The Great Recession and the Specificity of Labor

Steven Horwitz

Over at Marginal Revolution this morning, Tyler has some thoughts on the labor market, including this one:

If one wishes to consider state intervention, in light of recalculation, one is pointed in the direction of direct employment of the struggling workers; the recalculation argument does not rule out intervention per se.  This advice could potentially be useful and I don't see why even libertarians should consider it necessarily worse than $800 billion of fiscal stimulus.

This comment put me in mind of a conversation Roger Garrison and I had on Saturday morning as we were preparing to leave FEE.  Roger commented (and he may be working on an article on this) that one reason why "make work" programs were actually able to put people to work in the 30s (leaving aside whether it was a net addition to wealth - they actually did put people to work in a literal sense) is that labor was notably more homogeneous in the 1930s than it is today.  Work was predominantly physical and more adults were likely to have had experience doing the sorts of things that the WPA and the CCC were asking of them.  It was easier to just pay people to build and clean things in a world where the differences among types of labor were not all that great.

Fast forward to 2010.  The leaps and bounds of increased specialization and the division of labor that have taken place in the last 75 years has made labor much more heterogeneous and put greater "distance" between the different kinds of work people do.  More specifically, a far lower percentage of labor is physical in the ways that the WPA and CCC demanded.  And perhaps most importantly:  an economy that is orders of magnitude more complex than that of the 1930s will be one in which it's much harder to figure out what sorts of government make-work programs will match the human capital structure of the unemployed labor.  (One has this vision of the government paying people to design websites and databases then just delete them then design and delete a few more...)

So the question for Tyler is this:  which workers and what jobs?  The point of the recalculation argument is precisely that markets themselves need to sort out where the misallocated resources of the boom (including labor) now are most urgently needed.  When that labor is reasonably specific (and it's not just construction workers who are out of work, it's financial and managerial folks too), can governments actually know what sorts of jobs to create that will even remotely match what that labor is capable of doing? 

The argument above is also consistent with the scariest graph out there right now:

That the two most recent recessions had the longest period of labor market recovery is evidence in favor of the argument above.  Labor markets and unemployment are having trouble recovering perhaps because the very specificity of labor makes the recalculation process more protracted, and the variety of ways in which policy makers are complicating the process with stimulus spending and extended unemployment benefits is not helping.  My argument is consistent with the data that IF one is older and unemployed, education seems positively correlated with the length of that unemployment.

One implication is that booms of equivalent size will produce busts of different lengths and depths depending on how complex the economy is and how specific capital and labor resources are.  If we could conduct an experiment in which we generated the same amount of inflation, driving the market rate the same amount below the natural rate, for the same length of time in the US economy of the 1920s and that of the 2000s, whatever would be true of the size of the boom it generated (and I'd argue the boom would be bigger today than back then due to more sophisticated financial intermediation), I'm confident the bust today would be longer and deeper. 

As the boom pulls highly specific factors of production into the artificially stimulated areas, it will also increase the incentives for developing very specific sorts of human capital (one need think only of the high-end financial markets here).  Come the bust, this human capital may well have next best uses that are a far cry from the value it had been producing.  This is an even more Austrian version of Arnold Kling's "Labor is Capital" argument, and helps to explain the zero marginal product labor hypothesis.

Rather than a more complex, wealthier economy having greater dampening effects on the volatility of cycles, it may well be the case that an economy with a greater division of labor and specialization, and therefore more heterogeneous human and physical capital will actually suffer larger booms and longer and deeper busts from the same bout of inflation than would simpler economies.  There's a reason that red line is so deep for so long.

This also means that the recalculation process will both take more time to sort out and that it will be that much harder for governments to try goose aggregate demand with make work programs.  What exactly government is supposed to do to "directly employ" unemployed folks with these highly specific skills?  I really am waiting for someone to argue government should pay them to build websites and databases and then just delete what they've done.

Comments

[Daniel Kuehn]

I'm inclined to agree, but what evidence is there of greater labor specificity today? I only ask because you also hear about the growing importance of broader cognitive skills, non-specific human capital, and reduced job tenure (which of course is associated with non-specific human capital). Jobs may be more specialized today - I could believe that. But I don't know if workers are less homogenous today.

I imagine it largely depends on what industry or region the worker has been in.

This sort of argument is certainly not new, even before this jobless recovery:

http://www.ny.frb.org/research/current_issues/ci9-8/ci9-8.html

You allude to the prospect of some empirical analysis of all this. Do you know of any empirical work on Austrian economics that has been done specifically with these heterogeneity concerns in mind? I found a few that just (ironically) looked at aggregates but nothing specifically on capital heterogeneity.

