Sic Semper Tyrannis

The latest BLS employment situation report states 4.5% unemployment, and 167,000 jobs created in December.

Are you done doing backflips? Good (just kidding).

While anecdotally it is easy to see that there is indeed a shortage of "skilled" workers in the US today, it has struck me in recent months as a bit hasty to extrapolate this observation into the conclusion that "unemployment is low." Intuitively (to anyone with their eyes open, at least), one suspects that in a society as bifurcated as ours, the half with the "short end" of the stick doesn't have such bright employment prospects. You just get that kinda feelin', being out and about.

Suspicions that the headline numbers are somehow failing to reflect this reality find confirmation from a surprising place: the information put out by the government bureaus themselves.

Follow up:

First off, we learn here that the headline "unemployment" number we see every month isn't the unemployment number -- it's one of many, in a creative hierarchy introduced circa 1995. In fact, there are three more key unemployment numbers: one including marginally attached workers, one including discouraged workers, and one including compulsory (my term) part-time workers:

The explanation of these categories is :

Marginally attached workers are persons who currently are neither working nor looking for work but indicate that they want and are available for a job and have looked for work sometime in the recent past. Discouraged workers, a subset of the marginally attached, have given a job-market related reason for not currently looking for a job. Persons employed part time for economic reasons are those who want and are available for full-time work but have had to settle for a part-time schedule.

I don't know about you, but those strike me as really important things to track. Especially if, like, you subscribe to NAIRU and believe high employment is the source of inflation (here I'm thinking of fringe players like the Fed and Wall Street... and the mainstream economics profession. And finance.). Then wouldn't you want to know how high employment actually was, in the most comprehensive sense?

Maybe not, if you knew you might not like what you'd see.

So for one, it turns out all those college grads forced to take a few shifts at Starbucks to make ends meet show up as "employed" by the headline count, but tellingly aren't included in the compulsory part-time count. Because, well, in an important sense, these people aren't fully-employed.

What bothers me a little more, though, are those categories of "discouraged" and "marginally attached" workers. After enough failure at finding a job, the natural thing to do is give up looking. And the only real distinction I can divine between "marginally attached" and "discouraged" workers as that discouraged workers at least have a theory as to why they can't find a job. For that, they go even farther into the basement of statistical purgatory.

It's time for some numbers. What does the breakdown look like for this whole dichotomy? For December, 2006, We have:

Headline unemployment	4.5%
Unemployed+discouraged	4.7%
The above+marginal	5.3%
Above+part time	8.0%

Wow! 8%! That's scandalously close to "twice the headline number"!

As is probably not too shocking, the most widespread problem these days seems to be those "underemployed" part-time folks. I sure run into them a lot more than vagrants and deadbeat 20-somethings living back at home and the like -- but there's certainly no shortage of any of these demographics.

Interestingly, the data in the above breakdown isn't included in the employment situation report itself; that only gets the smaller, mostly-exclusive (headline) unemployment figure. How convenient.

If you aren't yet at least annoyed, don't fret: one can get even more creative from here. At one point not too long ago, I reflected on our nation's highest-in-world prison population which I had heard to be approximately 2 million, and wondered what the impact of that population is on unemployment record-keeping. Is the government sweeping an endemic unemployment problem under the rug, by doing things like locking up nonviolent drug offenders (rumored to be about half of the total) and vagrants?

The data is certainly suggestive.

Reporting from the Department of Justice's data, in 1980, about 1.5 million people were in the corrections system in some capacity (in jail or prison, on parole, or on probation). In 2005, that number had skyrocketed to about 7.5 million, with about 2.2 million of those physically in jail. As a sidebar, that is twice communist China's incarceration level, which compared to the 1.3 billion people of China represents a per capita incarceration rate over 8 times higher! Not something to be proud of.

Anyway, that means the ongoing corrections population in the US grew by 6 million in the past quarter century; or a growth of a factor of five (the national population itself has somewhat lagged this rate, racking up only a factor 1.3 in the same time span).

With numbers like that, we could be talking about some real distortion of employment figures. What if we were to actually count those people as unemployed -- which they are? (In fact, they're worse than unemployed -- we spend well over the median wage supporting each of them -- but we'll let that slide for now.) The picture might look like this:

Headline unemployment	4.5%
Headline+prison population	5.8%
Headline+all corrections	9.2%
All-unempl.+prison pop.	9.3%
All-unempl.+all corr.	12.7%

Almost 13%. That's a lot of the workforce that's actually not productively engaged at all! (Note: here I'm using 152 million as the baseline labor pool as per the BLS, and adding in 2 and 7.5 million from the corrections population as appropriate for comparison. I'm also implicitly assuming non-imprisoned people who are on parole or probation aren't very employable; but if you dispute this, you're welcome to use the 9.3% imprisonment-only number).

In some subpopulations, this is all rather noticeable -- better than 1 in 12 African American males are locked up.

I've been suspicious of other objective problems in the unemployment reporting, as well. For example, as someone involved in hiring in the IT area, I've noticed that our biggest employment competitor is... the government. And not just the government in general, but in specific, defense and intelligence as well as their private contractors. In sum, the homeland security complex. Given the amount of deficit spending supporting this massive employment complex, I have my doubts it will be around forever -- so we certainly shouldn't be relying on all those jobs it provides.

Another example is the approximately 2.5 million realtors counted as "employed." Are they, now? I bet a good half to a third of them aren't earning enough commissions to stay in the business anymore, now that we've hit a major slowdown. It's only a matter of months until quite a few of them (perhaps .5-1.5 million) hit the unemployment line. Ditto housing construction workers.

No amount of massaging is going to be able to hide all that.

And then there's health care; my arch-nemesis (ok, maybe banks are my arch-nemesis. Or maybe the Fed. But health care is at least on my "enemies" list). This industry has gone too far for too long, benefitting disproportionately from a broken tax structure. The double-digit annual expense growth in this sector for the past decade has naturally fuelled an employment bonanza. The spending in this area would surely be trimmed back by about 1/4 just bringing the US in line with the rest of the developed world; I suspect it could be cut back by 1/3 or more by applying actual "free markets" (I heard they're a cool thing invented by this "Friedman" guy...)

One day, the day of reckoning for the US health care complex will come (for some, the cold hand of death is already being felt upon the shoulder).

Coincidentally enough, BusinessWeek recently had a story where they analyzed the main sources of job growth in the US since 2001. Here's the chart that sums up the whole article:

Submitted for your consideration.

Please don't feed the statistigandists!

Update, 6 Jan. 2007, 6:30pm: I should also mention for those that want to understand the employment issue at an even less controversial level that the guys at Econbrowser have been doing a good job (no pun intended) of putting job growth in perspective. Notably, they frequently publish a comparison of employment growth in the "current" recovery with that in the last, revealing that the Bush "economic miracle" has been rather mediocre. However, it's charts like this one in a recent post that inspired this writeup: I find it quite ludicrous to suggest that we've dramatically reduced structural unemployment relative to the past, especially given the obvious and directly-observable societal trends highlighted above (among other factors). Sure, they're just republishing the government's data, but data like that should come with a warning label.