« Counter-cyclical asset of the day: Atlas Shrugged | Main | Markets in everything, New Orleans edition »
The unemployment rate as a measure of recessions
I wonder if it is as good a measure of economic severity as it used to be. The greater the heterogeneity of the labor force, the greater the potential for underemployment. Even if the downturn is bad, I am not sure unemployment will stay above ten percent for long. New search and matching technologies, such as found on the internet, might create quicker job pairings, albeit with continuing underemployment. Unemployment is of course important but let us not view this category in purely binary terms.
Posted by Tyler Cowen on February 27, 2009 at 12:10 PM in Economics | Permalink
Comments
Interesting... is there a metric for underemployment? Can we measure the number of people who have recently jumped jobs, and measure the percentage of them for whom the new job is significantly lower paying and/or a different line of work?
Posted by: mk at Feb 27, 2009 12:19:39 PM
Wow... That's a really good point. In the 20's labor mobility was probably much more limited than today - and the mechanism's for filling labor supply/demand were much more limited.
Nominal comparisons between the two periods are probably unfair entirely.
Posted by: Greg at Feb 27, 2009 12:21:09 PM
Collapsing housing market decreases worker mobility due to lack of buyer demand. Home ownership essentially becomes a trap of sorts.
Posted by: dude at Feb 27, 2009 12:25:26 PM
Except that it's the reason recessions are bad.
Posted by: TheophileEscargot at Feb 27, 2009 12:28:53 PM
I've been saying this for years, but being a humble accountant no one has been listening.
Underemployment is especially serious for former manufacturing workers, who are selling nuts and bolts at Wal-Mart and Home Depot, if they are lucky.
Posted by: save_the_rustbelt at Feb 27, 2009 12:37:17 PM
I wonder what we mean by underemployment. In Finland this statistic is calculated but it doesn't refer to people doing jobs they are over-qualified for. Rather, it refers to people not working as much as they would rather.
On the general point I agree with Theophile that recessions really aren't very unpleasant unless you become unemployed.
Posted by: Finnsense at Feb 27, 2009 1:08:46 PM
It's all about U-6
http://www.bls.gov/news.release/empsit.t12.htm
Posted by: Ano at Feb 27, 2009 1:19:53 PM
"Unemployment is of course important but let us not view this category in purely binary terms."
But of course the measure of (un)employment is itself binary. While I have been trying to politely but insistently explain to friends that the likelihood of stimulus funds to put unemployed resources (including labor) to work largely depended on how likely the <10% of unemployed people are being targeted (as opposed to those employed will just be shifted to higher paying alternatives), it is of course also relevant what the >90% employed are actually doing. They may have a job, but a graduate student working 40+ hours/week at Starbucks may not be a "full employment" of their labor.
Posted by: Saxdrop at Feb 27, 2009 1:42:46 PM
Although the risk of outright job loss declined from the early '80s to 2007 (during the "Great Moderation"), the volatility of household consumption actually increased a bit. It appears that employment has gone from the binary model to something more flexible, which implies a degree of underemployment in downturns: http://faculty.chicagobooth.edu/steven.davis/research/Interpreting_the_Great_Moderation,_NBER_version.pdf
Posted by: Virginia Postrel at Feb 27, 2009 2:23:06 PM
The decline in housing prices and the resulting negative equity may restrict the kind of labour mobility that has in the past helped reduce unemployment.
Posted by: Vladimir at Feb 27, 2009 2:23:35 PM
Any measure of underemployment beyond "employed part time for economic reasons" is going to be rather difficult to measure in an objective way. There is enough popular contention over the definition of unemployment and its objective measurement.
For instance, one of the most frequent concerns I hear concerning the measurement of unemployment is the "active search" requirement, which consists of questions looking as to whether a respondent who is not employed and is available for work has conducted some action which could directly result in a job offer over a given period of time. If the respondent isn't actively looking for a job (as explicitly defined), they are considered outside of the labor force, neither employed nor unemployed...
...but what if the respondent has stopped looking because of the perception that there are no jobs to be had? What if the respondent gives this excuse insincerely? I see real difficulties in objectively measuring perceptions and rationalizations of action, instead of the actions themselves (which would be costly enough to verify if we bothered doing so, which we don't). I personally think perceptions explain some regional variations in unemployment across the US, which are compensated by differing levels of labor-force participation.
This problem is only magnified in the case of underemployment, whose definition has not been operationalized at all, as far as I'm concerened. Are there concrete actions which indicate underemployment, or is this category separated from employment by perception alone? Should underemployment include persons who consider themselves underemployed because they overvalue their own skills? Is a bad artist working in a fast food restaurant underemployed, when even in good times this artist could not sell any work? Will the category of "not-underemployed" solely include people who consider themselves either overpaid or products of the Peter Principle?
Survey reticence is a thorny enough issue as it is.
Posted by: Garrett Schmitt at Feb 27, 2009 2:25:51 PM
How 'bout recessions as a measure of unemployment? Just because people have a job doesn't mean they are effectively employed.
We've had two in the last decade with competent Fed(s) and near infinite information availability. Maybe what they are doing very efficiently isn't very effective.
Posted by: Andrew at Feb 27, 2009 3:12:17 PM
Why is labor mobility better now? In the 1900s a lot of labor was basically physical. Now it relies much more on specific skills that aren't necessarily transferable, or as easily transferable.
I could see this going either way.
Certainly the speed of matching should have increased through technology.
