My lefty alter ego Tom Lee weighs in on the network neutrality debate. After an excessively generous hat tip toward my Cato paper, he focuses his criticism on Julian’s post. Tom is unimpressed by Julian’s concerns about the risks of case-by-case adjudication of network neutrality concerns.
Tom suggests that we’re looking at an unbridgeable ideological chasm, and maybe that’s true. But I want to (as Julian puts it) “go a little fractal” and highlight the structural similarity between the arguments on the two sides of the debate. As I pointed out yesterday, the Genachowski was quite right to argue that the key virtue of the Internet is that it “pushes decision-making and intelligence to the edge of the network,” which preserves “the freedom to innovate without permission.”
The most obvious problem with top-down, permission-based systems is that the decision-maker might just make bad high-level policy decisions. AT&T might, for example, decide to block VoIP traffic to protect its legacy telephony business. This is the kind of scenario that attracts most of the attention in the network neutrality debate.
But I think this isn’t the only, or even the most important, reason this kind of setup is a bad idea. Consider Larry Lessig’s work on the problems created by a “permission culture” in the copyright context. Lessig tells the story of a filmmaker who tried to get permission to show a few seconds of the Simpsons in the background of a shot in a documentary he was working on. After a bunch of calling around, he finally reached the right person at Fox, who demanded $10,000 for permission to include the clip. Else wound up digitally editing the Simpsons clip out of the scene.
The thing to note about this is that the senior leadership of Fox almost certainly did not make a conscious decision to start demanding outrageous amounts of money for the right to use trivial snippets of its content. This is probably a case of bureaucratic incompetence, not greed. Moreover, even if he’d gotten a reasonable answer, it still would have been a problem that he had to spend so much time on the phone.
The same points apply to the iPhone example I discussed in my previous post. The various problems with the iPhone app approval process aren’t (just) cases of Apple being greedy. Some of the decisions have been so transparently stupid that they can only be the result of incompetence on the part of individual Apple employees. In other cases, the problem seems to be that management guidance to rank-and-file Apple reviewers was vague, and so inconsistent results were reached. And in many cases, the problem isn’t that a bad decision was reached, but rather that it took an unreasonably long time for the answer. That’s problematic even if the ultimate answer is the “right” one.
The problem with a permission culture, then, isn’t just that the high-level policies might be bad (although of course they might). The more serious problem is the process of permission-seeking inevitably introduces unnecessary and often costly friction. Tom writes rather casually about ISPs “getting on the phone with Washington before they get on the phone with Cisco.” We could just as easily say that iPhone developers should “get on the phone with Cupertino” before they start writing a new application. For a variety of reasons, this doesn’t actually work very well. “Cupertino” may not have formulated a policy on the app you’re thinking about creating (indeed, this is especially likely if your app is unusually innovative). It may stall or give vague, evasive answers. It may give different answers to different firms with substantially similar products. And of course, an informal assurance from some Apple employee is no guarantee that the bureaucracy won’t change its mind after you’ve sunk thousands of hours into developing and tesitng your product.
All the same problems apply to a rule that says network providers need to manage their networks in a “reasonable” fashion, with the precise definition of reasonableness deferred to future case-by-case adjudication. Chairman Genachowski is not going to spend all his time taking calls from mid-level Verizon engineers seeking clarification on what this means. In practice, when a Verizon engineer wants to know what he’s allowed to do, he’s going to have to ask his boss’s boss to “get on the phone with Washington.” The Verizon executive will, in turn, wind up talking to some mid-level FCC bureaucrat, and may or may not succeed in clearly communicating what Verizon wants to do. And bureacrats—whether they work in Cupertino or Washington—have no particularly incentive to give prompt, clear answers that might come back to bite them later. So there’s likely to be a lengthy and inconclusive back-and-forth between the FCC bureaucrat and the Verizon one. By the time any sort of conclusion is reached, the engineer will probably have moved onto some other idea.
Now, I think Tom’s answer is that he basically views this as a bug rather than a feature:
If regulation is what it takes to convince Verizon that it’s selling a commodity, that’s fine by me. The market that’s going to grow up on top of that system is more important than the principle of avoiding any imposition on the ISPs. Put another way: it may be that the drab, regulated water system has cheated me of consumer-facing innovations like leased rootbeer faucets, or additives that would save me from cleaning the tub as often. But I don’t care about that: those gimmicks — any water utility gimmicks — are of miniscule importance compared to having a water system that works safely, predictably and efficiently. If Comcast needs to become a more boring place to work and a slightly less exciting business than it otherwise might be, it’s not going to bother me even a little bit.
