The Economics of Monopoly and Net Neutrality

Advocates of net neutrality have an inaccurate model of how monopolists behave. Scenarios are often suggested in which a media conglomerate partners with an ISP and throttles bandwidth from "anything that isn't us". I find these inconsistent with a basic economic analysis of monopoly behavior (not to mention a little paranoid). Note I'm not saying something like "Corporations aren't greedy" or "market competition solves all problems." I'm assuming for the sake of argument that corporations are greedy profit-maximizers, and I'll assume further that the ISP has a monopoly plus some kind of prospective partnership with a media company which could conceivably benefit from selective throttling.

I won't give you a full theoretical treatment of the economics of monopoly. That would require a chapter or so out of a good microeconomics book. (Do pick up Steven Landsburg's Price Theory if you're curious about this. Or David Friedman's book is free online; here is the chapter on monopoly and oligopoly.) But I'll start with a helpful illustration from an old post at TheBigQuestions.com.

Snidely Whiplash owns all the housing in an isolated region, the Yukon in this example. (Presumably Landsburg chose the Yukon to imply that it's costly, though not impossible, for the residents to move away. Thus Snidely's monopoly represents real market power.) All tenants are identical. Snidely eventually figures out how to extract all of the consumer surplus for himself. "Consumer Surplus" is the difference between what you'd be willing to pay and what you are actually charged. Say the tenants are willing to pay $10,000 max to live in the Yukon, but only have to pay $7,500. They reap a surplus of ($10,000 - $7,500 = ) $2,500, because they are paying $2,500 less than what they actually value their housing at.

If Snidely is a monopoly supplier of housing, he'll eventually capture all of that surplus. He can keep charging more until people almost move out, or he can use some kind of clever scheme to get people to reveal how much they value their housing. If he charges $10,001, everyone leaves. If he sets rent at $10,000, he captures all the surplus for himself and everyone stays. Snidely has a good thing going, assuming he can figure out this sweet spot.

Now Snidely monopolizes the grocery stores in the Yukon. Should he try to extract the consumer surplus using his grocery monopoly? The answer is a clear "No." Remember the residents are almost ready to leave, because he's already extracted all the consumer surplus from the housing market. If he tries to extract the consumer surplus from their grocery purchases, that will be the last straw and they will leave. He can get them to stay by compensating them for the surplus they've lost, but he must pay them more than he can extract with his grocery monopoly. (Do read the Landsburg post, which lays this all out very clearly.) If he manages to price-gouge at his grocery store and extract, say, $20 worth of extra payments from his customers, he'll have to turn around and pay them more than $20 to convince them to stay. 

Okay, now let's say instead of housing and groceries in the Yukon, Snidely monopolizes broadband internet. Suppose again he's serving a market of identical residents and extracting all the surplus. (We can relax these assumptions later. It's just a simplification so we can wrap our minds around the example; this argument doesn't hinge on it.) His internet customers value internet at $100/month, and that's what he charges. Any more and they'd all cancel their subscriptions. He's the internet service provider, but he's also the cable and phone company. One day he says, "I'm going to throttle Netflix so that people buy those $4.99 'watch instantly' rentals through their cable box. I'm also going to throttle Skype and Hangouts so that people will buy landline phone service or sign up for Snidecast's internet voice-and-video service." He gives his mustache a good twirling.

But wait! He was extracting the full surplus from his internet subscribers. He was charging $100 for something his customers value at exactly $100. If he throttles commonly used websites, he'll lower  the value of the internet to his subscribers (to, say, $90 or perhaps less) and they'll all quit en masse! If he throttles certain websites that his customers use, he'll have to give them a lower price. If throttling lowers the value to his users to $90, then that's the most he can charge. He'll have to be confident that he can pick up an additional $10/month by directing people to his brand name services, just to break even.

I think it's hard or impossible to make this work out for Snidely. The internet is one big bundled good. Any throttling of his competitors on the internet lowers the value of the overall bundle he's selling. This necessarily lowers what he can charge to his internet subscribers.

Suppose you relax the assumptions. Internet customers aren't identical. They all place different values on their internet service. That's fine. Monopolists can often engage in price discrimination, which means offering lower prices to customers with lower willingness-to-pay in order to attract those customers and offering higher prices to customers with higher willingness-to-pay in order to extract more surplus from them. Businesses can construct clever schemes to get people to reveal their willingness to pay. Cable companies often offer lower "new customer" rates to their subscribers, and they often extend this rate to repeat customers who are willing to call in and whine about their cable bill. Student and senior citizen discounts are other examples. Sometimes this is very subtle. I discuss other examples here. At any rate, relaxing the assumption of identical customers doesn't change anything if we suppose that Snidely can do some kind of price discrimination. Maybe there is one population willing to pay at most $50/month for internet service and another population willing to pay $100/month. The ISP can figure out who is who, through a combination of basic market research and clever "special offer" schemes. Supposing there are gentle gradations in "willingness to pay" rather than two discrete populations: some will pay $50, some $51, some $52... up to, say, $200. We can still suppose that the ISP has some clever method for price discrimination and can extract the consumer surplus for itself.

