externalities by niusheng11


									      Network Externalities
Sometimes it helps if everyone uses the same product
or service. This may drive us towards monopoly or
Somehow, this serious economic topic has gotten
bogged down in a lot of personalities.

One problem is that defining standards can both
encourage the spread of technology and limit the
development of new technology; so when to impose
standards is a difficult practical choice.
Network externalities are not the
  same as economies of scale
 “Economy of scale” means that making many copies of
 something is cheaper, per item, than making a few.
 “Network externalities” means that there are benefits if
 many people use the same thing.
 They both may encourage monopolies, but they are

 By the way, the phrase “network externalities” was
 coined by Jeff Rohlfs, once at Bell Labs.
          Specific examples

Having the only tennis racket in town is useless. The
more people that own tennis rackets, the more useful
they are. It doesn’t matter whether it’s cheaper to
make more rackets than fewer rackets. This is a
network externality.
Manufacturing many copies of the same digital camera
is cheaper, per camera, than making only a few. It
doesn’t matter whether people benefit from all using
the same camera. This is an economy of scale.
Solitaire doesn’t have any network externality. You
don’t gain because other people are playing it.
Bridge does. Being the only bridge player in town (at
least until we had computer programs to play against)
is pointless; and the more players there are, the more
useful it is for you to know how to play bridge.
In neither case is there any economy of scale. By
comparison, “pong” had enormous economies of scale
(the usual low cost of duplicating software) and no
network externalities. Multiplayer games, on the other
hand, do have network externalities.
       Ordinary externalities
In economics, the word “externalities” refers to costs
or benefits that fall outside the activity that is being
Since economists believe that if all costs and benefits
can be charged to the activity that creates them, the
market will produce an optimal allocation of resources,
they view externalities as tending to produce
misallocations, and thus as undesirable.
      Negative externalities
In the popular press, the most common use of
externalities is to talk about pollution. Pollution is a
negative externality: a cost which a business manages
to push onto society as a whole. If businesses had to
pay for their own pollution, then costs and activities
would reflect the “right” (or at least market-
determined) amount of pollution.
There are many references to this phenomenon. Few
seem to give real data.
              Pollution cost
An EU study (Mike Holland, Paul Watkiss) gives as an
example that the cost of putting a ton of sulfur dioxide
into the air in Stuttgart is about 36,000 euros,
representing the cost of extra hospital admissions.
If power plants in Stuttgart are using high-sulfur coal,
and not trying to clean up their emissions, they save
money buying cheaper coal, but they push a cost onto
society. In the USA, burning low-sulfur coal costs
$113/ton sulfur dioxide saved; scrubbing costs
$300/ton. So cleaning up power plants is a good deal
overall, but the power plant operator only sees a cost.
       Air pollution in China
The next two slides are almost unreadable. They
come from:
They show (1) years of life lost in areas of China from
air pollutants; note that the population in each square
is very large, so the thousands of years are divided
over a great many people; and (2) the costs
associated with a power plant’s emissions, both health
costs and decreases in agricultural yields.
      Many other examples
There are many kinds of negative externalities, some
of which get beyond economics…
Paper companies dumping acid into rivers may destroy
the fishery in that river.
Building a new house congests the roads and may
degrade the view from existing houses.
Smoking cigars may cause others to feel that it’s too
smelly to enjoy the location.
        Positive externalities
There are also positive externalities. You might do
something which helps others, but not get any of the
financial benefit.
For example, if you keep your lawn neat, the value of
your neighbor’s house goes up slightly. You won’t get
any of that increase when it is sold.
Or if you take the train to work, everyone else gets a
little less traffic; again, this isn’t reflected directly in
the price of your train ticket (although the government
