We’ve had a few days now to digest the FCC’s Open Internet Order (0). While this exercise will be ongoing for the foreseeable future, one thing has become abundantly clear: the Order is worse than expected. This is particularly true when it comes to the issue of paid prioritization. The FCC’s analysis is threadbare and conclusory throughout. It cites comments selectively and without serious effort to analyze substantive concerns. It ignores and mischaracterizes evidence in the record. It bases its conclusions on irrelevant and at times self-contradictory factors. The commission plays so fast and loose with the record and draws so many clearly erroneous conclusions of fact and law that the only questions left are these: how badly is the commission going to lose, and what will the long-term impact be for the commission’s reputation?
We can find examples of all of these problems by just looking at a few paragraphs of the Order – specifically, the analysis supporting the commission’s ban on paid prioritization. Given that concerns about paid prioritization were the main motivation for the commission to get into this quagmire, it is reasonable to expect the commission treated this discussion with some care. The key text is paragraph 126, which explains that “[t]he paid prioritization ban we adopt today is based on the record that has developed in this proceeding.” This discussion makes reference to concerns that allowing paid prioritization “will result in the bifurcating of the Internet into a ‘fast’ lane … and a ‘slow’ lane,” and that paid prioritization could hamper entry, deter innovation, harm competition and consumers, and give rise to a litany of other problems. The Order cites two examples from the comments (both expressing concerns about how prioritization could affect development of high-bandwidth applications and content like video). And the commission asserts that its “conclusion is supported by a well-established body of economic literature.”
The commission’s reasoning does not stand up to even cursory scrutiny. To support the assertion that paid prioritization could cause a bifurcation of the Internet into “fast” and “slow” lanes, and in particular that this could lead to “degraded performance” for non-prioritized traffic, the commission points to comments (1) from Mozilla (2) (arguing that prioritization is “zero sum”), and to comments from Sandvine (3) which show, as the commission incorrectly asserts, that bandwidth is finite, and that giving one application a greater share of that bandwidth reduces the bandwidth available for other applications.
Let’s start with this mischaracterization of the Sandvine comments (4). Sandvine’s discussion of paid prioritization starts by noting that “the FCC has put tremendous focus on Pay for Priority. We’re not quite sure why.” Sandvine goes on to explain why the commission should not be worried about paid prioritization, arguing that the bifurcation argument – the one that that the Commission cites to support banning paid prioritization – is likely “technically unsound.” The commission has taken the Sandvine comments out of context and misrepresented them as supporting its preferred policy conclusions. What is even worse, Sandvine’s actual argument, that concerns about paid prioritization are technically and economically unsound, directly contradicts the commission’s bifurcation concern. How does the commission resolve this contradiction? It doesn’t. The citation supporting the commission’s “bifurcation” concern only includes supporting comments – it simply ignores the concerns commenters like Sandvine have raised
And, lest there is any doubt as to Sandvine’s preferred policy outcome, it starts its comments by explaining that it “has seen firsthand how innovative service plans have increased adoption of the Internet around the world, enhanced competition, and given consumers more (and more affordable) choice.” And it spends the rest of its comments arguing that the commission should protect innovative business models and billing arrangements – contrary to the commission’s decision to ban or subject such arrangements to commission scrutiny. It is perverse, at best, for the commission to use Sandvine’s comments to make a point with which Sandvine disagrees, in support of a policy that Sandvine was arguing against.
Sadly, this is par for the course for the Order. I know, because my own comments were given a similar treatment.
In support of its “bifurcation” theory, the Commission primarily relies on Mozilla’s comments. In fact, they cite to Mozilla’s comments twice for the same proposition (that “Prioritization is inherently a zero-sum practice”). The only other support they offer for this proposition is the mischaracterization of Sandvine’s comments. The commission does cite both an ex parte notice (5) that I submitted and comments from ADTRAN as offering a contrary view. But it does nothing to consider, let alone rebut, these arguments beyond acknowledging them. In another set of comments (6), I explain in great detail why it is wrong to say that prioritization is “zero sum.” Contrast this with the Mozilla comments, which merely assert the zero-sum theory as true alongside non-technical hand-waving. My own hubris aside, senior commission officials have told me quite clearly that they understand the zero-sum argument is technically unsound.
But wait, there’s more! Not only does the commission’s justification for the ban on paid prioritization make no serious attempt at analysis and rely primarily on cherry-picked sentences from comments, but it also ignores the existing body of actual research on paid prioritization. Asserting that its “conclusion is supported by a well-established body of economic literature,” the Order cites a total of six academic articles: four articles on price discrimination from the 1980s and two “more recent” articles (both are from 2000, one was never published), all of which are almost entirely irrelevant to the discussion of paid prioritization. Meanwhile, the commission gallantly ignores the significant body of research conducted over the past decade on the precise question of paid prioritization – even though substantial portions of this literature were discussed in the record. To refresh our collective minds: the consistent conclusion of this literature, almost to the paper, is that paid prioritization can be beneficial to consumers and competition; there are also circumstances where it can be harmful. It is simply beyond belief – a violation of the most basic principles that govern agency decision-making – that the commission simply chose to ignore the past decade of research on the very topic it was considering.
Sadly, this is not the end of the problems the order poses. But it is the end of this post. I will conclude by noting one final expert opinion that the Order did not address: in May of last year, Tom Wheeler testified before Congress that “there is nothing in Title II that prohibits paid prioritization.” Yet here we are.