RE: [gnso-vi-feb10] First stab at objectives and a definition of VI

[Milton L Mueller <mueller@xxxxxxx>]

________________________________
From: owner-gnso-vi-feb10@xxxxxxxxx [mailto:owner-gnso-vi-feb10@xxxxxxxxx] On 
Behalf Of Jeff Eckhaus
Sent: Friday, February 05, 2010 1:40 PM
To: Gnso-vi-feb10@xxxxxxxxx
Subject: RE: [gnso-vi-feb10] First stab at objectives and a definition of VI


I am not sure that we can eliminate the cross-ownership piece of VI



Whoa. No one is talking about "eliminating the cross-ownership piece of VI." By 
definition, VI _includes_ co-ownership or joint control of separate stages of 
production. But the reverse is not true. Cross-ownership does not, by itself, 
constitute vertical integration. It's just a matter of keeping your labels and 
definitions straight.



Let me explain in a bit more detail why if you require equal access it is no 
longer VI.  If the .JUNK registry owns a registrar that can sell .JUNK but the 
registrar must be functionally separated and provide equal access to all other 
registrars, the consumer is confronted with a non-integrated industry 
structure. There is no vertical integration from the demand side. Consumers can 
buy .JUNK from multiple suppliers. The so-called vertical integration does not 
provide a registry with any of the classic leverage that true vertical 
integration provides. The lock in problem no longer exists. But if there is 
true VI - i.e., if you can ONLY buy .JUNK directly from the registry - the 
problems of lock in and opportunistic pricing may be more serious. But there 
may be other advantages or reasons for allowing that. No need to pre-judge the 
policy here.



While I agree with Milton's points below on what VI should be concerned with, I 
do not agree that if we allow for equal access than it is no longer VI. At 
least in the terms it has been discussed in ICANN, VI also concerns 
cross-ownership and does need to be resolved.



Again, you seem to be confusing two distinct things: 1) whether there are 
issues to be resolved; 2) whether it is vertical integration.



There may be many issues around cross ownership that need to be resolved. That 
doesn't mean VI and cross ownership are the same thing.



By insisting on different definitions I am not asserting that there are no 
issues associated with cross ownership. I am insisting only that cross 
ownership and vertical integration are distinct things and raise distinct 
policy issues. I am also concerned that your definitional approach obscures and 
defers the issues associated with REAL Vertical integration, issues which in 
the long run are far more interesting and serious than these small-potatoes 
cross ownership issues we fought over in Seoul.



Indeed, I think this definitional debate is really just a proxy for people 
concerned about Enom's cross ownership proposal. Some members seem to believe 
that if we call it VI we will somehow get more mileage out of it. I think 
that's a mistake, regardless of your policy position. In determining policy on 
both VI and CO, you must keep them clear and distinct. You will not solve cross 
ownership problems properly if you call it VI, and you will not solve VI 
problems by talking about cross ownership in the context of equal access.



The following is from the Issues Report on Vertical Integration



The issue for ICANN to resolve is whether the current practice of restricting 
cross-ownership should be extended to new gTLDs, and whether additional 
safeguards or rules are needed to minimize any anti-competitive effects of such 
cross- ownership.



Do you see the words "vertical integration" in there? I don't. That issues 
report was arguing that cross ownership was not really a policy issue and that 
the issue to be resolved was cross-ownership. It carefully avoided the term VI. 
Your citation supports my case.

While we may not all agree what constitutes true Vertical Integration the 
issues we have in front of us and this PDP concern not only VI but 
Cross-Ownership.

Now that's better! That statement distinguishes VI from CO and says that both 
may raise policy issues. That is all I am asking for.

You are saying that you want to expand the scope of the PDP to include Cross 
ownership. While I don't think CO is very important, I don't mind including a 
consideration of it in the PDP as long as we don't confuse with VI. The 
registrars here may feel differently. While I believe that CO was part of the 
intent of the original PDP motion, including CO in the scope does not mean that 
VI = CO. So let's try to be accurate with our definitions rather than political.

