Virulent Word of Mouse

May 12, 2010

The Simple Economics of broadband regulation

There is a simple economic rationale behind the FCC’s recent announcement, made last week by Chairman Julius Genachowski, on May 6th.The FCC had to act. The costs of not acting were too great.

Here is why. Broadband carriers have strong economic incentives to provide services that compete with the applications of others. Yet, those same broadband carriers carry the data of all those applications. These carriers face what is often called “mixed incentives”, and until recently all carriers were forbidden from acting on them. Genchowski wanted to keep things that way, but an appellate court made that hard to do.

Let me make the issue concrete. Just ask your neighbor what they would think of the following: Would they be angry if Comcast blocked Skype from operating and told all users they had to go through an approved vendor of IP telephony? Would your neighbor be unhappy if Google slowed down when the search concerned local car dealers because AT&T broadband had a local search service on which local car dealers advertised? Would your neighbor be frustrated if they could not go to Hulu, but instead had to go to the approved TV distributor who worked with Verizon?

Look, none of that has happened yet, but that is because it was forbidden by regulators until a few weeks ago. It is not anymore, and that illustrates what any sensible regulator should fear. They should fear slowly slouching towards a situation where the mixed incentives of carriers gets in the way of delivering services over the Internet. Households have gotten used to having unrestricted choice online of innovative services, and there would be a minor revolt if narrow firm self-interest got in the way of that.

I have no axe to grind with carriers, so let me say that more positively. The country has benefited from allowing carriers freedom to innovate and build, and that was so up to a point, painting a bright line in one clear place, limiting the discretion of carriers to interfere with anybody else’s innovative services.

More to the point, the present limitations have fostered an innovative ecosystem. Software vendors and Internet hosting companies have developed a range of innovative services in anticipation of (1) better infrastructure on which to run it, and (2) sending their applications across broadband lines that behave the same way everywhere.

In other words, application vendors do not worry about which carrier delivers the data to which homes because carriers are not allowed to act on their mixed incentives. Knowing that, application developers innovate in all sorts of ways that users enjoy.

Those simple economics explains quite a bit of regulatory action. If you look beyond all the posturing and public relations, it is not surprising the FCC feared what would happen if it permitted no legal restraint on carrier action.

To be sure, the FCC’s own explanation of its announcement is — *um*, how do I say this nicely? — confusing in its details. Many commentators have tried to figure out the legal subtleties, focusing on legal nuances about what the new regulations will do or will not do. I have yet to read a simple summary (and that is a worrisome).

More to the point, it is fine for commentators to gripe about legal definitions, but I think all the commentary must acknowledge the most primary economic motivation at the heart of the action: the FCC had to clearly signal that it would try to erect a new regulatory mechanisms for enforcing limitations on carriers’ mixed incentives.

That is the point of my post. There were two economic reasons to take action, and both point towards limiting mixed incentives. One has to do with the incentives to act on mixed incentives today, and the other has to do with the long term trends of the evolutions of the network, which reinforces incentives to act on mixed incentives tomorrow.

Review of recent events

To be honest, the history of US regulation of broadband firms is pretty darned confusing (and I will probably get some of the details wrong). Here is a brief history, just to explain why the FCC had to act now and not earlier.

In 2003 the FCC decisively moved towards light-handed regulation of facilities-based access providers. Michael Powell, the FCC chair at the time, made a deal with telephone broadband providers. The key provision was the the removal of any mandate to interconnect these facilities to others, putting them in essentially the same footing as the cable broadband providers, who were not regulated.

What were the telephone broadband firms supposed to do after this lifting of obligation to interconnect? They were expected to build broadband. That was a rather profitable assignment (as it turned out). That is not a bad deal if you can get it.

At the same time Powell did something else. He announced some basic restraints on the conduct of carriers — e.g., no blocking of any applications, equal access to all of sites for all users, and so on. Powell rephrased these in a positive tone, identifying four principles he intended to enforce, and these became known as the four freedoms.

Powell’s successor, Kevin Martin, continued with that deal in spirit, though the letter of it got changed by the legal staff. That got FCC-watchers all worried, and it might have mattered, but other events intervened first.

