Thursday, August 28, 2008

My Masters Thesis... done (download the PDF).

The Business of Broadband and the Public Interest:
Media Policy for the Network Society

[Update: Adam Thierer posted a response. My critiques of his past (and current) positions can be found at pages 8 and 72 of my thesis.]

Saturday, August 2, 2008

A Reply to Nachbar, Part 2

Last week, I posted a critique of Thomas Nachbar's latest paper. That post primarily addressed his general argument against "use"-based non-discrimination rules. In this in-progress excerpt from my thesis, I critique his specific argument against the openness conditions on the recent 700 MHz "C"-block spectrum auction.


The Commission’s final order struck a compromise on many of the issues at stake.[1] The “open devices”/”open applications” rules were adopted, while the more aggressive openness proposals were not incorporated. When bidding concluded months later, it was revealed that Google had indeed bid up to the reserve price on the “C” block but that Verizon had cast the winning bid. Telecommunications analyst Blair Levin quipped that because of the openness conditions on Verizon, “Google is the happy loser.”[2] Of course Google had a clear business interest in maintaining users ability to access their services. Likewise, the wireless carriers perceived a business interest in retaining the ability to control what users could do on their networks. While the news reports were dominated by analysis of which big company really “won,” many missed the more fundamental public interest issues at stake.

The openness conditions in many ways mirror traditional non-discrimination public interest safeguards. The conditions seek to preserve the freedom of users to use the network as they choose, and to access it with any device that did not cause harm to the network. The former resembles a weaker form of Computer Inquiries application non-discrimination, and the latter condition mirrors the Carterphone decision of 1968.[3] There are many potential loopholes in the rules. Indeed, no sooner had the rules been decided than Verizon began lobbying for a weak interpretation and Google began counter-lobbying.[4] I described earlier why I think that the combination of common carriage, the Computer Inquiries, and Carterphone were necessary for an environment that fostered the flourishing of early consumer internet access. In the wireless context, I believe that similar flexibility of use is essential to maintaining historical non-discriminatory access in this new medium, as well as preserving the internet ethos that has led to innovation and free speech online.

Thomas Nachbar believes that defining this use neutrality is too difficult, that regulators will tend to define in such a way that it constrains innovation, that the rules will not affect positive behavior anyway, and that the competitive market will better solve any concerns.[5] Undoubtedly, the “openness” conditions in the 700 MHz auction were defined at a high level, and were a result of political compromise. Of course, the Commission has long promulgated broad principles or rules to guide industry behavior and then specified particular guidelines or adjudicated on individual bases.[6] Nachbar goes on to claim that the rules were defined, “in a specific, technologically dependent formula,”[7] and that “imposing use neutrality requires addressing questions of design.”[8] This claim is hard to understand, given that the mandate to allow all devices and applications is clearly divorced from particular technologies and indeed is designed to open the possibility to unforeseen technologies. This is the heart of technology-agnostic network modularity. Nachbar would also have us believe that the rules represent only a weak form of Carterphone, which by itself will be ineffective.[9] This ignores the full implications of the open applications provision, which extends the non-discriminatory mandate into the network.[10] It appears that Nachbar and I agree that two-sided openness (user device and network access) would be necessary to encourage meaningful openness, but that we disagree as to whether this can be done through wireless use neutrality.[11] Nachbar instead sees promise in profit-motivated market actors. He makes much of the somewhat competitive wireless carrier market.[12] However, it is clear that carriers all share similar incentives to discriminate against content, and that there is no competitor that offers comparable non-discriminatory service. AT&T recently stated explicitly that its wireless network does not respect network non-discrimination, and that its terms of service – “which are similar to those of other wireless providers” – categorically prohibit all peer-to-peer use.[13]

Ultimately, Nachbar’s critique of the 700 MHz “openness” rules focuses almost entirely on competition-based analysis of the rules (which, even on its own terms, I consider to be deeply misguided). Missing from his analysis is any consideration of whether the 700 MHz use neutrality rules map to historical non-discrimination norms. This is odd, considering his masterful exposition of these norms earlier in his paper. Ultimately, the non-discriminatory considerations in the wireless space are parallel to the network neutrality debate overall, and my conclusions here are essentially the same as my conclusions there. As with wireline, wireless operators face genuine network congestion challenges. Content and application-based discrimination is one way of dealing with these challenges. There are many other approaches – including discrimination that is not content or application-based[14] – that do not present so directly threaten free speech, innovation, and established norms.

