What do we mean when we say "competition?" (was: Re: [Latest draft of Internet regulation bill])

Wed Nov 16 09:40:26 UTC 2005

--On November 15, 2005 11:02:18 PM -0800 David Schwartz 
<davids at webmaster.com> wrote:

>
>
>> --On November 15, 2005 8:14:38 PM -0800 David Schwartz
>> <davids at webmaster.com> wrote:
>
>> >> --On November 15, 2005 6:28:21 AM -0800 David Barak
>> >> <thegameiam at yahoo.com> wrote:
>
>> >> OK... Let me try this again... True competition requires
>> >> that it be PRACTICAL for multiple providers to enter the
>> >> market, including the creation of new providers to seize
>> >> opportunities being ignored by the existing ones.
>
>> > 	The worse the existing provider it is, the more practical it is to
>> > compete with them. If they are providing what people want at a
>> > reasonable
>> > price, there is no need for competition. If they are not, then the it
>> > becomes practical for multiple providers to enter the market. If you
>> > assume that the cost to develop existing infrastructure is not insanely
>> > less than the cost to develop new infrastructure, the isolation from
>> > competition comes directly from the investment.
>
>> 1.	The existing infrastructure is usually all that is needed for
>> 	many of the services in question.  Laying parallel copper
>> 	as a CLEC is not only prohibitively expensive, in most
>> 	areas, it's actually illegal.  Usually, municipalities
>> 	have granted franchise rights of access to right of
>> 	way to particular companies on an exclusive basis.  That
>> 	makes it pretty hard for a competitor to enter the market
>> 	if they can't get wholesale access to the existing copper.
>
> 	For now this may be true. But you'll set up another generation of the
> same problem if you continue to advocate subsidized infrastructure. At
> some point that infrastructure will be inadequate, and you will have done
> nothing to make it easier to build competitive new infrastructure. If
> munipalities granting monopolies is a problem, then stop such monopolies
> -- don't advocate them!
>
The problem is that because of cost and other factors of last-mile
deployment of terrestrial infrastructure, these are natural monopolies
whether you like it or not.  For example, how many streets from how
many different providers pass in front of your house?  How many different
telcos have copper to the junction box that could be used to provide
service to your home?  How many cable companies have fiber or co-ax
in your street?  For the vast majority of the united States, it is
very hard to answer anything other than 0 or 1 to any of these
questions.

The primary problem as I see it is not the monopoly of the infrastructure,
but, the inherent connection between the management of that monopoly
infrastructure and one of the competitors for the provision of services
over that infrastructure.

>> 2.	The existing copper was actually deployed (at least in most
>> 	of the united States) using public subsidies.  The taxpayers
>> 	actually paid for the network.  The physical infrastructure
>> 	should be the property of the people.  The ownership claim
>> 	of the telephone companies is almost as baseless as the
>> 	Verisign clame that they own the data in whois.
>
> 	It doesn't much matter and it can't be fixed. The static value of the
> infrastructure is basically depreciated to zero by now. The profits have
> been reaped. Don't justify future bad decisions on past inquities that
> can't be fixed anyway. Just start right from now on.
>
If I thought what I was suggesting was a bad decision, I wouldn't be
suggesting it.  However, I think there is much more justification for
this decision than past inequities.  As long as you have an area
that tends to create a natural monopoly and allow one competitor that
uses that infrastructure to also own said infrastructure, it creates
an unfair environment for other competitors.

Are you really advocating that the market is best served by multiple
providers laying last-mile fiber?  Doubling the cost of FTTH to
create just two FTTH providers in an area seems pretty stupid to
me.  OTOH, a 10% increase (probably much less) in the cost of FTTH
to facilitate a virtually unlimited number of service providers
being able to access said fiber on a consumer-choice basis
doesn't seem so stupid to me.

>> > 	For example, if Bill Gates took a few billion dollars out
>> > of his pocket
>> > and launched 80 satellites to provide wireless Internet access, it
>> > would be damn hard to compete with him if he wasn't trying to recover
>> > those few
>> > billion dollars. But if you spend a few billion, you get a few billion
>> > worth. Anyone else can spend the same amount and get the same
>> > advantage.
>
>> 3.	Except when you consider that there are only so many orbital
>> 	slots that can be maintained.  (see 1 above as well).  If Bill
>> 	manages to launch N satellites and N leaves N/2 orbital slots
>> 	available for other uses, then, it's pretty hard to launch
>> 	another N satellites at any cost.
>
> 	The present infrastructure in no way impedes the construction of future
> infrastructure. If it did, this would be a valid point. At best this just
> shows that the my analogy is not so good.
>
The present infrastructure in MANY ways impedes the construction of future
infrastructure in the last-mile arena.

