5.4.3. Democratic

Although the division between democratic and consensual models of Internet governance reform is somewhat arbitrary, for present purposes democratic models of Internet governance may be defined as those in which representation is applied as a key criterion. Most notable of these is the NTIA’s Green Paper on the future ICANN, that specified the four criteria of stability, competition, private sector coordination and representation.[1]

The high water mark in ICANN’s pursuit of the principle of representation was in the method by which five At-Large representatives on the ICANN board were selected in 2000,[2] prior to the development of the RALO model in 2002.[3] This experiment in large-scale online democracy followed the recommendations of the Berkman Center for Internet & Society at Harvard Law School, in a report to ICANN’s Membership Advisory Committee (MAC).

In response to that report, ICANN formed a Membership Implementation Task Force to build a broadly constituted base of at least 5000 individual members to participate in the inaugural elections. Registration of members required a name, email address and verification by postal mail (the cost of which was defrayed by a public grant). In the end, 143 806 members registered, 76 183 of whom were authenticated by postal mail, and 34 035 of whom actually voted.

Following a review by yet another ICANN-formed committee, the At-Large Study Committee, in 2001, this direct election model was abandoned, on grounds that “such an approach is administratively and financially unworkable on a global scale for a sizeable electorate, and fraught with potential dangers ranging from capture to outright fraud.”[4] In the place of direct election, the committee recommended a model that would have seen only the holders of domain names as at-large members of ICANN.

This recommendation was however trumped by the still more damning comments of ICANN’s then President M Stuart Lynn,

that the concept of At Large membership elections from a self-selected pool of unknown voters is not just flawed, but fatally flawed, and that continued devotion of ICANN’s very finite energy and resources down this path will very likely prevent the creation of an effective and viable institution.

However Karl Auerbach, the director chosen to represent North America in those 2000 elections, considers that democratic election remains a model worthy of serious consideration by Internet governance bodies. He states:

We could easily model elections for representatives on internet governance bodies on the elections that are held among shareholders of publicly held corporations. These are usually done over the internet or via paper mail. These are inexpensive and technically easy to administer.[5]

The problem is that the shareholders of a corporation are a fixed and ascertainable class of electors, but the stakeholders of a governance network are not. This opens the door to mischiefs such as the manipulation of a large bloc of voters who, although formally qualified, would not otherwise have voted, to support a particular candidate. This, and more rudimentary methods of election fraud such as multiple voting under false identities, were suspected of having occurred in the ICANN elections.[6]

The way in which CGI.br has tackled such problems in the elections for its constituencies is through the use of an electoral college constituted by relevant representative organisations. Although a better solution than individual direct election, this does have a distorting effect in that the votes of organisations with large memberships are not weighted differently to those of small organisations. Unless accountablity is strictly maintained, it also introduces the danger of democratic deficits emerging through the interpolation of layers of representation that distance the polity from the grass roots. Finally there is also still some potential for electoral fraud, illustrated by the fact that the use of GONGOs and other front NGOs within intergovernmental fora, including WSIS, is notorious.[7]

Despite these problems, at worst ICANN and CGI.br “can be seen as ‘pilot projects’ to explore the feasibility of new policy mechanisms which go beyond the traditional governmental top-down system,” utilising “new principles in global policy-making like bottom–up coordination, rough consensus, openness and transparency.”[8] Beyond adherence to these principles, all else that is really needed to constitute them as democratic governance networks is a structure that adequately fulfils the democratic principle of consent. In the absence of a sound method of representing stakeholders proportionally in their policy development structures, this means the development of adequate mechanisms for democratic deliberation within those structures.



NTIA, Request for Comments on the Registration and Administration of Internet Domain Names (1997). These distilled six “appropriate principles” earlier specified in the NTIA’s Notice of Inquiry.


See ICANN, Membership Implementation Task Force: Call for Expressions of Interest (1999).


ICANN, Evolution and Reform Committee’s Final Implementation Report and Recommendations (2002)


At-Large Study Committee, Final Report on ICANN At-Large Membership (2001)


Auerbach, Karl, Stakeholderism—The Wrong Road For Internet Governance (2006), 4


At-Large Study Committee, Final Report on ICANN At-Large Membership (2001)


Cukier, Kenneth N, The WSIS Wars: An Analysis of the Politicization of the Internet (2005), 161


Kleinwächter, Wolfgang, Global Governance in the Information Age: GBDe and ICANN as “Pilot Projects" for Co-regulation and a New Trilateral Policy? (2001), 3