Comment on Draft Initial Report on Criteria for Designating Subsequent .net Registry Operator

[Bill Adkinson <WAdkinson@xxxxxxx>]

From:  William F.  Adkinson, Jr.

Re:  Draft Initial Report on Criteria for Designating Subsequent .net
Registry Operator

Date:  June 18, 2004

A.  Introduction

The Progress & Freedom Foundation ("PFF") is a market-oriented, private
non-profit, non-partisan research institution established in 1993 to study
the digital revolution and its implications for public policy.  PFF's
underlying philosophy combines an appreciation for the positive impacts of
technology with a classically conservative view of the proper role of
government and its tendency to reach beyond its legitimate functions in ways
that harm consumers and slow progress.  PFF has previously described the
benefits of relying on competition rather than regulation for the governance
of domain name system (DNS) services, and questioned the need for regulation
by the Internet Corporation for Assigned Names and Numbers (ICANN).  

PFF hereby submits these comments on the draft report, titled "Dot net
subcommittee draft report version 6" (the "Report").  The report was
developed in response to a request by ICANN's vice president for Policy
Development to the Generic Names Supporting Organization (GNSO), seeking a
"consensus statement defining criteria and conditions to be applied in the
selection of a registry operator." The GNSO Council appointed a subcommittee
and charged it with "expediting" the report.  

B.  What Function Should Criteria Perform?

Ultimately, the criteria are to be used to "determine [which applicant] is
best qualified to perform the registry functions..."  Criteria thus should
be relevant to predicting how an applicant will in fact operate the
registry.  As the .net Registry Agreement states, "relevant experience" and
"demonstrated ability" of the applicant should be key considerations.
Indeed, evidence of direct experience should far outweigh planning
documents.  And since ICANN is seeking the best-qualified applicant, simply
setting minimum thresholds that the applicants must meet will not suffice.
Rather, ICANN must consider the fully demonstrated ability of each applicant
to serve user needs and ensure stability.

Moreover, ICANN must do this in a transparent and objective manner.  This
means at a minimum that ICANN must clearly state the decision criteria it
will use, and their relative importance, in specific and concrete terms.
This will provide a transparent process in which applicants can design their
proposals with full information on how they will be evaluated.  It will also
help make application of the criteria accurate and fair.

Of particular importance to this development of criteria is the fact that
this will be the first time that ICANN will be choosing a registry operator
where one of the applicants is the incumbent.  The criteria should strongly
favor the incumbent operator where the incumbent has performed well, and
disadvantage the incumbent where he has not.  Past operation of the registry
in question is uniquely relevant evidence of future operation.  Moreover, as
discussed below, favoring well-performing incumbents is very important to
provide adequate incentives for investment by registry operators.  In this
regard, the experience of the United States Federal Communication Commission
(FCC) in granting limited term broadcast licenses, and in particular its
policies regarding "renewal expectancy" to be afforded incumbents, may prove
helpful. 

B.  FCC Treatment of Incumbent Licensees in Comparative Hearings

In the United States, awards by the government of limited term licenses for
certain uses of spectrum, e.g. by television or radio broadcasters, have
long been common.  As a result, the FCC and the U.S. courts have often faced
situations where an existing license is terminating and the current licensee
is pitted against other applicants seeking the new license.  Both the FCC
and the courts have recognized the important public policy reasons for
factoring in a strong "renewal expectancy" where the incumbent has a good
record of service.   They have emphasized that this is not to benefit the
incumbent, but to benefit the public and the consumer. 

As the Court of Appeals for the District of Columbia Circuit explained in
Central Florida Enterprises, Inc. v. FCC, 683 F.2d 503 (D.C. Cir. 1982),
"[t]he reasons given by the Commission for factoring in some degree of
renewal expectancy are rooted in a concern that failure to do so would hurt
broadcast consumers."  The "renewal expectancy will be factored in for the
benefit of the public, not for incumbent broadcasters."  The court quoted
the FCC's presentation of the reasons for this policy with approval:

                The justification for a renewal expectancy is three-fold.
(1) There is no guarantee that a challenger's paper proposals will, in fact,
match the incumbent's proven performance. Thus, not only might replacing an
incumbent be entirely gratuitous, but it might even deprive the community of
an acceptable service and replace it with an inferior one. (2) Licensees
should be encouraged through the likelihood of renewal to make investments
to ensure quality service. Comparative renewal proceedings cannot function
as a "competitive spur" to licensees if their dedication to the community is
not rewarded. (3) Comparing incumbents and challengers as if they were both
new applicants could lead to a haphazard restructuring of the broadcast
industry especially considering the large number of group owners. We cannot
readily conclude that such a restructuring could serve the public interest.

