BLS Jobs Report Covering October: It looks a lot like August
The Bureau of Labor Statistics jobs report covering October is mildly positive. The number of jobs created was 80,000 with the private sector adding 104,000 and government losing 24,000. The number of jobs created for August was revised upward from 57,000 to 104,000 (an increase of 47,000) and September's figure increased from 103,000 to 158,000 (an increase of 55,000). While both numbers show improvement, both cast doubt on the reliability of the Establishment survey from which they are drawn. The September numbers were off by a third and will have another revision next month. The original jobs number for August was 0 so the entirety of the 104,000 jobs for that month came from subsequent revisions (0 > 57,000 > 104,000).
Turning to the Household data, in October the Civilian noninstitutional population over 16 (NIP, the potential labor force) increased by 198,000 to 240.269 million. The employment population ratio increased a tenth of a percent to 58.4%. Multiplying this ratio by the increase in the NIP gives us the number of jobs which needed to be created in October to keep up with population growth or 116,000. If the 80,000 jobs created were a number we could trust, and we can't, then we could say that the economy did not create enough jobs to keep up with population growth. As it is, the most we can say is that the economy is creating jobs at a rate that will have little effect on unemployment.
In October, the number of unemployed decreased by 95,000 to 13.897 million. The civilian labor force (the actual labor force) increased by 181,000 to 154.198 million. Dividing the one number into the other gives an unemployment rate of 9.0%, a decrease of a tenth of a percent (with rounding) from September. Additionally, those employed increased by 277,000 (the sum in the change of these two numbers) to 140.302 million.
Parenthetically, a full tenth of a percent (0.1%) change in the labor force, such as an increase or decrease in the unemployment rate would be a thousandth of the labor force number or 154,000. So to reduce unemployment by a full percentage point would entail the creation of 1.5 million jobs. However, these numbers are dynamic. Those 1.5 million jobs would need to be created tomorrow. If they were created over the course of a year, natural population growth would account for most of them. So something to keep in mind when hearing about proposals for jobs programs is that closer to 3 million jobs would be needed over the space of a year to decrease the unemployment rate a full percentage point, or 4.5 million to decrease it by 2%, or 6 million to decrease it by 3%.
However, the U-3 unemployment rate, as we all know, is not the whole story. In October, the U-6 measure of under and un- employment which spiked in September to 16.5% fell back to its August level of 16.2%. On the other hand, those marginally attached to the labor force (and part of those not counted in it) edged back up from 2.511 million to 2.555 million. Similarly, those employed part time (but who would wish to work fulltime) decreased 374,000 from 9.270 million to 8.896 million close to their August levels. The unemployed, the marginally attached, and involuntary part time workers are the 3 components of the U-6. In October, most of the change in the U-6 came from the decrease in part timers. The U-6 in October represented 25.348 million Americans.
Looking now at disemployment, the BLS showed an increase of 162,000 to 6.403 million in its category, Not in Labor Force Want a Job Now, which I take to be its measure of its undercount. Adding together the unemployed, forced part timers, and Not in Labor Force Want a Job Now gives 29.196 million, a decrease of 361,000 from September, for a BLS-based calculation of the disemployed, or a disemployment rate of 18.2%, down from 18.7% in September.
For my alternate calculation of the BLS undercount, I first calculate what the labor force should be if we were in a reasonably robust expansion: .67(NIP) and then subtract the current labor force from it. The difference is the undercount. My calculation for the undercount for October was 6.782 million, a decrease of 49,000 from September, and for disemployment it was 29.575 million, down 572,000, giving a disemployment rate of 18.4%, down from last month's 18.7%.
In other matters, the average work week was unchanged at 34.3 hours. Hourly wages increased 0.2% to $23.19. This continues to put it behind rises in the CPI-U measure of inflation (3.9% seasonally unadjusted) meaning that real wages continue to fall. The unemployment rate for white Americans remained at 8.0%. The unemployment rate for African Americans continued to decrease, although still high, to 15.1% from 16.0%. Teenage unemployment also discreased a half a percent to 24.1%. Long term unemployed (27 or more weeks) decreased 366,000 to 5.876 million, 42.4% of the unemployed.
All in all, the October report looks a lot like the August report with September being an outlier for all kinds of reasons. There look to be problems in both the Establishment survey with its large percentage revisions in job creation and in the Household survey with the September numbers. Was September the fluke with its spikes or October with its fall backs? Some static in the numbers is to be expected but not to this degree. It once again places in question the validity and value of the BLS numbers and models and vitiates the usefulness of making any predictions or projections based on them.