may subsidize rail travel for that reason).
Beekeepers get honey from beehives, and the bees also
pollinate the crops of the nearby farmers. In 1952 this was
quoted in an economics article (by Meade) as an example
of a positive externality: the farmers get their crops
pollinated as an external benefit of honey-making.
It appears that it was not until 1973 that Cheung published
an article observing that farmers pay beekeepers to move
their hives into orchards and fields. (See “The Fable of the
Bees”, J. Law & Economics, in JSTOR).
The extent to which economists appear to be ignorant of
the real world is appalling.
Lighthouse (Edward Hopper)
“since it is impossible that the ships at sea which are
benefited by a lighthouse, should be made to pay a toll on
the occasion of its use, no one would build lighthouses from
motives of personal interest, unless indemnified and
rewarded from a compulsory levy made by the state.” – John
Stuart Mill. Paul Samuelson, in his textbook, gives
lighthouses as an example of positive externalities.
Ronald Coase, a Nobel-prize winning economics, wrote a
1974 article pointing out that lighthouses in 18th century
England were privately owned.
But (see Bertrand, “The Coasean analysis of lighthouse
financing: myths and realities”) the government forced
shipowners to pay for the lighthouses, and the system was
very inefficient.
     Why these arguments?
These arguments are not really about beekeepers
and lighthouses. Paul Samuelson’s textbook, the
most prestigious and common introductory
economics book, is considered too liberal by the
more free-market economics groups, and they thus
look for opportunities to attack him and his book.
Samuelson has tried to revise his book to give more
attention to a wider variety of economic opinions,
but he’s never going to satisfy the more extreme
critics. (Anyway, he’s now 91 years old – he was
born in 1915).
         Network externalities
Some externalities have to do with a need for many people
to share something. When you buy a DVD player and
some DVDs, you encourage an industry, and thus more
DVDs are put on the market. When you throw out an 8-
track tape player, you discourage the industry, and if
enough people do that, no more 8-track tapes are sold.
In the early days of the telephone, you could only phone
up people who bought their service from the same
company. This doesn’t work out well: pretty soon the
companies must either interconnect or merge.
If there are strong benefits for everyone to use the same
product, how do we know the product will all work
together? Returning to the canonical example of tennis
rackets, it isn’t helpful if one person has a tennis racket,
another has a badminton racket, and another has a ping-
pong racket.
Standards can be established either by committees or by
the market. We can all exchange email because of IETF
standard RFC 822. (IETF: Internet Engineering Task
Force). We use Powerpoint because of Microsoft’s market
         When to standardize?
If we standardize too early, we may make a wrong
technology choice. If we standardize too late, we may
impede the development of a market.
The British chose early that railway track gauge should be
4 ft 8.5 in. Today many think that it was a mistake to use
that gauge when the carriages are 9 ft wide (10 ft or more
in most of the world); the Great Western tried 7 ft, but lost
out. (On the other construction is cheaper with a smaller
Today, failure to agree on Sony’s Blu-Ray vs. Toshiba’s H-D
formats for the next kind of DVD is holding up the market.
      Beta, VHS, and Krugman
Paul Krugman pointed to the Beta-VHS controversy as an
example of choosing a standard too early (“path
dependence”). He argued that Beta was a superior
technology but once most people had VHS players, most of
the movies came out on VHS and Beta disappeared.
His critics point out that Beta once outsold VHS, and that
VHS recorded for longer times, and thus Beta was not
superior and the market had chosen wisely.
Again, this has much more to do with Krugman’s status as
a liberal columnist than with technology.
Krugman also pointed to the typewriter keyboard as a
mistake made early; he suggested that a different
arrangement (the “Dvorak” keyboard) would have let us all
type faster.
Again, critics point out that no really good comparison
shows the Dvorak keyboard to be better; QWERTY had
become a standard long before people did experiments.
(What psychologist today, do you think, would want to run
experiments on whether light bulb sockets have the right
             Dvorak keyboard