The original infrastructure, in many cases, was trenched deep and laid in
all at once in reverse depth order (electric, gas, telco, cable).  Often,
this was done starting with bare earth, digging a trench to the depth
required by electric, putting in electric, then, filling in trench to gas
depth, putting in gas, fill to telco depth, put in telco, fill to cable
depth, etc.  This is a relatively quick and easy process compared to
retrofitting infrastructure into an existing service area.

While it is true that Comcast did not seem to view existing SBC services
as any form of impediment to their deployment of new infrastructure
in my area, I'm actually pretty unhappy about that and lodged several
complaints with the city, the PUC, SBC, and Comcast.  I got to know
most of the Cable Maintenance crews from SBC in my area over the
course of those two weeks because they spent a lot of time near
my house splicing the results of Comcast turned loose with a rock
wheel (or two).

Most responsible providers attempt to install new infrastructure in a
manner that does not involve destroying active existing infrastructure.
This takes significantly more time and money compared to starting with
bare earth or an area with no existing underground utilities.

True, pole-top deployment in an area with existing poles is less of
a barrier, but, there you run into the same limited capacity situation.
Do you really think that most neighborhood poles could support
fiber from 10 or 20 providers as well as electricity and possibly
copper phone service?  I tend to doubt it.

>> > 	If he already has the satellites and is providing the service other
>> > people want at a low price, then other competitors will lose.
>> > But so what?
>> > Consumers win. And competition doesn't exist to benefit the
>> > competitors.
>
>> 4.	But, what tends to happen instead is that Bill charges whatever
>> 	he can get to recoup his billions until someone else launches
>> 	their satellites (has expended the capital).  Then, when they
>> 	start to go after revenue, Bill drops his prices to something
>> 	they can't sustain because they don't have his bankroll and
>> 	have to recoup their costs.  They go out of business and Bill
>> 	either buys their satellites, or, they become space-junk.
>> 	Bill brings his prices back up to previous levels, and,
>> 	consumers lose and the competition loses too.
>
> 	This doesn't work in practice. It only does in theory. There are many,
> many reasons why. One is that service is often contracted for on a long
> term. Another is that spot competitors can compete in small areas when
> prices drop and you can't locally vary your prices forever because it's
> hard
> logistically.
>
It's hard to be a spot competitor with a satellite.  You chose the analogy.

> 	As soon as the prices go back up, the competitors come back. And the
> screwed customers don't.
>
Yes and no.  Once a competitor is fully dissolved through chapter 7, it's
pretty unlikely that they will be resurrected when prices rise again.

>> 	Even if Bill doesn't actually do this, the knowledge that he could
>> 	causes investors to view the new satellite company as a bad risk,
>> 	so, Bill's monopoly position prevents investment into competitive
>> 	entry into the market.
>
> 	Right, and this is appropriate. Large investments in infrastructure
> should *not* be made if there's already adequate service. Better to
> invest in places where there isn't.
>
Is that still true if the "adequate" service is being provided at a price
which is two to three times what it should be costing and the provider
is enjoying the ability to do this because nobody else is in the market
space?

>> 	Finally, since Bill doesn't have to worry about anyone else being
>> 	actually able to launch competing satellites, Bill has no reason
>> 	to innovate unless Bill can see a much higher profit margin
>> 	at the end of said innovation. (look at today's Telco as a prime
>> 	example of this form of complacency.  Actually, telco's are
>> 	very innovative, but, they focus on regulatory innovation instead
>> 	of technical innovation).
>
> 	If this happens, eventually there will be enough of an innovation gap
> that the superior products and services competitors could produce
> overcome the infrastructure advantage the incombent has. The better the
> incumbents prices and service, the less chance anyone will bother to
> compete with him. Virtual competition has the same effects as real
> competition.
>
Right... After 25 years, we're finally starting to see the beginnings
of recognition of that in American telecommunications services.
Generally speaking, I don't think the market is well served by
having to wait that long.

