Scott Sumner  

Supply and demand: Handle with care

A Lesson in Opportunity Cost... Anthony de Jasay on Pope Franc...

I recently did a post that discussed a "shortage" of skilled labor in Spain, despite 20% unemployment. Companies were having trouble filling positions such as nurses, computer programmers, etc. Several commenters suggested this made no sense; if there was a shortage then firms could simply raise wages until equilibrium was restored. I don't think it's quite that simple.

If the supply and demand model applies to individual labor markets, then my critics are certainly correct. But does it? See if these claims sound reasonable:

1. Companies don't do job interviews, as workers are identical. They simply hire workers as needed.
2. Companies can just as easily hire 1 worker or 1000 workers, all at the exact same wage rate.

That's because the S&D model assumes there is an infinitely elastic supply of identical workers facing any given firm.

Obviously, that's not how labor markets work. Instead, firms are monopsonists who face an upward sloping supply of labor. That's not to say that the S&D model might not be a useful approximation for many markets where it does not apply exactly. I am a Chicago economist, and one definition of Chicago economics is a belief that the S&D model is a useful approximation in most settings. I accept that. But it's not an exact fit, and the Spanish labor market is one example.

Let's consider a simple example of the problem faced by monopsonists. A firm has 100 computer programmers, all making $20/hour. They want to hire an additional programmer, but can't find anyone at that wage rate. The supply of workers facing an individual firm is upward sloping. Thus they must raise the wage to $21 dollars to get that additional worker. So how much would it actually cost to employ one extra worker?

It depends. In the standard monopsony model, the extra worker costs the firm $120/hour. That's because the standard model assumes all workers are paid the same. Everyone else's wage also has to be bumped up by $1/hour. One weakness of this model is that the firm might be able to hire a new worker at $21/hour, but continue to pay the existing workers $20/hour. Or there might be a third possibility. Hire the new worker at $20/hour, plus a $1000 signing bonus, to avoid disrupting the existing wage structure.

If we rule out the signing bonus, here are some reasons why it may be difficult to pay new workers more than existing workers:

1. Other companies hiring computer programmers would face the same dilemma, and hence existing programmers earning $20/hour would "jump ship" and go to competitors paying $21/hour to new hires.

2. Worker morale might suffer if new hires make more than existing workers.

3. There might be union rules that assure equal pay for a given job category.

The bottom line is that when a monopsonist faces an upward sloping supply of labor, the actual marginal cost of hiring one more worker may be far higher than the hourly wage rate. That still doesn't mean there is literally a "shortage", as the term is used in S&D analysis, but there is a situation that feels very much like a skilled worker shortage, from the perspective of the employer.

Employers observe different types of labor markets. They only assign the term 'shortage' to a subset of labor market conditions. The most reasonable response is not to scold them for misusing the term 'shortage', but rather to understand that they are now facing an unusually steeply sloping supply of skilled workers, perhaps because very few are still unemployed. And yes, it is quite possible for there to be relatively few unemployed nurses and computer programmers in an economy with 20% unemployment, if that economy has large numbers of unskilled workers (which seems to be the case in Spain.) And keep in mind that a college education does not mean that you are skilled enough for the actual jobs available in Spain

Comments and Sharing

COMMENTS (22 to date)
Ben H. writes:

Very interesting analysis, thanks.

Partisan writes:

Doesn't (1) undermine the monopsony assumption?

bill writes:

I like this angle. But it can only explain so much. In the short and maybe medium term, some or much (I really don't know) employment can be similar to monopsony.
I liked the two claims in the 3rd paragraph. Of course those aren't true for many jobs. But I think they do approximate reality for truck drivers. And there are many firms hiring in trucking, yet the perception of a shortage has continued for a long time.

Scott Sumner writes:

Thanks Ben.

Partisan, You asked:

"Doesn't (1) undermine the monopsony assumption?"

To some degree, but I was merely assuming some monopsony power, not a complete monopsony.

Bill, I don't know enough about trucking to comment. obviously it can't be perfectly competitive, or else firms could raise wages by one penny and hire workers away from other firms. But I agree that you'd think trucking was fairly competitive. Maybe firms are actually complaining about a leftward shift in the supply curve of truckers, as fewer people want those long hours on the road (compared to the past.)

Philo writes:

The firm might hire the extra worker at $21/hr., giving him/her a different job title, to mitigate problems ##2 and 3. (Not clear that anything can be done about problem #1, which points to a likely gradual rise in the wage rates for workers of this and similar categories. After all, the newly hired worker may well already be employed at another firm at $20/hr.)

