California’s net neutrality law, which a federal district court upheld in February, is already wading into the regulatory morass that brought down past regulatory regimes charged with maintaining neutrality in rail transport and energy. Specifically, the California law is likely to generate an endless and costly series of regulatory rulemakings on how to define categories of service that can be zero-rated—that is, used without counting against a data cap—and which applications must be included in those categories, without any evidence that it would generate offsetting benefits.
The opening kerfuffle stars telehealth and veterans. Politico recently wrote that the Veterans Administration (VA) questioned whether its telehealth app could continue to be zero-rated. The Wall Street Journal and others held this up as an example of how net neutrality can harm consumers by blocking zero-rating for telehealth services without evidence of counterbalancing benefits derived from disallowing zero-rating.
As Public Knowledge’s Harold Feld pointed out, however, the California law is more nuanced. It does not ban zero-rating, but rather, as the law states, prohibits ISPs from “Zero-rating some Internet content, applications, services, or devices in a category of Internet content, applications, services, or devices, but not the entire category.” Thus, carriers could zero rate the VA program as long as other programs that are part of that category get the same treatment.
This inclusiveness, which sounds nice, actually illustrates why the law can be so pernicious. It isn’t because the law might disallow one beneficial zero-rated service. It’s because the law creates a situation in which online business models, services, and applications are subject to regulatory approval and challenge.
Because the California law allows zero rating an entire category, but not individual applications, the regulator will have to define a relevant “category” and determine when any given app is part of such a “category.” Even with the VA controversy, the definition is not straightforward. Is the relevant category veterans’ health, or telemedicine more broadly?
Let’s suppose the regulator decides that telehealth broadly defined is a category. But how does the regulator then decide whether a service is part of the telehealth category or not? Are ISPs obligated to zero-rate FaceTime and Skype, since health care providers use them to offer telehealth services? Zero-rating distance learning, as well, may be popular. Will ISPs have to zero-rate classes at a reconstituted Trump University if legitimate community colleges want to zero-rate virtual classes to help their students? Or suppose an entrepreneur has a new idea that would benefit from zero rating. Will the regulator have to create a new category or approve the idea as part of an existing category?
That regulatory process is likely to expand beyond zero rating, since other attributes may fall under net neutrality rules. Telehealth is likely better not just as an unlimited service, but also as a guaranteed high-quality connection that does not sputter and freeze. Those issues may be related to factors like latency, jitter, and packet loss, and ISPs or health care providers may wish to offer a specific quality of connection. If California regulators decide a dedicated high-quality connection to certain services violates net neutrality, then they may want to extend the rules beyond zero rating to control these aspects of quality, creating multitudes of tariffs and categories.
Such regulatory creep is hypothetical, but not far-fetched. Connections with guaranteed quality measures sound a lot like “fast lanes,” which many net neutrality proponents have steadfastly opposed. It seems less likely that proponents of a net neutrality law that regulates zero-rating would allow unregulated targeted quality control.
Additionally, history shows that such classification problems do, in fact, arise and can distort industries and ultimately cause the regulatory regime to fail.
As I wrote several years ago, the Interstate Commerce Commission (ICC) enforced a net-neutrality-like regime on railroads in the early 1900s in which it was illegal to “make or give any undue or unreasonable preference or advantage to any particular person, company, firm, corporation, or locality, or any particular description of traffic, in any respect whatsoever…” Given the many different types of products being transported, a single price was not feasible. The ICC created a “classification committee” that ultimately set almost 229,000 different tariffs by the time the ICC was dismantled. Similarly, the energy regulator had 28 categories of natural gas by the time it ended.
Perhaps that red tape would be worthwhile if the rules generated some consumer benefits, but the evidence suggests they will not.
I do not dismiss outright the underlying concerns of some net neutrality proponents. In principle, an ISP with few competitors might be able use zero-rating to engage in anticompetitive practices, like foreclosing competing online services. For example, an ISP might try to benefit its own video streaming service over a competitors’. In practice, however, while ISPs have zero-rated their own services, those actions do not appear to have been particularly profitable or harmed competitors.
In 2012, then-Senator Al Franken worried that Comcast would use its on-demand Streampix service to foreclose competitors like Netflix. He believed that Comcast “will almost certainly drive consumers to Comcast’s Xfinity Streampix, rather than other internet video streaming services.” Presumably, Comcast was trying to compete with Netflix, but we know how that turned out. Netflix continues to grow, while most people would be hard-pressed to say what “Streampix” even is. If Comcast thought it would be able to foreclose Netflix by zero-rating its own service, it clearly thought wrong.
Similarly, in 2016 there were concerns that AT&T was unfairly benefiting DirecTV’s streaming service, which it owns, by zero-rating it on AT&T broadband and HBO on wireless. Again, this strategy did not appear to benefit AT&T or DirecTV. AT&T is now selling DirecTV after losing an enormous number of subscribers.
Even if we had ample reason to worry about anticompetitive behavior resulting from zero rating, the California law’s statutory language does not mention market power or competition. It simply regulates behavior it defines as violating net neutrality on the assumption that certain behaviors are bad, consistent with those who argue that zero-rating is wrong on its face. In 2013, for example, some objected in principle when T-Mobile, then the fourth-largest mobile cellular provider, began its “Binge-On” zero-rated video service. Similarly, opponents immediately challenged the zero-rated video service MetroPCS introduced in 2011, when it had all of 2.5% of mobile subscribers. T-Mobile and MetroPCS used zero-rating as a way to differentiate their service by offering innovative products to increase consumer benefits from their mobile services and make them more competitive.
California’s law would make it more difficult to compete in this way, with no demonstrated offsetting benefits to consumers. Worse, as in the examples above, it is smaller firms that could most benefit from such service differentiation but also have the fewest resources to create a new category of zero-rated service that satisfies the regulator.
It’s also possible that aspects of the law I criticized are features, not bugs. That is, perhaps those who drafted the law believe zero rating should never be allowed, but recognize that because zero-rated services, particularly those for video streaming, are so popular with consumers that an outright ban would generate opposition. A law that allows zero rating, but only after jumping through as-yet undefined hoops, might dissuade firms from trying such services in the first place.
Proponents of the California net neutrality law generally ignore the way these regulations are likely to play out over time and discount consumer benefits of zero-rating. The argument over the VA telehealth app is a small window into the regulatory battles we can expect to see if a regulator has to define categories of internet services and decide which applications deserve to be included in those categories. We’ve had those kinds of battles in the past. The enormous lobbying expenditures they generate will keep legions of lawyers employed and generate arcane rules that become more byzantine with each new type of service that regulators have to monitor and understand. As in the past, it won’t end well.