Let's start with October:
Unemployment drops from 6.1% to 6.0% according to the household survey.
According to the buisness survey only 137,000 new jobs were added.
Unemployment drops from 6.0% to 5.9% according to the houshold survey.
According to the buisness survey only 43,000 new jobs were added.
Now for December:
Unemployment drops from 5.9% to 5.7% according to the household survey.
The buisness survey has only 1,000 new jobs added.
Now I know, the household survey only accounts for those actively seeking employment, which can account for some of the drop in the unemployment rate. But still, the consumer confidence surveys are saying people are finding that jobs are increasingly easier to find. That's not a trend that says massive amonts of people are giving up looking for jobs.
(An interesting side note, 20 years ago economists thought that it was impossible for the unemployment rate to drop below 6%. So if you look at historical trends, we're actually at relitively low unemployment excluding the past couple of decades.)
Now the thing about the buisness survey is that, it's usually major buisnesses that get polled in it. So I think a lot of the drop isn't so much that people are getting out of the workforce, it's that people are going into buisness for themselves. Since there would be a lot of these startups, they would be off the grid for the buisness survey, but they would show up in the household survey.
Honestly, I think they need to come up with a method that combines both of the surveys. It would be more informative to know how many people actively gave up looking for work and how many people decided to go into buisness for themselves.
Personally, I think, we need a group to sit down and re-evaluate the surveys we use. Figure out the flaws and the strenghts of the various surveys and figure out some better ways of measuring the various indicators.
Of course, that then leads to the problem of the fact that there is no longer any baseline. If we come up with new methods, we can't compare it to historical data...