From: Mustafa Akgul (akgul@Bilkent.EDU.TR)
Date: Fri 04 Nov 2005 - 21:01:01 EET
Dunya bankasi infýdev wsis tartismalarindan ilginc bir yazi
Saygilarimla
Mustafa Akgul
---------
From: "Roland H. Alden" <ralden@ralden.com>
To: "infoDev Forum 2005@WSIS, Tunis" <infodev-wsis@dgroups.org>
[Note from infoDev: This posting is considerably longer than our normal limit of 300 words. However, we found it so interesting and provocative that we have decided to reproduce here in full.]
[Please also note: Many of the members of this list are accessing it over low-bandwidth connections. Please do not attach documents to your postings to this discussion list. If you would like to reference a paper, please just include a web address where people can download it themselves. Thanks!]
..
Thank you Sam. I think I understand the distinction/question you are
making now.
I will try to "think out loud" and make a couple of observations I think
you will agree with, and I hope will shed some more light on the topic.
First let me remind myself of the definition you set forth earlier "...
A "public good" is a good where consumption by one user does not
diminish access by another, or where the marginal cost of additional
supply is essentially zero."
You further noted that "information infrastructure" could mean both
"information" and "infrastructure". Clearly "information" meets the test
of being a "public good" in the sense that "consumption" of information
by one person does not reduce the supply of information that might be
available to others. Some forms of "information", such as works of
music, might not quite fit this principle, but I think we can agree that
there is a lot of information that is "owned" by "the public" and the
"consumption" of that information fits the "public good" "marginal cost
of zero" framework.
The fact that information has this quality of zero marginal cost is
arguably the driver of the exciting vision of an Internet that would
make such information available to all. If the information cost a lot of
money we would not be worrying so much about how we can distribute it to
poor people; the fact that information distribution potentially has a
marginal cost of zero, and the value of this information is such that it
can help lift people out of poverty, is what makes this equation so
interesting. Something that costs almost nothing and can have a powerful
impact is has high economic leverage when compared to other resources we
would like to have. For example, an AIDs vaccine would have a powerful
impact too, but depending on the nature of the vaccine the marginal cost
of delivery will not likely be $0; and for many complex health care
technologies the cost of delivery can be very high.
But we do have to make a distinction between "information" and
"infrastructure" and unfortunately most "infrastructure" does not tend
to have a marginal cost of zero. Network infrastructure can at times
appear to come very close to having that property, so we need to think
about this in detail and with great precision.
To illustrate this point let us consider a teacher in a classroom who
wants students to understand that the sum of the interior angles of any
triangle is 180 degrees. With only words, and perhaps a blackboard, this
concept can be taught to 10 students, and for essentially no extra
(marginal) cost it can be taught to 20. But, if we consider a textbook
to be part of the necessary "information infrastructure" for this lesson
then 20 textbooks cost something more than 10. Unlike the "information"
the "infrastructure" has a marginal cost that is large enough to matter.
Now, if we take this into the realm of digital wide area networking
there is another economic factor which could cause confusion if we don't
understand it carefully.
There are cases where digital networking technology has a "zero marginal
cost" characteristic for some services; there are other services where
the capital cost is high and the marginal cost is high as well.
For example, if I have a local area network using 100 mb/sec technology
and I have a nominal traffic flow of 50 mb/sec then the "marginal cost"
of connecting one more computer using 1 or 2 mb/sec of bandwidth might
be zero (if I have a spare port) or very cheap (if I have to buy a new
100 mb/sec port). This would be true in the case of a very "local" area
network.
However, suppose I recast the scenario to a satellite circuit. It is
true that I might have a satellite circuit with a "spare" capacity of 1
or 2 mb/sec going unused. But my marginal cost for hooking up one more
site to the satellite might be many 100's of dollars if not thousands,
depending on the technologies involved. Further, the notion that I might
have "spare" bandwidth on the satellite is very questionable. I might
have "unused" bandwidth but because a satellite, by its very nature, is
able to reach so many customers, in multiple countries, it is not really
possible to argue that I could not *sell* this capacity to someone,
somewhere, at some price. Thus, in my view the capacity may be unused
and therefore "free" but in economic terms there is an "opportunity
cost" associated with considering the bandwidth to be "excess" and
allowing it to be used for $0 consideration.
