Inappropriate multi-stakeholder structure for expansive public policy domain

[Dan Krimm <dan@xxxxxxxxxxxxxxxx>]

Note:  There are many people in the Internet governance community who
consider this RFC itself to be an example of poor governance performance by
ICANN, because it has a short time line and does not provide background
information to set context, as well as being opaque as to how these
comments will be used within ICANN itself.  (It's hard to formulate useful
comments on short notice without meaningful deliberative context.)

I take these criticisms seriously, yet there is a point to be made about
multi-stakeholder processes at ICANN that I think the folks at ICANN would
do well to consider, and so I am posting them here.

-----

With regard to:

>Q5. Have there been improvements in participation and in efficiency
>of the ICANN multi-stakeholder model? What more needs to be done?


I'm still a relative newcomer to this arena, but I can make a comment about
the current setup without regard to detailed comparisons with the past.

-----

If ICANN wants to address "Internet Governance" with regard to public
policy of an increasingly general nature, then I agree with the idea that
the standard to adhere to is not private hierarchies but rather public
governance structures (with a focus on systems that are genuinely
representative in nature as opposed to authoritarian regimes).

Public governance standards of representation require that constituencies
form themselves from the bottom up, not from the top down.  (The only
top-down segmentation is geographical, not "interest-oriented".)

But the ICANN setup of various advisory groups such as GAC and GNSO (and
the multi-constituency setup within the GNSO, for example) is set from the
top, not organically emerging from the bottom.

Even in principle, this could only work for well-defined (i.e., fixed)
policy topics where a particular (i.e., fixed) segmentation of stakeholders
is appropriate to that topic.  For example, there may have been some sense
to the GNSO constituency structure in the case of narrowly technical
operational matters.  But when policy domains expand to other general
topics such as personal privacy or access to knowledge, the stakeholder
balance is suddenly out of balance.

In fact, each different policy domain has a different set of stakeholders,
and policy domains are constantly evolving as well.

Thus a top-down fiat-based method of defining multi-stakeholder
participation is irrecoverably flawed from the very start.  Defining the
mix (and voting strength) of stakeholders prior to the determination of
policy domain imposes a profound skew on the representational process, and
there is no way to fix it by using similar top-down methods -- you can't
fix it by just tweaking the number of stakeholders and/or their number of
votes, because in an expansive policy domain there will always been a
majority of policies for which that breakdown is inaccurate, and thus
inappropriate and ineffective for balanced representation.


I don't know what the stakeholder setup was in the past, perhaps there were
fewer stakeholders participating.  But increasing the number of stakeholder
classes does not necessarily improve representation overall -- it could in
fact make things worse if it increases the skew in representation.

Also, if there is "policy mission creep" that is driving an increase in
stakeholders, what may be happening is that a feasibly narrow mission that
is reasonably addressed by a fixed top-down representational structure for
that well-defined policy domain is being stretched in a way that attempts
to address a much more complex and varied policy domain for which effective
representation cannot be handled in a top-down fashion.

In short, as the mission expands from a narrow technical policy domain to a
broad/complex public policy domain, the "elegant solution" of a fixed
stakeholder structure for the initial narrow mission is increasingly
recognized as systematically non-scalable to a broader policy domain.


I think that ICANN has basically two choices in order to return to
effective governance structure and processes:

 (1) Scale back to the narrow technical policy mission with the
narrowly-effective top-down stakeholder setup, or

 (2) Scrap the top-down (multi-)stakeholder setup entirely if ICANN wishes
to address the broader public policy domain that it seems to be endeavoring
to address these days, and instantiate a truly bottom-up representational
process that follows public governance standards and allows constituencies
to form themselves organically in response to specific topics.  In time
(perhaps relatively short order), this process should merge smoothly with
other public governance processes at large around the globe, because this
scenario suggests that the distinction between "Internet" governance and
"general public" governance is artificial at best.


Trying to patch the top-down definition of stakeholder structure to address
a consistently expanding public policy domain will simply not work.  The
greater the expansion of policy domain, the worse the representation will
be, because of the previously noted dependence of appropriateness of
stakeholder balance on the specific policy topic being addressed, and the
complex multiplicity of topics and stakeholder overlaps that grows as the
policy domain expands.

Coming from a policy background, this is the essential tension I see at
ICANN at present, and I don't think it can endure indefinitely.  Governance
structure and policy scope simply do not match.  One of these two
parameters has to change in order to find a good match once again.

If ICANN wants to address genuinely public policy matters effectively, it
should have a genuinely public governance structure that allows stakeholder
constituencies to emerge organically.  There's no quick fix to avoid the
hard work here, without distorting and damaging public accountability in a
fundamental and increasingly profound manner.  You can't square the circle,
or trisect the angle, and approximation methods increasingly fail at larger
scales.