The basic deal began to fall apart, because an appellate court told the FCC that it had no legal foundation for enforcing just about anything to do with regulating broadband, the four freedoms or anything else. Why did the court do that? The legal logic goes something like this: When Michael Powell moved to treat broadband firms lightly, he had the FCC classify the legal status of all broadband Internet access as an information service, putting it in an ancillary category (Title I). That was the regulatory trick to putting all carriers on the same footing.

The classification was challenged in 2005 at the US Supreme Court. A little legal nuance helps to understand what happened next. The Supremes took a narrow approach to the case, and did not decide whether Title I was a good choice or not. The Supremes said the FCC could use Title I, Title III or Titleist putters,  as far as the Supremes were concerned, who “deferred” to the expertise of the agency. The Supremes only did one thing; it agreed that the FCC could make a reclassification. It did not decide which reclassification made the most sense.

Five years later, as it turned out, Comcast decided to challenge the thing the Supremes had not decided, whether Title I would be a solid or problematic legal handle for regulating a specific act.

Why did Comcast do that? I am not sure, to be honest. This was an inherently risky legal strategy, and even if they won in court, as they did, it held the risk of leading to regulatory change, as it has. (It now appears Comcast would have been better off had they not brought the suit at all).

Here are the facts: In 2007 Comcast found itself in trouble with the FCC for throttling Bit Torrent without telling anyone, not even its own users or the industry software vendors. The FCC used its Title I authority to impose a regulatory mandate on Comcast, requiring them to treat all vendors equally, even Bit Torrent. Comcast disliked the ruling and sued in return.

In fact, what the Appeals Court said was quite sweeping, declaring that Title I is good for very little, and certainly not sufficient for regulating broadband firm conduct, such as Comcast’s throttling of Bit Torrent.

Got all that? Whether or not you like Comcast, Bit-Torrent, or Diana Ross and the Supremes, the real news is what the appeals court said: Title I supports very little broadband regulation. For all intent and purposes, that means Title I is good for nothing, none, nada. In short, one big part of Michael Powell’s 2003 quid-pro-quo with carriers has no legal foundation.

(It makes me wonder if Michael Powell got bad legal advise from his own staff back in 2003. Or, more speculatively, what would he have done differently if he had known this would not hold up? Ah, but I digress…)

Here is where it gets sticky: That court decision left the FCC in an uncomfortable position, and with no good option in the short run. If it did something, anything at all, it was bound to be imperfect. If they did nothing, leaving the status quo in place, there would be no regulatory limits on carriers. Period. No limits.

What would should a regulator do? It should not have taken long to recognize that carriers face mixed incentives, and have strong incentives to act on them. The costs of doing nothing were potentially high.

So that is my big point: the FCC chose not to stand still. There is a very sensible economic rationale for that. On the face of it, doing nothing seems worse than most imperfect regulatory actions.

Network economics and mixed incentives

Having said all that about mixed incentives, I think this decision also arose out of one other long term economic trend. The combination of economic determinants that led to prosperity for carriers are about to come to an end. That is worrisome, and it provides one additional reason to take action.

I realize that might sound confusing. Sorry. Here is an explanation.

While broadband networks involve a vast myriad number of actors, there are basically two tribes, infrastructure providers and applications providers. Infrastructure providers include broadband access providers (e.g., Comcast), backbone providers (e.g., Level 3), cache providers (e.g., Akamai) and CDN providers (e.g., Amazon Content Delivery Network), and others (e.g., Wireless ISPs). Application providers include big portals (e.g., Google, Yahoo!), big software hosting applications (e.g., Facebook, Flickr), e-commerce sites (e.g., Land’s End), and every blog (including this one), and lots more (e.g.,

Each tribe’s members must manage their own inputs and these must work together with others’ to produce a service that users recognize. Economists call this a complement.

Growth is also easy to explain. In the last two decades these inputs have steadily produced increases in traffic and revenue for the primary investors in each tribe.

Frankly, it was good business if you could get it. In a market where the size of the pie grows one year to the next, the slices gets larger for the dominant providers.

This growth will not last forever, nor could it. Three basic factors fueled growth in the past: diffusion, cannibalization, and innovative advertising. I believe only two of those will continue into the future, and that hurts carriers the most.

Diffusion: One of the oldest adages is high tech applied to broadband in the last decade: New users led to new revenue. As more households and more businesses went online there was more opportunity to charge little fees for services. Internet access fees alone account for over $45 billion in revenue in the US today, which is thirty billion more than a decade ago. Of course, there are also countless other little places where firms generate little fees (e.g., gaming, iPhone apps, etc) for services rendered online, but the broadband firms had the biggest gig of all. It was great to be a firm in that growing market.