Because of the nature of computer/telecom technology, other than
satellite circuits, it is very often the case that a particular network
reaches a limited group of potential customers and the theoretical
capacity of the network can be very high relative to what people are
actually using. For example, if one were to today build a fiber optic
network around a city, one would probably use either Gigabit Ethernet,
or possibly 10 Gigabit Ethernet. It would not make economic sense to use
100 Megabit Ethernet, and in fact it could cost more to do so than to
use Gigabit Ethernet. Such is the nature of digital technology; at any
transition point vast capacity costs the same, and then becomes cheaper
than, older technology. So it is not uncommon to find large amounts of
capacity under-utilized at times, and it is possible to argue that this
capacity has a "marginal cost of zero."
However, when this networking capacity is made available through free
market forces, it may be difficult to take advantage of this
characteristic of the technology. An example will help illustrate why.
Suppose there are 5 people that have enough money to buy a goat, and
there are 5 people that do not. And suppose that I have 10 goats. If I
define the "market" as one where I am going to sell the five goats to
the people that can pay, and then give away for free the 5 goats to the
people that need them, but can't pay, I will run into a problem.
First, let us understand that it might make perfect sense for me to give
away the five goats. If I don't give the goats away I will have to feed
them and they will cost me money. If there are truly only 5 people that
can afford to pay me for goats then you can argue that I should have
raised only 5 and not 10 goats. But you can make the same argument about
Gigabit Ethernet; you will likely be more successful telling goats what
to do :)
So, I am motivated to give away the 5 goats I cannot sell.
The problem is, the people that can afford to pay for 5 goats need only
stand at the end of the line, instead of the beginning, and they will
think they should be able to get a free goat, once I have sold 5. The
introduction of a "free" goat is an economic fiction; the 5 goats I am
going to give away will change the market to one where the price of a
goat falls to some average "market" value. As the price of a goat falls,
some more buyers will come into the market and pay something for a goat.
Now, if I have to travel to the market with my 10 goats and stay there
all day there is no reason to sell 5 goats at 9:00, give away five goats
at 10:00, and then sit down and do nothing until 17:00. I might as well
try to get some price for the 10 goats; a price that is somewhere
between what I could have sold 5 for quickly, and what the poorer people
who want a goat can afford to pay.
The owner of a gigabit network has the same problem. It might take a
year to build a city-wide or regional gigabit network; just as it takes
time to raise 10 goats and get them to market. If the owner of the
gigabit network sells a gigabit of capacity the day the network is
working then he is making an economic mistake compared to, for example,
selling 5 customers 100 mb/sec units of service quickly and then taking
some time to sell the other 5 100mb units. It makes great sense to keep
a certain amount of capacity in reserve as well. For example, a year
later a customer with a 100 mb/sec connection might say to the network
"I need 200 mb/sec service". In fact several customers may say this and
the network owner will have to plan to purchase 10-Gigabit Ethernet
equipment, and upgrade the network. This will cost money and take some
time to actually execute the upgrade. So it might be important to have
some capacity in "reserve" that can be given to customers with an urgent
need until the new technology can be installed and tested. Furthermore,
unlike goats, the capacity which is kept in reserve does not have to
eat. It does not cost hard money today to keep capacity in reserve.
So, I hope this makes clear that "excess capacity" with a "marginal cost
of zero" may not exist, as a practical economic matter, in a network
that is operating in a free market economy and is being actively managed
by a network operator that is forecasting customer need and planning on
investing capital in the future to meet customer needs.
Now, in order to build any network "infrastructure" capital investment
is required. I have tried to illustrate that a physical telecom network
cannot really have a marginal cost of zero quality that makes it a
"public good" but let us tolerate the notion that we can build networks
that come close to that vision; networks that can help deliver telecom
to a large number of people, and enable information, which can have a
marginal cost of zero, to be made available widely.
We know that private capital can build such networks but I have shown
why it is difficult for private capital to deliver service at anything
other than free market prices and why it is hard for a private network
to not keep some capacity off the market, which raises prices. However,
it is a fact that most telecom networks do not operate in a "free
market."
