I’ve remained rationally ignorant over the debate on net neutrality, as it’s not in my academic field. But with net neutrality being in the news a lot recently, I took to Twitter to ask if there were any good explainers or debates on the topic that gave both sides of the argument and could help me understand the competing perspectives. Some readers pointed to this debate from 2007 between Tim Wu and Chris Yoo, which seems like a good start.

In addition, professor Gus Hurwitz responded with a long email offering his overview of the debate. I thought his explanation was sufficiently interesting — at least to someone like me who has no background on net neutrality– that I asked him if he would let me publish his explainer here at the blog. He agreed.

What follows below is from Hurwitz, not from me, and it offers his effort to summarize the debates over net neutrality. Of course, it is only one person’s view, and Hurwitz clearly has his own perspective on things — especially near the end. But I thought it was worth offering this perspective, with the understanding that if others think this overview is wrong or misguided, it’s a big Internet and others can respond.

I’ll be happy to update this post with links at the bottom of this post to particularly helpful responses or criticisms of this overview. My goal is just to try to figure out what the debate is, rather than come to a particular view, so I’m particularly interested in perspective that disagree with this take about what the debate is. Just email me with the links to the responses. If there are any particularly good criticisms from people who seem well-informed on the debate, I would also consider posting them in a separate response post (with the authors’ prior permission, of course).

* * * * * *
To: Orin Kerr
From: Gus Hurwitz

“Net neutrality” refers generally to the idea that ISPs should not treat traffic traversing their networks differently based on its source, content, use, &c. That is, they should be passive conduits that treat all data the same. We’ve been debating whether the FCC should require such neutrality for the past 15 years – and this debate is actually an Internet-era extension of debates that extend back at least to the Federal Communications Commission (FCC) regulation of the historic AT&T monopoly.

I generally identify two main policy issues that animate these debates, which I’ll label competition vs. regulation in natural monopoly industries and the role of regulation in promoting investment and innovation. These are unpacked below. The reality today is that the policy discussion has broken down into pure tribalism – I don’t think there is room (or patience) left for serious policy discussion in this area, even among “serious people.” This is largely because “net neutrality” has become a token representing different, broader, social values that are quite separate from the technical, legal, economic, &c, issues implicated by regulation of how ISPs handle data traversing their networks. I discuss this, as well, as the third point below.

1. Competition vs. regulation in natural monopoly industries

The key background issue in net neutrality (as with almost all telecom and infrastructure discussions) is that last mile Internet access, like most communications networks, is often considered a natural monopoly. “Last mile” access refers to connection between consumers’ houses and their ISPs, and is generally the most expensive part of a network to build. “Natural monopolies” are industries where high entry costs make it economically undesirable to have a great deal of competition. For instance, it may cost $2,400 to connect a house to the Internet. Most consumers only need a single home Internet connection. If the capital costs of building the network are recovered over 10 years, adding a second last-mile connection (that is, ISP) doesn’t offer the consumer any greater functionality but on average increases the price of service from whichever ISP they buy service from by $20/mo.

Whether last-mile Internet access actually is a natural monopoly is often debated – most users have access to at least two high-speed wireline ISPs, many have access to more, and the role of high-speed wireless Internet access (both mobile and fixed) as a competitive alternative is likewise hotly debated. Regardless, it is fair to say that the market is not robustly competitive. This gives rise to concerns that ISPs will have more than a modicum of market power and that they may use this power anticompetitively in ways that harm consumers. The question, then, is what to do in order to protect against these potential harms.

The basic concern of net neutrality in particular is that ISPs, possibly having some market power, may be able to discriminate against data traversing their networks. For instance, an ISP may demand payments from a popular service in exchange for not slowing or blocking that service. An ISP may tell a start-up that it needs to pay more in order to be treated comparably to existing market participants (called paid prioritization). An ISP my disadvantage services that are competitive with the ISP’s own services. Practices such as these could, the concern is, increase costs to consumers or reduce the quality of their Internet experience.

The question of how to address these concerns bifurcates along several dimensions. For instance, whether this is a generalized problem that we should rely on the [Justice Department] or [Federal Trade Commission] to address, or is it a telecom issue that the FCC is better equipped to address (“equipped” in terms of technical expertise or insulation from public choice (FCC being more captured))? Do we want to rely on rules or standards to guide (im)permissible conduct? Should those rules/standards be more antitrust- or consumer-protection like? Do we want to rely on ex ante or ex post enforcement mechanisms for them?

