September 18, 2020

Greetings, and a Thought on Net Neutrality

Hello again, FTT readers. You may remember me as a guest blogger here at FTT, writing about anti-circumvention, the print media’s superiority (or lack thereof) to Wikipedia, and a variety of other topics.

I’m happy to report that I’ve moved to Princeton to join the university’s Center for Information Technology Policy as its new associate director. Working with Ed and others here on campus, I’ll be helping bring the Center into its own as a leading interdisciplinary venue for research and conversation about the social and political impact of information technology.

Over the next few months, I’ll be traveling the country to look at how other institutions approach this area, in order to develop a strategic plan for Princeton’s involvement in the field. As a first step toward understanding the world of tech policy, I’ve been doing a lot of reading lately.

One great source is The Creation of the Media by Princeton’s own Paul Starr. It’s carefully argued and highly readable, and I’ve found its content challenging. Conversations in tech policy often seem to stem from the premise that in the interaction between technology and society, the most important causal arrow points from the technologies into the social sphere. “Remix culture”, perhaps the leading example at the moment, is a major cultural shift that is argued to stem from inherent properties of digital media, such as the identity between a copy and an original of a digital work.

But Paul argues that politics usually dominates the effects of technology, not the other way around. For example, although cheap printing technologies helped make the early United States one of the most literate countries of its time, Paul argues that America’s real advantage was its postal system. Congress not only invested heavily in the postal service, but also gave a special discounted rate to printed material, effectively subsidizing publications of all kinds. As a result much more printed material was mailed in America than in, say, British Columbia at the same time.

One fascinating observation from Paul’s book (pages 180-181 in the hardcover edition, for those following along at home) concerns the telegraph. In Britain, the telegraph was nationalized in order to ensure that private network operators didn’t take advantage of the natural monopoly that they enjoyed (“natural” since once there was one set of telegraph wires leading to a place, it became hard to justify building a second set).

In the United States, there was a vociferous debate about whether or not to nationalize the telegraph system, which was controlled by Western Union, a private company:

[W]ithin the United States, Western Union continued to dominate the telegraph industry after its triumph in 1866 but faced two constraints that limited its ability to exploit its market power. First, the postal telegraph movement created a political environment that was, to some extent, a functional substitute for government regulation. Britain’s nationalization of the telegraph was widely discussed in America. Worried that the US government might follow suit, Western Union’s leaders at various times extended service or held rates in check to keep public opposition within manageable levels. (Concern about the postal telegraph movement also led the company to provide members of Congress with free telegraph service — in effect, making the private telegraph a post office for officeholders.) Public opinion was critical in confining Western Union to its core business. In 1866 and again in 1881, the company was on the verge of trying to muscle the Associated Press aside and take over the wire service business itself when it drew back, apparently out of concern that it could lose the battle over nationalization by alienating the most influential newspapers in the country. Western Union did, however, move into the distribution of commercial news and in 1871 acquired majority control of Gold and Stock, a pioneering financial information company that developed the stock ticker.

This situation–a dynamic equilibrium in which a private party polices its own behavior in order to stave off the threat of government intervention–strikes me as closely analogous to the net neutrality debate today. Network operators, although not subject to neutrality requirements, are more reluctant to exercise the options for traffic discrimination that are formally open to them, because they recognize that doing so might lead to regulation.


  1. Interesting. Of course, for such a regime to be workable, there has to be a credible threat of regulation, with fairly powerful forces aligned against the monopolists. The historical examples of the way that WU effectively bought off congress with free service and the newspapers with an independent AP don’t necessarily augur well for today’s situation. Reports seem to suggest that plenty of the established media outlets would be quite happy with a situation where they could pay a small (compared to overall revenues) surcharge to establish a barrier to entry for new competitors.

    The political clout of some of the newer players such as Google is unclear. And it’s not even clear that they will continue to believe that their longterm interests require a neutral medium.

    The only thing that is clear is that (contrary to the monopolist’s favorite argument) without a neutral medium, investment in network resources may well tend to be lower-than-optimal. If you make a big chunk of your profits by selling relief from bandwidth/latency/stability scarcity, then it behooves you to maintain that scarcity.

    (There may be an interesting lesson in the current credit crunch: in the past couple of years, it’s been reported that roughly half the profit of consumer lenders has come from fees and penalties for late payments, leading lenders to go after ever-riskier market segments and constraining the use of credit by more reliable customers. An internet built on projections of extra profits from a few key market segments may be similarly unstable.)

  2. Marvin Sirgu says:

    Paul Starr’s description of the battle between the AP and Western Union is incomplete. As Ithiel de Sola Pool noted in “Technologies of Freedom,” (Belknap Press, 1984), AP threatened to throw its support to competitors of Western Union if WU persisted in its efforts to enter the newsgathering business. This resulted in a agreement where WU agreed to get out of the news business, and AP agreed to use WU for its traffic. In the 1880’s, according to Pool, news reports constituted 1/6 of all telegraph traffic, giving AP a credible economic threat.

    Compare to the noises Google is making about getting into wireless…

  3. I isn’t an equivalent situation. The Internet world doesn’t (yet) have any real monopolists. At the client side, you can buy Internet access from a wide range of ISPs in any Western Democratic nation (obviously in Burma the situation is different). You can choose between technolgies (dial-up, ADSL, wireless, co-axial cable) depending on your budget and needs. You can also choose between grades of service (“business grade” is usually better than “consumer grade”, etc). Some ISPs will also tell you their contention ratios and throttling policies.

    On the server side, there are a huge range of hosting options — fat pipes, thin pipes, virtual machines, etc. There are also a great many independent suppliers.

    There’s no comparison to Western Union which was (for many people at the time) the only choice for high speed communications. You are taste-testing an apple as compared to a cannonball.

    I’ll also argue that a telegraph is pretty much the same no matter who is sending it and what the content is. Internet traffic is not like that… I’ll argue that we do not want “net neutrality”, we do want to be able to buy different grades of service. The only difficult question is how to organise the billing for a many-tiered service. We want VoIP traffic on a different QOS to email.

    On that note, I’ll point out that current billing is broken anyhow. If I buy an ADSL account that charges only on traffic from the net to the modem (not uncommon) and I sit there sending UDP packets (spoof the source IP for weak track covering, radomize the density for a bit more track covering) to some other ADSL user on the same plan, I can run up a large bill for them. Worse, if I find some poor unfortunate on a wireless device I can run them up a REALLY large bill at no cost to myself. This has been a problem for years and we don’t have campaign groups going round trying to fix the situation.