Almost every telecom network that serves customers is heavily regulated
by government and these regulations have much influence over how these
networks are built and operated and what their capital structure is. It
is not necessary for a government to have a direct ownership position in
a network in order to have an extremely high impact on how that network
operates.
Now, we should ask, why? What is the intellectual basis for so much
government intervention in telecom markets? We would not tolerate this
degree of influence over "IT Technology" in general. Governments don't
tell us what database software to use or what company we can buy it
from; as a rule. It is important to understand two things:
1. Why people think telecom is different; why it should be singled out
for heavy government regulation.
2. Why the views which support government regulation of telecom are
obsolete.
In answer to the first question we need to concern ourselves with only
one concept: right-of-way. All large telecom networks that serve "the
public" have been regulated by government for one single reason. They
are considered to "consume" a "public resource" which has "finite
(limited) capacity". To understand this clearly let us consider some
parallel examples of government regulation over resources.
Historically we have considered "air" to be a "public resource". All
citizens have a right to "air"; but because it is so abundant there has
not been much government regulation of air. Water, is a somewhat more
scarce resource, so historically governments have been very involved in
water regulation. For both air and water, government has a mandate to
try and deliver these to citizens at little or no cost.
Over the past century the view that "air" exists in unlimited abundance
and therefore need not be regulated by government has changed. We now
know that it is possible for some to use air in ways that have costs to
others.
Telecom has always been viewed as a consumer of a public resource;
consumption that has a negative impact on some and therefore must be
regulated to balance "public good" with consideration given to those
that are "harmed" by telecom. What "public resource" are we talking
about? Right-of-way. Telecom networks "consume" "right-of-way" in their
construction. In the case of a wired network, the physical cable
"consumes space" on utility poles, or it consumes space underground. If
it is routed along a public road, the public road is viewed as an asset
which belongs to "the people" and if a telecom company is able to
benefit from that asset by laying cables under the roadway there is a
tradition of asking the telecom company to pay.
In the case of wireless networks there is an even stronger set of
beliefs that the "electromagnetic spectrum" is finite and that
government must regulate the use of various frequencies involved in the
transmission of radio signals so as to insure that a "public asset"
(spectrum) is used wisely, and those who use it pay something for the
"privilege."
The price telecom networks pay for "right-of-way" is paid by all
customers of telecom networks. Some years ago "all customers" might mean
"some people". In developing countries it might mean "rich people",
multinationals, or "the government." However, in today's world, the
practical result of this burden is that it is paid by "all people." The
person who uses a cell phone to lift their income from less than $1 a
day to $3 a day is helping to pay for these regulatory burdens.
The principle seems perfectly sound. These companies are using scarce
resources that "belong" to "the public" and the resources can be
squandered by mismanagement, so there is some justification for
regulation in this area. However, while that all sounds reasonable, it
does not mean that it is true. We know that in the case of "air" the
assumption that the resource was abundant proved wrong and new
regulation is now being created. In the case of telecom we need to
really ask ourselves two questions:
1. Are right-of-way resources really scarce?
2. If they are "somewhat" scarce, or if they can be used in such
wasteful ways that the abusers impose costs on others, then perhaps some
level of regulation is justified. So the question then becomes: Is the
level of regulation we have today really justified?
To answer the question "is right-of-way scarce?" we must look at history
and the facts to get an understanding. Telecom used to occupy
"right-of-way" in a manner that imposed a large cost on the public.
Telephone poles were ugly and occupied space in the roadway and they
would blacken the skies with dozens of fat wires; they carried
electricity and were perceived as dangerous (and could be if they
crossed electric wires or were struck by lightening). Multiple telephone
companies building networks side-by-side created such intolerable
conditions in crowded downtown areas that the grant of a monopoly to a
single company became common. When a monopoly was granted, as was the
case in most countries by the turn of the 20th century, then the heavy
regulatory apparatus of government was fully justified; not because of
right-of-way issues, but because regulating a monopoly was, in fact, a
big job. But the original motivator of the monopoly was largely
right-of-way, and to a lesser degree, the desire for "universal service"
(the idea that a monopoly could be forced to build network "everywhere",
not just where there was high demand).
