So let’s see, consumer confidence has been revised downward for eight of the last nine months, not sure how you revise confidence downward, but the underlying components of the index are more objective. Given current persistent high prices and some level of continuing inflation, coupled with very high interest rates, it isn’t surprising that consumers are not feeling great about their finances. And then home prices continue to rise, despite those high interest rates, pushing home ownership out of reach of more buyers. But the big news of the day is what is referred to as the “JOLTS” report, put out by the US government and measuring changes in certain aspects of the labor market, such as number of open jobs, quits, firings, etc. As is usually the case with economic reports issued by the Bidementia administration, these reports have all kinds of subsequent revisions, usually in a negative respect, and other weird things going on. (BLS Release)
Today’s JOLTS report did have negative revisions, but also continued a trend from last month, where the only reason it looks like job openings aren’t dropping precipitously is a huge surge is government positions. In the private sector, in June compared to May, job openings dropped by 100,000 and by one million from June 2023. Hires in the private sector dropped by 300,000 month over month and 550,000 year over year. Meanwhile government job openings rise. If you look at the detail by industry, certainly looks like manufacturing and construction are struggling. But remember, this is all just made up, since response rates are down to about a third of all companies contacted, and even these made up numbers are subject to further revision in future months. So have fun trying to figure out what is really going on.
The jobs report requires statistically valid sampling. As response rates drop, the validity of the statistically sampling degenerates. The census in an effort to compensate in order to obtain sufficient data increases the sample size which has an even lower response rate. That makes the random sample even worse not better. .