Contractual Omnipotence

I’ve concluded that one of the central fault lines in the network neutrality debate is over the extent to which physical ownership of a data pipe gives an owner the practical ability to exert fine-grained control over the use of that pipe. There’s an implicit assumption on the pro-regulation side of the debate that if AT&T owns your DSL line, then it has the physical ability to, say, prohibit you from watching online videos or require you to use their email or VoIP services. Lessig and Lemley, for example, made this point repeatedly in their 2000 paper without ever explicitly justifying it. For example:

Under the design proposed by the cable broadband, AT&T and Time Warner affiliates would have the power to decide whether these particular services would be “permitted” on the cable broadband network. Cable has already exercised this power to discriminate against some services.

This is backed up by a footnote citing various restrictions mentioned in @Home’s terms of service. But as I noted previously, the fact that @Home’s terms of service formally prohibited some category of network activities did not mean that, as a practical matter, users were unable to take advantage of that service. To the contrary, it’s extremely common for users to use their network connections in ways explicitly prohibited in the terms of service, and ISPs have struggled to crack down on those who do so. The fundamental issue is that classifying traffic is a very hard problem, one that almost certainly can’t be solved in the general case. Which means that any automated filtering regime can be circumvented. And of course, having human beings monitor every user’s traffic and impose restrictions on those who violate the terms of service would be far too labor-intensive to be worth the trouble. So ISPs are forced to resort to extremely crude tactics to accomplish their filtering goals, and these tactics, in turn tend to produce both a lot of bad PR and the emergence of new, more sophisticated evasion tools. In the long run, it’s not at all clear to me that this is a battle ISPs could win, even if they had free rein to implement any policies they wanted without fear of regulatory intervention.

I think this belief in the omnipotence of paper contracts is the private-sector counterpart to the belief in the omnipotence of legislation in the political realm. Libertarians are constantly pointing out that laws restricting drugs, prostitution, “price-gouging,” file-sharing, immigration, “predatory lending,” campaign spending, gambling, etc won’t make those things go away. Rather, it will set off a complex chain of cause and effect in which the state spends resources trying to punish those activities while private citizens look for ways to evade the restrictions. When analyzing legislation, you have to not only ask whether the legislative goal is desirable, but also whether the legislature has the practical ability to mandate that goal, and what the costs and unintended consequences of doing so might be. In many cases, even if the result is desirable, the state may not have any practical means for achieving it.

Precisely the same applies to an ISP deciding whether to “legislate” a new routing policy for its network. Putting a clause in a TOS that says “no servers,” “no video,” “no VPNs,” or whatever doesn’t magically banish those things from it network. All it does is commit the companies to spend resources in an attempt to detect and/or block those activities from the network. Just as with legislative bans, users will react to this “legislation” by finding ways to circumvent it. As a result, enforcing contractual terms can be expensive, intrusive, and may produce unintended consequences. In many cases, a company may wish to prohibit some category of online activity but nevertheless conclude that there’s no practical manner of doing so, or that the costs of doing so (in labor/equipment, bad PR, lost customers, accidental blocking of legitimate traffic, regulatory intervention) aren’t worth the benefits.

As far as I can tell, no one on the pro-regulatory side has taken this limitation into account. They seem to blithely assume that absent government intervention, network operators will have fine-grained control over their users’ online activities, with the ability to block applications and content they disapprove of at minimal cost. As far as I can see, this is false as a general proposition, and that’s an important thing to keep in mind when evaluating arguments for new government regulations designed to protect users from their ISPs.

SHARE: 8 comments