In the case of wireless technology there is a long history of regulating
the radio waves "in the public interest". Most of that regulation has to
do with radio and television content and free speech issues, etc. but
digital data networking is not really considered a separate matter and
is heavily regulated as well; all based on the idea that radio spectrum
is "finite" since we "know" that no two transmitters and receivers can
share the same frequency without interference. So, we think a government
has to decide how frequencies (which are a form of right-of-way) will be
allocated to telecom companies.
This history ignores progress in science, technology and our
understanding of physics. It is simply a fact that radio and light
spectrum can be shared without interference and it is no longer possible
to argue that spectrum is a "scarce resource". In our world today it is
only misguided government regulations that make spectrum scarce by
denying the right to use it, except in ways codified in ancient and
obsolete laws which are inefficient.
Modern telecom networks no longer need to darken the skies with
thousands of copper wires carrying potentially dangerous electricity;
they are built with fiber optic cables that have so much capacity that a
six inch pipe buried under the street can carry, on behalf of multiple
competitive telecom companies, every phone call made all over the world
at the same time. In America, at the absolute peak of the Internet boom,
in the largest and richest cities, fiber optic carriers competed for
scarce right-of-way and angered citizens by digging up streets and
snarling traffic. While this would seem to argue for government
regulation of right-of-way, in fact, in America, like everywhere else,
right-of-way is highly regulated. But it is done so with gross
inefficiency and outright negligence. So, for example, a company with a
permit to dig a trench down a street in November is not informed that a
different company is going to dig a trench down that same street in
December, such that they could cooperate and dig one trench instead. The
cooperation is not "enabled" by telecom regulations; in fact the
regulators don't even have a database of planned construction activity.
Just as is the case everywhere else in the world, the purpose of the
regulations is to impose a tax on an industry, not to promote efficient
use of "scarce" resources.
In developing countries another "scarce" resource that laws can
"regulate" is foreign exchange. In many countries modern telecom
networks are built almost entirely out of foreign technology that must
be imported and paid for in "hard" currencies. Governments of poor
countries may have a legitimate need to regulate the expenditure of this
resource (that is a different debate) but it is a mistake to use telecom
regulation to govern that factor of the economy. Telecom regulation can,
and usually does, have the opposite effect. For example, a network
operator may be forced to build a satellite earth station using imported
technology paid for in US dollars, rather than build a fiber optic
network using the labor of unemployed local citizens. Yes, the fiber
optic cable itself would have to be imported using foreign exchange, but
the other costs of building a fiber optic network are local labor and
materials (like cement and PVC pipe).
But, you can't build a fiber optic network down a public road without
having someone from the government notice; you can build a small
satellite earth station and hide it behind a fence, or tell people you
are just watching TV. So, heavy government regulation of telecom can
force networks to be built using more costly technologies, which employ
fewer local people. This makes no sense.
The character of telecom regulation today has moved very far from any
legitimate concern for the efficient utilization of "public" resources;
instead it is the worst sort of "industrial policy" enforced by
government bureaucracy micro-managing a small segment of the economy;
one that could, and should, have a huge positive impact on all citizens,
but is instead is crippled by outmoded thinking. Developing country
governments really need to step back and understand that their PTTs,
where they are still parastatial enterprises, are basically economic
disasters; and the root of the problem is the much large issue of how
telecommunications is regulated.
> -----Original Message-----
> From: Sam Lanfranco [mailto:lanfran@yorku.ca]
> Sent: Wednesday, November 02, 2005 7:52 AM
> To: infoDev Forum 2005@WSIS, Tunis
> Subject: [infodev-wsis] RE: Is Information Infrastructure a Public
Good?
>
> I would like to make a brief response to Roland H. Alden's comment on
my
> posting regarding ICT as a public good. ... (Remainder deleted.)
Dgroups is a joint initiative of Bellanet, DFID, Hivos, ICA, IICD, OneWorld, UNAIDS and World Bank
--- You are currently subscribed to infodev-wsis as: akgul@Bilkent.EDU.TR
To unsubscribe send a blank email to leave-infodev-wsis-432210M@dgroups.org
_______________________________________________
Linux-sohbet mailing list
Linux-sohbet@liste.linux.org.tr
http://liste.linux.org.tr/mailman/listinfo/linux-sohbet