Prof. Randal C. Picker, University of Chicago Law School
March 16, 2015

On Feb. 26, 2015, the Federal Communications Commission announced its new Open Internet Rules (what the rest of the world calls network neutrality), but it didn’t actually release the rules until this past Thursday, March 12, 2015.  The rules themselves are quite short, only eight pages, but the full Open Internet Order is 313 pages and the download from the FCC website, with statements from the commissioners, runs an even 400 pages.

Net neutrality has taken on all of the trappings of a holy war.  One chunk of the world believes that the Internet has been a great success and that the legal regime that has applied while that has happened must be continued, so the FCC shouldn’t act.  Another chunk of the world believes that the Internet has been a great success and that the underlying rules that have applied while that has happened must be continued, so the FCC must act aggressively.  Given the depth of the disagreement, it is hard to find points of common ground.  It would be silly to try to address the full issue of net neutrality in a blog post, but I do want to try one simple example to frame a key issue considered in the Open Internet rules, namely, whether broadband providers should be permitted to create Internet fast lanes (also known as paid prioritization).

Early on in the process for creating the new rules, the FCC had suggested that it was open to paid prioritization, but the new order (at paragraph 18) flatly forbids this: “A person engaged in the provision of broadband Internet access service, insofar as such person is so engaged, shall not engage in paid prioritization.”  To further shrink the analysis, I will say upfront that I don’t know what I think about that.  I will say that multiple service levels are generally an important part of economic life and that they routinely produce better outcomes for everyone.

Take a simple example.  We are going to deliver two packages.  It costs $16 to create the infrastructure to deliver those two packages and for that expenditure, two different businesses could be built.  The one would deliver both packages at, call it, medium speed.  The alternative service would deliver one package quickly – faster than medium speed – and the second package more slowly (slower than medium speed).

We need some consumer values to assess this.  Assume that we have two consumers – one per package – and that each consumer values medium speed delivery at $10.  In our baseline, one level of service world, the firm delivers two packages at medium speed at a competitive price of $8 per package.  It covers its costs and consumers each get “extra” value – what an economist calls consumer surplus – of $2 per package, so we are $4 in value to the good from having medium service.

Would our little artificial society do better having two service levels?  That depends on consumer values. Assume that my first consumer is impatient and values fast delivery of the package at $15 and slow delivery at $0.  The second consumer is more patient.  She values fast delivery at $11 and slow delivery at $9.

Suppose that we have a two-tier delivery service that charges $12 for fast delivery and $4 for slow delivery.  How should we evaluate that?  The answer in this framework is clear: Everyone is better off than they were in the baseline case with one tier of service.  The impatient consumer pays $12 for a service valued at $15 and thus gets $3 in consumer surplus (before he got only $2 in consumer surplus). The patient consumer pays $4 for a service that she values at $9 and thus gets $5 in consumer surplus instead of the $2 she got before.  Costs are still covered, as the firm collects revenues of $16, but consumer surplus is now $8 instead of $4.

Both consumers are better off with the two-tier system.  The patient consumer is getting a service that is inferior to that in the one-tier system – she values slow delivery at $9 rather than her $10 value for medium delivery – but because the impatient consumer is now paying more of the fixed costs of running the delivery system, the patient consumer is actually better off than under the one-tier system. There was no good way to get the impatient consumer to bear a greater share of the fixed costs in the one-tier system, but that is straightforward to do in the two-tier system.  Again, both consumers are better off under the two-tier system.

There are many, many moving pieces on net neutrality.  My example assumes that the market for delivering packages is competitive and many people would say that the broadband market is far from competitive.  I don’t know where the FCC is on that exactly, as I am still reading the full version of the Open Internet Order.  Early on (at page 6 in footnote 12), the Order states that “this Order need not conclude that any specific market power exists in the hands of one or more broadband providers in order to create and enforce these rules.  Thus, these rules do not address, and are not designed to deal with, the acquisition or maintenance of market power or its abuse, real or potential.”

That seems to suggest that the FCC believes that paid prioritization should be barred even if the underlying broadband market is competitive.  I get, at some level, why the FCC might want to avoid basing its rules on the extent of competition in a particular market.  The path of the FCC’s rules for unbundled network elements – a key design point of the 1996 Telecommunications Act – was extraordinarily messy, and the FCC sought to avoid factually detailed individualized market determinations even when reviewing courts pushed it to do just that.  A generic rule is much easier in that regard.

And it is fair to say that there have been substantial amounts of investment inbroadband infrastructure without meaningful paid prioritization, though do note that a level down – at the level of content delivery networks and the like – the Internet hasn’t been neutral for some time as firms have been able to contract to improve the delivery of their content.  Concerns about paid prioritization in this space seem to reflect the idea that broadband providers will manipulate capacity choices to increase their bargaining power or to favor their own services, and that service levels will discourage the innovation that has made the Internet the rich space that it is.

Those are clearly important concerns, and the argument in favor of paid prioritization itself seems a little academic.  But capacity manipulation sounds like an antitrust issue and so the preference for a net neutrality rule depends a little on how you feel about untextured ex ante rules (the Open Internet Order) versus ex post targeted inquiries (the antitrust system).

And as to innovation, there is innovation all around, meaning that it would be a mistake not to think of paid prioritization as a potential source of innovation as well.  And the fact that we may not be able to point to what innovation that might generate misses the point, as that is the basic nature of decentralized innovation.

I will finish where I started.  I don’t know what I think about paid prioritization yet – still reading the FCC order and still thinking – but it is important to be clear on the general case for paid prioritization, to focus on the way in which the FCC seems to have elided the market power question, and to recognize the way in which paid prioritization might tie into other innovations.