Developing an analytical framework to talk about policies around net neutrality and why it is so hard.
By Soren Dayton
It is very easy to see telecommunications policy in local terms, as it is an industry that is literally grounded in a particular place. Indeed, speaking as someone who has been a consultant to US telecommunications companies, a remarkable amount of time and energy is spent, at least in the US, focused on issues like siting towers, getting access to particular places to lay cable, etc.
Given the salience of the net neutrality debate in India, it might be helpful to step back a little and try to consider it both in comparative and more abstract terms. One of the features of the net neutrality debate is that while the top-level messages have stayed the same at various times and places — “every bit is equal” versus “don’t over-regulate and restrict business models” — the underlying technical and policy questions have shifted with both jurisdiction and contemporary technology. In fact, the constancy of those messages in the various contexts, when you consider what the underlying policy differences are, should give you pause about them.
Rather than try to define net neutrality, I want to provide an analytical framework to allow us to talk about policies around net neutrality and why it is so hard. This should help explain why, in spite of there being no rules mandating net neutrality there also haven’t been significant violations. This is why US Federal Communications Commissioner Ajit Pai has said that net neutrality, “a solution in search of a problem”. But it also shows why this might change.
So how should we understand net neutrality? I think it is helpful to focus, in schematic terms, on the changes of two numbers over time. The first is the cost of transmitting a given amount of data at a given level of performance (speed, latency, minimum speed, etc. — I would note that limiting to the total set of variables is hard, so this is just a sample). Think of it as: C (data, latency, average speed, minimum speed etc.)
The important thing to realise is that due to technical and economic innovation by the carriers, this cost should fall over time for any set of variables, assuming that there aren’t severe problems in policy such as spectrum scarcity or in competition.
At the same time, innovation by the application providers should drive another number up: the economic value that can be extracted from that data, which we will call V: V (data, latency, average speed, minimum speed, etc.)
When the internet was created, the costs of data connectivity were very high and commercial value was very low, so C > V. Since then, costs have fallen and companies have created new ways of extracting value. At some point in the future (this is schematic, so we don’t know when) V > C, that is the economic value created by the application logic will outweigh the cost of transmitting that data.
The whole debate about zero-rating and ‘freemium’ business models occur when C is very close to V. Facebook is willing to pay for your data costs when the activity you generate on Facebook creates more value for Facebook than the cost of that data.
This framework also explains why net neutrality was an abstract debate, “a solution in search of a problem,” when C > V. Internet access is being consumed as a leisure good.
This framework also explains why it is suddenly an interesting debate, as C and V get close. Because internet access generally creates value, there is a potential additional source of revenue for the carriers. Both the carriers and application providers are planning for the future, where the carriers are eyeing a future source of revenue and the application providers are trying to plan for and control their future costs.
Part of the conceptual problem of the net neutrality debate is that we are making policy in an environment when C is roughly equal to V about a world in which not only will V > C, but V – C will get larger over time. We are trying to make policy about a world that are aren’t in and have trouble really foreseeing.
If you accept this framework, the policy question in the future should be: what input should policy have in the distribution of V – C, especially as it grows? Net neutrality, as a policy, is really about putting the thumb on the side of the application provider to limit the amount of V – C that can go to the carrier.
The position of the application providers is, not unreasonably, that the market power is of the carriers will mean that they pocket an outsized amount of V – C. While the carriers point to the increasing size of the markets on the application side (and in the case of mobile, the device side).
My own instinct based on experience in the policy debates in the United States is that the growth of Apple and Google suggests that they have more leverage over the carriers than they care to admit and that we generally perceive. Indeed, one of the most remarkable demonstrations of the changing balance of power has been the ability of Apple to force the iPhone down the throats of carriers around the world. Any rich world telecom executive will tell you that being able to serve the iPhone is a major driver of profit. And, indeed, Apple has found technological innovations in the design of radios in their phones that means that they have nearly universal devices, allowing the phones to work on nearly every network, turning the network into a commodity.
Another way of stating this problem is, as V – C grows, should something that has always been a one-sided market become a two-sided market, and how should that be regulated? This is why Niranjan Rajadhyaksha has said that the economics of net neutrality is uncertain, referring to the work of Nobel Prize winner Jean Tirole. That’s why my own policy preference would be to keep a sharp eye on the question, but be slow to make policy, recognising that it will be made into a rapidly changing environment in which we don’t actually have good instincts about how the markets will work or even what the markets will be.
Now turning to real-world policy, you can see how the net neutrality debate might work in different places. In a place like India, where internet penetration is low and a primary objective of policy might be to get more people online, carriers might argue, in a way that is both self-serving and may align with the public good, that some of V – C should be reserved to get more people online. This is a generous understanding of the argument that Internet.org and Facebook make.
In a place like the United States where penetration is high and the limiting factor for further penetration is typically not cost but perceived need, the question is probably different. It is how to distribute V – C in such a way that the cost is optimal for consumers. Here, antitrust regulation would seem to be the preferred tool to ensure that the carrier’s don’t exert too much market power. The arguments for net neutrality in the US turn on the idea that antitrust regulation is too slow and that regulators are subject to capture.
Soren Dayton is a public affairs and political consultant from Washington, DC, on sabbatical and living in New Delhi. The views expressed are personal.