In the few years that I have been more intensely looking at employment data, I have grown more skeptical about the accuracy of much of what the Bureau of Labor Statistics puts out. While I joke about political interference in data gathering and reporting, I would like to think it isn’t happening, that the run of constant unfavorable revisions is due to low survey response rates or some other methodological issue. That is hard to accept as an explanation when the revisions tend to run in only one direction and substantially increased during the Bidementia administration. So I don’t much trust them.
If you think big picture, the number of people employed on any given day in the US is an objective fact, subject to just a little waffling. The best indicator would likely be looking at payroll data–how many people did each employer give a check to each week (ignoring the cash economy). Social Security should provide some of this data–who were contributions being made for–which would include self-employed persons. Unemployment and workers compensation data is also helpful, employers typically have to make a contribution to these systems for every employee. Private companies like ADP and Paychex that administer payroll for firms also have data and ADP puts out its own monthly report on employment. Governments should know who they employ. So you would think that it wouldn’t be that hard to get accurate data.
The BLS relies on surveys–one of businesses and one of households–to get its employment data. The response rate to these surveys has become abysmal. The BLS data is regularly revised based on other measures, like the unemployment system records. In August we just saw a very large downward revision for 2023 based on this. Some labor statistics have other issues, like determining what the size of the labor force is–who could work and what percent actually do work. The data is critical because it gives a meaningful picture of general economic health and trends, but somehow we can’t seem to get a very accurate read.
And so it is with the August jobs report. First, note that as usual, the prior months were revised downward, July from 114,000 to 89,000 and June from 179,000 to 114,000. Then August comes in at 142,000 versus expectations of 165,000. So lower than expectations but up from the revised prior months, but given that almost every initial report this year has been revised downward, do you really think that August is accurate or is it going to be revised down to closer to 100,000 in subsequent months? The household survey, which has lagged the business survey in the past, this time showed 168,000 more employed persons. The unemployment rate, which comes from the household survey, declined from 4.3% to 4.2%, but as I said, determining the size of the population that could or wants to work is difficult.
We supposedly have 1.5 million people who have been unemployed for 27 weeks or more, which is just startling–in an economy in which hiring is supposed to be so hard, how do people stay unemployed that long–do they really want to work? We have another 7.1 million who could work, but apparently aren’t even trying. And why would you if you can get a government handout. Wages rose slightly month-over-month. Government supported sectors like health care and social services accounted for much of the gain, while manufacturing lost jobs. Part-time job-holders continue to rise, as do multiple job-holders. (BLS Data) (ZH Post)
