The federal government employment reports have been a particular source of weird numbers bouncing around from month to month. The JOLTS report released yesterday by the Bureau of Labor Statistics and covering a period through the end of October is no exception. The report in October, covering the month of September, had shown a large drop in job openings, an indication of a slowing economy. The current report reverses much of that decline, with openings rising by 372,000. Professional and business services and restaurant and hotel industries led the gains. Hires were down, which suggests a weaker job market, but quits were higher, an supposed indicator of a strong job market, as people quit when they either have or think they will easily get a new job. Who knows what to believe. The usual reminder that this data is two-thirds estimated, the survey response rate is only 33%. (BLS Report)
And then today, ADP, which handles payroll for many companies, issued its estimate of October employment. September’s number was revised down by 50,000 jobs. And October came in below estimates at 146,000 new jobs. Manufacturing jobs declined significantly. Health care and education, largely supported by government spending, rose. Wages rose at a good clip, presaging continued inflation pressure. And this is a lot of real data, based on actual business employment. On Friday we get the monthly employment report from BLS, and who knows what that will show and how believable it will be. (ZH Post)
As previously stated, when the random sampling is bad, then the statistical data derived from the random sampling is going to be very bad. 30% response rate and the statistical data is going to be worthless crap. Stat class 101