[Dave Prychitko]

There's an idea for another paper here that you and I might want to consider: labor specificity being ignored in the Post Keynesian call for the government to serve as Employer of Last Resort.

[Daniel Kuehn]

Note in the link that there are quite standard AD/Keynesian explanations as well that also don't in any way conflict with this labor specificity argument, contrary to Cowen's assertion that the nature of the jobless recovery gives reason to discount an AD story (granted, many people pointed this out to Cowen a couple weeks ago when he made the same assertion). This isn't really a test case to arbitrate between these approaches. In all likelihood, both processes are occuring. Which makes the empirical work untangling the two that much more important.

[Jerry O'Driscoll]

Earlier this week, the Wall Street Journal devoted a substantial part of its main section to the employment issue in all its complexity. One striking fact is how many firms are trying to expand and can't find workers with the skills they need (specificity).

One firm needs machinists and decided to train candidates themselves. I think 16 out of 24 candidates completed training. A couple older ones dropped out because they didn't want to invest in learning new skills at their age. (A separate story chronicles how many older workers are dropping out of the labor force and taking early social security.)

There are excess supplies of some labor and excess demands for others. The story fits Steve and Roger's narrative quite well.

A spearate but related story I've heard from employers. It is that many college graduates are unemployable when they graduate. They have no useful skills. One employer remarked it takes a year of training and he doesn't have the time. They seek out the schools that do a good job and only hire form them.

[Daniel Kuehn]

Jerry - there's a big push for career and technical education and apprenticeships to fill precisely this gap, as well as greater reliance on community colleges. We have something of a bimodal skills distribution in this country, and it's those middle technical skills that we're short on.

I think you're more likely to see these issues in specific industries, though - particularly advanced manufacturing and technician type jobs that do require asset-specific training. Generally speaking, the problem we're facing doesn't seem to be on the labor supply side, does it?

[Peter Lewin]

This is very good Steve.
Lachmann argued that the more complex the economy the vulnerable it is to macro-disturbances - like a house with many specialized servants, what happens when the cook gets sick?; versus one with a few jacks-of-all trades, hence redundancy.
But the hypothesis that increasing heterogeneity-complexity has lead to a higher amplitude in business fluctuations depends (as I have argued) not only on the extent of specialization (heterogeneity) of the productive structure; it depends also on the capacity of the economy to adjust to change. And I believe the latter has increased significantly (owing to improved information processing, education and training opportunities, increased mobility of human capital, labor-saving aids to learning and doing, etc.). In fact, it may be more accurate to view the extent of specialization-heterogeneity as endogenous dependent to some large degree of the capacity of the economy to adjust to circumstance. After all, the incentive to acquire specialized human capital depends on the risks involved.
What does the data say? Does it seems that the amplitude of fluctuations has increased. My casual impression is that the opposite is true.
That being said, fluctuations are, in essence, not really macro-phenomena. They are aggregations of micro-events and I do believe that more specialized human-capital experiences greater gains and losses - hence the higher returns.

[Steve Horwitz]

One comment on Peter's point: I wasn't arguing that fluctuations have greater amplitude per se. I was suggesting that the response of an economy to the same inflationary stimulus will be greater, hence the recovery will be longer. I'm not suggesting the inflations themselves have been larger. In other words, for a GIVEN rate of inflation, we should expect larger booms and busts, but observed cyclical variation is due to inflations of different amounts, so just looking at those variations doesn't address my argument, which would require an experiment we could never implement.

[Roger Koppl]

Steve,

According to the NYTimes story, unemployment is concentrated among the least educated workers. Are you saying that they have the more firm and industry specific human capital and are thus most at risk of relatively lengthy bouts of unemployment?

[liberty]

Two reactions:

1. To your question - "What exactly government is supposed to do to "directly employ" unemployed folks with these highly specific skills? "

What if government took the stimulus $ and invited the unemployed into jobs centers where they would TELL government what their previous jobs and experience had been (fill out a form provide a resume), then take unemployment benefits for a few weeks until an appropriate job was available. Government would collect forms and create jobs to order, essentially.

Although this is easiest to do directly when taking housing construction workers and using them for school renovation or road building, it would also be possible to take finance workers and get them jobs by creating a new macroeconomic forecasting arm of the CBO or something.

2. "it may well be the case that an economy with a greater division of labor and specialization, and therefore more heterogeneous human and physical capital will actually suffer larger booms and longer and deeper busts "

This has a distinct flavor... now what is it? ... Oh yeah, Marx.