The court went on to emphasize that each of these factors had been cited by
the Supreme Court in the FCC v. Nat'l Citizens Comm. for Broadcasting, 436
U.S. 775 (1978).

Of course, this "renewal expectancy" applied by the FCC neither creates a
right to renewal nor any certainty of renewal - the courts have insisted
that it not be raised into an "irrebuttable presumption."  Rather, the FCC's
implementation of this factor has emphasized the central importance of the
incumbent's service record:

                Where . . .  the incumbent rendered substantial but not
superior service, the "expectancy" takes the form of a comparative
preference weighed against [the] other factors . . . . An incumbent
performing in a superior manner would receive an even stronger preference.
An incumbent rendering minimal service would receive no preference.

Thus under U.S. law, incumbency is a double-edged sword.  It will greatly
benefit the incumbent who has built a strong record of service, but
substantially injure the changes of one who has built a poor one.  Or, as
the D.C. Circuit has repeatedly stated, "the only legitimate fear which
should move [incumbent] licensees is the fear of their own substandard
performance, and that would be all to the public good."

C.  Treatment of Incumbency in the .net Designation Criteria

These FCC policies identify three fundamental reasons why the criteria for
awarding the .net registry should strongly favor the incumbent who is
performing well.  First, actual performance in the role of running the
registry in question would provide a highly reliable indication of future
performance (either good or bad) relative to the "paper proposals" of other
applicants.  There could be other applicants with experience running a
registry, but the significance their past performance would depend on the
similarity between the registries.  Some applicants may have no registry
experience but will have experience in related businesses, which is likely
to have limited relevance.  Ultimately, demonstrated ability to provide the
service at issue is the best possible evidence of how an operator will
actually perform.  This must take precedence in ICANN's criteria.  

Second, the criteria should recognize that the treatment of the incumbent
will dramatically affect the incentives of current and future registry
operators.  Registry operators must make substantial investments in
facilities and personnel in order to provide service.  They would
under-invest if they believe that they are likely to lose the contract when
it is rebid.  This problem looms especially large where the contract is of
fairly short duration.  In this regard, the current term of the .net
registry agreement is limited to four years.  If registry operators must
assume that they are likely to lose their contract after four years, this
would dramatically reduce their willingness to make investments yielding
long-term returns.

Third, treating an incumbent .net registry operator as if it were simply
another new applicant could lead to "haphazard" changes in registry
operators, potentially reducing continuity and stability of operations.
Maintaining an existing operator avoids these potentially large costs, and
this should be considered in comparing an incumbent to another applicant.


Nor would criteria recognizing the value of the incumbent's proven track
record, strong investment incentives and continuity of operations compromise
the achievement of other policy goals.  The .net registry agreement does
provide that the incumbent registry "shall not acquire any right in the
Registry TLD by virtue of its operation of the Registry TLD..."  Recognizing
the full import of an incumbent's experience and demonstrated ability does
not violate this stricture.  Rather, failing to give the incumbent full
credit for the advantages he brings to the table advantages could violate
the .net Registry Agreement's requirement that ICANN not disadvantage the
incumbent.  Downplaying one applicant's advantages inherently disadvantages
it relative to the others.  

Similarly, recognizing the public interest in specifying criteria that
promote investment and product introduction by incumbents does not create
any property right in the operation of the registry.  The choice of the
successor will be ICANN's to make, and accordance with the various
contractual commitments it has made and the criteria adopted.  But ICANN
should use criteria that promote the public and user interests, and this
requires giving strong weight to good performance by incumbents.  

