Something Very Fishy About The BLS' September Unemployment Rate

by: TrimTabs

By Madeline Schnapp

The market was in a tizzy last Friday after the BLS reported a dramatic 0.3% drop in the September unemployment rate to 7.8% from the 8.1% rate reported in August. While a lot has already been said about whether or not the drop is real, we found something new this weekend that further confirms our suspicion that there is something very fishy with the September BLS unemployment data.

There are two BLS employment surveys. There is the larger BLS Establishment survey, which produces the widely quoted employment number. There is also the smaller BLS Household survey, which produces the widely quoted unemployment rate. In today's blog post, we are focusing on the unemployment rate derived from the BLS' smaller household survey.

The unemployment rate involves two numbers and is calculated by dividing the number of unemployed by the total number of people in the workforce. The total number of people in the workforce, in turn, is the sum of the number of employed plus the number of unemployed. In order for the unemployment rate to drop as much as was reported in September, two things must happen: first, the number of unemployed must fall; and second, the number of employed must increase. So what happened this past September?

First, the number of unemployed dropped by 456,000, the largest September one month decline in the history of BLS data dating back to 1948. In addition, the number of employed jumped by a stunning 873,000, the largest September one-month gain in the history of BLS data. How is this possible in an economy experiencing the weakest recovery since the Great Depression?

As usual, when we suspect a problem with the data, we compare the BLS seasonally adjusted data with the non-seasonally adjusted data. More often than not that is where the problems lie.

On a non-seasonally adjusted basis, what we discovered is that in a typical September, the number of employed declines as temporary summer seasonal hires are let go. According to the BLS, the September declines range from a 100,000 to over one million depending on the year and whether or not the economy is in recession.

Rarely in September do the number of employed increase. In fact, in the 64 Septembers since 1948, the number of employed has increased only six times. In four of those six instances, the increase in employed was less than a 100,000. But this September, the increase in the number of employed was a whopping 775,000, nearly five times greater than the previous high. Whoa! What is going on here? I smell a dead fish! Well according to the data, most of the increase was due to a highly unusual increase in the number of part-time workers that wanted full-time work but were unable to find full-time employment.

Source: Bureau of Labor Statistics

Also, on a non-seasonally adjusted basis, the number of unemployed dropped by the largest amount since 1948 by nearly a factor of two. Those smelly dead fish just got a lot stinkier!

Source: Bureau of Labor Statistics

That's not all. We found more dead fish in the non-seasonally adjusted part-time work BLS data as well. Since 1955, in every September except three, the number of part-time workers fell. Miraculously, however, this September, not only did the number of part-time workers increase but increased by a lot.

Source: Bureau of Labor Statistics

Is it possible that the BLS data is being manipulated for political reasons? I certainly don't have enough evidence to confirm my suspicions. What I can say is if it smells like a dead fish, there is always a possibility that it might actually be a dead fish. We leave it for you to decide.