Regulating the Social Network

January 5th, 2012  |  Published in Political Economy  |  4 Comments

[Tom Slee]( linked this [danah boyd]( post which posits that Facebook is a "social utility", and hence is likely to end up being regulated like a power company or a cable provider. Slee hopefully predicts that in 2012, "boyd's view that 'Facebook is a utility; utilities get regulated' will become mainstream".

This would definitely be a step in the right direction, and until recently I would have completely endorsed the sentiment, since I've thought this way about big Internet companies for a long time. Even before Facebook, it always seemed to me that Google's search engine, for example, was an immensely important and valuable social utility, and one that's too important to be left in the hands of a single private sector company. So my suggestion, only a bit joking, would be that we ought to nationalize Google and Facebook.

But I've started to wonder if this is the best way of characterizing what Facebook is, and what it does. Is it a company that delivers a necessary good to its customers, or is it an intermediary that facilitates transactions *between* its customers? Or to put it another way: is Facebook more like a utility, or is it more like a bank?

The most important distinguishing characteristic of utilities is that they tend to be natural monopolies. Because of the huge capital investment involved in laying down things like power or sewer lines, there are huge barriers to entry for competitors and one company tends to dominate. In order to prevent such utilities from extracting huge rents from their monopoly position, the government either has to heavily regulate them, or else take them over and run them directly.

Facebook doesn't have the kind of massive physical infrastructure of traditional utilities, but it has another kind of barrier to entry: since a social network is valuable in proportion to how many people you know who are already on it, Facebook's very size tends to draw in new members and dissuade people from jumping to alternative, smaller networks. As boyd points out in her post, a lot of people feel like they *need* to be on Facebook for social or professional reasons, even if they dislike the company's approach to to its users' data.

But it should already be clear from the above that while Facebook's position resembles that of the traditional utility monopolist in some ways, it isn't quite the same. Facebook's power comes not from controlling a lot of expensive infrastructure, but from the fact that people want to communicate and share with as many of their friends as they can, and they need to be on Facebook to do that. But what if it were possible to connect and share with someone on Facebook *without joining Facebook*?

To explain what I mean by that, I need a different analogy. Instead of a utility, compare Facebook to a bank. Banks operate as part of a larger national monetary system, in which the government regulates commerce, and the Federal Reserve creates money. The banks are repositories for that money, which they are allowed to take in as deposits and send out as loans. In order for the overall financial system to function, there have to be regulations that ensure that customers at different banks can easily deal with each other. For example, the [Uniform Commercial Code]( specifies that if I write a check to a someone and they deposit it at a different bank, my bank is obligated to transfer the funds. So Citigroup isn't allowed to say that I can only write checks to other Citigroup customers.

This is where the analogy with Facebook comes in. The "stuff" that flows through Facebook's network---the shares, connections, messages, etc.---is a sort of currency, less like the electricity that you get from the power company and more like the money that's exchanged between bank customers. An electric company actually *creates* electricity, but Facebook doesn't create its content any more than banks create money. (Yes, I know, banks kind of do [create money](, just bear with my simplification for a moment.) So the paradigm we need is not necessarily a single regulated or nationalized social network, but a tightly regulated social networking *system* that allows people to communicate across multiple social network institutions.

Imagine that there was a standard social networking protocol, which specified the general things you want to do on a social network: connect your account to the accounts of other people, post status updates, send private messages, post photos, send event invitations, and so on. Now suppose that you're on Facebook and I'm on [Diaspora*](, and you want to friend me. In a regulated social networking system, Facebook would be obligated to accept the connection between my Diaspora account and your Facebook account, and to deliver whatever messages or media I chose to share with you. Since the account wouldn't be on Facebook's servers, they wouldn't necessarily know anything about me, I wouldn't be subject to their ever-changing privacy settings, and the only information they would have about me would be whatever I chose to share with my Facebook-using friends.

Of course, being able to monitor the flow of sharing between Facebook and non-Facebook users still gives Facebook a lot of marketable data, and still raises privacy concerns. But this isn't an insurmountable regulatory problem, nor is it that different from what happens in the financial system. Credit card companies collect a lot of personal information about us too, after all, even though people seem less inclined to make a fuss about that than they are to fret about Facebook.

This is all a sort of half-formed thought, and it probably doesn't work for some reason I haven't thought of yet. And to turn this back around again, the way retail banking is currently structured may not even work very well for banking, much less as a model for social networking. As [Ashwin Parameswaran]( has argued, there's a good case for replacing federally-insured private banks with a public bank to take retail deposits, something like the [postal banking system]( So maybe we're back to nationalizing Facebook after all. I certainly wouldn't be sorry to see Mark Zuckerberg expropriated.


  1. Ralph Haygood says:

    January 5th, 2012 at 6:29 pm (#)

    “Imagine that there was a standard social networking protocol…”: the jargon for what you’re talking about is “federation”; cf. another post that appeared today:

    I expect Facebook will publicly disregard federation efforts until one gets big, at least ten million users, at which point they’ll embrace it, both to avoid disgruntling their users who want to interact with friends not on Facebook and because, as you pointed out, they could still get lots of valuable data from such transactions.

  2. Anonymous says:

    January 5th, 2012 at 8:52 pm (#)

    So instead of regulating the company you mandate a set of APIs? I do think that as any area of software grows there is a natural tussle between standardized interfaces and innovation (which demands uniqueness). I can see that mandating a core set of APIs would be an interesting way to enable more sharing – although the privacy implications may be difficult to work through.

  3. Linked and Loaded Friday | Punditocracy says:

    January 6th, 2012 at 11:04 am (#)

    […] a half-formed thought, this Peter Frase piece on the idea of Facebook as public utility is pretty good: This would definitely be a step in the […]

  4. AP² says:

    January 18th, 2012 at 7:49 pm (#)

    A working model already exists – it’s email! It’s a federated system composed of multiple providers which communicate through a standard protocol (SMTP).

    And like email, this proposal suffers from a big problem: spam. Because you see, there are methods that websites use to reduce the impact of automated spamming systems (Captchas, for example) which are simply unavailable when one is required to accept messages from anyone – which is why 80-90% of all email is spam, and a huge amount of computer time and energy is consumed every day to filter it, and it still lets them through and has false positives.

Leave a Response