There is one kind of market inefficiency that is quite common but is dealt with in very different ways. In general, a “monopoly” is bad because if only one actor can be in a market, it can abuse its position, raising prices or offering a bad product or not being efficient or not innovating. Monopolies can be created by law, such as a state monopoly or a legal monopoly. Monopolies can also be created by large players using their clout and capital to push out or buy up smaller competitors. However, I want to talk about the kind of monopoly that sortof just appears due to the nature of the thing that they do. Perhaps the most tangible example is facebook.
- Your friends join facebook, maybe because it’s new and decently good.
- You join facebook because your friends are on facebook.
- Your friends don’t leave facebook because now you are on facebook.
- You all stay on facebook.
- Facebook gets worse
- You still don’t leave facebook, because everyone is on facebook.
This is a “network effect” because facebook has the advantage that the whole network of friends is on facebook, and for any one person to leave would kindof suck for that person. It requires a significant change in quality, perhaps even a generational transition, for a meaningful competitor to facebook to become popular.
However, “network effect” is a great name for this because, well, it’s the case for basically anything that could be called a “network”. Here’s more examples:
- The power grid: A network of wires and substations, it would be awfully inefficient to have two of them
- A railway network: Cmon, “network” is in the name!
- The internet: A series of tubes
- An auction site: for homes, for stuff, is a network because you are connecting many sellers with many buyers
- A stock exchange: same as 4
- Payment systems: you have a card, the shop has a card reader, they better be compatible
- Operating systems: If one app needs to talk to another app, the OS is in charge of how that can happen.
- A road network: Once again, it’s in the name
- A dating app: The more people are on it, the more chances you might get to find the one (if the app developers cared about that, and not just showing you exactly what would keep you on the app for longest and spending money)
In general, there are two ways to deal with these natural monopolies and network effects.
- Codify them as state monopolies, or regulate them (sometimes better than nothing!)
- Slice out just the network as best you can, have that be state run or regulated or standardised, and let market actors mess around on top. Sometimes, the market just does 2 because it makes sense, but surprisingly rarely.
I obviously prefer 2, and we already do 2 in a lot of good cases. Here are some examples:
- Water: keep the whole thing public. Scottish Water provides households with an integrated public service. Businesses, however, can choose competing retailers that buy their underlying water services from Scottish Water. Even within one system, we draw the line differently. Sources: public provision, retail competition.
- Electricity: share the wires, choose the supplier. In Sweden, you choose who sells you electricity, but your local grid remains a regulated monopoly. Competition solves part of the problem; someone still has to keep the network operator accountable. Source.
- Broadband: public fibre, competing services. Stockholm’s Stokab leases fibre to operators that install equipment and sell services over it. A new provider can rent infrastructure instead of digging up the city again. Source.
- Railways: separation needs coordination. Different operators can use shared tracks, but timetables and maintenance still need to fit together. Britain’s 2018 timetable disaster exposed gaps in responsibility for the whole system. Slicing things apart does not automatically make them work together. Source.
- Payments: shared infrastructure can still have a powerful owner. Brazil’s Pix combines public payment infrastructure with competing providers. Requiring major banks to participate helped overcome the empty-network problem. Visa and Mastercard also connect competing businesses, yet a UK review found ineffective competition and rising network fees. Interoperability alone isn’t enough. Pix study, card-fee review.
What can we learn?
The shared part needs fair access, sustainable funding, and someone responsible when things go wrong. It also needs participants: an excellent standard used by three empty services won’t threaten an incumbent.
And separation should produce meaningful choices. Five companies sending differently coloured bills for essentially the same service may not add much.
So where could we take this further?
- Social media: separate relationships and communication from the application and feed. Let me change providers while continuing to interact with my friends. Mastodon already demonstrates parts of this, although moving an account still has limitations. Source.
- Dating: let consenting users discover and contact people across services, while apps compete on matching and experience. Visibility, blocking, and responsibility for abuse would need to work across providers too.
- Auctions and marketplaces: share authorised listings across competing storefronts. Keep one authoritative auction or sale record so an item cannot acquire two winners. Let sellers carry verifiable transaction histories between providers.
What these have in common is that they all have some data that really is the users data. It would make sense that any service could request any data, but any data they get from the user they have to put into a pool that any other service can access, some portion that is always shared, and some only when the user consents to sharing it (e.g. phone number versus house listing. The former you share when you want to, the latter is why you are here, to sell your house).
Then the market test is fairly simple: can someone build a better service, and can I choose it without persuading everyone else to come with me?