>> > 	If he already has the satellites but is not providing the
>> service other
>> > people want or isn't charging a reasonable price, or both, then anyone
>> > else can make the same infrastructure investment for a comparable cost.
>> > If he's not satisfying demand, the demand is still there, and he's just
>> > losing some of the benefits his infrastructure could be giving him.
>>
>> 5.	But, if you want this analogy to match the current copper plant
>> 	in the ground in most of the US, then, you have to also account
>> 	for the fact that Bill received 30-45 of his 60 billion in
>> 	investment in the form of public subsidies.  Are you going to
>> 	give all comers the same public subsidy (blank check)?  Instead,
>> 	you end up with exactly what we have today in the telcos.
>> 	Semi-regulated monopolies that think they own an infrastructure
>> 	built with taxpayer money. (see also 2 above)
>
> 	I would write off the money badly spent as just that and stop using it to
> justify repeating the same errors over and over. This is much better than
> repeating the same error so that 50 years from now we have the same
> problem with a new generation of inadequate publicly-financed
> infrastructure.
>
I don't want to repeat that error.  I want to make the public infrastructure
paid for by the public accessible to the public on the public's terms 
instead
of on the terms of some monopoly business.  As stated, the company that gets
the contract to maintain the infrastructure in my scenario is contracted to
do so in the public interest, NOT in the service of any particular service
provider.

> 	The government can't do a good job of picking winners and losers, so stop
> letting it.
>
Agreed.  So, let the government do what it does well... Manage the things
that tend to be natural monopolies and keep it out of everything else
as much as possible.  Are you arguing that there should be competition
for the provision of highways, for example?  How would you see that
working?  Do we stack six copies of I-80 on top of each other, or, do
we allocate multiple semi-parallel routes for freeways and let different
companies build different routes and charge what they want for each of
them?  How do you see that working on a neighborhood level?  Even
if you think it works at the highway level, we're really talking about
the neighborhood streets here.

The last-mile infrastructure for terrestrial services looks a lot more
like a local roadways inside a neighborhood than any other analogy
I can think of.  I have yet to see any environment where this has
been accomplished on anything other than a monopoly basis.

>> >> No... Actually, the lack of market forces in the beginning
>> >> is what allows the incumbent providers to have an advantage.
>> >
>> > 	There is only a lack of market forces if the incumbent is
>> meeting the
>> > needs of the consumers. And if they are, there is no need for a
>> > competitor.
>>
>> Nope.  There is a lack of market forces for several other reasons.
>>
>> +	Lack of access to right of way
>> +	Burdensome regulatory environment requiring huge up-front
>> 	overhead -- you can bet that AT&T didn't start with 5
>> 	full time lawyers.  It's pretty hard to run a CLEC
>> 	in most states with anything less today.  Even if you
>> 	only want to serve your neighborhood.
>
> 	Yes, but these are hardly arguments for more public subsidy or control of
> infrastructure.
>
>> +	Public subsidies for the ILEC's initial (and in some cases
>> 	subsequent) buildout which is not available to CLECs.
>
> 	Yes, so stop the subsidies. More subsidies to fix the past subsidies
> won't work. "Fair" subsidies won't work because you can't subsidize
> everything.
>
I don't want to subsidize anything, but, there are certain limited portions
of the network which tend to create natural monopolies no matter how hard
you try to prevent them.  For those situations, decoupling said monopoly
from as many of the "uses" for said infrastructure is a necessary
step to prevent the monopoly from expanding beyond that scope.

>> The list goes on, but, believe me when I tell you that there are
>> plenty of consumers in California that do not feel that SBC
>> is meeting their needs, but, they don't have access to a real
>> CLEC.
>
> 	Oh, I know that story, believe me.
>
So, do you really think that if SBC had the same terms for access to
the MDF<->MPOE leg that any competitor had this would not actually
change or would get worse?  I don't.  I think it would actually
solve many of the current problems and encourage many of the CLECs
to re-enter the market.

>> Huh?  How does this favor one set of business models?  What it
>> does is take
>> the portion of the infrastructure that was built with taxpayer money and
>> put it back in the hands of the taxpayer so that whatever carrier the
>> tax payer wants to buy service from has equal access to the
>> infrastructure.
>
> 	And what about a carrier that needs different infrastructure to provide
> the type of service it wants to provide?
>
They can build it, and, if they get a competitor that wants equal access
to it, they get reimbursed for the build.

>> The current bill actually has a pretty good chance of levelling
>> this playing
>> field by giving not only access to infrastructure, but, also access to
>> right of way to all comers on a non-discriminatory basis.  I think that's
>> a good thing.
>
> 	And repeating the same problem 50 years from now when innovative services
> can't compete with the maintained subsidized infrastructure.
>
I don't see that.  First, the current bill doesn't do some of the things
I'm saying need to be done here.  It, instead, requires what you want in
terms of everyone lays their own copy of infrastructure into the
public rights of way duplicating and multiplying the cost of the
last-mile infrastructure by the number of competitors.  Yes, that's
a level playing field, but, it's _NOT_ an efficient one.  I suppose
we have to decide which "mistake" is the better one to make.

