The US unemployment rate measure is deceptive | Hacker News Reader

The US unemployment rate measure is deceptive

qz.com

19 pointsprostoalex9 comments

How the Government Measures Unemployment https://www.bls.gov/cps/cps_htgm.htm

Beyond the unemployed there are separate numbers published by BLS for those not in the labor force. This number is at about 95 Million.

Who is not in the labor force? https://www.bls.gov/cps/cps_htgm.htm#nilf

The numbers https://www.bls.gov/news.release/empsit.t01.htm

As long as the metrics are consistent, then for most applications it doesn't matter whether they precisely measure actual unemployment, only that they track it. The second graph indicates that this generally holds true, with some divergence due to the Great Recession. Public policy initiatives are pretty blunt instruments, and it is usually enough to say that they are working when the unemployment index is going down and not working when it's going up. Anything beyond that is relying on a level of precision that neither the measurements nor the policies justify.

That's not entirely true. Consider the metric "Yearly Income" which has been consistently measured for decades. It is a proxy for "Total Compensation" but isn't exactly the same because the latter includes health benefits, vacation pay, etc.

In the last decade, total compensation has risen drastically while income has not. If you only measure income, then it will appear as though the lowest paid fulltime workers did not see as great of gains as the highest-paid workers, when in fact the difference is much smaller when you include health benefits.

So, income is an example of a consistently-measured metric that does not really corelate with the attribute that most people would care about (total compensation) and thus does not act as a good proxy for the measured attribute.

But in the employment case, the proposed 'improved' metric tracks very closely, with some divergence around the recession. Obviously, if that divergence continued indefinitely so that the metrics didn't track together anymore, your argument would hold.

As a side note, the increase in income inequality is not "much smaller" when benefits are included. See for example https://ourworldindata.org/incomes-across-the-distribution/. The gap is almost entirely due to soaring executive compensation, and since there is no correlation between performance and executive compensation (some studies even show an inverse correlation: https://cooleypubco.com/2016/07/25/new-study-shows-inverse-c...), any natural brake on the spiraling salaries seems to have failed.

"According to the November 2016 data, over 5.5 million Americans said they want a job, but don’t have one, and are not considered a part of the labor force. If these people were included in the unemployment rate, it would jump to 8.2%. (a rate 72% higher)"

3 million of these 5 million did not look for job in the last year. Hard to get a job if you don't search for one.

Those did not report to EDD that they were looking for work during that census period, Not the same thing.

Another way that the number is fudged is that those who are eligible to work under DAPA are not counted in the population until/unless they are employed (in an actual W-2 or 1099 job).

"want a job, but don’t have one" may be because they have other commitments such as child care or looking after elderly relatives, rather than that they can't find one. Arbitrarily lumping them in with the unemployed is completely unwarranted and missleading.

The philosophy behind the change is that they are now publishing 'structural unemployment': the amount of people who will be unemployed no matter what (because of an inherent lack of jobs).

This is opposed to the old use 'transient unemployment' which means the amount of people who do not currently have a job, perhaps because they are searching.

Reply on news.ycombinator.com