What do we want broadband competition to accomplish?

On October 30, Blair Levin delivered a speech, Achieving Bandwidth Abundance: The Three Policy Levers for Intensifying Broadband Competition, to an Internet Society conference. This week, we're providing excerpts of the speech. Blair Levin oversaw the development of a National Broadband Plan. He is a nonresident senior fellow with the Brookings Institute's Metropolitan Policy Program. Levin serves as the executive director of Gig.U: The Next Generation Network Innovation Project, an initiative of three dozen leading research university communities seeking to support educational and economic development by accelerating the deployment of next generation networks.

The highest priority for government broadband competition policy ought to be to lower input costs for adjacent market competition and network upgrades

Much has been written on competition and broadband. My comments represent a progress report and work in progress from the field, derived from game theory and lessons I learned in the government with both the Telecommunications Act of 1996 and the National Broadband Plan, as well as working with broadband competition initiatives, such as Gig.U and Republic Wireless.

The trial and many errors of my own work have led me to believe in the following bottom line: that the highest priority for government broadband competition policy ought to be to lower input costs for adjacent market competition and network upgrades.(1) I believe the greatest opportunity is to create a virtuous cycle of upgraded mobile stimulating low-end broadband to upgrade, which in turn causes an upgrade of high-end broadband which, by using its assets to enter mobile, accelerates the need for mobile to accelerate its upgrade further.

My purpose is not so much to convince you that I am right as to move the broadband competition discussion away from the emptiness of much of what is written to the reality of how enterprises have incentives to invest in the faster, cheaper, better delivery of bits.

And if someone has a better bottom line, great.

But let me start the discussion by telling you how I got to mine.

I start with three questions:

  1. Where does broadband competition come from?
  2. Given the current market, what are the appropriate government levers to intensify competition at this part of the cycle?


Competition can optimize for multiple factors according to what customers want more adroitly than a policy process

Competition is generally thought of as the means, not the ends, of improving consumer welfare. That is, we believe competition is the most likely means to deliver the optimal goods and services.

In the debate leading up to and in the implementation of the Telecommunications Act of 1996, the vision was increased competition in all communications markets but most of the debate focused on the voice market. The outcome sought was clear: lower prices.(2)

Broadband is different. There are a number of variables we wish competition to deliver. The two most prominent are lower prices and improved performance,(3) but ubiquity, security, privacy protection, and providing a platform for free and diverse speech, among others, are also desired outcomes.

Optimizing for multiple factors makes policy calls more complex than when aiming for a single goal.(4) Different policies can deliver better outcomes on some metrics and worse outcomes on others, requiring decisions about priorities and trade-offs for which there may be no “right” answer. This makes competition more important, as competition can optimize for multiple factors according to what customers want more adroitly than a policy process.

We want competition over who can best deploy bandwidth abundance

My own view of what we want competition to deliver at this point in the cycle is the elimination of bandwidth as a constraint to innovation, economic growth and social progress.(5) As the global economy moves from being primarily about the manipulation and transportation of atoms to being primarily about knowledge exchange, bandwidth becomes our commons of collaboration and bandwidth constraints would present a major obstacle to economic and social progress.(6)

Further, I believe that goal is likely to be achieved when there are at least two next generation networks capable of delivering all foreseeable needs for the next decade and with a viable upgrade path. With only one such network, economic forces will likely price the marginal use of bandwidth at a level that constrains growth and progress. Thus, we need multiple networks to upgrade to next generation networks.

In short, we want competition to help move us from today’s world, where the dominant business model is how to allocate bandwidth scarcity, to the world we need, which is competition over who can best deploy bandwidth abundance.(7)

In Blair's next post, he'll discuss where broadband competition come from.
  1. Some might argue that closing the adoption gap, sometimes referred to as the digital divide, should be a higher priority for broadband policy. While I agree that it ought to be a high priority for the policy, I am focused here on competition. While bringing more customers to the market will help with the competition issues, it will not, in and of itself, drive the network upgrades that I believe necessary.
  2. In both vision and specifics, it succeeded, but not necessarily in a way that reflected the most heavily-debated provisions, the 14 point check-list for Local Exchange entry into long-distance. Wireless and VoIP entry, as discussed below, were the bigger factors.
  3. This is generally expressed in terms of greater bandwidth, and often illustrated by the time it would take to download an HD movie. History will probably regard this as the least important use of next generation networks, recalling Henry Ford’s comment that before he produced his cars, his customers, if asked, would have said they wanted “faster horses.”
  4. I personally encountered this when I was involved in cable rate regulation, as called for in the 1992 Cable Act. To the extent the law sought to lower prices, that was relatively easy and the February 1994 decision did so initially. But the law also, correctly in my view, wanted the cable industry to be able to continue to invest in more and better programming. The initial price cuts were then reversed by the “going forward” rules, which allowed such investments. Optimizing for both proved difficult, if not impossible, for regulation. The difficulties of that effort are well described in Reed Hundt’s memoir of his FCC’s Chairmanship “You Say You Want a Revolution: A Story of Information Age Politics.” (2011)
  5. Expressed this way, the vision captures a number of different variables, including affordability, ubiquity, performance and others.
  6. There are a number of important government initiatives, including the reform of the E-Rate and Lifeline programs and ConnectedHome, which also are part of the effort to remove bandwidth constraints. As they are not directed toward changing the current mass-market competitive market structure, they are beyond the scope of this paper. Nonetheless, issues of adoption and anchor institution connectivity are critical to the vision that animates the framework I present here.
  7. For a more complete discussion of the transition from moving from bandwidth scarcity to bandwidth abundance, see “The North Star of Bandwidth Abundance”, at http://www.gig-u.org/the-north-star-of-bandwidth-abundance/. I should note that the goal of bandwidth abundance might strike an economist as encouraging an overproduction of bandwidth, not justified by actual consumer demand, and that that goal could lead to stranded investment. This is unlikely, in my view, for a variety of reasons, some of which are discussed in the speech. The principle point is that given the transition to the information economy, abundance is a good in and of itself, which drives new use and consumer surplus. I would also add that unlike cyclical industries, where demand goes up and down, the use of bandwidth seems to go up and up. Of course the timing of such investments can lead to financial losses, as occurred in the early years of this century, but the assets were not stranded but rather picked up by a number of enterprises, like Google, to accelerate their own network operations.

By Blair Levin.