Comparing Presidents, The Employment to Population Ratio

This post is on unemployment… well, kind of – really, its on the civilian employment to population ratio. I don’t like the unemployment rate as a figure… even disregarding issues of transparency (what are the parameters of the ARIMA model the BLS uses for forecasting – it matters, because results of ARIMAs will fluctuate quite a bit depending on one’s choice of parameters), unemployment figures depend on counting those seeking jobs. But the “discouraged worker” effect is notorious – these are people who either have given up looking for work, or aren’t proving to the satisfaction of the unemployment office that they’re looking for working. Since after a while unemployment compensation runs out… how many people that are unemployed for long periods are going to keep going down to the unemployment office every week to show that they’re still looking if there’s no gain to doing so?

Instead, I want to use civilian employment to civilian population. The BLS provides the data… and civilian employment and population is for all civilians 16 and over. Civilian employment, of course, is just an estimate, as is population, but with less to estimate, and no need to guess whether people who don’t have a job are actually looking for one, there’s less room for error.

So here’s the civilian employment to population ratio for people over 16…

Now, it could be that this figure is very dependent on women entering the workforce, or teenagers getting jobs. So let’s narrow down a bit… to the civilian employment to population ratio for males age 20+ which is also provided by the BLS. Presumably, most of these people (and I wish the data allowed us to separate out retirees!) would have needed jobs before and after women’s lib and whatever cultural changes might have occurred in the population.

So what do we learn from this… well, there is a greater share of the population employed over time. But males 20+ are having a harder time finding employment, or in my opinion far less likely, mooching off their wives/girlfriends and enjoying leisure time. It seems some of the employment obtained by teenagers and women has come at the expense of men’s employment prospects. I’m really surprised… maybe I haven’t considered this carefully enough, but offhand it seems to me it lends a bit of support to the “lump of labor” concept, the notion that there are only so many jobs at any one time. On the other hand, it could alternatively mean that with more people to pick from, employers are no longer having to accept the least desirable men into the workforce. Also, we all know that many of the jobs traditionally considered to be men’s jobs (assembly line workers, lumberjacks, milkmen, gas station attendants, programmers, barbers, etc.) have mostly disappeared and/or been outsourced and/or have morphed and are no longer considered “men’s” jobs.

Now a summary for each series…

Overall civilian employment to population rose in the Carter, Reagan, Clinton, and JFK/LBJ administrations, which were also the administrations with the fastest growth rate of real GDP per capita, though not in the same order.

For males 20+, the ratio only increased during the Clinton administration, though declines were small during Reagan, JFK/LBJ, and Carter.

I note… GW’s 2003 – 2006 performance puts him behind Clinton and above JFK/LBJ overall, and in first place for Males 20+. Cherry-picking sometimes works! However, it does seem that both series may have hit a local peak in December of 2006… monthly figures have been declining since then. Perhaps GW won’t look as good when measured from 2003 to the rest of his term. We don’t hear much out of him about employment since 2003 any more, but maybe that’s because he isn’t campaigning for anything any more. All that said… has there been an increase in employment of traditionally male jobs?

FWIW… a means test for each series rejects the null hypothesis that the growth rates are the same under Dem as under Rep administrations at the 95% level for both series. If the assumptions of the t-test are acceptable, we would conclude that growth rates of the civilian employment to population ratio (both series) are faster under Dem administrations.

That’s about all the time I have right now. I think I should revisit this as soon as possible and see how these series tie in with other things (changes in real GDP per capita, inflation, etc.)


Update… Originally I forgot to list the BLS site from which I took the data.