>> If the people have control over the "last-mile" infrastructure and it is
>> operated in a carrier neutral fashion, that creates a level playing field
>> for innovation in everything outside of "last-mile" infrastructure.
>
> 	That's not enough. The last mile infrastructure is as important as every
> other part. It may even be more so.
>
Agreed, but, it's also an area that tends to be a natural monopoly.  Leaving
that natural monopoly in the hands of a single provider that is also
providing services on top of that infrastructure creates an advantage for
that provider.

Requiring a separation between LMI operator and service provider (or at
least the ability to create such a separation when it is in the public
interest), OTOH, might just be an improvement.

> 	The present system was made by subisidizing infrastructure that was felt
> to be good enough at the time it was designed. This is what created all
> the flaws in it that bother you.
>
Not true.  Most of the flaws that bother me in the current system
have to do with the fact that, for example, SBC is both LMI infrastructure
operator _AND_ service provider.  This makes it very difficult for
anyone outside of SBC to have a truly competitive offering because they
are dependent on SBC for their LMI whether they like it or not.

OTOH, if the shared LMI was operated by a neutral third party
and leased to SBC and any other competitor at the same price for
the same component, that would resolve most of what is
bothering me about the current system.  It would allow me
to buy phone service without giving money to SBC.  Today,
I can't do that unless I go to VOIP over WISP which has its
own set of tradeoffs.

>> If the people also make sure that right-of-way is managed in a
>> carrier-neutral fashion (no more exclusive franchises to single
>> carriers), that creates a level playing field for innovation in
>> infrastructure.  All providers have the same difficulty getting
>> stuff into the right-of-way and the same inherent costs.  No
>> carrier gets favorable treatment over another.
>
>
> 	The same would be true without subsidized or government owned last miles,
> wouldn't it?
>
Partially, but, do you really think that the market is there
to support duplicate copies of last mile infrastructure?  That's
a huge cost at a multiple of 1.  If you want two providers in
a given area each with 50% market share, that's huge*2.  If you
want real competition, say 5-10 providers, that's huge*5 to
hughe*10.

I don't know how realistic this number is, but, I bet if I'm off
by much, I'm way too low.  Let's assume it costs $1,000 to bring
some infrastructure component (fiber, copper, etc.) to each
house in a city.

Let's look at San Jose as a typical city example.  According
to 
<http://www.sanjoseca.gov/planning/Census/Citywide_dp_pdf/housing_char_2000-4.pdf>
(the 2000 US Census), there are 281,706 housing units in San Jose.

So, that's a total cost to deploy San Jose once of 281,706,000.
Assuming a standard 20 year payback on infrastructure, that's 
$50/yr/household
that each company needs to make, just to cover the deployment.

If we have 10 providers, then, each theoretically needs to make
$500/yr/household.  I don't think there's even $250/yr/household
on the table for deployment if you really think it through.

That's why I believe last mile deployment of terrestrial resources
creates a natural monopoly.  Whoever deploys first gets a commanding
lead in controlling the infrastructure and there is strong disincentive
to invest in competing infrastructure absent some certainty of
gaining a substantial portion (more than 50%) of the existing
competitors customers.  Sure, if the existing competitor CAN'T
serve their customers better, better service and/or lower price
might make that possible.  However, if the existing competitor
can do so, they have a serious advantage being after customer
retention instead of trying to gain new customers.

Multiple studies estimate that it is between 5 and 20 times
as expensive to gain a new customer as to preserve an existing
one.  This is essentially "marketing 101".

>> Today, nobody can put CATV infrastructure anywhere in San Jose
>> if their name isn't Comcast.  Period.  The city sold us out to
>> an exclusive franchise deal.  The current bill proposed eliminates
>> that.  That's a good thing.
>
> 	No, it does not. It does precisely what you are complaining about -- it
> grants the city a monopoly on the last mile! And still nobody can build
> infrastructure if there name isn't San Jose.
>
No, the proposed bill doesn't, and, neither does my proposal.
The proposed bill does exactly what you claim you want.
My proposal says that anyone who does has to be willing to sell
the LMI portion to the city if the city feels it is in the public
interest for the price quoted for buildout prior to the start
of the buildout.

IOW, you can build whatever you want/need, but, if a competitor wants
equal access to it and you aren't providing that, you have the risk
of the city coming to you and making that possible for your
competitor whether you like it or not.  You haven't lost anything
because the city has reimbursed you the cost of building said
infrastructure.  You still have the same access to that
infrastructure as your competitors, and, your customers don't
see any disruption in service, but, your competitors gain
access to the LMI without having to do a parallel build.