[JP Koning]

Hi Steve, good article. Would you say that the threat of discoordination (sustained idleness) is higher in a complex economy – assume here a free market characterized by monetary equilibrium - than a simple one (also free and in monetary equilibrium) because the process of recalculation (given some real shock) is more protracted in the former?

[Steve Horwitz]

@Roger: there's two facts at work here. One is that the least educated are most likely to be unemployed. I suspect those are folks in construction and manufacturing type jobs. The former took the hit from housing and the other from the resulting downturn across the board.

The other is that *among the unemployed* older and more educated workers are most likely to be unemployed for the longest.

It may well be that the least educated will have the most difficulty shifting to those areas of the economy where the jobs are still around: services and technology. The housing boom made jobs requiring relatively less formal education look more common and remunerative than was justified by long-run fundamentals. It's a mismatch of human capital with the underlying variables. The manufacturing problem has been ongoing. These folks have fairly general human capital in a world where more specific human capital is now required.

At the other end, there were the well-educated folks who got caught in the undertow, perhaps in the financial sector. They might well have somewhat overly specific HC and are finding it difficult to adjust to where the jobs are now.

As others have suggested, more detail on the empirics of the unemployed would be really helpful here. If I were writing a real article, that's where I'd be going. The blog post was meant to be suggestive.

[Steve Horwitz]

@JP: I think that's what I'm saying, yes. The more complex the economy, the more scope there is for mischief in malinvestment in human and physical capital due to increased specificity (Peter L's house example) and therefore the more protracted is the adjustment after the boom turns to bust.

[Peter Lewin]

@Steve: Given your clarification I understand you to say that the hypothesis is - given the size of the (credit-induced, mal-investment-producing) boom, the larger will be the bust, the more complexly specialized the economy. Worth considering and investigating.

[Steve Horwitz]

Yup, that's it Peter.

[Jerry O'Driscoll]

Some Points:

(1) Labor specificty should not be equated with technical skills. On-the-job experience makes every worker have a higher value in that specific job in that specific firm.

(2) Unskilled and skilled workers may be complementary. How many unskilled or semi-skilled workers lose their job for every loss of a skilled workers' job? (Construction is a good example.) By numbers, the job losses may be concentrated among unskilled labor. But those losses are a function of the loss of capital values and skilled workers.

(3) We need more micro data.

[Current]

The problem I have with this explanation is that employment has jumped in the UK recently. That may be a statistical anomaly certainly.

My view though is that welfare and employment costs are more important. There have been no rises in unemployment benefit in the UK or Ireland. Rates in Ireland have been cut slightly from a high starting point.

[JP Koning]

Thanks Steve.

I may be repeating what Peter Lewin said earlier but if a complex free market economy has developed (ie. one with greater division of labor and therefore more heterogeneous human capital), then somewhere along the way private institutions will have emerged to cope with this complexity, specifically those specializing in the mechanisms for reallocating capital and speeding up recalculation. Here I am thinking head hunters, internet job sites, work training through private schools, private unions providing temporary funds, NGOs specializing in unemployment etc.

Perhaps a complex free market will be better suited to deal with real shocks than a complex mixed economy because the above institutions (those specialized in aiding recalculation) operate in a competitive environment and therefore have incentives to do their job well, whereas in a mixed economy many of those institutions are government-run or -sanctioned monopolies without the same incentives.

[FC]

1. The CCC was in part a skills-training program for young men.

2. Of course many more people were farmers during the GD. Farming requires skills in everything from finance to ditch-digging, so perhaps people really were more willing and able to adjust to changing demand for labor.

3. Training and education take longer and cost more. To move from web designer to biologist, for example, would take ~10 years and hundreds of thousands of dollars. To learn machining at my local community college takes two years. Who can predict conditions then?

[Roger Koppl]

Steve,

You had both those points in your post and yet I wasn't quite glomming onto the one. Thanks for the repetition; I'm afraid I needed it!

Overall, I confess, I don't see a clear and clean story here. My apologies if I"m just being obtuse. It makes sense that older people would take longer to get re-situated. But I don't see why the same would be true of the more educated unless it is because they are richer and can afford a longer search. Age and education are both correlated to wealth, so it might just be that you look longer when you're rich.

In some sense complexity is a prerequisite for a trade cycle. Robinson Crusoe cannot suffer cyclical unemployment. But it kind of goes against the grain to say that growth makes the trade cycle worse. Facts are facts and I'll take it if that what the evidence says. But it violates my intuition to think "the market" would grow worse at self repair the richer a society is. If I were working on the topic I might be more inclined to look at the net equity of the unemployed, transfer payments, number of two-earner households, and so on.

