The BLS sample survey of business and government Establishments shows 31 states, D.C. and the U.S. as a whole added (seasonally adjusted) jobs in July. The more reliable two-month totals shows 34 states, D.C. and the U.S. added jobs over the two months of June and July with Indiana, Illinois and Virginia suffering the worst losses while Alaska, North and South Dakota enjoyed the strongest percentage gains.
The BLS Establishment survey has shown overall job growth to have been quite weak over the past three months but continuing to grow overall.
However, the BLS sample survey of Households—the survey that produces the unemployment rate—suggests a very different and quite disturbing story. As I’ve pointed out for many months now, the overall U.S. unemployment rate would be over 15% rather than the current, official rate of 9.1% in July if the number of people counted in the labor force had continued to grow with population as usual rather than decline for the first time on record over the past three years.
In July the (seasonally adjusted) number of people counted in the labor force as either working or looking for work DECLINED by 193,000, and this was after a decline of 272,000 people in June—a two-month decline of 465,000 people. My analysis of today’s new state data show that the number of people counted in the labor force also declined in 44 states and D.C. in July and in 45 states and D.C. over the past two months, limiting the rise in the unemployment rates that nonetheless rose in 28 states and D.C. in July, fell in nine states and the U.S., and remained unchanged in 13 states.
The July unemployment rates in Texas and Wisconsin, for example rose from 8.2% to 8.4% and from 7.6% to 7.8%, respectively, but these jobless rates would have risen even more were it not for for sharp declines in the number of people counted in each state.
But my most disturbing finding in the July state data is that the number of people claiming to be employed (including self-employed) declined in all 50 states, D.C. and for the entire U.S. Over the two months of June and July employment declined in the U.S. and in 49 states and D.C. Only in Alaska did 19—yes, 19—more people indicate that they were employed in July than in May.
Although both surveys have suffered from smaller sample sizes and other issues in recent years, readers of this note know that I routinely emphasize the establishment jobs data over the household employment data. But there is no question but that the economy has grown much more slowly this year than many (including at the BLS) expected, which almost certainly led them to overestimate in their birth/death model the number of net new businesses and jobs in recent months.
This raises two very important questions for understanding the condition and prospects for the economy now and in the near future: by how much is BLS overestimating the number of jobs being created and how severely is the slowdown affecting entrepreneurs and the self-employed? The answers to these questions are very different for each state, region and metro area, but today’s data make the answers important to pursue.