>> > 	Competition in business models, infrastructure technology,
>> > and the like
>> > is just as important (if not more so) as competition in price
>> > and services
>> > within a given model.
>
>> Yes... By taking away the subsidized monopoly infrastructure from the
>> ILEC and making them compete on a level field with other LECs, they can
>> do just that.  The existing infrastructure was built with taxpayer money
>> and should be equally accessible to all service providers.
>
> 	Just write it off. In 50 years it will be worthless anyway. Let's set the
> rules for the future rather than trying to fix the past. I don't want a 20
> year solution that screws us over for 100 years. That's what we have now,
> and that's what you're proposing.
>
Not exactly.

Try to stop looking at what I'm proposing through whatever filter you've
applied and look at the whole proposal:

1.	ANY provider may build infrastructure in public right of way.

2.	ANY provider doing so must provide a price quote for said
	buildout to the local governing body.  The governing body
	and provider must come to agreement in good faith as to the
	quote being reasonable and just (e.g. provider isn't quoting
	10x cost and city isn't trying to claim 0.1x cost).

3.	Provider retains title to infrastructure until such time
	as the local governing body makes a finding of fact that
	public ownership of said infrastructure is in the public
	interest and buys it from provider at above quoted price.

4.	City retains, in perpetuity, right to purchase said
	infrastructure at quoted price.

5.	For any such infrastructure owned by the city, city shall
	operate or contract for the operation and maintenance of
	said infrastructure in a non-discriminatory fashion
	open to all service providers on an equal cost basis
	and to the benefit of the public.  Such costs shall
	reflect the actual cost of operations and maintenance
	for the previous year, appropriate contingency and
	liability funds, overhead, etc.

I think this is win-win for everyone except a monopoly-oriented
service provider.

Consumers get easy access to competition for any form of service that
can be delivered by existing or future infrastructure.

Providers have no disincentive from building new infrastructure as
long as they do so in a model that assumes they will have
competition and cannot count on LMI monopoly as an advantage.

Providers which sell wholesale at fair and reasonable prices with
comparable service levels to their wholesale customers are unlikely
to lose control of their infrastructure (it is unlikely
to be in the public interest to buy it in that case).

>> Sure, FFTH is another roll-out.  One way that could happen is when
>> a new provider comes along and wants FFTH to provide their service,
>> they are asked to provide a quote for the roll-out cost.  Then, after
>> roll-out, if another provider wants access, the people have the option
>> of buying said infrastructure and rolling it into the existing management
>> system for the price quoted.  That way, the provider didn't pay for
>> infrastructure being used by someone else and the people aren't
>> forced to pay for infrastructure they don't want.  Market forces
>> still dictate which infrastructures get deployed, but, there's
>> no early-deployment monopoly as a result.
>
> 	The problem is, the early-deployment monopoly may be what justifies the
> cost of the rollout. Take that away, and you just don't get the new
> technology. So long as there is equal access to do a roll out, there is no
> problem. Yet your plan *reduces* such access.
>
Nope.  It doesn't.  It just prevents the early-deployment monopoly from
becoming a long term competitive advantage.  Re-read my detailed 5 point
explanation above.

However, please don't confuse this part of my proposal with the current
bill. The current bill does exactly what you describe as desirable,
but, I still think at least that is preferable to the current
situation.

>> > 	You can't "correct" the damage. It's not possible. All you
>> > can do is pick
>> > winners and losers *again*. The previous chosen winners and losers
>> > don't really exist anymore in their previous form -- all you can do is
>> > more damage.
>
Perhaps correct the damage is the wrong term.  I want to avoid creating
the damage all over again.

>> Actually, once AT&T and SBC merge, it will be pretty close to the
>> original form of AT&T prior to Judge Greene's decision.  Sure, there
>> will always be winners and losers.  The question is the selection
>> method.  Today, the selection method is the government selecting the
>> incumbent carriers as winners and the ratepayers are the losers.  Under
>> the proposed bill, that changes and I think the ratepayers at least get
>> more input into the selection process.
>
> 	But it helps innovation not one bit and does nothing to ensure
> competition in the future. It's another small step to preserve what's
> here at the cost of what could be.
>
Nope... The current HR does exactly what you say is desirable.  What
you claim not to like about my plan is NOT part of the current HR,
just something I think we need if we are to have viable competition
for last-mile services in the long run.

Owen

-- 
If this message was not signed with gpg key 0FE2AA3D, it's probably
a forgery.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 186 bytes
Desc: not available
URL: <http://mailman.nanog.org/pipermail/nanog/attachments/20051116/8221f49e/attachment.sig>