2006-2016: A “net neutrality” retrospective

Skeptical scrutiny is the means, in both science and religion, by which deep thoughts can be winnowed from deep nonsense. – Carl Sagan

I gave the following speech at the Freedom to Connect conference in Washington DC on April 3rd, 2006. I was on stage following Tim Wu, who coined the term “net neutrality”, and former FCC chairman Reed Hundt. The speech is reproduced here with only corrections to spelling and grammar; it is not verbatim, but an edited copy of my notes. At the end I have written a short epilogue to reflect on what I said, and what has changed.

Within the current funding and construction approach to networks, I believe a network neutrality law is a tactical, practical, strategic and philosophical error.  It takes us further away from Freedom to Connect.

A tactical error

As a tactical proposition, it supposedly is to solve a problem.  So what’s the problem it might be solving?

Well, we certainly have a consumer protection and disclosure issue. Consumers are buying a product that has ‘broadband internet’ on the tin can, but they don’t really know what’s inside.  The terms are often too obscure, hidden in obtuse language in the terms and conditions. That means we need disclosure along the lines that exist (in the US) for credit cards, where the key terms must be presented in a standard format. Furthermore, I’d suggest that there be regulated terms like “full internet access” and “partial internet access” to make it crystal clear.

Then there’s the issue of monopoly and unfair competition.  But there already are laws to deal with this. Sherman Act, RICO, and so on. The problem is process and organisation; getting the FTC, not the FCC, to enforce them.

The unbundling regime in the US has failed. “We need to stop the monopoly rents”, they exclaim. All network neutrality does is take the pain away. But morphine doesn’t cure cancer. Price discrimination in competitive markets is, on average, beneficial to consumers. This is well established economic fact. By eliminating price discrimination, you’re throwing out the baby with the bathwater.

Is there a “moral right” to neutrality and traditional common carriage rules, given that the networks were (supposedly) built using public money? Sorry, no. The elected representatives in the US made a bunch of terrible deals on behalf of the public. Telcos got great deals in exchange for empty, unenforceable promises. The capital has all been transferred into private hands. Gone!

There’s a frequent complaint that “the Net needs us”, and is under attack. But it’s never been healthier. We’ve never had so many people so well connected. It’s an emergent outcome of individual actors expressing their preferences via voluntary exchange. And they, by and large, demand an open connection! It’s not sacred, an object of worship. We can think of better Internetwork architectures.

A practical error

Network neutrality can’t be made to stick. Telcos will evade whatever definition you put up; it’s easier than fighting UNE-P unbundling rules. It’s easy to create a tilted playing field.

There are so many blocking options, ways of setting aside reserved capacity, creating gateways, proxies, and private subnets. How about special peering agreements, unusual terms of service, locked-down edge devices supplied by the telco, different price policies, router queueing algorithms, topologies, hard-to-change defaults and settings, varying network symmetry and private IP address ranges. Good business for consultants like me in helping them evade the rules, but bad for the public.

And when to ‘design-in’ neutrality? Do you really want courts involved in network design? Won’t new entrants need to set aside as much money for lawyers as for network engineers?

A strategic error

Network neutrality makes competition and consumer welfare dependent on law and lobbying, not natural competition. So you’ve chosen the area in which the telcos are strongest on which to fight!

It creates great uncertainty for new entrants. Remember the CLECs? Well, I wouldn’t want to be building infrastructure based on neutrality rules. It inhibits the very price discrimination that makes new access networks more economically feasible, particularly wireless ones.

We actively want un-neutral nets. For example, we’d all benefit if there were local peer-to-peer networks, where local traffic was free, reflecting the low cost of exchanging local traffic. Perhaps we’d benefit if Google could subsidise connectivity, so the person with the 2 Mbit/s connection gets their video at 10 Mbit/s, because the video is subsidised by ads.

A philosophical error

As noted by others, the Internet isn’t a thing, it’s a tangle of agreements to pass packets based on a standard interface. Those agreements are contracts that were freely entered into. You have to interpose yourself into millions of those agreements. This is a bad precedent, whereby the certainty of contract is overturned.

It also surprises me the number of people who violently object to the use of eminent domain to expropriate private property in the Kelo decision are also the same people who would happily see private capital being partially appropriated for quasi-public use.