Posted by: ah at Feb 27, 2009 3:18:54 PM
to me, this is one of the main reasons using any measure that way. not that i think they're intentionally wrong, but i take the numbers with a grain of salt. sure, they could estimate, somehow, underemployment, but how accurate can it be? you might have all kinds of human capital from working the last 20 years in auto manufacturing, so when your company goes bankrupt you now work somewhere else using a fraction of that capital. are you underemployed? certainly. but what does that tell us? not much. do we artificially create demand so you're no longer underemployed? no. any kind of measure of underemployment would be subject to not only personal preferences (i've got a phd in linguistics, so i consider myself underemployed b/c i'm teaching high school english) but also broader market preferences (how much does the market value the human capital in which you've invested).
i see similar problems with gdp. i think some people, is using gdp as a measure of the health of the economy, forget it is merely a proxy for economic well-being and not a foolproof measure of economic well-being. just because measured gdp goes up (after the stimulus, i'm sure it will) doesn't mean the economy as a whole is any better off.
Posted by: hutch at Feb 27, 2009 3:44:11 PM
Is it really underemployed if you have advanced skills that no one needs?
Posted by: JH at Feb 27, 2009 3:47:43 PM
@JH
"Is it really underemployed if you have advanced skills that no one needs?"
Exactly. E.g., a worker who's heavily invested in operating complex risk assestment models (that have turned out to be worse than useless) and no other marketable skills, and is now flipping burgers and going to community college to learn a real trade. Is this really underemployment or just a sunk cost of malinvestment in highly specific (human) capital that occured during the boom and has now been revealed?
Posted by: Jim at Feb 27, 2009 4:28:23 PM
Tyler argues that changes in the nature of the economy mean that the unemployment rate understates the severity of the recession. I would tend to think it was just the opposite. A 10% unemployment rate (or whatever) means a lot more in a society where each household typically has a single breadwinner than it does in a society where most women work outside the home.
Posted by: Blackadder at Feb 27, 2009 4:46:19 PM
Calculated risk has been making this point regarding under-employment for quite some time.
Posted by: diogenes at Feb 27, 2009 4:56:37 PM
Just thinking out loud here, but I wonder whether the "what's the best measure of recessions?" question depends on what aspect of recessions we are trying to measure. I could imagine scenarios where the fall in GDP is the best measure of the loss in living standards
(for reasons Tyler mentions.) However unemployment might be the best metric of the need for
policies to stimulate demand. Think of a Central American economy devastated by a hurricane where the public is all hard at work clearly up the damage. Real output falls (with the
loss of capital) but pumping money into the economy to boost employment would be pointless. Alternatively, if unexpected deflation raised real wages and increased unemployment, then the unemployment rate might be the best measure of the economic slack that needed to be addressed with demand stimulus. Still other scenarios are also possible.
Posted by: Scott Sumner at Feb 27, 2009 5:14:07 PM
The problem with underemployment is we all know a case when we have intimate familiarity with a
person, we just can't measure it from afar.
In the 1950's thousands of boilermakers, pipefitters and locomotive mechanics lost high paying jobs as
US railroads converted to diesels. Many had invested years in learning their craft and were forced to
accept lesser paying industrial positions. Were they "under-employed" or permanently displaced?
Mike, a current-day degreedoffice worker, loses a specialized office position that formerly
paid $55G/yr and is now working as an accounts receivable clerk make $15/hr until he finds
something else- he gets an offer with pay & benes comparable to his old position, but
its in Houston which is 1500 miles away. He decides to stay, because of the wife's relatively
safe job as a nurse manager. Is he underemployed or geographically displaced?
There's a myriad of variations like this? How do you define and measure underemployed?
Posted by: Superheater at Feb 27, 2009 5:19:17 PM
Unrelated:
Tyler, would comment on this:
http://www.bloomberg.com/apps/news?pid=20601109&sid=ajz1hV_afuSQ&refer=home
Yale's Tobin Guides Obama From Grave as Modern Keynesians Eclipse Friedman
Posted by: bl at Feb 27, 2009 6:53:03 PM
@Jim (4:28)
I hope all those risk assessors out on the job market were able to apply their principles to their personal finances.
But there may still be some room for cyclical underemployment--there may be room for a buggy whip crafter if the market grows a little--but if the market's oversupplied already, then it seems better for everyone if you lay off leatherwork for now and dig those ditches.
Given that it's tough to tell whether your skills are worthless until you're looking for work, it seems even tougher to craft a government policy to address it. And some forceful policy seems to be where the discussions of "underemployment" are headed.
Posted by: JH at Feb 27, 2009 8:14:41 PM
It's impossible to have a legitimate discussion on unemployment and resultant underemployment when we're dealing with government-supplied unemployment figures. Officially, unemployment has hit anywhere between 10%-12% in California and Ohio. In reality, the true figures are closer to 20%.
Posted by: Henry Blankett at Feb 27, 2009 10:07:20 PM
@Henry Blankett: OK, I'll bite. How can you tell that the true figures are closer to 20%?
Are you using a measure like U-6 from the BLS? [link above, repeated here: http://www.bls.gov/news.release/empsit.t12.htm ] Or have you some separate measure?
Posted by: ZBicyclist at Feb 27, 2009 10:37:32 PM
@Save_the_rustbelt, writing "Underemployment is especially serious for former manufacturing workers, who are selling nuts and bolts at Wal-Mart and Home Depot, if they are lucky."
Is that "underemployment", or less pay? Those aren't the same thing. A former journeyman plumber working at Wal-mart is underemployed. A former unionized laborer working at Wal-mart is lowly paid.
Posted by: ZBicyclist at Feb 27, 2009 10:41:57 PM
