The Berkman Center research staff is analyzing alternative proposals for ICANN's membership structure according to the proposals' conformity with the White Paper principles and the general structure outlined by the ICANN By-Laws. These principles are a standard against which any proposal for ICANN's membership structure must be considered. Further, ICANN's membership structure should conform to the other sections of its By-Laws. Ideally, no section should confuse the interpretation of another section. The RCS Group has been focusing its analysis on three discussion alternatives. None of these is fully developed, and they certainly do not represent the entire range of options. But taking a preliminary look at them through the lens of the White Paper principles and the ICANN By-Laws may highlight key areas for further development and research.
In response to this situation, the U.S. government issued the Green Paper (available at http://www.ntia.doc.gov/ntiahome/domainname/022098fedreg.htm). After receiving comments and criticism from many quarter, the Clinton Administration issued the revised White Paper proposal. The White Paper calls for the creation of a private, non-profit corporation to handle four primary functions in a coordinated, centralized manner:
The White Paper also sets forth four key principles to guide proposals for the new corporation:
These four White Paper principles have informed the drafting of ICANN's By-Laws and Articles of Incorporation, and must be considered when formulating proposals for ICANN's at-large membership structure.
B. Stability: The WP stresses the value of continued Internet reliability. A standards-making body like the ICANN Board will, presumably, set better policies and adopt better standards when parties who understand the underlying technology are diligent participants. A successful plan will provide a significant incentive for infrastructure providers (telcos, ISPs, networking equipment makers, etc.) to get involved in the ICANN process. Plans that allow providers to vote will provide this incentive in greater measure than those that only allow providers nonvoting participation; plans that allow nonvoting participation are more likely to draw infrastructure providers than plans that exclude them altogether. Thus the "Organization" plan is more likely to draw large organizations into the dialogue than the "Individual" plan.
Supporting Organizations do provide explicitly for input from networking professionals and technical experts. So, are the SOs enough? As it stands, the equal numbers of At-Large and SO-appointed Directors could produce compromise or deadlock.
Any structure that prevents this deadlock, however, will allow either At-Large Directors or SO-appointed directors more control over ICANN policy. The White Paper offers guidance only in the principle that "The Board of Directors for the new corporation should be balanced to equitably represent the interests of IP number registries, domain name registries, domain name registrars, the technical community, [ISPs] , and Internet users (commercial, not-for-profit, and individuals) from around the world." Voice: Most corporations and associations may not qualify for membership in an SO, though many would argue that they do not need group representation, because they can influence individual voting decisions. The RCS "Individual" model is predicated on this assumption. Yet this means of group representation, even if effective, does not necessarily assure that corporations' and associations' distinct "voice" will be heard. The White Paper does use the word "equitable," rather than equal. The "Individual" model's principle of "one individual, one vote; one company, no vote" may not provide sufficiently direct representation to be "equitable." The "Organization" represents the other end of the spectrum; there, corporations may vote as a class and retain the opportunity to influence individual behavior.
C. Private, Bottom-Up Coordination
Funding: The WP notes that "the new corporation's activities would need to be open to all persons who are directly affected by the entity, with no undue financial barriers to participation or unreasonable restrictions on participation based on technical or other such requirements." The first part of this requirement, "undue financial barriers," raises the question of how high the membership fee may be without being "undue." For example, the "Organization" model's $5000 minimum membership fee for corporations might exclude many smaller e-commerce businesses, who would nonetheless have a financial stake in ICANN decisions.
The second part of that WP provision, "restrictions...based on technical or other such requirements," opens further discussion. If ICANN requires even minimal online access of its members, it may exclude stakeholders who, though they do not use the Internet, are still "directly affected by the entity." Under even a more modest interpretation, this phrase in the WP demands that we consider the level of online accessibility required. The graphical, forms-capable browser that would facilitate online voting is not yet universal. The Berkman Center welcomes any input on the appropriate threshold of technical capability for members, Director nominees, and Board members.