Cannibalization of offline activities: A tremendous amount of the revenue generated by the on-line economy is simply a restructured — and better delivered and more efficient — version of activities previously performed offline. Much online retailing is a better and cheaper version of what consumers used to buy without the Internet, whether it involved music, books, plane tickets, or shoes. It is great to be a pure play in this situation (e.g., Orbitz or Amazon), but everybody else just tries to survive. They invest because they have to (e.g., Catalog firms, such as LLBean).

Innovative Advertising: Let’s give Google its due. They invented a new advertising market for key words to support search. It supports billions worth of advertising (approx $15B for Google and counting). They also invented a way to bring such affiliate networks to blogs (approx $7B and counting). This revenue partly cannibalizes what used to go into magazines and newspapers and television advertising revenue, but it is also partly something new. It supports a wide set of firms online. It is good for Google, and it is good for many players in the Internet ecosystem, such as blogs, who would have otherwise had no equivalent way to raise advertising revenue.

No sensible observers thinks the first category, diffusion, will grow much more in the near term in the US, while every sensible observer remains optimistic about the potential for growth in the latter two categories, cannibalization and innovative advertising. Got that? The size of the pie will grow in some respects, and as it has for the last decade, but not in all respects.

And that also is the source of the potential long term problem for the FCC. Infrastructure firms, carriers included, have largely benefited from the first factor, diffusion, while the other tribes largely benefited from all three, but especially the latter two, cannibalization and innovative advertising.

In short, I believe the pie was growing for all the main tribes over the last two decades, but that will soon stop for broadband firms. If a regulator believed that too, then that is another reason to impose regulatory restraint as soon as possible.

When one firm’s gain comes at the expense of another

What happens when one major actor in an ecosystem sees its revenue growth slow down while the other actors do not? I cannot forecast the future in the Internet, but I can explain what happened in the PC industry under slightly different circumstances: the ambitious tribe began to aspire to gain the revenue in the other areas in which it previously had not offered service. If the same thing happens in the Internet, then it will produce fireworks. Such fireworks will invite regulatory intervention.

More concretely, in the 1990s the PC industry became quite a contentious place because an ambitious operating system company and occasional word processing company (e.g., Microsoft) began to try to improve its word processors, expand into CD-ROM encyclopedias, presentation software, expand the functionality of its spread sheets, enter into browser markets, and provide just about every piece of softwarefound on every PC. Similarly, an ambitious microprocessor firm (e.g., Intel) began to try to expand into mother board design and production, memory chip design and production, wireless antennae design and production, and cell phone design and production, and just about every other major hardware device.

Look, ambitious firms will expand if they can. That is not a statement about Bill Gates or Andy Grove or any other executive. Expansion is what stock holders want from their firms. Capitalism works this way, and it always has.

More to the point, it begs the question: Who are the potential Microsofts and Intels of the Internet ecosystem today? Remarkably, it is not either Microsoft or Intel, who are having their struggles.

Rather, good candidates seem to be broadband providers, such as Comcast, AT&T, and Verizon. Their management is ambitious. The diffusion of the first generation of broadband provision is coming to a close in the next few years. Thus, the primary source of revenue growth for those firms will naturally slow. Yet, at the same time, those firms must be eying other parts of the Internet, where revenue continues to grow.

Look, I am not saying anything that many other analysts have not noticed. Comcast’s management is ambitious. So is AT&T’s, and so is Verizon’s. They must be considering how a carrier can develop its own television service (instead of relying on Hulu), or get a piece of revenue from selling online movies (instead of letting Netflix collect all the revenue), or get a piece of online advertising for local services (instead of letting Google collect all the revenue).

This is an old lesson in regulatory economics. Commercial ambition from a dominant firm is a good thing when it fuels competitive conduct, innovative services, and invading of new service territories. Ambition from dominant firms is usually not such a good thing when it motivates such those firms to block a rival’s access to channels, when it leads dominant firms to refuse to deal with potential rivals, and when it leads dominant firms to raise a rival’s cost.

(BTW, other large firms, such as Google, also face such mixed incentives. But Google does not offer broadband access, so the FCC cannot regulate Google’s conduct directly. It is for another government actor, if any, to do something, such as the FTC or DOJ antitrust authority. But I digress.)

