Comments on An Economic Framework for the Analysis of the Expansion of Generic Top-Level Domain Names

[Olivier MJ Crepin-Leblond <ocl@xxxxxxx>]

Dear Kurt Pritz,

This contribution contains comments about the recent Study entitled: "An
Economic Framework for the Analysis of the Expansion of Generic
Top-Level Domain Names" by Michael L. Katz, Gregory L. Rosston and
Theresa Sullivan, June 2010. downloaded from:
http://www.icann.org/en/topics/new-gtlds/economic-analysis-of-new-gtlds-16jun10-en.pdf

I am making these comments in a purely personal capacity, although a
several people from the At Large community wrote to support a longer
version of my comments which I had posted to the At Large gTLD working
group.

For the sake of clarity, I'll quote the headers (" ") used in the
Economic Framework document when commenting about specific sections.

Abstract and general comments:

No abstract.
No conclusion.
No high level summary.

"II. Background"

A long background section about the creation of gTLDs is provided.

Alas, this information contains an error: the initial Y1985 gTLDs
included .NATO and did not include .INT which was only created in March
1988.
The NATO gTLD was superceded by .INT and was eventually deleted in July
1996 or possibly earlier.
Ref: http://www.nsrc.org/oclb/msg00024.html

Also - .EDU, although being initially designed for US educational
establishments, was used elsewhere too. Acceptable Use Policy was not
strictly enforced.
Ditto for .GOV (which was also briefly used in Canada).
Even .NET had an initial AUP restricting it to Network Operators, and
.ORG for non-profit organisations. Things were not black/white in those
days, but rather in tones of grey.

I leave it to others more familiar with ICANN's history to review the
ICANN part of the story (successive waves of gTLD creation etc.), but I
find it unwise/unhelpful to have a history of gTLDs when the document is
aimed at stakeholders already having an extensive and intrinsic
knowledge of such history.

"III. Theoretical Framework"

Two formulae are proposed by the authors:

- one for net private benefits  
- one for net social benefits

This supports the proposal of two basic categories of new gTLDs:
.    gTLDs with private benefits only
.    gTLDs which bring social benefit

which, I point out, is something that ALAC has been emphasizing for a
long time.

Further paragraphs point specifically in this direction:

Paragraph 23: In theory, ICANN could address the potential divergence
between private and social benefits by adjusting the fees assessed on
new gTLDs.
There is even a suggestion for a negative fee - ie. subsidy for some new
social gTLDs.

Paragraph 24. In practice this is not possible: ICANN would have to
develop accurate quantification of social costs & benefit.
Furthermore, this is not ICANN policy which is purely to use fees to
recover its costs.

Paragraph 26. ICANN could structure approval processes to favour new
gTLDs with high social benefit. The two category scenario still stands.
It could structure its rules to increase benefit & decrease expected
costs to third parties which generate social benefit.


"III. Survey of Existing Studies"
(actually this section should be numbered IV. This mis-numbering
reflects badly on the minutiae with which the report was collated)

This section contains references to various studies which have already
been undertaken previously. I am surprised that such a large section of
the current survey re-hashes older studies.
The very fact that this report was commissioned shows that there was a
feeling that previous studies were not enough, so I don't understand how
repeating the findings of the previous studies would enhance the worth
of this report in this round. Basically, there is nothing new here.
Please ignore this comment if this section was designed as a mission
brief from ICANN.

Furthermore, the reviews do not include any commentary or analysis from
the authors: they merely copy the findings of the publications reviewed,
and I find this disappointing.

"IV. Potential Projects"
(should be numbered V)

This section, as its name implies, provides suggestions for further work.
Some of these projects are interesting but I would have hoped that by
this point in time, we would be looking at results rather than potential
project descriptions. Nonetheless, please disregard this comment if this
section was part of the mission brief from ICANN.

One of these suggested projects did raise my eyebrow:
"2: Analysis of domain name resale price"
This would be based on data from SEDO and was ultimately considered to
be low priority by the authors because of different business models.
However, I question the wisdom of basing ICANN decisions on "trading"
(buy/[re-]selling) of domain names. Are Domain Names supposed to be
primarily used as unique identifiers on the Internet, or as a
speculative trading stock?

