiple jobholders declined 55,000 without affecting either employment or unemployment. It could though have retarded job growth in that jobs could have increased 169,000 but the elimination of 55,000 ""second"" jobs would have resulted in the (net) increase of 114,000 jobs reported by the BLS.
The last time the unemployment rate fell by as much as 0.3 percentage points in one month was in January 2011 and, guess what? Payroll jobs increased by only 110,000. (The first preliminary estimate of payroll job showed an increase of just 36,000, increased to 63,000 in the second estimate before the final revision produced a gain of a still weak 110,000 jobs.)
In January 2011, the number of multiple jobholders fell 49,000 which means each of the newly created jobs â€" whether 36,000, 63,000 or 110,000 -- went to reducing unemployment and thus contributed to lowering the unemployment rate. In the same month, the number of self-employed increased 139,000, reducing unemployment without showing up in the payroll report.
The September Employment Situation report was 38 pages long, crammed with important information explaining nuances of the economy. Economists, journalists and analysts have diverse constituencies. To dwell on two numbers without understanding them to the exclusion of all the others is a disservice.
To be sure, 114,000 new jobs in one month is nothing to cheer about when the over-16 population is growing at the rate of more than 300,000 a month. The unemployment rate was under 7.8 percent for the first 95 months of the 96-month George W. Bush administration, but given that the rate was over 8 percent for 43 straight months, 7.8 percent is a move in the right direction.
The irony though is the rate could easily pop back over 8.0 percent in October as individuals who were on the sidelines start to believe they have better chances of getting jobs and re-enter the labor force as unemployed. Arithmetically, if both the numerator and denominator of a fraction increase equally, the percentage value of the fraction will increase. It would be a shame if critics deny this basic arithmetic truth and instead claim they forced the BLS to correct itself.