ANALYSIS OF INTERNET BY-LAW PROPOSALS

["Hans K. Klein" <hans.klein@pubpolicy.gatech.edu>]

                     INTERNET BY-LAWS:  
         AN ANALYSIS AND CRITIQUE OF THREE PROPOSALS

                            by 
                        Hans Klein

     Computer Professionals for Social Responsibility (CPSR)
  Public Policy Track Co-chair, INET-95, INET-96, INET-97, INET-98
     School of Public Policy, Georgia Institute of Technology

                     September 26, 1998


As the Internet governance stakeholder process approaches its September 30
deadline, we have seen three different versions of by-laws proposed for the
new corporation ("Newco.") These three proposals are "IANA-NSI" (the August
17 joint proposal of those two players), the "Boston Group" (a group of
technical lawyers out of Boston), and "EFF" (the Electronic Frontier
Foundation).  (Their web addresses are at the end of this document.) These
by-laws are difficult to understand, and the implications of their different
features are not obvious.  In what follows I analyze and contrast them to
make them understandable, and I critique them on the basis of the values
embodied in CPSR's principles for Internet governance.

The three proposals can be understood in terms of how they achieve basic
functions of governance.  Governance functions are similar for both nation
states and private firms, so they should be familiar to most
newspaper-reading citizens.  The governance functions considered below
concern: 1) representation, 2)transparency, 3) rights, and 4) finances.  I
compare how the different by-law proposals perform these different functions.

Representation
==============
Newco's Board of Directors can be compared to the Congress of the United
States.  Both are both the top authority for their respective "societies."
A lot of thought has to go into ensuring that the top authority effectively
represents those who are governed.  Representation in the US government is
accorded by certain categories.  These include: 1) geography (seats are
reserved for representatives from different localities,) and 2) size
(Congress has two chambers, one favoring populous states and another in
which little states have equal power.)  These categories of representation
ensure that different kinds of states all receive representation.  Different
Newco proposals also have different categories for representation.

The IANA-NSI proposes three categories of representation: function,
geography, and membership.  Functional representation is accorded to Board
members by reserving Board seats for three Supporting Organizations, the
Address SO, the Domain SO, and the protocol SO. Much like you will find the
"California delegation" in the US Congress, you would find the "Address
delegation" on Newco's Board.  Each SO would control three Board seats, so 9
(3+3+3=9) of 19 seats are functionally-defined.

Geographic representation is facilitated (but not guaranteed) by the
definition of 6 "geographic regions."  Really, IANA-NSI only guards again
excessive concentration:  no one region can have more than 11 of the 19
Board seats (2+2+2+1+4=11).  That doesn't ensure that places like Africa or
Central America will have any representation, but it does ensure that a
region like North America will not control two-thirds of the Board.

Membership representation is not explicitly defined, but is expected to be
defined by Newco's first, Interim Board.  So-called "At Large" Board members
will be elected by some kind of membership (where the definition of
membership remains undefined and is perhaps problematic.)  If we think of
the SO's Board members as "insiders" with a financial interest in Newco (as
many stakeholders seem to think) the At-Large Board members will represent
more "outsiders" or end users.  There will be 9 At-Large Directors.  (The
president/top officer of Newco will also be on the Board, so we get
3+3+3+9+1=19 Directors.)

So that is how IANA-NSI proposes to handle representation.  What about the
others?  Interestingly, EFF proposes exactly the same thing.  EFF has not
concerned itself too much with issues of representation.  Its concern is
with rights (below).

In contrast, for the Boston Group representation is the main thing.  They
are especially concerned that Newco "insiders" -- the Supporting
Organizations and the President -- will attain too much influence.  To
prevent this the Boston Group's proposal eliminates those 9+1 insider Board
members, leaving only the 9 At-Large Directors elected by the membership.
(This small Interim Board could be increased later.)  In the Boston Group's
version the Board represents only users.  Thus this proposal does away with
functional representation by the SOs entirely.  The SOs have only advisory
power.

The Boston group also proposes a different mode of geographic
representation.  Instead of setting an upper limit on any one region's
representation, they endorse a lower limit, which would guarantee that the
Board include at least one director from each of the 6 geographic regions.
On a 9-member Board, that means that no region could have more than 4 of
nine seats.  Were the Board to expand, however, those 6 seats would become
less significant  (asymptotically approaching zero in the hypothetical case
of an infinite Board.) Ultimately, greater geographic concentration might
become possible with this plan than with the IANA-NSI proposal.

So, to summarize, when it comes to representation the choices so far are
between an insider/functional model and an outsider/user model.  How much
democracy do stakeholders want?

