Tim Wu professor law at Columbia University
Is this a fad?
In answer to that I want to talk about some of the issues of how the underlying issue of discrimination has existed in telecommunications history.
Third I want to comment on legislative proposals.
The label network neutrality is not that important. In fact the legislation, which we support, is not that important. What is not going away is the underlying ideas.
It’s a label to capture a series of intuitions and views that have been away ever since anyone began debating between Ethernet and ATM. It’s the choice between centralization and decentralization. It is the choice between planning and unplanning. It is a choice between innovation coming from the center or the edges.
Time and again networks have proven that they innovate better and serve human needs better when they’re perceived as being networks of general purpose. That’s the central gospel of this conference. What is a network for?
The idea that the value of a network is what it makes possible means you want the most generally useful network.
So the big question is discrimination rules. A rule that shapes or effects how general purpose the network is determines its social value.
Let me give you an example of how this has evolved in history.
The telegraph was the nation’s first national network, the great revolution of its time. It was also the first one to be monopolized, the first one to have a problem of discrimination.
In 1866, Western Union put together its monopoly, acquiring other companies. One of their most important customers was the Associated Press. You could beat your competitors by 24 hours. It was an issue of latency. Western Union offered AP an exclusive. It offered to give it preferable rates provided they never used any other telegraph service. AP and Western Union made this deal, and Western Union effectively never allowed another news organization to be sent on its wires. It never blocked services. It just charged such low rates to AP that no one else could compete. Western Union also put into this agreement that the newspapers would never criticize Western Union nor support its competitors. They used the power of the press.
The consequences of this were political. The AP was close to the Republican Party of its time. Their strategy to support the party was not to offer coverage to Democrats, who had no access to the wires, for about 20 years. The Democrats had to rely on other forms of communication. The high point of this was the 1876 election in which the popular vote was won by the Democrat, Samuel Tilden. Republicans were about to concede, but the Democrats made it known through the wire that they were concerned about the results, and this was passed on via AP to the Republicans. The Republicans contested elections in South Carolina, Louisiana and Florida. A commission was appointed, all on the basis of the wire service story. Throughout the process the AP leaked secret telegraphs from the Democrats to the Republicans. And eventually the commission chose the Republicans.
What is now known is the role a network monopoly played.
The names change. The issues change. But it’s all the same question, namely what are the duties of these carriers of information.
Common carriage originally had nothing to do with market power. It had to do with the public duty infused in private actors. If a seaport gives preference to one ship over another it’s bad for the economy. If the roads work bettter for Fords than Toyota that’s bad for the economy. That’s a powerful intuition and it’s right.
People who put themselves in a position of common carriage. are in a position to distort other parts of the economy. So it has been the public policy to regulate input makers for the national interest. There are some businesses in a position to be inputs into other business. When you’re dealing with input businesses the more they are allowed to distort the market the worse it is for the antional economy.
You might think companies in a position of carriage will naturally arrive at the conclusion that the only thing that makes sense is to carry everyone neutrally. But there are many examples of companies which don’t act that way.
We saw it in the 1950s and 1960s. It was probably in the interest of hotels to serve black people as well as whites. But for various reasons they didn’t. Supreme Court decisions on this issue often relate to the commerce clause – Congress’ power to write laws banning private discrimination.
The same things are happening in network neutrality.
At certain times, through sharp rules that break off transport from other markets, government has managed to give birth to entire industries and markets.
The clearest example is the network attachment rules. The Part 68 rules. These rules say users have a right to use the telephone as they like, the carriers are forbidden from banning devices. This led to the fax machine, the modem, the answering machine, the possibility of mass personal use of the Internet, all because you had an absolute non-discrimination rule.
There is a complete break between transport and application. This is the wisdom of the 19th century. It is the wisdom that neutral inputs serve the economy best.
The greatest value of the Internet isn’t any site, or video on demand. It’s probably e-mail, which is the hardest thing to make money on. The most valuable uses of a network infrastructure are the hardest to commodify. And trying to commodify one is going to be the worst thing you can do to the economy.
There are three issues here:
- Blocking
- Transparency
- Tiering
It is not without hesitancy I’ve supported the Wyden legislation. There are times when these laws are perverted. But the right, minimalistic rules, modeled on Part 68, can make a big difference.
There is an internal debate on how networks should be designed, what companies should do. There are engineers who look at these value-added business models and see a disaster. There are finance people who look at these and say these are a disaster. I don’t think it’s worth giving up on these companies. I think there are elements within them that are interested in closed networks and elements interested in open networks.
Some companies see their highest path to growth as selling a general purpose tool.
The first issue is blocking.
The industry and the FCC have conceded that blocking is bad. Only a few years ago cable companies were sending notices to people saying their VPN service could not be used – they were blocking individual services. The user agreements said the use of WiFi might be a crime. AT&T’s original agreements said WiFi could be prosecuted. There has been a major concession here, and Michael Powell deserves credit for it.
The second issue is transparency.
For the network to be useful as an input people need to know what that input is. Application designers need to know what they’re dealing with. The market only works when it has information. None of the reasons the carriers give for not giving information make any sense at all. There’s this instinctive holding of information, a collective action problem, which legislation can solve.
The last and biggest issue is tiering.
No one thinks it’s a problem of charging for different amounts of bandwidth. It is discrimination, but it’s a good form of it. The hard question is what happens when carriers are offering one auction site better access, one movie delivery service better access than another. The best answer is user tiering only. No application side tiering because of its inherent temptation to distort.
I was looking at AT&T’s design for Lightspeed. It’s a 25 Mbps service, 19 Mbps of which are reserved for TV delivery whether or not your TV is on. Inside the remaining 6 MBps there are preferential service agreements. This leaves neutral users to whatever is left over. The idea that 76% of the future is TV service is …people dig their own graves.
Any form of tiering runs the risk of overdistortion of applications.
If there is to be tiering the best approach is to demand more neutral bandwidth. People in these companies think that consumers want them. Consumers don’t want them. They want to get to other things. They don’t want your stuff.
The only basis to offer this is on an MFN basis. If you offer video speed basis they must be offered equally – everyone gets to sign up for tiering or no one does. that’s the compromise position. The best position is a non-tiered one.