Partisan writes:

By (1) I meant "Other companies hiring computer programmers would face the same dilemma..."; I failed to notice there were multiple numbered lists here.

My thought is that if employees "jumping ship" is a concern, then we're not really talking about monopsony anymore, right?

MikeP writes:

The situation for computer programmers would be particularly grim. Their productivity is wildly different and correlates with interest, aptitude, and experience -- all of which favor those already educated and employed.

You cannot pull someone out of some other line of work and expect them to be productive programmers just by dangling more money in front of them. Indeed, you are wildly overpaying your last hired worker and wildly underpaying your actually productive workers if you do so.

Scott Sumner writes:

Philo, There are many possible techniques, but none work perfectly.

Partisan, Yes, that's how I understood your question, and my answer still applies.

MikeP, I agree.

john hare writes:

Skilled workers are in short supply in a number of trades here in Florida. At the same time there is an abundance of non-skilled bodies that i can't even dignify as help. I am quite annoyed with the ones that won't do the things that are required to make good money, and yet want to demand more money because they 'need' it.

James Alexander writes:

US small employers cite skilled labor shortages as a big issue, but are doing little net hiring, failure rates in hiring are high:
"LABOR MARKETS . Fifty-six percent reported hiring or trying to hire (up 3 points), but 48 percent reported few or no qualified applicants for the positions they were trying to fill. Hiring activity increased substantially, but apparently the “failure rate” also rose as more owners found it hard to identify qualified applicants. Thirteen percent of owners cited the difficulty of finding qualified workers as their Single Most Important Business Problem, #4 on the Most Important Problem list. Twenty-seven percent of all owners reported job openings they could not fill in the current period, down 2 points, but historically strong. Fifteen percent reported using temporary workers, up 2 points from April, 5 points from March. Overall, it appears that labor markets are tightening. A seasonally adjusted net 12 percent plan to create new jobs, up 1 point from March."

dajeeps writes:

The situation being overlooked is that if there were truly a "shortage" of needed skills and lots of demand, that would ultimately be remedied by the employers themselves, because one does not have to pay the going rate to a capable worker who needs training. It worked that way back in the 90's in the US for tech. I have many friends in the tech industry who had been in other careers and crossed over into the tech world when many employers were willing to train in order to get the job done without having to pay the going rate for "experts".

I don't really want to call it cherry picking, but two professions that one at least needs to have some background in before one could actually become employed doing it were chosen to be highlighted. Certainly one couldn't just be grabbed off the street and taught to be a nurse. But the original story that was posted wasn't about nurses.

For programmers, when one's been in tech as long as I have, there is the knowledge that comes with experience that having an IT degree only means potential competency. Almost nobody can walk right into an unfamiliar IT environment and start doing without experiencing a crisis rather quickly because potential competency varies with the complexity. It is almost preferable to hire someone who is smart and can learn the environment from the ground up rather than someone who believes they know what they are doing, but in reality hasn't a clue. It's just a matter of content and intensity of the training between the two.

I don't agree that out of the millions of unemployed in Spain, not even a handful them can be trained while workers are being imported from abroad by the truck and boatload. My guess is that arguing that point does nothing to identify the root cause, whether it be either demand or supply side policies or some combination of both, and to remedy it.

Gordon writes:


I found a 2015 news report from Bloomberg that talked about the flight of young skilled workers from Spain. Here's the closing paragraph from the article:

"In other words, Spanish production prices compared with those of its main European partners need to drop. The logical solution for Spain is also the impossible one -- devaluation. This is indeed something Spain did regularly back when it still had the peseta. Today the balancing mechanism is the steady hollowing out of its human capital."

If the article is correct, Spain may be in a Catch 22 situation. Here's the link to the full article:

Benjamin Cole writes:

Well, nice post but how does this explain multi-decade complaints of labor shortages in huge national markets for farm workers, truck drivers and computer programmers?

They are not monopsonies. America has been deunionized. Wages are often private, and there are non-wage filips. Free memberships in a gym, or gift cards, job titles, work schedules etc.

The idea of a leftward shift in supply has merit and that has to be accommodated by higher wages---or importing labor.

I can remember when a George Romney (Mitt's father) or a Ronald Reagan used to brag about how high wages were in America...that is not done anymore.

Times change .

Thomas Sewell writes:

I'm a unique individual with a unique combination of skills and experience.

Therefore, I'm a monopoly supplier of my work. How far does that get me in these models?

