I’m in the midst of a big writing project on network neutrality, and so I’m going to do a series of posts on papers I’ve been reading. Some of the material in these posts may find its way into the forthcoming paper. I’m going to start with “A Coasian Alternative to Pigovian Regulation of Network Interconnection,” a paper by two FCC economists, that purports to offer an alternative to the FCC’s current inter-carrier compensation regime whereby long-distance firms pay local exchange carriers to terminate calls to the LEC’s subscribers. I’m not specifically interested in telephone regulation, but Atkinson and Barnekov suggest their arguments apply to other networks as well, and they’re cited by others (including Kevin Werbach, whom I’ll discuss in a future post) in the network neutrality debate, so I thought it was worth reading.
It seems to have become trendy to label one’s policy prescriptions “Coasian,” and that’s how Atkinson and Barnekov frame their analysis. They argue that the FCC’s current compensation regime is “Pigouvian” because a government bureaucrat dictates the prices that network owners must pay each other for the privilege of interconnection. Under Atkinson and Barnekov’s alternative, the FCC would… dictate the prices that network owners must pay each other for the privilege of interconnection. But they think they have a formula that is less arbitrary than the formula currently being used, and would therefore better approach the Coasean ideal of clearly-defined property rights.
In a nutshell, when one network owner wished to connect with another network owner, Atkinson and Barnekov would have them calculate the total cost of interconnection and then split it down the middle. This total cost would not just include the costs of interconnection at the edge of the network (say, stringing fiber between their facilities) but also the increased cost imposed inside each network, such as the additional capacity one network would need to carry the other’s traffic. This total cost would be computed, it would be divided by two, and then one party would pay the other so that each bore half the total cost.
In their paper, they have a beautifully simplified model involving networks of 4-6 telephones and a dozen or so links among them. In this idealized model, it’s trivial to count up the number of additional links required to accomplish the necessary throughput and determine how much one network provider owes the other for interconnection. But of course, real-world networks have more than 6 nodes on them, and more than a dozen links. The whole issue with mandatory-interconnection schemes is determining what the fair rate should be.
As policy wonks a little older than me remember all too well, we’ve tried this before. The long, painful fight over local loop unbundling was fought over precisely this issue: calculating how much it cost an incumbent like Verizon to interconnect with a competitive local exchange carrier. The formula the FCC came up with was known as TELRIC, and it was the subject of years of litigation as incumbents fought to raise the price and CLECs fought to lower it. As my fellow TLFer Adam Thierer argued at the time, there are a host of uncertainties in computing these costs in the absence of a competitive market. I was blissfully ignorant of these issues at the time, so I don’t have any strong opinion on whether TELRIC-derived rates were too high or too low. But I do know that any claim that such costs can be computed in a straightforward manner is clearly false.
And that, ultimately, is why the Atkinson-Barnekov approach can’t be plausibly called “Coasean.” The touchstone of Coasean regulation is to clearly define property rights and then allow the parties to negotiate a price among themselves. Atkinson and Barnekov, in contrast, requires that a government bureaucracy set interconnection prices based on what would end up being an extremely complicated and subjective government formula. Property rights rhetoric aside, there’s nothing Coasean about that. The ironic thing is that we do have a Coasean interconnection scheme to use as our guide: the Internet. Its defining characteristic is that each carrier has property rights in its own network, and can choose not to interconnect with any network it feels hasn’t made a sufficiently attractive offer. So far, that approach has worked splendidly, and it hasn’t required the FCC to get involved in rate-setting at all.
This isn’t to minimize the difficulty of the problem Atkinson and Barnekov are trying to solve. The Internet is a relatively easy case because it’s already a competitive market (at least at the backbone level, which is where they focus their analysis). The challenge for the legacy telephone network is that it’s not currently a competitive market, and it’s not obvious how one would bootstrap it into one. As a result, the FCC is forced to adopt second-best regulatory approaches like unbundling and government-determined interconnection policies. Atkinson and Barnekov’s proposal is best understood as being in this vein. It may be a more elegant way to set prices than what the FCC is doing now, but I don’t see how it’s any more “Coasean” or property-rights-based.
The Technology Liberation Front is the tech policy blog dedicated to keeping politicians' hands off the 'net and everything else related to technology.