Proponents of net neutrality generally favor strict ex ante rules enforced by the FCC; critics generally favor flexible ex post standards enforced by the FTC, DOJ, and state attorneys general (possibly with an assist by the FCC in investigating problematic conduct).

Informing these worldviews are priors about error costs and, in particular, the costs and benefits of paid prioritization. Proponents generally think that nonneutral conduct by ISPs, and especially paid prioritization, is overwhelmingly likely to be bad for consumers – pure rent extraction by ISPs that will increase consumer prices and harm entry by edge providers. The extent to which some of this conduct may be good for consumers is so diminishingly small that the benefits of reduced enforcement costs, clear rules, and certainty for consumers and application developers outweighs them.

We have the exact opposite view on the other side: the critics think that the likelihood of harmful conduct is small, where it happens existing legal frameworks are likely sufficient to address it, the political economy of consumer concern about and outrage over discriminatory practices will serve as a substantial check on problematic ISP conduct, and where nonneutral practices are implemented it will be because there are significant gains to trade that will ultimately benefit consumers. (In the interest of disclosure, I am in the latter camp, primarily because, as I read it, the economic literature overwhelmingly shows that non-neutrality has potentially substantial benefits for consumers and slight risk of harm – and where that harm has occurred in the past we have been able to address it without recourse to strict ex ante rules.)

Related to these issues is the headline question of “classification” – whether Internet access is better classified as an “Information Service” under Title I of the Communications Act or as a “Telecommunications Service” under Title II. In recent years this issue has become synonymous with net neutrality for many, because Title II classification triggers broad regulatory authority under which the FCC can easily implement strict ex ante rules. It is also symbolically important because Title II is a framework for traditional public utilities regulation, which appeals to those who believe that the Internet should be regulated as a public utility.

In terms of policy, however, this question is mostly instrumental, determining what legal authority the FCC uses to implement whatever rules it does – it doesn’t tell us why or what those rules should be. This is a seemingly important question because in 2010 and 2014 the FCC lost cases in which it attempted to enforce net neutrality rules under non-Title II authority. The 2014 case, however, made clear that the FCC could implement net neutrality rules under non-Title II authority, albeit with some limitations.

2. The role of regulation in promoting investment and innovation

The secondary policy argument deals with investment in building out more, faster, networks and, to a lesser extent, encouraging innovation at “the edge” (that is, by application developers/Silicon Valley).

This issue is principally important as a legal hook for FCC regulation. The Communications Act charges the FCC (in hortatory terms) with ensuring that communications services are deployed to all Americans on a timely basis, and Section 706 of the Telecom Act can be read as giving the FCC regulatory authority if it finds that “advanced telecommunications capabilities” are not sufficiently being timely deployed. (Over the years the FCC has read it both as an independent grant of regulatory authority and not, and the courts have deferred to both readings.)

This brings us to the commission’s “virtuous circle,” or “triple bank shot,” theory of investment and innovation. The theory is that net neutrality rules will promote application development; users will want to make use of these applications and will need higher performance networks to do so; and this demand will drive investment by ISPs in better and faster networks. Thus, through a chain of proximate causation insufficient to satisfy Palsgraf, net neutrality rules facilitate the FCC’s mission of encouraging deployment of such networks.

Taking this a step further, net neutrality proponents also believe that investment will be harmed without these rules. ISPs will seek to extract rents from their existing networks rather than upgrading them and, in fact, will view network degradation as a tool to force edge providers to pay for prioritization. And the lack of certainty over how ISPs will discourage application development, reducing innovation (a direct consumer harm), reducing demand for better and faster networks.

The flip-side of this argument is basically that the virtuous circle can go in two directions: net neutrality opponents are concerned that the rules will increase regulatory costs and decrease the profitability of networks, decreasing investment, harming consumers and edge providers alike. And, relating back to the discussion above, the nonneutral practices may, in fact, facilitate certain types of innovation that make otherwise impractical applications possible.

Much of the current debate, and what is in the new order from the FCC, relates to the investment effects of the 2015 order. The data is thin – only two years with a lot of noise – but suggests that network investment is down since the 2015 order was adopted. On the margin, this weighs against the 2015 order, which we would have expected (if the virtuous circle theory was correct) would increase investment. Realistically, this is mostly reading tea leaves. But agencies are supposed to consider relevant facts in policymaking, and this data tends to contradict the 2015 order and support the proposed new order.