I also query the response provided by the authors to the next topic:
"3. Switching costs and behaviour"
This research topic would consist in an examination of the costs to
consumers in switching from one gTLD to another, and whether this is
feasible.
It is considered to be not recommended by the authors because it is
unnecessary in their view.
Question: It is "unnecessary" because:
- nobody will switch? or
- because they didn't want to switch or because it is too expensive? or
indeed,
- because it is so inexpensive that everybody will be able to switch
easily?
This is unclear.


"B. Projected External Costs of New gTLDs

1. Costs of increased registration, monitoring, and enforcement of
trademarks across multiple gTLDs"

This would use simple mathematical formulae to attempt the modelling of
costs and benefits of a domain name registration.

The trouble I find with this type of analysis is that the "benefit" of a
registration is not absolutely quantifiable, even if split as
"affirmative and defensive benefit". Furthermore, the complexity of this
task is probably underestimated since the proposed algorithms only
simulate the first year scenario whilst a domain name registration
incurs a recurrent yearly cost. For this reason, I doubt that this
simplified estimate will yield significant results if yearly
extrapolations are not used -- and the complexity of such a task could
easily spiral out of control.
The authors of the study think otherwise and appear to promote this type
of analysis.

"2. Costs for consumers from increased confusion or fragmentation of the
Internet"

The authors of this report consider this to be impossible due to the
complexity of this modelling which involves variables such as the
proportion of users using search engines, the various ways in finding a
destination site, as well as the potential for a radical change in
navigating patterns brought through the introduction of a large number
of gTLDs. The suggestion by the authors is for ICANN to develop some
kind of methodology to measure consumer confusion as new gTLDs are
rolled out over time.
This suggestion is very helpful. (I do have to give credit where credit
is due!)

"C. Case Studies to help Project Expected Net Benefits from New gTLDs"

Recommendation by the authors to have case studies to examine likely
costs and benefits in broad categories of gTLDs, including enhancement
of net social benefits:
- Study of two currently competing gTLDs - how one has affected the other.
- Study of two non-competing gTLDs -- how one has affected the other.
- Case study of ccTLDs marketed as gTLDs.

In strategy, case studies provide an obvious first step to a new project
and I would have hoped that the present document already included them
at this point in time. The fact that this stage has not been reached yet
is disappointing.

"D. Using the Introduction of New gTLDs to generate additional information"

Suggests continuing practice of introducing new gTLDs in discrete
limited rounds and to monitor effects on other gTLDs so as to build a
historical database.
This is recommended by the authors since it appears that no such data
exists from the previous rounds of new gTLD introduction.
Whilst I fully support this suggestion, I am particularly surprised that
no such data currently exists.


The report ends abruptly, as though the writers had run out of time.
No conclusion is present in the published study! Will there be one for
Part 2?


CONCLUSION

In short, I am *unsure* whether this document provides adequate or
sufficient input as a document entitled "An Economic Framework for the
Analysis of the Expansion of Generic Top-Level Domain Names". That said,
it appears that the authors responded to what was asked of them by the
ICANN mission brief (quoting from the document):
- Survey published studies and resources that describe the potential
impacts of new gTLD introduction;
- Examine theoretical arguments about the benefits and costs of an
increased number of gTLDs; and
- Consider and propose new empirical studies that could help assess
costs and benefits of new gTLDs. The studies should be planned and
structured to address open questions and to provide information about
how best to structure rules for new gTLDs.

In this respect, it has fulfilled its contract, but this document has
raised more questions than answers in my mind.

With no conclusion, neither abstract nor clear overall vision emanating
from the prose, it is hard to draw a clear set of conclusions.
Nonetheless, the conclusions which I sketch from the document are:

* The support by the authors for a proposal of two basic categories of
new gTLDs:
   - gTLDs with private benefits only; and
   - gTLDs which bring social benefit;
which is something that ALAC has been emphasizing for a long time.

* Not that many studies on costs and benefits of new gTLDs have been
published so far.

* The authors recommend mathematical modelling of Costs of increased
registration, monitoring, and enforcement of trademarks across multiple
gTLDs. I differ in viewpoint because believe their proposed model is
over-simplified, uses too many assumptions, and risks either wasting
valuable time or will lead to incomplete and flawed results.

* Case studies based on historical data of how a new gTLD introduction
has affected another gTLD are recommended by the authors. I think that
this is long overdue and I am baffled as to why this has not yet been
undertaken in-house.

Kind regards,

Olivier
 

-- 
Olivier MJ Crépin-Leblond, PhD
http://www.gih.com/ocl.html