Two unemployment puzzles

by on October 14, 2010 at 7:15 am in Economics | Permalink

The first puzzle about unemployment when thought about from within the search-matching framework is that unemployment rates are highest among the least skilled and most homogeneous skills, i.e. among those worker/jobs with the easiest matches.  It's hard to believe that it takes a year to match a construction worker to a job.  

Closely related is the issue of how much uncertainty is holding back employment. The case for uncertainty is that hiring a worker is like exercising an option–once you hire, there are sunk costs of hiring (and potentially firing) that go beyond the wage such as administrative and training costs.

Note that you may not need a lot of uncertainty (i.e. you may not need regime uncertainty) to reduce hiring because you don't have to explain why firms aren't hiring only why they aren't hiring this day.  Even if we assume, for example, that hiring would be profitable, all else equal, it doesn't take much uncertainty to make it worthwhile to delay hiring a little bit, to wait and see.  It's precisely when sales are low and unemployment is high that firms don't mind waiting because uncertainty may resolve in due course and the workers aren't going away.

Ok, that's the positive case for uncertainty but the second puzzle is that uncertainty should matter most when hiring and firing costs are high and once again these costs are lowest for those workers with the greatest unemployment rates.  It's one thing not to hire when you can't fire but when firing is easy what's the risk? Moreover, unemployment has increased more in the United States than in Europe even though hiring and firing costs are higher in Europe.

Search-matching models of unemployment have a lot to add but don't necessarily overthrow the importance of AD shocks, sticky wages and prices and other sources of unemployment.

1 Josh October 14, 2010 at 3:37 am

re the first puzzle, this ( seems germane: construction workers are often nowadays married to women with highly institutionalized, hence immobile, jobs, and have mortgages which are underwater. This means they can't move to find a new job themselves.

2 Slocum October 14, 2010 at 4:00 am

construction workers are often nowadays married to women with highly institutionalized, hence immobile…

Yep, not sticky wages, but sticky lives. It's not that they won't accept less, it's that they can't move. High rates of home ownership and two-career families mean low flexibility. And lower urgency, too — one of the couple still has a job, so the family can scrape by for much longer (possibly indefinitely).

3 Mario Rizzo October 14, 2010 at 4:19 am

Are construction workers really unskilled?

4 Geoge McCandless October 14, 2010 at 4:32 am

One needs a good theory of the firm to explain this. Firms are

organizations that require very specific skills at some levels and more

general skills at others. The specific skills can be very difficult to

replace so they are kept on during downturns (often with changes in

their work assignments) and those with general skills are let go because

they can be easily replaced when the downturn ends. Searching for the

very firm specific skills is very costly, so the downturn will have to

be very bad before you let these workers go.

5 Andrew Edwards October 14, 2010 at 5:14 am

George says it the nice way. The tougher way to say it is that we should always think about agency issues. In particular, the people choosing who to hire and who to fire are unlikely to be from the low-skill group and likely to be from the high-skill group.

They are also likely to experience a feeling of solidarity with or of empathy for other high-skilled people, less likely to experience such feelings with respect to lower-skilled people.

In terms of search models, my thought would be:

If all firms simultaneously lay off low-skill workers for whatever reason (say, an AD shock), search costs could be high – it may be easy to align any one low-skill worker with a single low-skill job, but it is hard to sift through 1000 applicants to get the one you want, and it is extremely challenging to be the 1 among 1000 that gets picked. Even if there are 1000 jobs, you need to "guess right" at which job you will be the one best among the large pool of applicants – no one can apply to 1000 jobs.

6 DanC October 14, 2010 at 6:05 am

Construction work is often regional and seasonal. A construction worker needs strong incentives before they will relocate Vs waiting for improvement locally. Uncertainty can constrain the supply of workers.

Employers don't want to make job promises, or spend much on employee searches, when they are uncertain about future cash flows.

During uncertain times the expected marginal revenue from any project is lower (due to increased competition) and potential employers must lower their marginal costs i.e the workers they do hire must be highly productive (or have very low wages). If it is difficult to spot high productivity workers so you retain your high productivity workers (allowing you to submit low bids) and don't take a risk on unknown workers. The increased competition and lower margins combine to restrict hiring.

Ignoring government interventions that distort hiring people, if you are in a labor intensive field then the relative productivity of workers decides profitability and employment. It can be difficult to discover these traits in some jobs.

Europe may just have less intense competition. Firms don't bid against each other as much. Unions can constrain competition, and keep wages from falling. This reduces total employment in the field and will restrict employment to highly productive employees. In some markets that may give an advantage to union workers who have invested in human capital. You have a highly productive workforce but employment is lower. Which is the type of workforce you may want in tough times.

