Wednesday, October 30, 2013

Monopsony and expectations about the minimum wage

There are two understandings of monopsony power - each quite different. The first is a raw market power model, comparable to monopoly. This is the one most people are familiar with (if they're familiar with monopsony at all). The second is based on fixed costs associated with labor. In the case of the low wage labor force these are typically hiring and firing costs, although for higher wage labor forces training costs can be substantial. These fixed costs are taken into account in the long run but they are considered sunk in the short run, driving a wedge between marginal revenue product and the wage rate.

This is important in considering what to expect from minimum wages because it's not clear that the two cases will have the same effect (and of course in reality any given employer/employee relationship is going to have a mix of sheer market power and fixed costs in play).

The market power case is fairly simple - a minimum wage is good for wages and good for employment. We're just eating into the monopsonist's surplus, but she has no reason to pull out of the market unless the minimum wage is set too high (which of course is a very real possibility, particularly when the federal government is setting it for a broad array of occupations and jurisdictions).

The turnover cost case is trickier.

At a first approximation, we should not be so sanguine about the minimum wage if most monopsony power comes from turnover costs. In the short run, yes, firms will increase employment in the form of reducing separation rates (assuming most of the costs are on the hiring end) and raises will rise and everything will be great. But in the long run there's a much greater incentive for reducing employment - maybe.

There is also an incentive to reduce turnover or invest in human capital to make up the difference. That's the thing about fixed costs - you don't have to cut jobs. Since they're associated with hiring and firing you can also amortize them over longer periods of time. So in this monopsony model we have the potential for something bad - reducing employment in the long-run as a result of the minimum wage, or something really good - increasing employment and job tenure for low wage workers.

So monopsony makes the minimum wage waters murkier than a lot of people appreciate. And the monopsony argument is more complex than a lot of people appreciate.

This makes testing on multiple outcomes - long term employment effects, turnover rates, on the job training rates, etc. very important.


  1. Are hiring and firing costs even significant for minimum wage workers? Back in the day, hiring was basically 15 minutes of the supervisors time and firing was a phone call telling the person he's fired and when to get his final check.

    I remember one minimum wage hike of 20 cents an hour resulted in labor being reduced from two to one person for the last two hours of the workday at the gas station I worked at. It was usually slow for that time period, but the change increased safety risk. It barely impacted average hours worked, but significantly reduced the hours of the guys who did not have the trust of the manager to do a full close down.

    1. Usually in low-wage labor markets there's also concern about initial productivity losses which, because they are increasing in hiring rates and not hours worked, are considered hiring costs.

      These are all much less significant than in higher wage labor markets, which have more monopsony as a result of turnover costs. But then again minimum wages aren't binding on higher wage labor markets so it doesn't matter.

      You'd expect the raw market power to be somewhat more significant for lower wage labor markets (unless you get into very specialized professions), because low skill workers have less of a market for their labor for mid and high skill workers (who are usually eligible for most low skill jobs in addition to jobs at their skill level).

    2. Thanks for the explanation.

  2. It seems pretty obvious to me that the vast majority of labor markets are markets where the employers (a) form a cartel to set wages and (b) exercise severe market power. Heck, this was going on, on a systematic basis, early enough that Adam Smith wrote about it, saying that businessmen organized in a cartel fashion as soon as they had the chance.

    I'm sure there are some exceptional markets where this isn't the case, but they seem to account for very few jobs, and they're all high-wage markets, anyway. Accordingly, a *general* minimum wage (as opposed to, say, a special minimum wage for professional basketball players or for Congressmen), can clearly be made quite high and will definitively benefit everyone.

    The question of how high "quite high" is, well, that's the question of *how large the profit* being made by the monopsony employers is. If you actually raise the minimum wages in a given business to the point where the employers aren't making a profit, or where they decide to liquidate and go into another line of business, then you will have some rather dramatic effects.

    I would not be surprised if gas stations were exercising relatively little monopsony power and were mainly looking at turnover costs, but that's because they are atypical -- low-profit-margin small businesses. The oil companies who they buy the gasoline from are a different matter, and are far more typical employers.