Now, on the technical question here I think I’m closer to Tom than Julian is. I want Comcast to provide me with a cheap, fast, “dumb” pipe, and I can’t think of any “premium” services I’m particularly itching for Comcast to provide to me. Moreover, I suspect AT&T and Verizon will be able to take care of themselves in the regulatory arena: the process may be wasteful, but if they really want to do something, they can probably persuade the FCC to let them do it.
But we’re not just talking about regulating cable and telco incumbents. We’re talking about rules that will apply to wireless in addition to wired infrastructure, and to new firms as well as incumbents. On the wireless side, carriers are facing genuine and serious bandwidth constraints. There are a number of ways to deal with bandwidth limitations: you can meter by the bit, impose a cap on total bandwidth, deliberately throttle each user below the full capacity of the network, or some combination of the three. You can also use a variety of indirect mechanisms to limit bandwidth consumption. The iPhone is an example of this: Apple basically conserves bandwidth by declining to approve the most bandwidth-hogging applications.
I honestly have no idea what “reasonable network management” means for a product like the iPhone. I think you could make the argument that consistency requires the FCC to order Apple to re-design the iPhone to make it an open platform. That strikes me as a bad idea for a variety of reasons that are probably best left for a future post. But the more fundamental point is that I don’t want a future in which Apple has to “get on the phone with Washington” before it makes architectural changes to the iPhone. Even if we assume the FCC will craft a perfectly rational policy, the very process of asking permission is going to impede the rate of progress.
Finally, I think it’s naive to assume that network providers will always be big, lumbering companies, or that network neutrality regulations will never be turned into a weapon against challengers. I’ve written at length about the story of MCI, whose entry into the long distance market the FCC delayed by a decade. Presumably, Congress could not have foreseen in the 1930s that the rules they were enacting would have an anti-competitive effect a quarter century down the road. But the fact that some industry is currently dominated by a handful of huge corporations does not mean that it will always be so, and we should therefore be cautious about the danger that such an assumption could become a self-fulfilling prophesy.
The “permission culture” frame is an interesting way, but it cuts both ways: in a world where network operators are free to treat applications in unpredictable ways, the authors of envelope-pushing network applications must essentially seek permission from every network operator that their app may encounter. I think that this is a vastly larger potential drag on innovation than telling network operators that envelope-pushing network management must be cleared with a single central authority. And I think you may be overstating the number of these check-ins — it’s likely to be classes of interventions that would need to be cleared with the FCC, not individual instances. This sort of pre-action contact between rule-making agencies and the organizations they regulate isn’t at all unusual, and in most cases seems to work perfectly fine.
To back Tom up a bit, and modify or enhance the “permission culture” framework, I would add “truth in advertising” to the list of concerns about effective network regulation. I think the Internet so far has fostered a useful consumer assumption: that all data packets are created equal. The two existing natural-monopoly incumbents in any region (phone and cable) have benefited from tighter control of data priority in the past, and they would like to return to that past. Telcos want to go back to the days when /their/ voice data took priority over any other type of traffic. Cable wants to go back to the days when /their/ pre-ordained video data took priority over all other traffic… that is, all traffic was one-way from only pre-chosen content providers to pre-paid consumers, with no real choice given to the consumer outside a few “bundle” rates. Only one side in these data transactions (the distributor) has any real control or choice, and control-monopoly rents are still extracted from each end.
We still let the incumbents have their old monopolistic methods of data control, but they have to be truthful and say one type of data is “phone,” another type is “TV,” and the new type is “Internet” — all 3 are thus managed and billed as separate items. We consumers accept that they have obvious controls over phone and TV service types, that they don’t have over the Internet. They want to have control over all 3 and still call it all “the Internet,” but I contend that would be false advertising. They can now take this control surreptitiously, as proven by the Comcast-BitTorrent case (among others, and we can’t always depend on non-profits like EFF to suss these cheats out). Some *transparency regulation is required* to keep them from lying about this. If they’re not treating my chosen packets as equal to theirs, I don’t think they should be allowed to claim they’re selling me “Internet access.” Sure it’s network access, but it’s not *true Internet access.*
Doing nothing guarantees they would get away with this false advertising, with no penalty whatsoever. Without equal regulation on all ISPs, the most sneaky and unethical ISP wins the profit-motive-only/screw-dumb-customers game every time.