[Barkley Rosser]

I think that there is something to this basic story over time in the US. However, there are problems when we compare internationally. In particular, Germany had as sharp a decline in GDP as the US did, but had barely any decline in employment, thus tending to keep workers in those jobs for which they are well trained by experience. They had very different approaches to dealing with that than we did. Maybe theirs will hurt them in the long run (and they have had higher unemployment rates than the US in recent decades), but they have certainly done far better on the employment front than has the US.

[Jonathan Finegold Catalán]

Some points, and I apologize if these have already been brought up (I will scour the comments once I return from work):

1. Structural unemployment issues regarding the ability of government to employ the unemployed seems a little simplistic to me. Even assuming that most of the unemployed were employed in a type of white-collar labor, this doesn't mean that they lose their ability to clean streets, dig holes, or operate machinery.

Furthermore, even if it was true, this would only provide greater incentive to the government. Indeed, think of all the stimulus it could do by simply training these unemployed! This would, in effect, be no different from the idea of subsidizing education.

In fact, I seriously doubt that the government, so far, has put much effort into directly employing the unemployed (apart from programs such as the census). There is also the question of whether or not government jobs are desirable for the unemployed at this point. An ex-teacher may not be willing to turn into a highway worker, just yet (the situation is not that dire, especially with extensive social safety nets in place).

2. I would say that the labor market is having trouble recovering because of wage disequilibrium and stagnant productivity. Again, I also question to what extent the government has directly employed the unemployed (or tried to).

3. It's true that in a bust in a more specialized labor market it may be more complex to employ the unemployed. Nevertheless, in a more extensive division of labor the level of entrepreneurship is also more complex. The possibilities of investors are much wider. Furthermore, I argue that (even if at a lower wage) in a growing economy employers are willing to train future employees. The problem, again, is a lack of productivity.

4. I think it's worth looking at the labor markets in two cases. First, the great reallocation of labor from agriculture to manufacturing during the so-called Long Depression (1873 to 1894). Second, the post-war years, where according to Higgs around 40% of the population was essentially unemployed until 1946 (these include those in the armed services and those working for the military indirectly, including munitions manufacturers - these skills were not directly transferable to the post-war private sector).

The evidence seems to support the thesis that the problem is a lack of productivity, not any inflexibility in the structure of the market.

[Jeff Oxman]

The structural complexity of employment is important, no doubt.

The other issue the article quoted by Prof. O'Driscoll brought up is high indebtedness, especially connected to housing, and how that prevents people from moving for work.

So, if jobs with your skills are far away geographically, it is a lengthier process to match up work and people since the avenue of moving for work is closed.

[Pietro Poggi-Corradini]

So, if I understand, from the employers point-of-view labor is just capital (organizational capital in one formulation) which suggest a certain 'liquidity', but from the employee's point-of-view specificity of skills makes recalculation hard. In other words, when saying that 'labor is capital' one might have to switch between 'capital' as 'liquid asset' and 'capital' as 'specific machinery or technology not easily reconfigured'.

[Troy Camplin]

You might want to check out this Beveridge Curve, which shows that jobs are in fact being created, but the unemployment rates remains at ~9.5%, suggesting a structural change in the economy. If this level of unemployment remains, what we are in fact looking at is the emergence of the economy (as complex adaptive system) into a new level of complexity. This new level isn't necessarily more complex, however. This kind of structural change is what one would expect in a complex adaptive system if simplifying rules are added to it: the system would simplify until it reached a critical point, where it leaped into a less complex level. This recession may have been precisely that discontiunous leap. I hope that's not the case, but I fear it may be. If it's not this one, I predict it will nevertheless come soon enough.

[Troy Camplin]

Didn't like the way I tried to make the link for the Beveridge Curve:

http://macroblog.typepad.com/macroblog/2010/07/a-curious-unemployment-picture-gets-more-curious.html

[Jerry O'Driscoll]

Barkley makes a good point. Notwithstanding one's view of government's role in these matters, the US approach to jobs is counter-productive. It alters the calculation of an unemployed worker in favor of delaying taking a job.

The German approach is to the alter the firm's calculation in favor of retaining workers. An alternative approach was articulated by Bob Litan years ago, and it tilts an unemployed worker's decision in favor of accepting a job earlier rather than later.

No jobs policy can avoid unintended consequences. On the surface, however, both alternatives make more sense to me than the current US policy.

Again, I leave open the issue of the federal government's proper role.

[supra skytop]

i thought this was funny. Natalie Dee's comic shows what happens between egg friends when one them becomes a chicken. As Serious Eats put it.

[Employment Genius]

Glad I drop by! This is a great site and contains fantastic contents. Thanks for sharing this to us.