Where the jobs went: Dec 2007 snapshot

The private sector's job-creation rate is trending downward, and the unemployment rate ticked upward.  That's the "bad" news.  The good news: the private sector nonetheless created 1.5 million new jobs in the last 12 months, which more than offset the half million jobs that went away, for a net increase of one million new jobs.  More good news: one million (of the 1.5 million) new jobs pay more than the half million jobs that went away. 

I'm guessing that should please Lou Dobbs (...if only he would drop by this blog and take a look at the chart).  Apparently he doesn't, though; he always has either a frown or a sneer on his face whenever I check, which is at least twice a month. 

Here's the job quality chart.  Click to enlarge.

And here's the breakdown by job category.  Click to enlarge.

This month, I added a bonus chart.  This one is specifically for the commenters here who have apparently fallen in love with a statistic called the "Employment to Population Ratio" (EPR).  I chose to chart the Labor Force Participation Rate (LFPR) instead of the EPR, for reasons explained below.  Here's the bonus chart, showing the private sector's downward-trending job creation rate (red line), and the steady-as-she-goes Labor Force Participation Rate.  Click to enlarge.

Why did I choose the LFPR instead of the EPR?  If you're really serious about understanding the details, I suggest reading chapters 4-7 of Gene Epstein's 2006 book, Econospinning; here's an excerpt from chapter 7:

This short chapter is the fourth to reinforce the argument: The relatively low unemployment rate of the past few years has been approximately right, notwithstanding the fantasies of Krugman and others about long-term unemployment, labor force "dropouts," and labor force participation rates already covered in previous chapters.  This chapter deals with the often-mentioned EPR, which turns out to be old wine in new bottles.  The EPR is essentially a mixture of two other Household Survey indicators that have already told their stories: the labor force participation rate and the unemployment rate itself ... The EPR can tell no coherent story until we've distinguished one effect from the other.

Again, here's the link to Epstein's book, which helped me a lot.  (Warning: Not recommended for anyone who has already made an ideological commitment to the Employment-to-Population Ratio.) 

Comments

[Steve Dalton]

Just think how good the result would have been if those decent paying construction jobs had held their own.

We are seeing construction incomes begin to slowly slide, a result of too few jobs available and way too many guys out of work in that industry during this residential downturn. This may help builders bring prices down and get even more aggressive as we all await the market's return.

[Ironman]

The report is also interesting in that it identifies who the biggest job losers are: teenagers. December's much higher rate of unemployment for this one group compared to others might have a lot to do with employers moving to take their youngest employees off the books ahead of the higher minimum wage taking effect in January.

Here are a number of charts related to the minimum wage, showing proportion of full-time vs part time workers, age group, etc.

http://tinyurl.com/22od6k

Here are the most recent unemployment rates by age - check the top lines since these are for the youngest workers:

http://tinyurl.com/228tm4

And finally, here's a tool for measuring the impact of a minimum wage increase on a small business:

http://tinyurl.com/jcdjw

[kmfurr]

marmico, perhaps you should look at your link again. EPR is not at all a "leading indicator" of recession. Just the reverses: recession is a leading indicator of declining EPR.

Do you know why that is, marmaco? It is because a recession causes more job losses than suicides, and so the EPR falls as a result.

[Kevin]

marmico, since you didn't even address what I said in my previous post about which thing is the leading indicator, I take it you are are bowing to my superior ability to read charts from left to right.

[Pete Storm]

Kevin,

If we look for flat spots in the EPR, instead of the obvious downturns, the EPR has correctly predicted 9 of the last 4 recessions, and only missed one. That's a pretty good record!

PS

[Counter Revolutionary]

PS, are you saying the EPR over predicts (9 of the last 4 recessions)?

[Pete Storm]

CR,

Yes. It's an old joke about using the stock market to predict recessions. Updated for the new metric!

PS

[Andy]

Steve Dalton, regarding construction, in Oklahoma, labor costs & housing pricing are going up and completion rates delayed as significant numbers of illegal immigrants have exited the labor force.

In a mirco sense, we're seeing what happens when employers become responsible for not hiring illegals.

On one hand, we see the illegals "deporting" themselves -- so much for the canard that the govt couldn't deport 12 million due to costs. Consequently, the labor rates go up as Americans -- who formerly "wouldn't" do the work -- pick up the slack.

The lesson here is that illegal immigrants drive down labor rates. However I think the correlation between them and unemployment rates can't directly linked.

[Big Daddy Matty]

I generally agree with Steve's analysis (including past analyses of this chart), but I'm having a hard time buying in this time. If the LFPR is the important metric, and it has stayed constant, it would stand to reason that the _average_ wage of newly created jobs would have to match the average wage of those lost in order for overall average wages to stay constant. (In fact, it would have to be slightly higher so as to offset the effects of inflation.) But the average wage of the new jobs this time around is only $17.09, roughly 10% less than the average wage of the lost jobs. It seems to me that breaking the new jobs into two raw number segments and pointing out that a certain _number_ of high-wage jobs were created means very little because it ignores the growth in the raw number of people in the labor force. It's like the utterly ignorant conclusion that Sean Hannity and others drew after the 2004 election that President Bush "had a mandate" because he had received more votes in one election than any presidential candidate in history, never mind that more people voted than any other election in history.

[Bob]

"It's like the utterly ignorant conclusion that Sean Hannity and others drew after the 2004 election that President Bush "had a mandate" because he had received more votes in one election than any presidential candidate in history, never mind that more people voted than any other election in history."

Didn't hear the Hannity conclusion nor do I know the vote participation and candidate counts

If Bush did in fact receive more votes do to a record voter turnout that only legitimizes the statement. Now I'm not sure what constitutes a mandate; perhaps there is some entity somewhere that calculates percentages and bestows such an honor.

Was it the "mandate" part that bothered you or the notion that a record turnout somehow favored the incumbent?

[Big Daddy Matty]

Bob, my point (and in the interests of disclosure/context, I am a Bush supporter) is that saying that a candidate got more _votes_ than any other candidate is meaningless in terms of determining comparative popularity (which is the apparent determinant behind a "mandate"). Sure, 2004 GWB got more votes than 1996 Clinton or 1984 Reagan, but that isn't because Bush was more relatively popular than either of those candidates, but rather because the voting-eligible population was larger in 2004 than in either of the other two years. (Additionally, Clinton ran against two other major candidates rather than one, rendering apples-to-apples comparisons even more difficult.)

The comparison I was trying to draw is that raw numbers mean very little without context, usually in the form of percentages and ratios.

For the record, here are the relevant numbers on the three elections:
1984: 92.7 million votes/174.5 million eligible (53.1% turnout); Reagan: 54.5 million votes; Ratio of Reagan's votes to 2nd place candidate's votes: 58.77%/40.56% (1.45)

1996: 96.3 million votes/196.5 million eligible (49.0% turnout); Clinton: 47.4 million votes; Ratio of Clinton's votes to 2nd place candidate's votes: 49.23%/40.72% (1.21)

2004: 122.3 million votes/217.8 million eligible (56.2% turnout); Bush: 62.0 million votes; Ratio of Bush's votes to 2nd place candidate's votes: 50.73%/48.27% (1.05)

Bush, the candidate who got the most votes, was clearly the least popular relative to his main opponent.