Furthermore, The WP suggests that, "Once established, the new corporation could be funded by domain name registries, regional IP registries, or other entities identified by the Board." NSI and its counterparts are quite likely to continue charging customers for their service, so some might argue that Internet users should not be charged by two different arms of the Internet administration organization. However, the wording ("could be") makes this particular clause just a suggestion, rather than a hard requirement for ICANN. ICANN could be funded solely through a percentage of domain registration fees, channeled through the SOs; however, the organizations who register domain names would bear the costs of decisions made by At-Large board members.
Support of Elections: Article V, Section 9(e) of the By-Laws states, "Resources of the Corporation will not be expended in support of any campaign of any nominee for the Board." This could be interpreted to mean that ICANN cannot provide Web space as a "soapbox" to nominees, although such a provision might be a logical way to facilitate Director elections. The Board might plausibly argue that a modest apportionment of HTML server space or Webcasting time is still in the spirit of this section so long as all nominees get an equal share. The Internet community benefits when the candidates have some means of presenting their views (webcast debates, moderated email discussions, etc.), all of which require some use of impartial resources; ideally, Section 9(e) would not prevent ICANN from providing those resources.
D. Representation
Geographic Diversity vs. SO Representation: None of the Center's three models yet includes a mechanism for assuring that the elections fulfill the requirements of Article V, Section 6: that "(1) at least one citizen of a country located in each of the geographic regions listed in this Section 6 shall serve on the Board (other than the initial Board) at all times; (2) no more than 1/2 of the total number of At-Large directors serving at any given time shall be citizens of countries located in any one Geographic Region, and (3) no more than one-half of the total number of Directors, in the aggregate, elected after nomination by the SOs shall be citizens of countries located in any one Geographic Region."
Article V, Section 2 gives substantial latitude to the SOs in choosing their representatives to the Board, however: "Immediately upon the recognition of a Supporting Organization by the Board pursuant to Section 3(b) of Article VI, the Board shall request that such supporting Organization nominate three persons to be directors. Upon receipt of such nominations, the Board shall elect such persons as members of the Initial Board." Section 11 states that "If an SO votes to remove their appointed Board Member, the "Board shall vote to remove such Director." At no point do the By-Laws provide for the Board to overrule or decline a nomination from a SO. Thus four At-Large Directors from a single region would preclude the nomination of a SO-representative Director from that region, but provide no means of enforcing the Section 6 prohibition. The same problem would occur if the only Director from a given region was a SO-representative Director, and his/her SO elected a Director from another region in his/her stead.
Election Oversight Committee: Each of the three models contemplates a large number -- perhaps hundreds of thousands -- of potential nominees. One suggestion within the RCS Group for dealing with this level of complexity is to have a committee of the ICANN Board winnow the candidate pool before elections. The procedure for winnowing is, obviously, a matter that would itself have a large impact on the election outcomes; the RCS Group has not yet examined the possible procedures in any depth. This "electoral college" structure is formally a part of the "Individual" model, although it could be an option in any model.
The vetting provision creates a significant tension between democracy and efficiency. The electoral function could occupy an enormous portion of the Board's time each year, diverting the Directors from their primary duties of Internet administration. Does that prohibit granting to a nomination committee any ability to "filter"?
Capture: The White Paper is fairly adamant about making ICANN's procedures capture-resistant: "The new corporation's processes should be fair...protecting against capture by a narrow group of stakeholders.... Super-majority or even consensus requirements may be useful to protect against capture by a self-interested faction."
The White Paper sets a difficult task with this requirement. Any membership model can be subverted by special interests given the right circumstances; the best that ICANN may do is craft careful and vigilant protections into the model it eventually chooses. The "Organization" proposal allows a vote to pass when only four-ninths ((2/3+2/3+0/3)=4/9) of the total membership has actually acceded to it. The "Individual" model also could be captured through a "machine politics" technique similar to that found in large U.S. city governments if organizations funded their employees' membership fees in return for a contract that the individuals would vote in the corporate interest. The "Open" model does not contemplate any controls on who may register to be a member, so groups might capture a vote through ballot-stuffing.
Capture Through Collective Apathy: Special interests might capture ICANN more easily if the less focused parties simply lost interest and dropped out of the process. ICANN should strive to have a large and diverse membership in order to keep itself truly representative. Nothing in the By-Laws suggests a minimum number of members or nominees, and it would be possible to institute a provision requiring a certain "quorum" for membership votes to be valid.