Although there are no decent studies by modern
standards, perhaps it suffices to say that some typing
competitions banned the Dvorak keyboard as unfair.
On the other hand, people willing to do text messaging on
telephone pads have no business complaining about
QWERTY. (And a 93-year old beat a 13-year at text
messaging; he used Morse code).
                   Being first
“Pioneers are the guys with the arrows in their backs”
It isn’t always important to be first. It’s not just VHS
overtaking Beta. There were once more cars in the US
driven by steam than by internal combustion engines,
and there were once 12-inch laser videodisks, and so on.
In the Internet, Lycos was once the leading search
engine, then it was AltaVista, and then it was Google.
Traditionally, companies like Sears, Roebuck observed
that most of the money is made as a product moves
from 5% adoption to 95% adoption, and they aimed at
that growth, not the stage which goes from 0% to 5%.
                   Getting big
There are those who do try to get market share early
and hang on. The Japanese were known for the strategy
of market share first, profit later. Amazon.com lost
money for years establishing itself as the main source for
online books, driving Borders out of the market.
The larger the effect of network externalities, the more
important it is to be the first big player. And depending
on whether you can maintain control, you may as a
result get a big profit.
A great many web pages about network externalities talk
about the benefits of people sharing a word processor or
calendar or email system and get involved in the
Microsoft antitrust issues.
Needless to say, you can find many pages arguing
whether having only one standard word processor is
good for consumers (because economies of scale reduce
the price and network externalities increase the benefit),
or bad for consumers (because Microsoft’s monopoly
position lets it raise prices for a bigger profit than, say,
Intel or Dell per sale).
In today’s politics, people will defend unregulated
monopoly but they will not defend regulated monopoly.
The Microsoft Zune music player has 802.11 wireless,
and if you are near someone else with a Zune player, you
can lend them a song for 3 days or 3 plays.
This is marketing for both Microsoft and the music
companies. Unless lots of people buy the Zune, the
feature is of little value; you have to be physically near
the other machine.
Presumably as a result of pressure from the recorded
music companies, the person who receives a loaned song
can not then lend it out again (unless they buy it).
So what does this mean for information? Are there
advantages to sharing the same information? Besides
the economy of scale in making many copies, perhaps
people are more likely to go into a bookstore or library if
they are confident they will find the same things they
saw advertised or heard about.
Isn’t this just the “long tail” discussion from the other
side? Yes, as far as content is concerned.
But there are also questions about systems, interfaces,
etc. Would it help if libraries shared borrowing facilities,
organization, and the like?
Once upon a time many major libraries had their own
classification systems. Eventually they all joined either
Dewey or LC.
Mostly: “economy of scale” – it’s cheaper to classify
books only once than to do it in each library.
But also: “network externality” – the users gain by not
having to learn a different shelving system in each
library, assuming that most of them will use several
different libraries in the course of their studies.
Could we imagine a major change to either the Dewey or
LC standards today?
Perhaps the most dominant network externality is
language. If you write a web page in Welsh or Inuktitut,
not many people can read it. If your only audience is
that community, you may be OK with that. But if a
general essay on physics or Shakespeare was written in a
less-common language, it’s not likely to get attention.
Before telecommunications, this was less of a problem; if
you’re limited to face-to-face conversation, and you
happen to be in Pond Inlet, Nunavut, Canada, you are
better off with Inuktitut. But now the Web may be
threatening not just the uses of Inuktitut, but even
languages like Danish or Greek. European scientific
journals, for example, are now overwhelmingly in English.
                 Library users
If recommendation systems (social filtering) are
important, then library users will want to find the same
books in each library.
If most library users are remote, rather than local, the
advantages of gathering local resources may be limited.
Effectively, the switch to online means that the user
population is now some large, geographically dispersed,
and loosely coupled group. Any advantages will come
from membership in very large external networks, not
from small ones. (If there were no long-distance calls, it
wouldn’t matter if different cities had incompatible phone
          Back to railway gauges
Isolated railway lines can pretty much use any railway
gauge (either track gauge or loading gauge) that they
want. But if they want to interchange trains or cars with
other railways, they need to agree on the gauges.
So originally, the fact that British railways had 9-ft wide
carriages while European railways had 10.2-ft wide
wagons didn’t matter.
Then the Channel Tunnel was dug, and it’s now
inconvenient that continental-size trains don’t fit through
British stations. Britain is thinking of spending a lot of
money to partially fix this.
Same problem with info: the Internet makes compatibility
more important.
   Libraries and externalities
There is some advantage to common catalog systems
across libraries, but whether they should share books
is less clear. If the “long tail” argument is right,
libraries should actually spend more effort than they
now do selecting materials specifically for the
community they serve. But if everyone is going to
move around and deal with all libraries at once, then
one big system is going to have advantages.
I’m fairly pessimistic; I don’t see why independent
libraries should be more viable than independent
bookstores, or than independent radio stations.

To top