Now suppose we relax the assumption that the monopolist is already extracting the full surplus. "Okay, now there's some room to play!" Snidely rubs his hands together. Now he has some leeway to gain by throttling his competitors...or he could just simply raise the price for internet subscription. He could concoct some elaborate scheme to direct internet subscribers to his phone and movie services. Or he could just charge a few more dollars for internet service and call it a day. Too many things have to go right for the "throttle competitors" strategy to actually work. If the monopolist throttles Skype and Hangouts, maybe internet users just use the third best option or use their cell phones to make actual phone calls. If they throttle Netflix and Hulu, maybe internet users go to Youtube or watch TV (which doesn't do the cable company a whole lot of good, given the fixed monthly rate). Or maybe they break out their physical DVD movies and box sets, or rediscover Family Video and Redbox. (Hmm...there used to be a service that would mail you DVDs and you'd mail them back when you were done watching them. It had a much better selection than that movie-streaming service that everyone's using now.) Maybe they even decide to keep watching their slightly slower Netflix and slightly less reliable Skype. Throttled internet users may end up spending money to get what they were previously getting out of their internet connection, but there is no guarantee that they'll be spending those dollars on the monopolist's preferred provider. It makes a lot more sense to simply raise the price of an internet connection. "Throttle my competitor's services so they will buy mine" is an incredibly precarious strategy. It depends on too many things going right. For the strategy "Throttle Netflix so customer will rent more Comcast movies" to work, Comcast has to be the customer's second most favored option after Netflix. That may well be true for some of the ISP's customers, but fat chance it's true for most of them. Rather than foreseeing and selectively throttling all of the customer's many alternatives, the ISP would do better to just raise their monthly rates. This could entail a general rate hike or raising rates selectively by getting better at price discrimination. And to get the highest possible price, the ISP will want to maximize the value of an internet connection to its customers. That probably does not entail throttling popular websites.

The basic problem here is monopoly per se, not the implausible scenarios illustrated by neutrality advocates. If ISPs are actually monopolists, then they can extract most of the surplus from their subscribers. Granted that's a problem if you're an internet user. However, they're more likely to exert that power in the form of higher monthly rates than sinister self-serving throttling of competitors.

I want to reiterate that I don't think ISPs are monopolists in any meaningful sense. This is important because ISPs can't do all of those nasty non-neutral things if they aren't a monopoly. You can get a pretty good data plan through your smartphone. You can get the non-internet versions of things: newspapers, DVDs and books (for free) from your local library, Walmart instead of Amazon, talking to people on the phone instead of over Skype. It's slightly less convenient, but not exactly devastating. Alleged monopolies are always too narrowly defined. A "monopoly" on broadband internet competes with other kinds of physical media; if one gets too expensive people will substitute in others.

_____________________________________________

Some unconnected thoughts on this topic below.

I think David Friedman is trying to make this same point here in a Slate Star Codex thread. Another commenter totally misses the point. Excerpts:

DF: They could give their movie firm an advantage with their ISP customers by charging high prices to their competitors, but why should they?

Anonymous commenter: Because they could easily dominate the online movie market and make way more money that way?

Ugh. Sometimes you have those "why even bother" moments. The conversation is instructive, anyway.

I have quite a lot more to say about net neutrality here, with a lot of overlap in this post you're reading now.

Recall there are historical examples of people offering an exclusive, non-neutral internet connection and the market breaking them down. ISPs eventually learn that customers want a neutral internet, and it's easier to maximize the value of the internet subscription and just charge some fixed rate for it.

The Landburg post really is brilliant. If it's unclear on an initial reading, read the comments. Someone was probably as confused as you are and asked your question. This comment by Landsburg hit the mark, at least for me:

Yes, someone with a double (vertical) monopoly has no more power to abuse the market than someone with one.

The easiest way to see this is to consider a monopolist who owns two toll booths, one at each end of a bridge. You’ve got to go through both booths to get to the other side.

This monopolist is no richer than a monopolist who owns just one toll booth on an identical bridge. Either of them will charge you (in total) what you’re willing to pay to get to the other side.

Riffing on the monopoly bridge-owner. Imagine someone who monopolizes the bridges to some island, which contains an awesome shopping center. Will they try to get people to buy Bridge Corp brand clothing and books and groceries once they're in the shopping center? Will they strategically sabotage other brands, staining their clothing, running magnets over their computers, relegating them to a dark corner of the shopping mall that never gets cleaned or checked by security? Probably not. They'll probably say, "We're good at building bridges. That's our comparative advantage. Let's maximize our revenues from the toll booth. Let Starbucks, the Gap, and the Apple store specialize in coffee, clothing, and computers." Everything they do to hamper their competitors (or the competitors of a company they hold a sinister partnership with) decreases the overall value of the shopping center, and thus reduces the toll they can collect at the bridge. ISPs are essentially monopolists on the bridge to the internet (at least in the world of net neutrality).  Lots of stores carry their own brands for sale; they usually don't make the other brands hard to find because doing so would lower the value of the overall shopping experience. They could make more revenue by selling more of the store brand product if all else were equal, but all else is not equal.

It's worth reiterating that some "non-neutral" behaviors are good for customers. Identifying and throttling the big bandwidth hogs on my network is good for me, even if it means I'm occasionally the one who gets throttled. Offering tiers of service and internet fast lanes is also good for consumers. It gives the ISPs incentive to build new infrastructure. They can reap a temporary premium for their "fast lanes" until the fast lanes become the normal-speed lanes, at which point they build another fast lane. Despite all the griping about potential throttling, internet connection speeds are increasing exponentially. Anyway, it's important to distinguish between "throttling bandwidth hogs because they're slowing down the network" and "throttling Netflix because I have a deal with Hulu."