Even a lawyer can get the summary. One of the biggest participants in this ecosystem, broadband carriers, face mixed incentives to compete — and not cooperate — with application providers. The shadow of those incentives will grow longer as the revenue growth begins to slow.

Public policy?

Time to finish. Without assessing the merits of what the FCC chose to do, and without having any inside information about their thinking, it is possible to say this: it is no surprise that the FCC felt compelled to act. The appeals court left the FCC in a position where it could not do anything to a broadband firm taking action against an application, as their mixed incentives would tempt them to do.

Mixed incentives are a very old problem in communications access regulation, and easily understood by every regulator on the planet. Every regulator can sense the danger of letting carriers move down the slippery slope of not cooperating with another participant in the Internet ecosystem. That bright line had to be protected, if only for the sake of preserving the innovative ecosystem that has driven the Internet forward in these last two decades.

Let’s hope the ecosystem continues to be innovative in the next decade with these new set of rules.


  1. You are incorrect about the sequence of events. Broadband has never been regulated as a Title II common carrier. You also fail to point out that the FCC went out of its way to make the argument that the information service component of broadband is inexorably linked with the transmission component, that in theory could be regulated as a common carrier. The FCC will be sued for this reclassification and it will be hilarious to watch them argue the complete opposite of what they said 5 years ago in the Brand X case. The FCC is sure to lose, either at the Appeals level if it goes to say the DC Circuit, or if the FCC gets lucky and say the 9th Circuit gets the appeal, then SCOTUS will slap down the FCC. And rightly so. It’s unconscionable to change the regulatory construct AFTER carriers have spent hundreds of billions of dollars building their networks. Surely no right minded management team, and certainly not their investors, would have sunk capital into a network if they had known ex-ante it was to be regulated as a common carrier.

    Comment by joe — May 13, 2010 @ 11:05 am | Reply

    • Thanks for the correction on legal history. As I said, it is confusing. Appreciate getting the facts straight. However, I am not sure you are correct about investment. Cable firms were doing plenty of investing in broadband before 2003. These paid off relative quickly, so they would have continued to do plenty without Powell’s 2003 declaration.

      Comment by Shane Greenstein — May 13, 2010 @ 9:47 pm | Reply

  2. Just want to say what a great blog you got here!
    I’ve been around for quite a lot of time, but finally decided to show my appreciation of your work!

    Thumbs up, and keep it going!


    Comment by Dardrymnmar — May 22, 2010 @ 6:33 am | Reply

  3. “Surely no right minded management team, and certainly not their investors, would have sunk capital into a network if they had known ex-ante it was to be regulated as a common carrier.”

    Surely no right-minded management team would have sunk capital into anything on the assumption ex-ante that the Bush Administration and its peculiar approach to regulation — Micheal Powell?? — were a permanent manifestation. At some point the regulatory authority, and simple common sense, was bound to catch up with the imaginary difference between voice and data upon which the earlier misguided FCC decision rested.

    Comment by leo — June 12, 2010 @ 2:47 pm | Reply

  4. P.S. If you want an idea of how an unregulated information highway would operate just look at how AT&T rigs its hardware — particularly smartphones. In the case of the HTC Fuse, AT&T completely rearranged the standard qwerty keyboard to accommodate buttons that link directly to its own fee-based propriety services.

    Just imagine if they had had this much control over the network at an earlier stage in the development of the Internet — just imagine what the computer keyboard you’re currently bent over would look like today.

    Comment by leo — June 12, 2010 @ 3:14 pm | Reply

  5. Minor nit: The decision to put telco broadband on the same footing as cable modem service was made in 2005, not 2003, following the 2005 Brand-X decision which upheld the 2002 declaratory ruling that cable modem service would be unregulated.

    What did happen in 2003 is that the FCC said that companies that build fiber-to-the-home (think Verizon FIOS) did not have to provide unbundled access to others as they do with copper.

    Comment by Marvin Sirbu — June 12, 2010 @ 9:55 pm | Reply

  6. There’s also an old lesson in telecomm: the telecomm industry produces lousy content because the business models of production and distribution are very different.

    …but they want to be in pictures…

    Comment by unsecured small business loans — October 25, 2010 @ 7:44 am | Reply

RSS feed for comments on this post.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Blog at

%d bloggers like this: