An anthropological exploration of why decentralization expedites progress.

I Find Karma (
Sun, 15 Aug 1999 15:03:39 -0700 (PDT)

First, the punchline:
> We can extract from human history a couple of principles. First, the
> principle that really isolated groups are at a disadvantage, because
> most groups get most of their ideas and innovations from the outside.
> Second, I also derive the principle of intermediate fragmentation: you
> don't want excessive unity and you don't want excessive fragmentation;
> instead, you want your human society or business to be broken up into a
> number of groups which compete with each other but which also maintain
> relatively free communication with each other. And those I see as the
> overall principles of how to organize a business...

If this interests you, the following is definitely worth a read. I'd
like to think of FoRK as intermediate fragmentation, with neither
excessive unity nor excessive fragmentation, and whose members maintain
a relatively free communication with each other (and the world, since
our archives are openly available on the web). Not a bad way to
organize a community, especially since, as the talk suggests, most ideas
and innovations come from the outside.

Somewhat off-topic is a riddle/challenge to the member of FoRKers to
compute with 95% confidence how long the human race is expected to last.
You can give the answer in the form of a range of number of years.
Readers of The New Yorker are disqualified from playing this game.

Okay, now here's why decentralization expedites progress...


Jared Diamond was in New York several weeks ago and we had an early
dinner across the street from the Museum of Natural History where he was
scheduled to speak later in the evening. Jared first visited the Museum
in 1963, when he was 25 years old, preparing to go to New Guinea on his
first expedition to study New Guinea birds. Subsequently he analyzed his
bird collections in the museum where he is on the staff of the Museum's
Department of Ornithology in addition to his position at UCLA.

Jared noted that "probably most lectures one hears at the museum are on
fascinating but impractical subjects: namely, they don't help you to get
rich. This evening I plan to redress the balance and talk about the
natural history of becoming rich."

A Talk by Jared Diamond

In Guns, Germs, and Steel I asked why history has unfolded differently
over the last 13,000 years in Eurasia, in the Americas, in sub-Saharan
Africa, and in Aboriginal Australia, with the result that within the
last 500 years Europeans were the ones who conquered Native Americans
and Aboriginal Australians and sub-Saharan Africans, rather than vice

Most of that book, was concerned with comparing the peoples of different
continents, but I knew that I couldn't publish a book comparing the
histories of different continents and considering Eurasia as a unit
without saying something about the fascinating problem of the
differences of history within Eurasia. Why, within Eurasia, was it
Europeans who conquered the world and colonized other people, rather
than the Chinese or the people of India or the Middle East? I devoted
seven pages to that subject at the end of Guns, Germs, and Steel, and I
think I arrived at the correct solution. Nevertheless, since the
publication of Guns, Germs, and Steel, I've received a lot of feedback,
and the most interesting feedback has been about the implications of
that comparative analysis of the histories of China, Europe, India, and
the Middle East.

In particular, in addition to the review of my book by Bill Gates, I've
received a lot of correspondence from economists and business people,
who pointed out to me possible parallels between the histories of entire
human societies and histories of smaller groups. This correspondence
from economists and business people has to do with the following big
question: what is the best way to organize human groups and human
organizations and businesses so as to maximize productivity, creativity,
innovation, and wealth? Should your human group have a centralized
direction, in the extreme having a dictator, or should there be diffuse
or even anarchical organization? Should your collection of people be
organized into a single group, or broken off into a number of groups, or
broken off into a lot of groups? Should you maintain open communication
between your groups, or erect walls between them, with groups working
more secretly? Should you erect protectionist tariff walls against the
outside, or should you expose your business or government to free

These questions about group organization arise at many different levels
and for many types of groups. They arise, of course, about the
organization of entire governments or countries: what is the best way to
govern a country? Remember the classic arguments about whether the best
government is a benign dictatorship, or a federal system, or an
anarchical free-for-all. The same questions also rise about the
organization of different companies within the same industry. How can
you account for the fact that Microsoft has been so successful recently,
and that IBM, which was formerly successful, fell behind but then
drastically changed its organization over the last four years and
improved its success? How can we explain the different successes of what
we call different industrial belts? When I was a boy growing up in
Boston, Route 128, the industrial belt around Boston, led the industrial
world in scientific creativity and imagination. But Route 128 has fallen
behind, and now Silicon Valley is the center of innovation. And the
relations of businesses to each other in Silicon Valley and Route 128
are very different, possibly resulting in those different outcomes.

Of course there are also the famous differences between the
productivities of the economies of different countries: the differing
national average productivities of Japan and the United States and
France and Germany. Actually, though, there are differences between the
productivities and wealths of different business sectors within the same
country. For example, the German metal-working industry has a
productivity rivaling that of the United States, so the Germans are
certainly capable of organizing industries well, but the German
beer-brewing industry is less than half as productive as the American
beer-brewing industry. Or take Japan -- we Americans are paranoid about
the supposed efficiency of Japanese business, and the fact is that the
Japanese steel industry is 45% more productive than the American steel
industry. Why is it, then, that the Japanese food-producing industry is
less than 1/3 as productive and efficient as the American
food-processing industry? Still another example: in Korea, the steel
industry is equal in efficiency to American steel making, but all other
Korean industries lag behind the United States. What is it about the
different organization of the German beer brewers and the German metal
workers, or the different organization of the Japanese food processors
and the Japanese car manufacturers, that accounts for the different
productivities of these sectors within a given country?

Obviously, the answers to these questions about the different success of
organizations partly depend upon idiosyncracies of individuals. The
success of Microsoft must have something to do with Bill Gates. If an
idiot were in command of Microsoft, then however superior Microsoft's
organization, Microsoft would be unlikely to be a successful business.
But nevertheless one can still ask , all other things being equal, or
else in the long run, or else on the average, what form of organization
of human groups is best? I'm sure that there are many of you here who
are involved with businesses that would like to know the answer to that

I propose to try to learn from human history. Human history over the
last 13,000 years comprises tens of thousands of different experiments.
Each human society represents a different natural experiment in
organizing human groups. Human societies have been organized very
differently, and the outcomes have been very different. Some societies
have been much more productive and innovative than others. What can we
learn from these natural experiments of history that will help us all
get rich? I propose to go over two batches of natural experiments that
will give you insights into how to get rich.

The first batch of natural experiments concerns understanding the
effects of isolation and of group size and of communication with other
groups on the productivity of human societies. Let's learn from the
extreme examples of isolation of human societies. If isolation has any
effect on human societies, the places we're most likely to see that
effect are the histories of those two islands off southeastern Australia
called Tasmania and Flinders Island. They lie about 200 miles off the
southeast coast of Australia and are separated today from Australia by
Bass Straits, but those straits are relatively shallow, so their floor
lay above sea level at glacial times of low sea level up to about 10,000
years ago. The Bass Straits between Tasmania and Australia were then dry
land, and Tasmania was part of the Australian mainland, just as Britain
used to be part of the European mainland. When the glaciers melted, sea
level rose and cut off Tasmania from the Australian mainland. So when
Tasmania and Flinders were part of the Australian mainland, Australian
Aborigines walked down to Tasmania and Flinders from the mainland.

And then 10,000 years ago the glaciers melted, sea level rose, and
Tasmania became cut off from mainland Australia by Bass Straits, which
are really rough waters. In addition, the watercraft of the Tasmanians
were wash-through rafts that got waterlogged and sank after about a
dozen hours. The result was that the boats of the Tasmanians could not
reach Australia, and the boats of the mainland Aboriginal Australians
could not reach Tasmania.

Thus, for the last 10,000 years the Tasmanians represented a study of
isolation unprecedented in human history except in science fiction
novels. Here were 4,000 Aboriginal Australians cut off on an island, and
they remained totally cut off from any other people in the world until
the year 1642, when Europeans "discovered" Tasmania. What happened
during those 10,000 years to that isolated 4,000-person society? And
what about nearby Flinders Island, which originally supported a
population of 200 cut-off Aboriginal Australians? -- what happened to
that tiny isolated society of 200 people during those 10,000 years?

When Europeans discovered Tasmania in the 17th century, it had
technologically the simplest, most "primitive" human society of any
society in the modern world. Native Tasmanians could not light a fire
from scratch, they did not have bone tools, they did not have
multi-piece stone tools, they did not have axes with handles, they did
not have spear-throwers, they did not have boomerangs, and they did not
even know how to fish. What accounts for this extreme simplicity of
Tasmania society? Part of the explanation is that during the 10,000
years of isolation, the Aboriginal Australians, who numbered about
250,000, were inventing things that the isolated 4,000 Tasmanians were
not inventing, such as boomerangs. Incredibly, though, archeological
investigations have shown one other thing: during those 10,000 years of
isolation, the Tasmanians actually lost some technologies that they had
carried from the Australian mainland to Tasmania. Notably, the
Tasmanians arrived in Tasmania with bone tools, and bone tools disappear
from archeological record about 3,000 years ago. That's incredible,
because with bone tools you can have needles, and with needles you can
have warm clothing. Tasmania is at the latitude of Vladivostok and
Chicago: it's snowy in the winter, and yet the Tasmanians went about
either naked or just with a cape thrown over the shoulder.

How do we account for these cultural losses and non-inventions of
Tasmanian society? Flinders Island was even more extreme -- that tiny
society of 200 people on Flinders Island went extinct several millenia
ago. Evidently, there is something about a small, totally isolated human
society that causes either very slow innovation or else actual loss of
existing inventions. That result applies not just to Tasmania and
Flinders, but to other very isolated human societies. There are other
examples. The Torres Strait islanders between Australia and New Guinea
abandoned canoes. Most Polynesian societies lost bows and arrows, and
lost pottery. The Polar Eskimos lost the kayak, Dorset Eskimos lost dogs
and bow drills, and Japan lost guns.

To understand these losses in extreme isolation, the easiest case to
understand is Japan, because the loss of firearms in Japan was witnessed
and described. It took place in a literate society. Guns arrived in
Japan around 1543 with two Portuguese adventurers who stepped ashore,
pulled out a gun, and shot a duck on the wings. A Japanese nobleman
happened to be there, was very impressed, bought these two guns for
$10,000, and had his sword-maker imitate them. Within a decade, Japan
had more guns per capita than any other country in the world, and by the
year 1600 Japan had the best guns of any country in the world. And then,
over the course of the next century, Japan gradually abandoned guns.

What happened was that the Samurai, the warrior class in Japan, had been
used to fighting by standing up in front of their armies and making a
graceful speech, the other opposing Samurai made an answering graceful
speech, and then they had one-on-one combat. The Samurai discovered that
the peasants with their guns would shoot the Samurai while the Samurai
were making their graceful speeches. So the Samurai realized that guns
were a danger because they were such an equalizer. The Samurai first
restricted the licensing of gun factories to a hundred factories, and
then they licensed fewer factories, and then they said that only three
factories could repair guns, and then they said that those three
factories could make only a hundred guns a year, then ten guns a year,
then three guns a year, until by the 1840s when Commodore Perry came to
Japan, Japan no longer had any guns. That represents the loss of a very
powerful technology.

This loss was possible only in Japan because of its isolation; there
were no other neighbors threatening Japan. When firearms arrived in
Europe, there were European princes who similarly banned firearms, and
there were European princes who banned printing, but you can guess what
happened. When a prince in the middle of Europe banned firearms, within
a short time the prince next door who did not ban firearms either walked
in and conquered, or else the prince who banned firearms quickly
realized his or her mistake and reacquired firearms from next door. The
banning of the guns could work only in isolated Japan, where there were
no neighbors as a threat, and where there were no neighbors from whom to
reacquire the technology.

So these stories of isolated societies illustrate two general principles
about relations between human group size and innovation or creativity.
First, in any society except a totally isolated society, most
innovations come in from the outside, rather than being conceived within
that society. And secondly, any society undergoes local fads. By fads I
mean a custom that does not make economic sense. Societies either adopt
practices that are not profitable or for whatever reasons abandon
practices that are profitable. But usually those fads are reversed, as a
result of the societies next door without the fads out-competing the
society with the fad, or else as a result of the society with the fad,
like those European princes who gave up the guns, realizing they're
making a big mistake and reacquiring the fad. In short, competition
between human societies that are in contact with each other is what
and those differences are very instructive. Let me give you two
examples from case studies carried out by the McKinsey Corporation, an
economics study industry based in Washington. These two examples involve
the German beer industry and the Japanese food-processing industry.

What about the German beer industry? Well, the Germans are very
efficient in some of their industries. The German metal-working industry
and the German steel industry are equal in productivity to those of the
United States, but the German beer-producing industry has a productivity
only 43% that of the United States. And it's not that the Germans make
bad beer; the Germans make wonderful beer. Whenever my wife and I go to
Germany, we take along an extra suitcase specifically for the purpose of
filling it up with bottles of German beer, which we take back and dole
out to ourselves for the year after each of our trips to Germany. Why,
then, since the Germans make such great beer, and since their industrial
organization works so successfully for steel and metal, can't they
achieve a successful industrial organization for beer?

It turns out that the German beer industry suffers from small-scale
production. There are 1,000 little local beer companies in Germany,
shielded from competition with each other because each German brewery
has virtually a local monopoly, and shielded from competition with
imports. The United States has 67 major beer breweries, producing 23
billion liters of beer per year. Germany has 1,000 major beer breweries,
producing only half as much beer per year as the United States. That's
to say that the average brewery in the U.S. produces 31 times more beer
than the average brewery in Germany.

That fact results from German local tastes and German government
policies. German beer drinkers are fiercely loyal to their local brand
of beer. And so there is no national brand of beer in Germany, analogous
to Budweiser or Miller or Coors in the United States. Instead, most
German beer is consumed within 30 miles of the place where it is brewed.
And any of you who have been in Germany know that Germans love their
local beer and loathe the beer that comes from next door. The result is
that the German beer industry cannot profit from economies of scale. In
the beer industry, as in other industries, production costs decrease
greatly with size. The bigger the refrigerator unit for making the beer,
and the longer the bottle-filling line, the cheaper is the cost of
brewing beer. So these tiny German beer industries are relatively
inefficient. There's no competition; there are just 1,000 local

That outcome, of Germans having their local beer loyalties, is
reinforced by German government law. The German government makes it hard
for foreign beers to compete on the German market. The German government
has so-called beer purity laws. The German government specifies exactly
what can go into beer, and not surprisingly what can go into beer is
what German breweries put into beer, and it's not what American, French,
and Swedish breweries like to put into beer. So it's difficult for
foreign breweries to compete on the German beer market. The result is
that German beer is not exported very much. Any of you who like to buy
Lowenbrau in the U.S. should look at the label in the supermarket: your
U.S.-bought Lowenbrau is not brewed in Germany, it's brewed on license
in the United States with American productivity and American
efficiencies of scale.

The same inefficiency turns out to characterize some other German
industries. The German soap industry and the German consumer electronics
industry are also inefficient; their companies are not exposed to
competition with each other, nor are they exposed to foreign
competition, and so they do not acquire the best practices of
international industry. But that disadvantage is not true for the German
metal-producing industry or steel industry. There, big German companies
compete with each other and they compete internationally, and therefore
they are forced to acquire best international practices through

There you have an example from the German beer industry about the
disadvantages of having lots of small groups that are secretive and
don't compete with each other. The other example that I want to tell you
about is the Japanese food-processing industry. I mentioned that we
Americans are virtually paranoid about the efficiency of the Japanese,
and it's true for some Japanese industries, but not for their
food-processing industry. Japanese processed food is produced with an
efficiency 32% of American processed foods. There are 67,000 food
processing companies in Japan; there are only 21,000 in the United
States, although the U.S. has double Japan's population, so the average
food-processing company in the United States is six times bigger than
its Japanese counterpart. What is the reason why the Japanese
food-processing industry, like German beer industries, consists of small
companies with local monopolies?

It turns out to be basically the same two reasons as with German beer:
namely, local tastes creating local monopolies, and government policies.
The Japanese are fanatics for fresh foods. Any of you who have been to
Japan, as my wife and I were in October, will remember what it says on
Japanese containers. In the United States, when you go to the
supermarket, there's one date on the container, the date by which you're
supposed to throw away that bottle of milk. In Japan there are three
dates on the container: there's the date when the milk was manufactured,
and there's the date when the milk arrived at the supermarket, and then
there's the date when the milk should be thrown away, and these dates
are in big letters; the Japanese really care about the dates. So the
result is that milk production in Japan always starts at one minute past
midnight, so that the milk that goes to market that morning is today's
milk. If milk had been produced at 11:59 p.m., the milk company would
have to stamp on its container that this milk was made yesterday, and no
Japanese person would buy it. The result is again that Japanese
food-processing industries enjoy local monopolies. Obviously, a milk
producer up in Hokkaido, northern Japan, is not going to be able to
compete in Kyushu, in southern Japan, with a Kyushu producer, because of
the several days in transit from Hokkaido. By the time a carton arrives
in Kyushu, the people will read on the container that this milk is three
days old, and no Japanese person would buy it.

So that's one thing that creates local monopolies for food production in
Japan: Japanese fanaticism about really fresh food. And the second thing
is Japanese government policy, which reinforces these local monopolies.
The Japanese government obstructs the import of foreign processed food
by slapping on a ten-day quarantine. And because the Japanese care about
food that was produced that very day, naturally by the time that
American beef, chicken, or whatever arrives at the supermarket and the
date says ten days old, the Japanese are not very enthusiastic about
buying those American products. And there are other restrictions that
the Japanese government imposes on foreign imports.

The result is that Japanese food-processing industries are not exposed
to domestic competition, they're all local monopolies, they're not
exposed to foreign competition, and they don't learn the best methods in
the international trade for producing food. And the result is that, in
Japan, Japanese beef costs $200 a pound. My wife and I had heard about
that before we went to Japan, but what we did not realize until we were
brought into a supermarket by my wife's Japanese cousin is that chicken
in Japan costs $25 a pound. The reason the Japanese can get away with
that is that Japanese chicken producers are not exposed to competition
with super-efficient American chicken producers.

Now all those features are not true for some other Japanese industries.
The Japanese steel industry, the Japanese metal industry, the Japanese
car industry, their car-part industry, and their electronic industries
have productivities greater than our American counterparts. But the
Japanese soap industry, and the Japanese beer industry, and the Japanese
computer industry, like the Japanese food-processing industry, are not
exposed to competition, do not apply the best practices, and so have
ended up with productivities below those of corresponding industries in
the United States.

Now let's finally apply these lessons to comparing different industries
or industrial belts within the United States. I mentioned that when I
was growing up, Route 128 outside of Boston led the world in
productivity for an industrial belt, but Route 128 has now fallen behind
Silicon Valley. Since my book "Guns, Germs, and Steel" was published,
I've spent a lot of time talking with people from Silicon Valley and
some from Route 128, and they tell me that the corporate ethos in these
two industrial belts is quite different. Silicon Valley consists of lots
of companies that are fiercely competitive with each other, but
nevertheless there's a lot of collaboration, and despite the competition
there is a free flow of ideas and a free flow of people and a free flow
of information between these companies that compete with each other. In
contrast, I'm told that the business of Route 128 are much more
secretive, and insulated from each other like Japanese milk-producing

Or again, what about the contrast between Microsoft and IBM? Again,
since my book was published, I've acquired friends at Microsoft, and
I've learned about Microsoft's organization, which is quite distinctive.
Microsoft has lots of units, with free communication between units, and
each of those units may have five to ten people working in them, but the
units are not micro-managed, they are allowed a great deal of freedom in
pursuing their own ideas. That unusual organization at Microsoft, broken
up in to a lot of semi-independent units competing within the same
company, contrasts with the organization at IBM, which until four years
ago had much more insulated groups. A month ago, when I was talking in
the industrial belt of North Carolina, the Raleigh-Durham area
industrial belt, I met someone who is on the board of directors of IBM,
and that person told me, Jared, what you say about IBM was quite true
until four years ago: IBM did have this secretive organization which
resulted in IBM's loss of competitive ability, but then IBM acquired a
new CEO who changed things drastically, and IBM now has a more
Microsoft-like organization, and you can see it, I'm told, in the
improvement in IBM's innovativeness.

So what this suggests is that we can extract from human history a couple
of principles. First, the principle that really isolated groups are at a
disadvantage, because most groups get most of their ideas and
innovations from the outside. Second, I also derive the principle of
intermediate fragmentation: you don't want excessive unity and you don't
want excessive fragmentation; instead, you want your human society or
business to be broken up into a number of groups which compete with each
other but which also maintain relatively free communication with each
other. And those I see as the overall principles of how to organize a
business and get rich.

But, let me conclude by emphasizing some obvious restrictions. I'm sure
all of you are already thinking to yourselves, "But, but, but, he's
forgot... but, but, but..." Yes, let's go back to those but-but-buts.
One restriction is, I mentioned at the beginning, "all other things
being equal". Obviously the best organization is not going to help with
an idiot as a CEO, and the success of Microsoft certainly depends, at
least in part, on the unusual qualities of Bill Gates, as well as on the
unusual organization of Microsoft.

In addition, I've been talking about conditions to maximize productivity
and creativity and moneymaking ability. There are other considerations
in organized human groups, and there are conditions under which
productivity is not the thing you're most interested in. There are
conditions where more centralization may be appropriate. For example,
during a war, you do not want your air force, army, and navy to be
fiercely competing with each other, but instead you want during a war
more centralized control than you do in peace time. And there are also
human groups for which productivity and differential money-making
ability are not the overriding consideration. I don't want you to go
home tonight and each of you to say to your spouse or significant other,
"Darling, I've just heard this guy Jared Diamond, who says that within
human groups competition is what spurs productivity and innovation, and
so I think we need to follow his advice in our household. For the next
month let's see which of us earns a bigger income, and at the end of the
month the bigger income-producer will keep on with the job, and the one
of us who has lower income and is less efficient can turn to scrubbing
the floors and shopping at the supermarkets." That just illustrates:
there are other considerations in a marriage than optimizing

Again, I don't want you to go home to your several children, and say,
"Sweetie-pies, I heard this talk today by this guy Jared Diamond who
enunciated some principles that I think would be really good for rearing
children. We're going to see what your grades are at mid-term, and based
on those grades, whichever one of you comes closer to getting all A's,
that one we will support to the hilt, private schools, college, whatever
you need, whereas those of you who get poor grades can start jobs as a
shoe-shine boy or girl" -- No! In a family, and in some other human
groups, productivity is not the appropriate consideration for judging
the best organization of the group.

Nevertheless there are some human groups where productivity is indeed a
significant consideration. And that certainly includes businesses,
industrial belts, and to a considerable degree, countries. In order to
understand how to organize these businesses, we could perform natural
experiments. We could set up, if we were rich enough, a hundred
businesses, organized a hundred different ways, see which businesses
went bankrupt, and after 20 years figure that we now have the correct
industrial organization. But that's an inefficient way to do it. We can
instead learn from the comparative approach, by looking to natural
experiments of history. I hope that some of you will be able to apply
these lessons to acquiring the wealth that has so far eluded me.


The biggest things are always the easiest to do because there is no
-- William Van Horne