Why do the red and blue lines move so closely together?

by on April 10, 2013 at 4:03 pm in Economics | Permalink


That is from Casey Mulligan.  What is actually the binding constraint in this labor market?  Mulligan himself offers one hypothesis:

A significant part of the recent reductions in the unemployment rate may reflect movements of people between safety net programs rather than any significant change in their job-finding prospects.

yang April 10, 2013 at 4:16 pm

People are now trapped in long term unemployment because they can no longer lie about their job histories to hide large gaps in employment. Employers correctly realize those with large gaps are inferior workers due to physical or mental defects.
Blame employment agencies, credit reporting agencies, and the ease of verifying work history. Even low level retail jobs now require background checks.
The days of padding your resume with fake jobs are over. The wheat is more easily separated from the chaff.
The chaff now goes on disability.

bluto April 10, 2013 at 4:50 pm

Sounds an awful lot like ZMP employees.

mulp April 10, 2013 at 7:11 pm

Don’t economists believe that lack of demand for ZMP workers will result in them disappearing from existence?

Or is the problem the failure of government to provide corporate ownership rights to ZMP workers, allowing them to be scrapped and reallocated to supplying body parts or making soap?

Clearly, these ZMP workers can’t even be turned into slaves because feeding, clothing, and housing them would mean losing money to their owners.

DM April 10, 2013 at 7:43 pm

Mulp: “Don’t economists believe that lack of demand for ZMP workers will result in them disappearing from existence?”

Hm? Lack of demand at the price means they can’t sell their labor, aka being unemployed.

babar April 10, 2013 at 5:21 pm

um, maybe stay in a job until after they’ve lined up a new one?

wilkins April 10, 2013 at 5:33 pm

“Why do the red and blue lines move so closely together ?”


….why can’t our legions of highly-paid, professional economists definitively answer such a relatively simple question ?

(.. instead we get hypotheses & guesses)

David Wright April 10, 2013 at 8:02 pm

What legions of highly-paid, professional economists? Most highly paid people with degrees in economics are working on strategies for high-frequency trading operations. Even among academic economists, only a minority work on macroeconomics. And we don’t let those people do actual experiments, but instead leave them to pore over rather inaccurate time series polluted by all sorts of confounding effects.

Claudia April 10, 2013 at 8:56 pm

Let me introduce you to some economists: http://www.federalreserve.gov/econresdata/theeconomists.htm We are paid comparable to academic economists and well below our Wall Street counterparts. All our work is in support of monetary policy and regulation, so there are many macroeconomists and several microeconomists too. I work in the same group as the ones tasked with examining the labor market. I can assure wilkins that there is nothing “simple” about the interpretation of this chart.

Not my specialty but I am suspicious of where Mulligan ends. There may be some shifting to support programs (govt or family) but why is that? The long term unemployed have been a serious concern in this cycle and their exits now may say a lot more about the crappy recovery (or long-term stagnation) than the generosity of safety net programs.

Harry Truman April 10, 2013 at 9:53 pm

Can you introduce me to a one-armed economist? I’ve been dying to meet one.

dirk April 10, 2013 at 10:23 pm

“I can assure wilkins that there is nothing “simple” about the interpretation of this chart.”

Why? Isn’t the blue line a function of the red line (at least in the short term)? The out-of-labor force population for this age group isn’t much higher than the unemployed population for this age group, so an increase in the out-of-labor force population, ceteris paribus, *should* result in a proportional decline in unemployment. No?

Max April 11, 2013 at 3:41 am

I think it gets even more clear when we plot what age groups leave the work force. Retirement ages come back to the Labour force and entry level youth drop out. This is really worrisome.

Claudia April 11, 2013 at 6:24 am

dirk, the question was *why* the lined move together … those why questions are almost always tough. I haven’t done it but I suspect these two lines don’t move so closely at other times. note there are flows from unemployment to employment too.

mark April 11, 2013 at 4:32 pm

I have a lot of respect for the Fed economists. In fact, I have a family member among them. And I agree the chart is hard to interpret. But Mulligan is right about part of what’s going on. Unemployment comp is partly state funded. Disability is wholly federal funded. So states have an incentive to move unemployed out of that status and into disability. And incentive matter! So some of the people leaving the unemployed ranks are just moving into the “getting disability” ranks, a move within the safety net as Mulligan writes. Google: “disability industrial complex” to read about this further.

mw April 10, 2013 at 5:39 pm

What do we do with all the job creation numbers from ADP and BLS?

kebko April 10, 2013 at 5:45 pm

I believe that the labor force flows data are useful here. Unfortunately, I don’t know how to paste their tables into a comment.

If you look at the “Not in LF to Not in LF” flow, it is a very stable long term trend upward. There is very little indication of any business cycle effect in this flow. This seems very clearly to me to represent long term demographic and cultural factors, and it explains a loss of about 10 million workers since the beginning of the recession.
In the flows between Employed and Not in Labor Force, the flows in both directions declined during the recession, and are now recovering. There is a trend of long term increases, which I suspect is the product of baby boomers and students engaged in non-critical employment.
In the flows between Unemployed and Not in Labor Force, the flows in both directions increased greatly at the height of the recession. Their levels are very high, but they are following the same path as other recent recessions of very slowly decreasing over time as the economy recovers.

I don’t see any sign in this data of newly discouraged workers piling out of the labor force. Comparing the numbers of permanently Not in the Labor Force to the levels at the start of the recession, I estimate that the reduction in the labor force from cyclical issues is, maybe, 1 million people, and this number is declining, not increasing.

ziel April 10, 2013 at 7:13 pm

No, you can’t post a table in the comments – but neither do you know how to post a link, or even a useful title for what you’re looking at? Doesn’t give a lot of confidence in your analytical abiiities.

ziel April 10, 2013 at 7:40 pm

Hers’s a link to FRED Not in Labor Force data.