--MM





-----Original Message-----
From: owner-gnso-vi-feb10@xxxxxxxxx [mailto:owner-gnso-vi-feb10@xxxxxxxxx] On 
Behalf Of Milton L Mueller
Sent: Friday, February 05, 2010 5:54 AM
To: Stéphane Van Gelder
Cc: Gnso-vi-feb10@xxxxxxxxx
Subject: RE: [gnso-vi-feb10] First stab at objectives and a definition of VI





From: Stéphane Van Gelder [stephane.vangelder@xxxxxxxxx]

>I have to admit don't understand why the equal access part is in the proposed 
>definition. If a registry is >vertically integrated with a registrar, it can 
>still be required to provide full and equal access to its TLD to all 
>>accredited registrars. For me, any VI definition need only describe the 
>relationship between the integrated >parties, not any theoretical relationship 
>between them and other parties which may be governed by other >circumstances 
>(such as an ICANN contract for example).



>Avri, Milton, as you say you have already discussed the subject at some 
>length, perhaps you can enlighten me >as to why the equal access part was 
>deemed necessary.



Yes, we did discuss that at length. Its basically a semantic distinction, but 
one that I think avoids potentially serious confusion about policy.



If there are regulations that require a bottleneck provider to provide equal 
access to downstream suppliers, then market structure can no longer be said to 
be "vertically integrated." Many if not most of the technical economies of 
scope and scale made possible by VI are gone, and all you are left with is 
cross ownership. The distinction between a cross-owned but fully separated 
subsidiary subject to regulations governing its relationship with competing 
suppliers (on the one hand) and true vertical integration (on the other) is 
well-recognized in regulatory economics and has important implications for how 
firms behave and are regulated.



If the effects of VI are "regulated away" via equal access, and uniform 
contractual obligations, then there is no point in talking about or worrying 
about VI anymore. Because it isn't VI.



Thus, I have trouble understanding how you can say, "For me, any VI definition 
need only describe the relationship between the integrated parties, not any 
theoretical relationship between them and other parties."

Vertical integration is of interest to policy makers ONLY because of what it 
does or might do to other parties; i.e., other suppliers and consumers. Does it 
foreclose competitors? Does it facilitate lock-in of consumers? Does it permit 
innovations or make a supplier more efficient, giving it a competitive 
advantage over another? If VI simply describes what a company does in isolation 
it is of no interest to us, it is nothing more than an internal organizational 
issue. Why would we care? What do we need a PDP for?



If allowing vertical integration is a real change in the policy regime, which I 
am convinced it is, then it is likely to bring both potential benefits and 
potential harms.



--MM

________________________________________





Thanks,



Stéphane



Le 4 févr. 2010 à 21:03, Milton L Mueller a écrit :



>

>

>> -----Original Message-----

>> From: owner-gnso-vi-feb10@xxxxxxxxx [mailto:owner-gnso-vi-feb10@xxxxxxxxx]

>

>>> So just to make sure I got this right, the definition that is being

>> proposed at this stage is the following?

>>>

>>> Vertical integration is defined as a business structure in which there

>> is no separation between the registry operator and the registrar; they are

>> owned and operated by the same company and the domain name supplier is not

>> required to provide access to independent firms to sell names under its

>> TLD.

>

> Yes, that's a perfectly fine definition. But the good news is that Avri's 
> proposed definition really has very few, if any, substantive differences from 
> a correct one:

>

>> Vertical integration is defined as a business structure in which there is

>> no separation between the registry operator and the registrar in relation

>> to a particular gTLD; they are either owned and operated by the same

>> company or have another contractual affiliation that covers the specific

>> gTLD, and the domain name supplier is not required to provide full and

>> equal access to independent firms to sell names under its TLD.

>

> Other than the phrase " or have another contractual affiliation that covers 
> the specific gTLD", which is a bit vague, I have no problem with this 
> definition. This is in fact not what our difference was within NCUC. It is 
> the equal access issue that didn't seem to be understood at the time, but 
> this definition includes that.

>

> I still think it wise to include the economic reference and the official 
> definition used therein, so that policy experts and real economists who 
> observe our work at least know we did our homework. I do not consider 
> Wikipedia to be an authoritative source on anything, and typically no subject 
> matter expert does.

>

> --MM

>

>