These two modes of representation are not above criticism.  Both of them
ignore some categories of representation that were proposed during the last
year.  In 1997 the Domain Name Rights Coalition proposed that the Board
represent "all major groups on the Internet, including individuals, small
businesses and entrepreneurs, political and community organizations, and
large companies (Press Release, 1 May 1997).  Computer Professionals for
Social Responsibility (CPSR) proposed that the Board "should include at
least 3 members [of 15] who represent the 'Internet Community' at large ...
[and who] should be free from direct financial interests..." (CPSR
Proposals, 30 June 1998).  These suggestions are not incorporated into any
of the three proposals. The Boston Group's emphasis on users/members goes in
this direction, but does not reserve a specific number of Board seats for
any type of user.

The by-laws could define categories for Board members, such as "for profit"
and "educational/non-profit."  Language could be included to ensure that a
certain number of Board members correspond to each category. (More on this
below.)

Transparency
============
It is important that people know what their "governors" are doing.  For this
reason many states, notably California, have "sunshine laws" that prevent
public business from being conducted in secrecy.  Newco proposals differ in
their emphasis on transparency.

The original IANA-NSI proposal favored corporate confidentiality over
transparency.  Some meetings were to be closed, minutes would be distributed
with some delay, etc.  Overall, Newco could operate with a fairly high
degree of discretion.

Both the Boston Group and EFF reacted strongly against this, proposing
procedures for a high degree of transparency.  EFF included the most
specific language, adapting language from the California "sunshine laws" for
Newco's charter.  Thus "every portion of every meeting of every [Newco
body]" shall be open to public attendance, and "minutes ... shall be made
publicly available within one day."  In this way Newco's activities are
known to all.

Rights
======
The Constitution of the U.S. defines a number rights that the Congress may
not abridge, most notably the First Amendment right to free speech.  As a
private entity Newco would not be as strictly constrained by the First
Amendment as is the U.S. government (and its current agents, NSI and IANA).
Therefore, in the transition to Newco some protections of speech could be lost.

EFF, in keeping with its civil liberties orientation, makes this a
centerpiece of its proposal.  Rather than referring to U.S. law, EFF
incorporates a reference to international law, The U.N. Universal
Declaration of Human Rights.  This reads: "Everyone has the right to freedom
of opinion and expression; this right includes freedom to hold opinions
without interference and to seek, receive and impart information and ideas
through any media and regardless of frontiers."  Neither IANA-NSI nor the
Boston Group proposed special measures to preserve freedom of speech.

It is my guess that the U.N. declaration provides even stronger protections
than does the U.S. First Amendment.  If that is the case then under the EFF
proposal communication rights would be even stronger than under the current
U.S. regime.

Finances
========
Money plays a big part in governance.  To prevent abusive taxation, the U.S.
founding fathers placed the power of taxation in that house of Congress that
was the most representative of the citizens, maximizing the relationship
between taxation and representation.  Today money remains a big issue. A
majority of Congress current supports campaign finance reform that would
limit the influence wealthy groups to "buy" or at least influence elections.
The regulation of campaigns is intended to reinforce the principle of "one
[person] one vote" and weaken the principle of "one dollar one vote."
(Campaign finance reform in the U.S. failed, by the way!)

All Newco by-law proposals also address money issues, and as far as I can
see, they are all adopt the language proposed by IANA-NSI.  Concerning fees
("taxation," if you will,) Newco will charge enough to cover its expenses
and keep some money in the bank.  Undoubtedly, the big business players who
will pay these fees will keep a close eye on them.

One allowable expense for Newco is to pay the costs that Directors incur
from Board service.  This means that personal or organizational wealth is
not an absolute prerequisite for Board service, a set-up that makes it
easier for non-profit organizations to participate on the Board. 

All three Board proposals contain the provision that the "resources of the
Corporation will not be expended in support of any nominee's campaign."
However, there are no other restrictions on campaign financing.  This
provision could favor wealthier individuals and members who can expend
considerable funds in election campaigns.  That is what happens in public
elections.

The final by-laws might want to include funding and expense provisions that
level the playing field for rich and poor Board members.  Allowable expenses
could include reimbursement of the value of a Board members' time expended
on Newco business.  Campaign spending could also be limited.  An upper
spending limit might be defined, or candidates could be limited to
campaigning via a common web location on the Newco web site.

Finally, start-up funding for Newco could present tricky issues.  The money
needed for start-up could be substantial.  Currently some particularly
wealthy stakeholders have offered to donate money to cover its initial
costs.  At the very least such funding would create the strong impression of
industrial dominance of Newco through this financial dependency.  If Newco
is to get off to a clean start, a neutral start-up funding source should be
identified (e.g. a bank loan against future revenues.)


Odds and Ends
=============
Pre-existing contracts: The IANA-NSI proposal contained a controversial
provision that would bind Newco to pre-existing contracts.  The EFF proposal
does away with this.

Location: All proposals foresee the corporation being incorporated in Los
Angeles County.  This may contribute to continuity of staffing in the
IANA-Newco transition, in efffect "bringing the mountain to Mohammed."  On
the other hand, a new location could help make a clean break with the past.
Some have also said that California's laws of incorporation provide fewer
free speech protections than do those of Deleware.