Nor would this undercut the "promotion of competition and maximization of
consumer choice."  First, a well-functioning registry is vital to providing
consumers with real choice of service - picking a less experienced operator
in the interests of diversity will not promote consumers' interests.
Second, while competition among registries is important, considerable
competition already exists due to the addition of new gTLDs with a diverse
set of registry operators as well as competition from operators of ccTLDs.
If further diversity is desired, more gTLDs could be added.  This is the
most appropriate mechanism at ICANN disposal for increasing diversity.  

D.  Implementing Criteria  

The Draft Procedure for Designating Subsequent .net Registry Operator, also
currently out for comment, explains that ICANN will develop and issue a
request for proposals (RFP) based in part on the criteria, applicants will
submit proposals, a team of advisors will "evaluate the proposals against
the criteria ... in the RFP" and make a recommendation, and the ICANN Board
will vote on the recommendation.  This underscores the central importance of
designing criteria that are transparent to the potential applicants, and
that can be directly applied in evaluating applications.  In particular
criteria must be transparent and operational.

        *       Criteria must be set forth clearly in the RFP, so that the
applicants can use them effectively in developing their proposals.  The RFP
must make clear how the evaluators will measure or assess the extent to
which each applicant meets each of the criteria.  In addition, it must
indicate how the relative importance of the criteria.

        *       Similarly, if the decision process is to be open and
transparent, the advisors must implement them in an open and transparent
manner.  The criteria must be sufficiently clear and unambiguous, so that
the reviewers are not left with broad discretion in measuring and applying
them.  Perhaps most importantly, the weight to be assigned to different
criteria should be clearly stated.  

        *       In order to meet these goals, criteria should be readily
operational and measurable, in terms of explicit performance
characteristics.  Because the search is for the best registry operator, the
criteria should not simply set minimum levels of performance, but also
indicate how evidence of performance above this minimum will be treated.

Endnotes

1 William F. Adkinson, Jr. is Senior Policy Counsel at The Progress &
Freedom Foundation. The views expressed here are his own and do not
necessarily reflect those of The Progress & Freedom Foundation, its officers
or Board of Directors.
2 Additional information about PFF can be found on its web site at
www.pff.org.  
3 William F. Adkinson, Jr., Domain Name Services:  Let Competition, Not
ICANN, Rule (Progress on Point, Release no. 9.21 (September 2002); Letter
from Jeffrey A. Eisenach, Ph.D., President, The Progress & Freedom
Foundation to Members of the Board of Directors, Internet Corporation for
Assigned Names and Numbers (October 27, 1999).
4 Report at 1.
5 .net Registry Agreement § 5.2.4.
6 .net Registry Agreement § 5.2.4.
7 The report indicates that it will "tak[e] account of any elements from the
dot org re-assignment where relevant."  The .org reassignment does not offer
relevant guidance to applying criteria to an incumbent, and so will not be
helpful in this regard.  
8 Central Florida Enterprises, 685 F.2d at 507 (emphasis in original).
9 Id. at 507
10 Central Florida Enterprises, 683 F.2d at 507 (quoting Cowles
Broadcasting, Inc. 86 F.C.C.2d 993, 1013 (1981)) (emphasis in Central
Florida Enterprises)
11 Cowles Broadcasting, Inc. 86 F.C.C.2d 993, 1012 (1981).
12 Central Florida Enterprises, Inc. v. FCC, 683 F.2d at 508.
13 See .net Registry Agreement § 5.1.4.
14 The .net Registry Agreement's provides that "neither the procedure
established in accordance with subsection 5.2.1 nor the fact that 'Registry
Operator' is the incumbent shall disadvantage Registry Operator in
comparison to other entities seeking to serve as the successor Registry."
See .net Registry Agreement § 5.2.2.
15 See .net Registry Agreement s 5.2.5.

                Respectfully submitted,

                William F.  Adkinson, Jr.
                Senior Policy Counsel
The Progress & Freedom Foundation
1401 H Street, NW, Suite 1075 
Washington, D.C. 20005 
Phone: 202-969-2945 
        Email:  wadkinson@xxxxxxx