Monopolies don't make enough money
Monopolies are inefficient because they don't extract enough of the surplus they generate for society
Building on last week's theme, thanks Josh, I want us to think about counterfactuals today—that dreaded, "compared to what?" we economists harp on.
Instead of focusing on the first model taught in undergrad, I want to jump all the way to the second model: monopoly. 🤯
Things are moving quickly here at Economic Forces University (trademark pending).
Everyone who has taken economics 101 knows that a monopoly is inefficient. Well, a monopoly can be inefficient. Declaring something inefficient implies an implicit counterfactual: there is another feasible outcome that is somehow “more efficient”, in whatever form of efficiency you mean: Pareto, Kaldor-Hicks, whatever.
After finding the possible inefficiency, we then dig deeper and see what is driving the problem. Only then, maybe, we can find a mechanism for improving outcomes.
Consider externalities. We first show students that the outcome is inefficient; too much is consumed in the case of a negative externality. But why? The price mechanism seemed to work before. What is different with externalities?
That is when we bring up the distinction between private and social benefits. In the case of a positive externality, the social benefits outweigh the private benefits. If I manufacture electric cars (in a world without subsidies), I do not get all of the benefits of those cars. There is a social benefit (through reduced emissions) that I do not receive.
As a result, I am underpaid and so I under produce electric cars, again compared to some counterfactual benchmark. Then the way to improve the outcome becomes clearer: how can we properly reward people for making electric cars? (Lots of policy implementation issues are swept under the rug here, but the exercise is still valuable.)
Wait, wasn't this supposed to be a newsletter about monopoly?
Yes. It is. Let's go back to the monopoly problem.
What's the analog in the case of the monopoly? We first see that the monopoly quantity is below the efficient quantity, just as in the case of the electric car. The problem with a monopoly is that not enough of the good gets consumed and produced.
The logic is just as before, and this is where most people make the mistake. If we follow the electric car example, the monopoly seller is unable to receive all of the benefits that it generates.
A famous study (unpublished, so take that for what it’s worth) by William Nordhaus estimated that only around 2% of the gains from innovation go to the innovators. If that is even remotely close to the true number, innovators, aka monopolists, do not make enough money.
The monopolist is paid too little! If the monopolist was paid too much, the monopolist would produce too much. Marginal cost curves slope upward.
Raise your hand if you thought the problem with monopoly was that the price was too damn high and the monopolist is paid too much. Don't be shy. Get it up there. 🙋🙋🙋 I did too for a while. It's a forgivable sin; unlike claiming demand curves slope up.
This is not some clever trick; it is right there in the textbooks, if people squint properly. Efficiency improves when the monopolist makes more money.
Consider the counterfactual where the monopolist now price discriminates and charges different prices to different consumers. In that case, the monopolists would make more money, consumers could be better off, and overall efficiency improves. The welfare-maximizing scenario is perfect price discrimination.
Of course, perfect price discrimination does not make consumers better off in the simple model, if that’s what you care about. It is an extreme case. But it is simple to show how price discrimination does make consumers better off. For example, I have a paper that shows that with two firms perfect price discrimination is consumer optimal. (n.b. If you ever referee this paper, it’s wonderful. Accept without revisions 😂)
In terms of econ 101 models, we can show how consumers benefit as well. First, consider the simple example, shown in the graph below, of a coupon that gives a discount to those who are willing to spend the time cutting coupons.
Coupons are a classic way to price discriminate. Coupons improve efficiency. Not only is the monopolist always better off with price discrimination, so is the marginal consumer. To the extent the marginal consumer tends to be poorer than the inframarginal, we should not ignore this benefit. The one-percenters aren't cutting coupons. They lead to higher profit, higher quantity sold, and lower prices for those coupon cutters!
Three cheers for price discrimination!
Even more important, price discrimination promotes innovation! To see this suppose there is a fixed cost to develop a new product. Suppose the cost is bigger than the uniform price profit (the left rectangle above) but smaller than the profit with price discrimination (the two rectangles combined).
Without price discrimination, the profit from selling the good is not large enough to justify investing in the new product, so innovation does not occur. But price discrimination makes it profitable to innovate. Price discrimination is a causal factor for dynamic innovation in this model, and I would argue in the world.
Take each new generation of the iPhone. Apple uses price discrimination through time. At first, the price is high and sold to people who are really into the latest tech. Over time, the price drops. By price discriminating across time, more people can afford an iPhone when they wouldn’t be able to in a world where either the invention did not happen or if it did the price would be fixed at a single higher price.
Think of how amazing this process is. In just a few years, a technology that did not exist before is now so cheap that it gets thrown in with a cell phone contract. If you look across technological innovations, you see a similar pattern of dropping prices (at least according to one theorist’s empirical expertise).
But is this efficiency improvement technologically feasible? I don’t know. I mean, we are comparing a world where the firm must set some price to one where they can set different prices. How does that data come about? Does it just fall from the sky? If collecting data for price discrimination is costly, then perfect price discrimination is not efficient.
To see the connection between price discrimination and innovation, consider a simple dynamic model. Each additional innovation becomes more and more costly. Over time, since firms need to make a profit to invest in innovation, all else equal, innovation will slow down as firms need to save profits over multiple periods to “buy” the innovation.
Now add another technological choice: the firm can invest in better price-discriminating technology. For example, they can invest in collecting data on consumers. Again, this has increasing costs in quality. Then the firm will choose to invest in these data technologies. That’s the dynamically efficient outcome.
It’s not just idle speculation by this weird theorist. Real economists find evidence for this story using eBay data. If that’s the world, the dynamically efficient outcome will feature a decrease in innovation and an increase in data collection and price discrimination, and maybe even an increase in measured markups. Do we see those things in the United States? 🤔
Let's finish by returning to the 101 model—before I get too out of my depth talking about data. The missing punchline from the price discrimination discussion is that welfare improves from price discrimination precisely because the monopolist gets paid more.
The form of price discrimination that always leads to efficiency is first-degree price discrimination, where the monopolist gets all of the surplus. Flipping these results around, the inefficiency in the original case is because the monopolist did not get paid enough.
Or, and this time repeat after me, monopolies don't make enough money.
Thanks for reading 👍
Have I slipped in my logic? Let me know below 🍻
Tap the ❤️ and share with your friends (or enemies) 👇
also, further explanation of the nordhaustt 2% value capture would be fascinating...have read that stat multiple times, but never seen great explanation. Although is has massive implications, any place that you or others explore that would be fascinating!