Conclusions and Recommendations
===============================
If these three proposals express the range of stakeholder opinion in the
Internet community (and this is a big "if"!), then consensus may not be far
away.  The three proposals have more in common than they differ.
Reconciling them into one set of by-laws that can garner consensual support
may well be possible.

The Boston Group and EFF's proposed safeguards on transparency could be
adopted.  These seem likely to be preferred by most stakeholders, who will
ultimately be outsiders looking in at Newco's activities.  These provisions
are also consistent with the June 1998 White Paper's original principles.

EFF's proposal on rights also seems likely to attract wide support.  It is
consistent with the libertarian philosophy of many stakeholders
participating in the White Paper process.  Even those stakeholders concerned
exclusively about management issues might approve of this provision: a
non-negotiable commitment to non-interference in communications might help
keep Newco out of disputes over content control.  On the other hand, free
speech protection may be anathema to trademark holders, who own many of the
words we use.  (Don't have a cow!)

The fundamental difference between the three versions is the definition of
representation and the associated problem of membership.  IANA-NSI and EFF's
proposals for functional representation give considerable power to
"insiders" in the SOs.  If this is something to be avoided, something like
the Boston Group's proposal is needed.

As described above, the Boston Group makes Newco completely accountable to
users/outsiders, granting the SOs only advisory status.  But this proposal
suffers from problems of representation.

The first problem of representation concerns geographic equity.  The Boston
Group's proposal only guarantees one seat per region -- even though the
Board size may vary.  As the Board size increases, the significance of a
single seat declines.  Instead of guaranteeing one seat to each region, the
by-laws could set minimum and maximum percentages of Board seats.  For
example, the by-laws could mandate that no region have less than 10% of
Board seats (1 out of 9 seats, 3 out of 30) or more than 45% (4 out of nine
seats, 13 out of 30).  Alternatively, were the Board size to be fixed,
distribution could be guaranteed with specific numbers (as in the numbers
just mentioned.)  Currently, a guarantee of one fixed seat is unsatisfactory.

The second problem of representation concerns diversity.  The Boston Group's
Board size of 9 is too small to allow diversity (although the final Board
size is not specified.)  quantitative changes in Board size have qualitative
impacts:  a bigger Board has more seats with which to allocate, so it can
accommodate more diverse constellations of user types.  On a small Board, a
single Director may have to represent all telecommunication firms combined
or all non-profits combined or all of Europe combined.  On a larger Board
Directors might represent more distinct user types.  For a body that hopes
to represent all the world's Internet users a larger Board might well be
appropriate.  What number? The Internet Society's governing body has 20
members.  For Newco a bigger number, say 30, could be appropriate.

The final problem of representation concerns sectoral equity.  If no
differentiation is made between user types, then one type may come to
dominate the Board.  This type would probably be for-profit firms who have
an immediate financial interest in Newco's activities.  To prevent this the
by-laws could define user-type categories.  For example, the by-laws of the
Internet Society include three categories:  industry, education/non-profit,
and government.  Since Newco is to be a private firm, its Board categories
could be just industry and education/non-profit.

Each Board category could have a percentage of seats reserved for it.  This
has a long tradition in Board design.  For example, back in the 1960s when
the COMSAT Corporation was created, 20% of its Board seats were reserved for
representatives accountable to the American people (these Board members were
appointed by the President.)  Other Board members were accountable to their
respective company stockholders.  What percentages could be reserved on
Newco?  Like COMSAT, should it be 20%? Or is the Internet more of a
"people's technology" than a satellite?  Maybe one-third of the Board could
be reserved for non-profit and educational institutions.

Even if the IANA-NSI and EFF "insider" model were preferred over the Boston
group's "outsider" model, these Board categories could apply.  Percentages
of At-Large members and SO-based members could be reserved for industry and
for educational/non-profits.

In summary, the three by-law proposals currently in circulating could be
combined into a single consensual set of by-laws.  The combined set of
by-laws could adopt the highest standards for transparency.  They could
guarantee individual rights.  And they could include categories to ensure
that both industry and non-profits receive representation in sufficient
number to ensure diversity.

        Boston Working Group	http://www.cavebear.com/bwg/
        IANA and NSI 		http://www.iana.org/intro-coop.html
        EFF			http://www.eff.org/
        CPSR			http://www.cpsr.org/onenet/

* * * * * * * * 




===========================================================================
  Hans K. Klein
  Assistant Professor               Phone: 404-894-2258
  School of Public Policy, MC:0345  Fax:   404-894-0535
  D.M. Smith Building, 313          Email: hans.klein@pubpolicy.gatech.edu
  Georgia Tech
  Atlanta, GA 30332-0345
===========================================================================