A related, though tangential, point is the effects that this has on broadband deployment. One of the FCC’s core goals, both statutorily and as a matter of current policy, is broadband deployment (“closing the digital divide”): getting Internet service built out in un- and underserved communities. If it is true that investment is decreasing, those decreases will most likely come from areas that are higher-cost to build and generate the least revenue – that is, from un- and underserved areas (high cost/low revenue is why they tend to be un- and underserved). This suggests that net neutrality rules could actively frustrate the FCC’s mission.

There is a related argument that I won’t get into here that the rules also have the effect of subsidizing the highest-cost uses of the network, which are primarily luxury uses like watching lots of high-definition video and playing video games, which has the effect of increasing the cost of the highest-value services, such as accessing basic news & information, using government websites, searching for jobs, &c.

3. The social meaning of “net neutrality”

The most confounding aspect of the contemporary net neutrality discussion to me is the social meanings that the concept has taken on. These meanings are entirely detached from the substance of the debate, but have come to define popular conceptions of what net neutrality means. They are, as best I can tell, wholly unassailable, in the sense that one cannot engage with them. This is probably the most important and intellectually interesting aspect of the debate – it raises important questions about the nature of regulation and the administrative state in complex technical settings.

The most notable aspect is that net neutrality has become a social justice cause. Progressive activist groups of all stripes have come to believe that net neutrality is essential to and allied with their causes. I do not know how this happened – but it is frustrating, because net neutrality is likely adverse to many of their interests. One hears lots of stories about how activists are using the Internet, and how small, typically minority-owned, businesses rely on the Internet and therefore on net neutrality. The reality is that there is exceptionally little reason to believe that any ISP would ever do anything to hurt these users. The general animating fear of net neutrality is that ISPs want to move to a “pay-to-play” model, charging those putting content online to deliver that content to users.

But there simply isn’t enough revenue to be generated from charging most of these users to make it worth the ISPs time (not the mention the political costs that ISPs would face if they did). Moreover most of these users do not deliver their own content, but instead rely on third party providers. There are many stories of people producing video content who are worried that net neutrality will make it impossible for them to reach users – despite the fact that most of them use YouTube to deliver the content. And, as I said above, if net neutrality rules decrease investment they hamper efforts to close the digital divide, harming already marginalized and disadvantaged communities.

The most frustrating of these stories, which relates back to the innovation theories above, is that there is a pervasive belief that net neutrality is needed in order for entrepreneurs to enter the market. The concern is that ISPs would charge start-ups a prohibitive amount in order to get access to ISPs’ customers on terms sufficient for start-ups to compete with firms such as Google and Netflix. This is tragic because almost all of today’s big content providers – the Googles and Netflixes – have invested massively in “content delivery networks.” These are networks that allow their content to bypass almost the entire Internet, dramatically improving performance. In other words, they have already paid for prioritization – they just haven’t paid ISPs for that competitive benefit. Start-ups don’t have access to (or need to pay a pretty penny for access to) services such as these. Allowing ISPs to charge for prioritized performance would give start-ups a lower-cost alternative to CDNs that could give them a competitive advantage (or help remedy a competitive disadvantage). The reality is that most start-ups don’t need that sort of performance assistance (so would not be harmed) – but those offering complex-enough services would have access to a valuable new offering.

It also must be noted that for many (most?) people net neutrality is about regulating ISPs. ISPs are reviled. They are known for bad customer service, they always seem to be too slow, they cost too much (especially bad since they’re mere middlemen, not providing anything of value – just access to content providers’ things of value), and when anything on the Internet isn’t working it’s their fault (e.g., if Netflix is down, users blame their ISP, not Netflix). In this context, “net neutrality” is about “having the FCC regulate Comcast so it will have better customer service and I’ll have someone to complain to when Comcast raises its rates.” Never mind that net neutrality has nothing to do with this. This is why many people are in favor of Title II, utility-style, regulation for ISPs. We don’t like ISPs so we should regulate them; end of story.

The last comment that I will make is how I think about this entire issue: it’s just the latest example of a fight between bilateral media oligopolists. “Big content” and “big distribution” have always fought over how to split the rents they extract from consumers, users have always distrusted distributors, and content providers have always used this to their advantage. From this perspective, the net neutrality rules are pure rent seeking by a content/edge industry that had largely captured the previous FCC.