  3. "The market power case is fairly simple - a minimum wage is good for wages and good for employment. " ... up to some point (obviously). That point being the competitive wage (i.e. wage that would prevail in absence of monopsony). Or (1+e)/e times the monopsony wage, where e is elasticity of labor supply. Replace by N*e, with N=number of hiring firms in the market if it's a oligopsony.

    On the empirical/anecdotal side. For min wage employers the fixed/hiring cost is about a single shift, so min wage*6 hrs or something. Fixed, but not independent of min wage. That's a low ball estimate. That's just my guess of how long it takes a newly hired worker to get the ropes enough so that they can be included in *some* production. It takes quite a longer for them to reach a level of productivity of a tenured worker.

    Nathanael, my expectation and experience is that the case is exactly the opposite that you imagine. I seriously doubt that employers in low wage markets get form cartels or that they have much market power. The goods and the nature of the job, are pretty homogenous, assymetric information isn't much of an issue (you know what you're getting with a min wage job) and - what's mostly being ignored whenever these search-monopsony arguments come up - the search for min wage jobs is low intensity. You walk out of a Burger King and into a McDonald across the street, they ask you if you have a felony, if no, you're hired, if yes, depends on how desperate they are. They don't usually even care that much why you left the Burger King five minutes ago.

    On the other hand, most high wage markets involve very differentiated labor (within a particular job), lots of unobserved ability, imperfect information as to the nature of the job, and (for those reasons) very intensive and costly search process. So if there's market power it's likely it's for high wage labor, though I have no strong priors on which way it goes (it's potentially a case of a double monopsony, or close to it).

    1. "I seriously doubt that employers in low wage markets get form cartels or that they have much market power."

      Adam Smith disagreed with you, and you might want to look at some of the coordination which *has been documented* among (for example) McDonalds, Burger King, Wendys. I'm not making this stuff up; businessmen form cartels whenever they can get away with it. Why *woudn't* they? It's only rational.

  4. Two things - the search for minimum wage jobs is not as low intensity as YouNotSneaky suggests at all times and in all places - there are quite a few Burger Kings that are not hiring at all, many jobs able to demand a high school diploma given the state of the labor market, many towns where there are only limited employment opportunities in a given distance and essentially no public transportation. So far many, attempting to get a minimum wage job is not as easy as some make it sound.

    Second, I think the cost of hiring and firing, even in minimum wage jobs, is more than one shift. In many jobs, it takes several shifts for a new employee to be a net positive (as in, they accomplish more work than the effort they siphon from their co-workers who are training them), and weeks to be able to perform all the tasks of an experienced worker. There is an assumption that minimum wage jobs are simpler than others, but in many cases they involve a very wide array of different tasks because they are not very specialized, and it takes time to learn the proper corporate procedures for each. More importantly, in my experience most turnover happens in the first week or two of employment, and in that period turnover can be substantial (I'd venture 50% in a labor market near full employment, for particularly demanding or unpleasant tasks). So, losing an experienced employee and replacing him or her with a new one costs not only training time and lost experience, but the very real possibility that the time and energy spent recruiting and training a new worker will be wasted when he or she doesn't stay on the job, either by his or her initiative or that of the company.

    1. You point out how the state of labor markets can result in higher skill requirements for minimum wage jobs. I've seen that a lot. In downturns, minimum wage jobs often have higher skilled workers. Assuming layoffs do not occur, increases in wages could be useful to those people. However, it also suggests that low skilled workers may be screened out by minimum wages. I would contend that part of the effect of increasing wages is to change who gets those jobs. Raise wages enough and more of those jobs will go to college students.

  5. Matthew,
    The "search intensity" here is in relative terms. Search intensity for fast food job vs. search intensity for, say, a physician or an engineer.

    I agree with the second paragraph mostly. The one shift was meant as sort of a lower bound.


All anonymous comments will be deleted. Consistent pseudonyms are fine.