7 To October 14, 2010 at 6:58 am

Are construction workers really unskilled?

Good question 🙂

8 Finch October 14, 2010 at 8:11 am

Hypothesis: Construction workers are unemployed in Southern California. They're doing fine in North Dakota. It's hard to match the construction worker in Cali to the job in ND. After all, he can't, or doesn't want to, sell his house. This ought to be testable, and of course, might be wrong.

It's not just the skills that matter. Or maybe the relevant skill is "being in North Dakota."

9 Balun S October 14, 2010 at 8:55 am

The job market is broken. Badly broken.

@andrew of course you can apply for 1000 jobs, 5 resumes a day every weekday is only 100 applications a month. You'd apply for 1000 jobs in only 10 months.

But don't forget a lot of skilled people are also looking for work. The job group that I meet with last week had a CFO who's been looking for over a year and a VP-sales who's been looking for 6 months. I'm an engineer who's been looking for 19 months. Even my alternates (tech writing and web design) have been going nowhere. Almost all of the engineers in my network have changed careers going into fields like; nursing, retail, even beekeeping.

Job seeker can't find work. Managers complain that they can't find qualified workers. HR is doing its level best to weed thru the 1000's of resume it gets every day. Companies are looking for "perfect fit" candidates.

And I do mean perfect fit. Look at the job descriptions that say they want 5 years experience with 12 programming languages on 5 different platforms. Skilled workers can pick up a related skill in just a few weeks. But sometimes they also ask for the impossible, like 10 years experience in iPad programming (hint: its only been out for 6 months and the iOS only out for 4 years)

So the job market is broken and the government couldn't fix it even if they wanted to. Even if the Congress flipped completely next month, and cut spending and dialed back regulations and all the other business friendly stuff they yammer about. It might bring jobs back a little but the job market itself is broken and can't be fixed.

10 Bill October 14, 2010 at 10:10 am

Agree with all of the comments above concerning housing as an impediment to moves and with double incomes being a family safety net for the right job.

But, where I disagree is with the comment in the main post that it is easy to lay off a new hire–that is true, but that iooking at it from only the employers perspective in a two sided market.

It is costly for the prospective hire to pick the wrong employer, because if the new employee quits, after having been on unemployment, he/she will not get unemployment if they quit.

Looking at a market from one side ignores the risk of the other part.

11 Floccina October 14, 2010 at 10:50 am

A long training period could cause an employer to hold on to employees that he is losing money on.

12 Morgan S. Warstler October 14, 2010 at 12:22 pm

The matching issue we need to confront is this hypothetical:

A small online textile business will gladly hire seamstresses at $9 per hour as long as they can operate a computerized industrial serger and a fabric laser… because they now sell customized Diaper Bags online for $110 with shipping included… and if they can scale that and drop prices to $75 they'll increase demand by 300%

The problem is all the current textile workers cant use either piece of equipment.

The machinery is expensive, in fact the only machines in local market are owned by the start-up.


1. UI benefits needs to be tied to registration online for training.
2. Businesses need to be able to write off the costs of training… as long as they hire say 25% of trainees.

Suddenly, to get UI the unemployed are required to show up at the shop from 7pm-11pm 4 nights a week, and spend 6 weeks learning to become semi-expert at both machines.

The business to get the right off needs to hire 25% who finish the task.

BUT NOW we also have unemployed who know how to USE THE MACHINES.

13 Jon October 14, 2010 at 2:08 pm

Morgan Warstler: I would amend your proposal to say that only long-term (>6 months) UI benefits should be tied to training. There are plenty of people already up on their skills, but who may also be laid off in a downturn (or even in good times if the company they work for goes under). Such people will generally find a new job in a fairly brief period of time and so do not need to be retrained for anything, but may still need UI to bridge them between jobs.

Otherwise I like your idea.

14 bunkerbrown October 14, 2010 at 4:05 pm

@Brock-what do you do?

15 Mr. E October 14, 2010 at 7:22 pm

North Dakota only has 650K people. It would take only a 20K workers to saturate the market there. Orange county has 3M people.

16 Kevin Postlewaite October 15, 2010 at 12:01 pm

Perhaps in the case where search costs are low, substitutes may be found in lower-wage countries. This doesn't work so well where the employee production cannot easily be geographically, e.g. for construction workers, but it may explain other situations.

17 Mens Abercrombie Cla November 22, 2010 at 7:27 pm

However, this kind of experience brought a lot of thoughts for me and made me awaken that I was only a student who had few abilities and no working experiences.

Comments on this entry are closed.

Previous post:

Next post: