Kathy G has been repeatedly taking me to task for not linking her post on monopsony. I don't like to make a policy of linking to people who whine that I am not linking to them--indeed, rather the opposite. (Other new bloggers take note). But in this case, I actually tried to link to her; I just forgot to paste in the link, which I do all too often, a careless habit that I am trying to curtail. So, fair enough.

Here are all Kathy G.'s posts on monopsony, at least as far as I am able to determine.

Another note to new bloggers: you've heard that "blogging is a conversation". Unfortunately too true. By the time you've been doing this for six years or so, you will often sort of forget that the conversation you are having now is not with the same people you were talking to four years ago. You will thus find yourself accidentally referring to past arguments as if everyone reading the current post had lived through them with you. Just as you do in a bar at 2 am when everyone's been arguing for a while, you will, without quite realizing it, make references that are only really comprehensible to a few long-time readers.

Case in point: in a post on the academic job market, I made a side reference to the minimum wage.

I hear economists on the left endorse a monopsony model of minimum wage employment that sounds frankly ludicrous to me, and should to anyone who has worked in fast food or retail--how could employers in industries that fragmented, with turnover rates well over 100%, possibly collude? On the other hand, it's a pretty plausible model for academic environments where a squillion graduate students are all chasing three jobs.

When I wrote this I was thinking of a very long-ago debate that raged through blog comments on my very first blog, the one with the eye-searing candy-colored pastel boxes, using an extremely buggy comment system that I don't even think exists any more. Back then, we didn't have all these bells and whistles that you ungrateful kids take for granted--why, I had to carry water for the comments uphill three miles, and then boil it over an open fire . . .

I'm sorry, where was I? Right, the minimum wage.

The funny thing is that I had lunch with Tyler Cowen the other day, and he actually remembered the debate, and caught the reference when I made it the first time--blogging is a long-term conversation with at least a few readers. But Kathy G. can be forgiven for not having had the exquisite good taste to read my blog way back when.

For those who haven't following along at home, this was an argument over whether the minimum wage market is appropriately described by a monopsonistic model of labor employment. Monopsony, if you fell asleep in Micro 101, is monopoly's evil twin brother: it's what happens when you have a single buyer rather than a single seller. As with monopoly, much un-hilarity ensues.

The reason we were arguing about this is a very famous (to econowonks) study done by Card and Krueger on minimum wage employment in Pennsylvania and New Jersey following a minimum wage increase in New Jersey. Their counterintuitive finding was that employment actually rose in New Jersey.

There are several possible explanations for this divergence from the standard finding that demand curves are downwards-sloping.

1. It was a statistical outlier
2. There was some swamping effect in New Jersey, like a local economic boom.
3. The study was not well done
4. Fast-food employment is monopsonistic, at least in Pennsylvania and New Jersey.

Ignore the first two; it would be fiendishly difficult to establish the truth of either proposition. Let's tackle number four first, because this is the heart of the debate that I was having with a number of people four years ago, and with Kathy G. now. Consider a classic model of minimum wage employment in a simple model where all workers are paid the same wage:


When you institute a price floor, the wage goes up, but the number of people employed goes down.

Under monopsony, things look different. Competitive employers have to offer a market wage, so they maximize production where the market wage is equal to the marginal product of a new worker. Monopsony employers, on the other hand, are not price takers; they can set the wage to maximize their profit. This means that they face an upward sloping marginal cost curve for workers, because raising the wage in order to hire another worker also means raising the wages they pay to their existing workers.


As I hope is easy to see from the graph above, which I shamelessly ripped off from Wikipedia, either a market wage or a minimum wage essentially gives you a horizontal marginal cost curve. Those employers simply pay the wage where MPR=S. But if they are monopsonists who control the wage, then they can probably maximize their profit with fewer workers, because they are only competing with staring home and staring at the wall. They maximize profit not by setting wages where MRP=S, but where MRP=Marginal Cost. That means a smaller workforce.