[1] Second Report and Order. FCC 07-132. (Rel. August 10, 2007).

[2] "Verizon and AT&T Win Big in Auction of Spectrum," New York Times, March 21, 2008. by Saul Hansell.

[3] Tim Wu, Wireless Carterfone. International Journal of Communication, Vol. 1, p. 389, 2007
Available at SSRN:

[4] Letter from Richard S. Whitt, Google, WT Docket No. 06-150, (October 1, 2007).

[5] Nachbar, 80-89.

[6] Elsewhere, Nachbar endorses precisely this approach. (at. 90)

[7] Nachbar, at 81

[8] Nachbar, at 88

[9] “The rules adopt a version of what has become known as ‘Wireless Carterfone’.” Nachbar, at 81.

[10] To be sure, whether or not this is the case could be disputed. Nachbar’s narrow interpretation is that the provision only limits “the ability of carriers to prevent consumers from loading and running third applications on those openly accessible devices.” (at 81). Even if the 700 MHz rules as adopted did not effectively mandate use neutrality in the network, this does not mean that the approach should be abandoned altogether but rather that such rules should perhaps be more explicitly defined. I am considerably more hopeful that it is possible to do this than is Nachbar. This is essentially the same question that plays out in the broader network neutrality debate, which I discussed earlier.

[11] Nachbar states that, “from a consumer standpoint, the product is the combination of device (or application) and carriage.” (at 82) I agree. I remain confused, however, about why he sings the praises of the Computer Inquiries while maintaining that use neutrality is categorically a bad idea. I am not persuaded by the argument that the IP environment fundamentally different from the circuit-switched environment in such a way that use neutrality is impossible or undesirable.

[12] “But the wireless markets of today are not like the wireline market that AT&T operated in years ago. Today’s wireless carriers face 2 competitors in over 90% of their markets, and therefore have far less market power than AT&T did.” (Nachbar at 82) Nachbar goes on to perform an analysis of the market incentives of wireless operators that I believe is fundamentally flawed on several accounts. He begins by noting the “internalizing complementary efficiencies” phenomenon and claiming that, “If wireless carriers actually do have market power, then opening device and application markets to competition will have no effect on their ability to charge monopoly rents.” (at 82.) Of course, neither of us thinks that wireless carriers are strict monopolists, and thus the ICE exception is irrelevant. On the other hand, these similarly situated companies sometimes resemble an oligopoly, with strong incentives to leverage market power into adjacent markets. Because they face potential competition on price, speed, and device exclusivity, they are motivated to increase switching costs and customer lock-in. I am puzzled by Nachbar’s assertion that "carriers are selling a commodified, undifferentiated service (carriage)," given the ample evidence that carriers are in fact differentiating between content, and Nachbar’s own claim that in IP carriers are motivated to differentiate in a way that they were not in the circuit-switched environment. Nachbar then claims that any market power being exercised is likely coming from the device manufacturers instead, citing the iPhone-AT&T tie-up and the fact that the iPhone has lured many customers to AT&T. Of course, one might just as easily conclude that it was precisely the distorted wireless carrier market that motivated Apple to strike the exclusive deal. In any event, despite the perennial appearance of blockbuster devices, the device market is far more diverse and competitive than the carrier market. Furthermore, the device market continues to move toward open platforms of its own accord, with device juggernaut Nokia announcing the open-sourcing of its operating system on the eve of the launch of Google’s own free and open source “Android” mobile operating system (Nokia. Press Release (June 24, 2008). "Nokia to acquire Symbian Limited to enable evolution of the leading open mobile platform" Open Handset Alliance Press Release (November 5, 2007). "Industry Leaders Announce Open Platform for Mobile Devices" What’s more, it is hard to imagine any strong device leveraging as general-purpose computers increasingly become one of the key devices using wireless internet. Despite such developments, if carriers insist on discriminatory practices, the same bottleneck to innovation remains: use neutrality of government-granted spectrum. Critics of non-discrimination mandates on wireless spectrum raise myriad concerns that such requirements restrict possible business plans. They undoubtedly do. The relevant question is whether or not this benefits or harms overall innovation, growth, and public interest.

[13] Letter from Robert Quinn, AT&T, WT Docket No. 06-150, (July 25, 2008).

[14] Geoffrey Goodell, Allan Friedman, Scott Bradner. "Scarcity, Discrimination, and Transparency: Understanding Network Management" (Paper to be presented at TPRC 2008 Conference, Saturday September 27, 2008).