It doesn't prevent me from being competed with by substitutes who are "close enough" to fulfill the more general requirements of marginal supply and demand fr various specific jobs.

Scott's right that workers aren't identical, but then, neither are jobs. It's an informational/matching problem on both sides, which delays/adds friction and costs to meeting the right supply with the right demand, but they do match up over time.

The problem* in Spain is that there are specific sets of skills which employees can more valuably employ elsewhere, so as they match up with their preferences, many of them leave for a better deal, while the individual new entrants looking for experience are slow to replace them, leading to a draining equilibrium. Hence why specific high skill industries are impacted by lack of qualified applicants, but there is no shortage of laborers (with a much higher percentage supply available) for hire.

*We don't call this a problem when the exact same situation applies to a much smaller geographic location, like say a single hiring firm. We just say they aren't offering a competitive wage for the skills they want to hire. The same applies between countries where labor is relatively free to move around.

Brandon Berg writes:

Another problem with this is that it assumes a large supply of people who would do these jobs but choose not to because they don't pay enough. When the going rate for a worker doing a particular kind of work is already near the top of the income distribution, this starts to get implausible.

For example, I keep hearing that US employers need to pay software engineers more rather than lobbying for H1-B visas. As a software engineer, I admit a certain appeal to this, but software engineering is already about the best-paying job that doesn't require years of post-graduate training. An even bigger pay gap won't turn career retail workers into programmers, and you'd have to pay a lot more to get doctors to switch. And even if they did, that would just exacerbate the shortage of doctors.

Higher pay doesn't create skilled workers; it just shuffles them around. Which is a good thing insofar as it optimizes the use of skilled labor, but that can only take us so far. Ultimately we need more skilled workers.

Brandon Berg writes:

I can remember when a George Romney (Mitt's father) or a Ronald Reagan used to brag about how high wages were in America...that is not done anymore.

Which is weird, because in the top half of the distribution (not just the top 1% or top 10%, but the whole top half), wages in the US are pretty much the best in the world.

Matthew Waters writes:

"Worker morale might suffer if new hires make more than existing workers."

I think there are many studies showing this is how labor markets and adjust. I don't mean new graduates out of college, but experienced workers hired laterally get higher income.

Of course this is a bit of an open secret at many companies. Leave and come back to make a lot more money at the same job. I have also seen companies be more aggressive in giving raises to existing employees, to prevent the turnover.

I think S&D is enough of a approximation to not give credence to years-long or decades-long shortages. For a very hot industry, it can be like housing in ND during the fracking boom. The market just couldn't clear fast enough. But over many years, shortage is not a proper explanation for a tight labor market.

Benjamin Cole writes:

Here is one to ponder.

Okay, we have many editorials against the minimum wage. In theory, I agree there should be no minimum wage.

But! No one cobbles such an op-ed with the observation that, "Of course, with no minimum wage, the Fed would have to shoot for 0% unemployment."

To call for the Fed to target 5% unemployment and to call for no minimum wage is essentially to call for the Fed to keep tightening in perpetuity. Every time unemployment approached 5% the Fed would tighten. 5% unemployment and no minimum wage is a null set.

I would also like to see calls for eliminating the minimum wage be coupled with calls for eliminating the routine criminalization of push-cart vending, and the ending of retail property zoning.

Funny, all I ever read is the first half of the sorry---bash the minimum wage.

MikeP writes:

But! No one cobbles such an op-ed with the observation that, "Of course, with no minimum wage, the Fed would have to shoot for 0% unemployment."

The unemployment rate is not the percentage of people who are not employed: it's the percentage of people who are looking for work who are not employed.

As the minimum wage drops and finally vanishes, more people can look for work -- and more people can and will be employed.

There is no reason to believe that the natural rate of unemployment with no minimum wage isn't still 4 to 5%, representing both churn and the marginal worker. The margin is simply at a different place.

john hare writes:

Regarding unemployment and minimum wage. I would like to see more focus on workforce participation rate. I would also like to see more work on why there are so many jobs begging for people at the same time that there are millions either prevented from working, or not even trying to be productive.

My experience is that most low skilled people are also low go-git-it types. I would be glad to hire non-skilled that were truly willing to change their status by paying the dues to get there.

Mark V Anderson writes:

[Comment removed. Please consult our comment policies and check your email for explanation.--Econlib Ed.]

Scott Sumner writes:

Ben, If they are not monopsonies, then what is your explanation?

Comments for this entry have been closed
Return to top