This looks sort of counterintuitive--why would the marginal cost be higher than the wage? But that's because they have to raise the wages of all the other people as well--thus, marginal cost increases faster than the wage. That's why the company can be better off having a smaller workforce at a lower wage than a bigger workforce at a higher wage, even if the marginal revenue produced by the additional worker is higher than their wages.

And that's why, if you set the minimum wage right, giving them a horizontal cost curve, you could increase both employment and wages (though if you set it too high, you'll still see disemployment effects, as you do in the graph above.)

That is one possible explanation. But I find it unconvincing, for several reasons.

First, there's pretty good evidence for explanation three. The original study was a phone study; when another study asked for actual payroll records, they found the same result the standard model would predict: fast food employment dropped in New Jersey. Additionally, as Kevin Murphy has pointed out, the survey started long after employers knew that a minimum wage hike was coming--he compares it to assessing a midnight curfew by comparing the number of teenagers on the street at 11:59 to the number on the street at 12:30.

But second, I have trouble arriving at a reasonable model for monopsony in minimum wage markets. The retail and fast food industries that employ most minimum wage workers have extremely low search costs on both sides of the market. The burger joints are all right there, next to the one you're already working at--if you can get to one fast food outlet, you can get to ten. Ditto retail stores. When they're hiring, they put signs right there in your window where you can see them. On the other side, the firms wait for someone to walk through the door, and put a hat on them. They don't spend six weeks calling your references and checking out your scholarly work.

Kathy G. likes heterogenous preferences between workers, which is to say workers liking different kinds of workplaces. This could also explain it. Except . . . in the fast food market? This seems unlikely. It's not like you're taking a lower wage at Wendy's because they have a great dental plan and they let you use the pool. The labor is unskilled, the wages are undifferentiated, and the benefits are nonexistent. Maybe there are some people out there who love Wendy's food, or Gap clothes, so much that they never want to consume anything else, making the employee discount super valuable. But I cannot believe that this group is sufficiently large to be driving the market. Besides, there's empirical evidence against this; most people do not stay in only one industry, much less only one firm.

There are a couple other small quibbles, each less convincing than the last; by the end of the debate, we were stuck on collusion, the most plausible thesis. Except it's not plausible. Fast food and retail are extremely fragmented, and also extremely competitive on extremely thin margins. Maybe this happens in some small towns, but they'd have to be awful small. My grandmother's town of 10,000 has dozens of minimum wage employers. Large cartels without legally enforceable deals almost always break down; the incentive to cheat is too high. They might not raise wages, but the perks would get nicer and nicer.

But more broadly, minimum wage markets don't look like we would expect a monopsony labor market to look. Retail and fast food employment is pretty sensitive to even pretty small changes in relative wages, which shouldn't be true if firms really have that kind of control over their labor force. Also, there's decent evidence that raising the minimum wage lowers employment among the groups most likely to be making the minimum wage, especially teenagers. Again, if monopsony were persistent in these labor markets, this is not what we would expect to see--at least, not if the government is doing a good job setting the minimum wage near the 'competitive' wage.

This long argument, which culminated in the (I hope not too boring) analysis above, is what I was thinking about when I wrote a two-sentence throwaway line. Kathy G. can certainly be forgiven for not knowing that; I hope that I can be forgiven for acting as if blogging were a conversation, rather than a lecture.

There's another lesson for new bloggers here, though, one that most of us had to learn the (very) hard way: when arguing with people you disagree with, it's better to leave yourself room to back down. Just off the top of my head, before declaring that the other person is an idiot who has no idea what they are talking about, you should be very, very sure that there is absolutely no possibility that they do know what they are talking about. You're coining a fact--Person X doesn't understand Theory Y. If it turns out that your analysis is wrong, my own hard-won experience is that you do more damage to your reputation than you could ever have hoped to do to theirs--I'd say the ratio is 3 or 4 to 1. So unless you're pretty confident that your opponent cannot demonstrate the basic familiarity you have confidently declared they lack, this one is almost never a win.