If you download the data, you can calculate the slope of the change over time. The slope has doubled during this recovery compared to the recovery from the 2001 recession (which in turn was double the rate in the 90′s). Had the growth rate of “Not in LF” been the same this recovery as the last one, there would be 4 million less in that category today.

Of course that number will rise with population, but since the Dot.com bubble burst, there is a notable increase in the rate, and since the Great Recession a further acceleration.

ziel April 10, 2013 at 8:15 pm

but they are following the same path as other recent recessions of very slowly decreasing over time as the economy recovers.

And that’s also not true either – the increase in “Not in LF” for the last 6 months are above the average for the recovery.

kebko April 10, 2013 at 10:07 pm

Well, the data are noisy, but specifically what I’m talking about is the flow from unemployed to Not in LF. It is trending down, very slowly, but more or less typically. The flow from employed to Not in LF is trending up. So, while the total effect, when added to other long term factors, gives a stagnant or shrinking LFP, the movements of the individual components are fairly typical for a recovery period.

bmcburney April 10, 2013 at 8:30 pm


But see “Undesirable trends in the labor force participation rate” posted at this site on April 8. “Baby Boomers” are the only demographic group in the US which has increased it labor force participation rate since 2008.

The overall decline in labor force participation is NOT a demographic event. It is the result of current economic policy. In fact, the net decline in labor force participation would be greater but for the fact that Baby Boomers are fighting with grim determination to get and hold on to jobs (probably anticipating that the social safety net will soon be unavailable for them).

kebko April 10, 2013 at 9:10 pm

Here is the site where you can pick labor force flow data:


The chart for Not in the Labor Force to Not in the Labor Force is the one that has the most pronounced long term trend. I don’t think you can look at that chart and deny that the vast majority of LFP movement is secular. The data here are noisy, but this group generally comprises 92-92.5% of all non-LF population, which shrinks to 91.5-92% during recessions. It shrank to that level this recession, and now has moved back to the 92-92.5% range.

Partly, what I think is happening here is that from 2004-2007, it looks like LFP is stable at around 66%. Part of that is recovery from 2002-3, but demographics and culture should have been moving it down toward 65%, so part of what looks like a collapse in 2009 is just pulling back to what is clearly a long term trend, although it’s hard to know exactly where the trendline should fall, I think it’s hard to deny that a very large part of LFP movements are secular, especially since any time before 2004 or after 2010.

kebko April 10, 2013 at 9:49 pm

And, this isn’t just about baby boomers. Here is the site for employment stats:


Even in the 35-44 year old category, there is a decades long trend among males of around 2% per decade in decreasing LFP, and women in this age group peaked in the nineties, and also now are showing about a 2% per decade decline. So, compared to the dot com bust, subtract 2% from the target LFP rate before you even figure in the baby boomer effect.

ziel April 10, 2013 at 10:24 pm

Why do you think that the category “Not in Labor Force to Not in Labor Force flow” is the gold standard in labor-force participation analysis? Why don’t you just look at the total Not In Labor Force (see my comments above)? But even looking at the data you’re looking at, you’d have to blur your eyes real hard not to see the acceleration in the trend since the last recession. Your numbers (the ones you quote above that you “analyzed”) show ~1% growth in the 90′s until the 2001 recession, 2% from 01-04 (during the notorious “jobless recovery”), back to ~1% in 05-06, then ~3% since 2008.

bmcburney April 11, 2013 at 4:40 pm

I would be willing to grant that the mid-2000s slow down and/or slight decline in labor force participation could be the result of demographics. Times were (relatively) good then and baby boomers were retiring early and often.

Things are different now. The post-2008 decline is not caused by demography or baby boomers. It can’t be. The baby boomers are increasing their labor force participation. The decline in participation is most severe in the 20-25 group.

dirk April 10, 2013 at 6:05 pm

The blue line is a function of the red line. Why the need for a hypothesis?

Edward Lambert April 10, 2013 at 6:47 pm

Workers are being marginalized from the economy. it is the same dynamics seen in Latin American countries. here is the post on what is called “non-inclusive” growth model. Non-inclusive refers to labor not participating in the real GDP gains over time.


maguro April 10, 2013 at 7:42 pm

Why “offset” and not “caused”?

John April 10, 2013 at 7:48 pm

It wasn’t that long ago that the American Ideal was a single wage earner in the household — that was sufficient income to cover the expenses, own a car and house, put the kids through college if they could get in…

While we’re clearly not at any point that could be called optimal perhaps we’re simply moving back in that direction. I don’t see how we can believe that the level or participation in the LF during the bubble decades was not influenced by the bubble. The difficulty is in the transition, some households have two wage earners making very good money and others have none or one with a part-time job.

Andrew' April 11, 2013 at 6:30 am

I thought that was THE reason unemployment rate was a flawed indicator.

loveactuary April 11, 2013 at 7:41 am

Can we even compare these two rates? The denominators seem to be very different, and thus the absolute value of the changes would be very different.

Unemployment-Reduction (say from 8% to 7%) would have a denominator of the Employable Population, which might be something like the [Labor Force Participation + Unemployed Population], say around 70%.

The Out-of-Labor-Force-Increase has a denominator of about: (1 – Labor Force Participation), or maybe around 36%.

So a 1% change in each of these results implies very different absolute results.

This doesn’t mean that they are not related, but may make us think about which matters more.

Comments on this entry are closed.

Previous post:

Next post: