Clay Shirky's Writings About the Internet
Economics and Culture, Media and Community, Open Source

Half the World

Version 1.03 | September 3, 2002 | Version History
First published on the 'Networks, Economics, and Culture' mailing list

A good deal has been written about the digital divide, the technological gap that exists between the developed and developing world. If you wanted a striking illustration of the problem, you could turn to Thabo Mbeki's speech at the Information Society and Development Conference in 1996, where he told the delegates "Half of humanity has not yet made a phone call."

Or, if you prefer, Kofi Annan's 2000 speech to the Australian Press Club, where he said "Half the world's population has never made or received a phone call." Or Thomas Homer-Dixon's version, from a speech in 2001: "...half the people on the planet have never made a phone call." Greg LeVert of MCI said it at a telecom conference in 1994; Richard Klugman of PaineWebber said it in The Economist in 1995; Jeffery Sachs and Al Gore both said it in 1998; Reed Hundt and Steve Case both said it 2000; Michael Moore and Newt Gingrich both said it in 2001, as did Carly Fiorina and Melinda Gates; and not content with merely half, Tatang Razak of Indonesia told the UN's Committee on Information "After all, most of the people in the world have never made a phone call...", in speech from April of this year.

The phrase "Half the world has never made a phone call" or some variation thereof has become an urban legend, a widely believed but unsubstantiated story about the nature of the world. It has appeared countless times over the last decade, in essentially the same form and always without attribution. Where did that phrase come from? How did it take on such a life of its own? And, most importantly, why has it gotten so much airtime in the debate over the digital divide when it is so obviously wrong?

You Can't Keep A Good Factoid Down

The Phrase is such a serious and important statistic that only the boorish would question its accuracy. There is a kind of magical resonance in advancing arguments on behalf of half the world's people, and it allows the speaker to chide the listener for harboring any lingering techno-optimism ("You think there's a revolution going on? Half the world has never made a phone call!") The Phrase establishes telecommunications access as a Big Problem and, by extension, validates the speaker as a Thinker-About-Big-Problems.

But saying "Half the world has never made a phone call" makes no more sense than saying "My car goes from 0 to 60" or "It rained 15 inches." Without including the element of time, you cannot talk about rate, and it is rate that matters in dynamic systems. Half the world had never made a phone call on what date? And what has the rate of telecom growth been since that date? Because it is that calculation and only that calculation which could tell us anything important about the digital divide.

Static Statements About Dynamic Systems

Virginia Postrel, in her book "The Future and Its Enemies", (ISBN: 0684862697) suggests that the old distinctions of right and left are now less important than a distinction between stasists and dynamists. Stasists are people who believe that the world either is or should be a controlled, predictable place. Dynamists, by contrast, see the world as a set of dynamic processes. This distinction is the key to The Phrase. Anyone who uses it is affirming the stasist point of view, even if unconsciously, because they are treating telecommunications infrastructure as if it were frozen in time.

To think about rate, we need three things: original time, elapsed time, and speed of change. The figure first appeared in print in late 1994, when the Toronto Sun quoted it as part of Greg LeVert's speech at TeleCon '94. (Mr. LeVert was no stasist -- though he seems to have accidentally bequeathed us The Phrase, no one now remembers that he introduced it as a way of dramatizing the magnitude of the coming change, and went on to predict a billion new phones by 2000.) Therefore we can restore the question of rate by asking what has happened to the number of telephones in the world between the beginning of 1995 and now.

Restoring Rate

The ITU estimates that there were approximately 689 million land lines at the beginning of 1995, and a little over 1 billion by the end of 2000, the last year for which they have figures available. This is an average annual growth rate of just over 7%, and a cumulative improvement in that period of over 50%, meaning that the first two-thirds of the world's phone lines were run between 1876 and 1994, and the remaining third were run between 1995 and 2000. Put another way, half again as many land lines were run in the last 6 years of the 20th century as were run in the whole previous history of telephony. So much for stasis.

Of course not all of this growth touches the problem at hand -- a new phone line in a teenager's room may increase several sorts of telecom statistics, but the number of people making their first phone call isn't one of them. Since The Phrase concerns the digital divide, we should concentrate on telecom growth in the less developed world.

From the beginning of 1995 to the end of 2000, 8 countries achieved compound average growth rates of 25% or more for land lines per 100 people (against a world average of 7%), meaning they at least tripled the number of land lines over the whole period. They were, in order of rate, Sudan (which improved six-fold), Albania, China, Sri Lanka, Viet Nam, Ghana, Nepal, and Cambodia -- not exactly the G8. China alone went from 41 million land lines to 179 million in those 6 years. And there were 35 additional countries, including India, Indonesia, and Brazil, with annual growth of between 10 and 20% from 1995 to 2000, meaning they at least doubled the number of land lines in that period.

And mobile telephony makes the change in land lines look tectonic. In 1995, there were roughly 91 million cellular subscribers. By 2000, the number had risen to 946 million, a ten-fold increase. Twenty-seven countries had growth rates of over 100% annually, meaning that, at a minimum, they doubled and doubled again, 6 times, achieving better than sixty-fold cumulative growth (not 60%, but a factor of 60), and 22 of those had better than hundred-fold growth. Senegal went from around 100 subscribers (not 100 thousand subscribers, 100 subscribers) to 390 thousand. Egypt went from 7 thousand to almost 3 million. Romania went from 9 thousand to almost 4 million. An additional 44 nations with no measurable wireless penetration in 1995 had acquired wireless subscribers by 2001.

Because wireless infrastructure does not require painstaking building-by-building connectivity, nor is it as hampered by state-owned monopolies, it offers a way for a country to increase its telephone penetration extremely quickly. By the end of 2000, there were 25 countries where cell phone users made up between two-thirds and nine-tenths of the connected populace. In these countries, none of them wealthy, a new telecommunications infrastructure was deployed from scratch, during the same years that keynote speakers and commencement invitees were busily and erroneously informing their listeners that half the world had never made a phone call.

Two Answers

So, in 2002, what can we conclude about the percentage of the world that has made a phone call?

The first, and less important answer to that question goes like this: Between 1995 and 2000, the world's population rose by about 8%. Meanwhile, the number of land lines rose by 50%, and the number of cellular subscribers by over 1000%. Contrary to the hopelessness conveyed by The Phrase, telephone penetration is growing much faster than population. It is also growing faster in the developing world than in the developed world. Outside the OECD, growth was about 130% for land lines and over 2,300% for cellular phones -- 14 million subscribers at the beginning of 1995 and 342 million by the end of 2000. If we assume that LeVert's original guess of half was right in 1994 (a big if), the new figure would be "Around two-thirds and still rising."

There is another answer to that question though, which is much more important: It doesn't matter. No snapshot of telephone penetration matters, because the issue is not amount but rate. If you care about the digital divide, and you believe that access to communications can help poor countries to grow, then pontificating about who has or hasn't made a phone call is worse than a waste of time, it actively distorts your view of the possible solutions because it emphasizes a stasist attitude.

Though a one-time improvement of 5% is better in the short run than a change that improves annual growth by 1%, the latter solution is better in the medium run, and much better in the long run. As as everything from investment theory to Moore's Law has shown, it's hard to beat compound growth, and improving compound growth in the spread of telephones requires reducing the barriers between demand and supply. Some countries have had enormously favorable rates of growth in the last decade, so we should ask what has gone right in those countries. It turns out that money is less important than you might expect.

Examples from the Real World

In 1995, Brunei had almost twice as many land lines and 50 times as many cell phones per capita as Poland, not to mention more than twice the per capita GDP. By the end of 2000, Poland exceeded Brunei in land lines, and had even matched it in cell phone penetration. Brunei is smaller, easier to wire, and much much richer, but Poland has something that all Brunei's money can't buy -- a dynamic economy. Ireland similarly outgrew Denmark and Egypt outgrew Kuwait, even though the faster growing countries were the poorer ones.

The Democratic Republic of Congo lost 16 thousand of its 36 thousand fixed lines but grew overall, because it added 140 thousand cell phones. Its unsurprising that people would abandon the state as a provider of telephones in a country riven by civil war. What is surprising is that Venezuela had the same pattern. Venezuela, with a monopoly telecom provider, saw its per capita penetration of land lines fall by 0.3% annually, while cell phone use exploded. In both cases, the state was an obstacle to telephone usage, but the presence of a private alternative meant that telephone penetration could nevertheless increase.

Bangladesh, with a per capita GDP of $1,570, has had annual cellular growth of nearly 150%, in part because of programs like the Grameen Bank's Village Phone, which loans a phone, collateral free, to women in Bangladeshi villages, who in turn resell the service to their neighbors. This is double leverage, as it not only increases the number of phones in use, but also increases the number of users per phone.

These examples demonstrate what makes The Phrase so pernicious. Something incredibly good is happening in parts of the world with dynamic economies, and that is what people concerned with the digital divide should be thinking about. If the world's poor are to be served by better telecommunications infrastructure, there are obvious things to be done. Make sure individuals have access to a market for telephone service. Privatize state telecom companies, introduce competition, and reduce corruption. And perhaps most importantly, help stamp out static thinking about telecommunications wherever it appears. Economic dynamism is a far better tool for improving telephone use than any amount of erroneous and incomplete assertions on behalf of half the world's population, because while The Phrase has remained static for the last decade or so, the world hasn't.

First published on the 'Networks, Economics, and Culture' mailing list
Subscribe to the list

Version History: This is version 1.03 of this essay, from September 3, 2002. Version 1.0 was published June 30, 2002, 1.01 appeared on July 3, and 1.02 appeared July 19th.

Changes from 1.0 to 1.01: I realized after publishing the original essay that the ITU statistics run from January 1 to January 1, meaning that the 1995-2001 period is 6 elapsed years, not 7. To make this clear, I re-wrote the sections "Restoring Rate", "Two Answers", and the first paragraph of "Lessons from the Real World" to refer to "the end of 2000" or "2000" as the stop date for the growth figures, and to list the elapsed time as 6 years.

This means that the changes described here happend in less time, i.e. at a faster rate, than I originally suggested. The only place it materially affects the numerical conclusions is in 30 countries which had in excess of 100% annual growth. Only 26 of these countries had 100-fold growth (i.e. a compound average growth rate of above 116%). The remaining 4 grew between 64 and 100-fold.

Changes from 1.01 to 1.02: Re-wrote the second paragraph. I had incorrectly located Kofi Annan's speech at the ITU, rather than at the Australian Press club (though he has said it in several other venues as well, including his Millenium Report, and the World Economic Forum in Davos in 2001.) Fiorina's use of The Phrase dates from 2001, not 1999. I also added the references to Steve Case and Melinda Gates. The paragraph in its original form can be found at the link to version 1.0, above.

Changes from 1.02 to 1.03: Noted that the 8 countries listed as having the highest percentage of wired telecom growth, in the paragraph beginning "From the beginning of 1995 to the end of 2000...", all had growth rates in excess of 25%, not merely 20%. (Thanks to Jennifer Weaver of Wired, who edited a version of this article to appear there.)

Wired telecom penetration in the developing world 2001 was 230% of what it was in 1995, meaning it grew 130%. Changed the 230% figure to 130% in the paragraph beginning "The first, and less important answer to that question..." to reflect the growth rate, rather than absolute penetration. Likewise changed the 2,400% figure for wireless to 2,300%, for the same reason.

Changed the wireless figures, in the paragraph beginning "And mobile telephony makes the change in land lines..." to include the 44 countries that created divide-by-zero errors when using the ITU statistics, because their measurable wireless penetration was 0 in 1995.

NOTES: The statistics on teledensity used here are drawn from the International Telecommunication Union ( The statistics page is at, and the documents concerning compound and overall growth in main telephone lines and cellular subscribers between 1995 and 2001 are at and respectively.

The estimates for the developing world were derived by treating the membership of the Organization for Economic Co-operation and Development (OECD, as a proxy for the developed world. Growth in the developing world was then derived by recalculating the totals from the ITU documents after removing the 30 countries in the OECD.

The figure for population growth was derived from the US Census Bureau's estimates of world population in 1995 and 2001.

The original Toronto Sun piece, from the Business section of October 13th of 1994) read:

  So you think the world is wired? Half the world's population -- an
  astounding three billion people -- has never made a phone call, a
  telecommunications conference was told Wednesday.
  "Most people on Earth live more than two hours from a telephone,"
  Greg LeVert, president of U.S. giant MCI's Integrated Client
  Services Division, told delegates to TeleCon '94 in Toronto.
  Things are changing fast, though.
  "Nearly a billion more people will have access to a telephone by
  2000," LeVert said. Clay Shirky's Writings About the Internet
Economics and Culture, Media and Community, Open Source