And where does it stop? Should any networked industry involved in any form of carriage be subject to such rules?

But there’s a deeper reason to be suspicious of neutrality rules that reinforce the status quo. The Internet is not “neutral”; it’s baked in with all sorts of assumptions. It chose to ignore national boundaries, for example, in the structure of IP addresses. It chose to make IP addresses reverse traceable to ISP, but not to individual users, cramping the network’s identity infrastructure by making anonymity the default. Good or bad? Doesn’t matter; the existence of alternatives means we shouldn’t prevent other internetworks emerging by assuming the one we have is the only way of doing it.

I will concede that there’s a free speech issue. But to speak you need access at affordable prices, just like before to write you needed to have access to a printing press. If Yahoo! wanted to do a $10/month bundle deal with SBC where you could only access Yahoo! content, and consumers buy it, where’s the problem? Outlawing it hurts the most price-sensitive (read: poorest) customers.

A better approach

I started off by listing the problems for which network neutrality is a purported solution, and hinted that there was a framing problem. So it falls to me to not just throw stones at network neutrality, but to suggest a better way.

The fundamental problem in networks is “less is more”; stupid transport has more option value than smart transport. The conduit of information goods has unique properties. Ownership of a network is a kind of Tolkein’s ring, where the temptation to capture the value of the bits slowly corrupts and drives mad the owner.

We need models that better align the interests of network owners and users. This involves moving away from the feudal model of telecom and move to a new funding model that captures the positive externalities from a broader pool of beneficiaries. We need a “third way” other than monopoly and crypto-nationalisation via unbundling and price controls.

The focus of telecom will move to innovation in pricing, financing, ownership and purchasing of networks. We already see the initial steps, where the first hop may be over a wi-fi connection. In future, we’ll see user-owned mesh networks. There are plenty of other models, be they local utilities, open public LANs, clawback of property price increases, muni networks, etc.

We have a coordination problem. We can’t easily collectively get together to commission connectivity over which we have local control (as opposed to the feudal overlord). If necessary localities should be able to secede from the telcos by buying back the infrastructure at a reasonable replacement cost.

Most importantly of all, the natural unit of purchase of connectivity isn’t necessarily the individual or household. The marketing and billing costs come to dominate the actual underlying service costs. Look at the US long-distance network, where the players all lost money bribing people to switch by using frequent flyer miles.

Look at what the Wikipedia technology and constitution did for collective information gathering. We need same community mechanisms for networks.

Other progressive actions are “abundance” policies: spectrum liberalisation, “layer 0” rights of way, eliminating franchising rules, opening up phone numbers to new services via ENUM, etc.


An open, free net is an emergent outcome, not an a-priori input to be legislated into existence. We need to capture and accelerate the experiments in how networks are built, financed and sold; and protect those experiments from incumbent wrath until the results are in.

But most critically, don’t fossilize the network in 2006 by adopting network neutrality.

2016: A net neutrality retrospective
I got a lot of things right in 2006: emergent outcomes, the absence of the service definition, the ineffectiveness of regulating for some technically undefined “neutral” outcome, the potential for new architectures, the feeding festival for lawyers that would result.I got some things wrong: my hope had been we would quickly transcend the problem by having empowered communities in control over their own network infrastructure destiny. In some ways we’re still stuck arguing over the prison food, rather than wondering why we’re in ISP jail in the first place. I had underestimated the technical and political barriers to escaping vertical integration. The stupid vs smart network issue is also a false dichotomy. What matters is how the resource trades are exposed and priced.

What I had missed is the struggle between two economic models for quality: rationing and market pricing. I had also failed to grasp the critical importance of stochastics, and the consequences of the science being missing from the policy debate. The nature of the fairness outcome we wanted was something I hadn’t yet considered.

My hope looking forward is for some kind of reconciliation between the polarised viewpoints. We do need a way to constrain arbitrary ISP power, but regulating traffic management (incompetently) isn’t it. We should focus on clear science, clear ethical outcomes, and how these are related. Then we can move away from “deep nonsense” back to towards “deep thoughts”.

The “net neutrality” debate is over. Not because anyone won, but because it’s the wrong debate.

Image © 2006 Phil Wolff via Flickr, CC BY-SA 2.0 license, original link.

For the latest fresh thinking on telecommunications, please sign up for the free Geddes newsletter.