University of Southern California Dissertations and Theses

Mapping out the transition toward information societies: social nature, growth, and policies

(USC Thesis Other)

Mapping out the transition toward information societies: social nature, growth, and policies

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content MAPPING OUT THE TRANSITION 1
MAPPING OUT THE TRANSITION TOWARD INFORMATION
SOCIETIES:
SOCIAL NATURE, GROWTH, AND POLICIES

by

Martin Hilbert

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Communication)

December 2012

Copyright 2012 Martin Hilbert
Running head: MAPPING OUT THE TRANSITION

MAPPING OUT THE TRANSITION 2
Table of Contents
List of Tables 4
List of Figures 5
Abstract 7
Chapter One: Introduction 11
Theoretical Background: Societies in Transition 12
Chapter Two: The Social Nature of the Transition 15
The End Justifies the Definition: the Manifold Outlooks on the Digital Divide and
their Practical Usefulness for Policy-Making
,
16
Introduction 17
Theoretical background: Diffusion of innovations 19
How to define the digital divide analytically? 24
Who defines the digital divide in practice? 50
Conclusion: Impact over analytical coherence 63
When is Cheap, Cheap Enough to Bridge the Digital Divide? Modeling Income
Related Structural Challenges of Technology Diffusion in Latin America 67
Introduction 67
How to define the digital divide? 71
How important is income distribution for the digital divide? 75
How much resources are available for ICT? 79
How much does ICT access currently cost? 83
How does income distribution create ICT diffusion patterns? 89
What does our model tell us about future scenarios? 94
What are the odds for the digitally excluded? 106
Digital Gender Divide or Technologically Empowered Women in Developing
Countries? A Typical Case of Lies, Damned Lies, and Statistics
,
114
What is the digital divide? 117
What does the data say? 124
A small change in mindset can sometimes make a large difference 138
Chapter Three: Magnitude and Growth of the Transition 142
Information Societies or “ICT Equipment Societies”? Measuring the Digital
Information Processing Capacity of a Society in Bits and Bytes
,
143
Multiple dimensions of technology diffusion 144
The closing digital equipment divide 148
The three subsystems of information processing 153
The amount of digital information 156
The digital divide as a moving target 162
Limitations and corresponding research challenges 171
MAPPING OUT THE TRANSITION 3
The world’s Technological Capacity to Store, Communicate and Compute
Information
,
176
Previous work 177
Scope of our exercise 178
Information, not hardware with redundant data 181
Storage 183
Communication 185
Computation 189
Comparisons and growth rates 193
Perspectives 195
Chapter Four: Policy Actions for the Transition 197
Foresight Tools for Participative Policy-Making in Inter-Governmental Processes in
Developing Countries: Lessons Learned from the eLAC Policy Priorities Delphi 198
Regional agenda-building for digital development in developing countries 203
The eLAC Policy Priority Delphi 210
The five consecutive Delphi rounds 218
Evaluation of the effectiveness of the Delphi: the acceptance by the inter-
governmental process 225
Conclusions and lessons learned 232
References 239

MAPPING OUT THE TRANSITION 4
List of Tables
Table 1 Governmental spending on ICT in Chile, 2003, in percent of a total of US$ 205
million ............................................................................................................................... 55
Table 2 Results of discriminant analysis of household Internet access for the individual 78
Table 3 Annual ICT prices per inhabitant in US$, national average and per income
quintile .............................................................................................................................. 88
Table 4 ICT penetration rates per inhabitant; Affordable portion of ‘ICT access packet’;
national average and per income quintile ......................................................................... 91
Table 5 Required price reductions and subsidies for scenarios 1, 2 and 3; personal
communication spending; comparison to other national expenditures; weekly minutes in
public access center; all as national average and per income quintile .............................. 99
Table 6 Correlations and controlled correlations of gender with ICT usage in Latin
America; working and studying populations by men and women. ................................. 128
Table 7 Percentage of man/women that use the Internet and own a mobile phone in Latin
America; place of Internet usage, Internet use frequency. .............................................. 130
Table 8 Percentage of man/women that use the Internet and own a mobile phone; literacy,
working and income inequalities; in Africa 2007/08. ..................................................... 134
Table 9 Frequency of Internet usage; online service used by men and women in Latin
America. .......................................................................................................................... 136
Table 10 ICT equipment diffusion per 100 inhabitants in OECD and Latin America and
the Caribbean, 1996-2006 ............................................................................................... 150
Table 11 Price decline and performance increase in the technological frontier of all three
ICT subsystems, 1980-2005............................................................................................ 156
Table 12 Worldwide installed capacity to compute, communicate and store digital
information ...................................................................................................................... 162
Table 13Evolution of the world’s capacity to store, communicate and compute
information, absolute per capita, compound annual growth rate (CAGR), and percentage
in digital format............................................................................................................... 195

MAPPING OUT THE TRANSITION 5
List of Figures
Figure 1 Overview: “Mapping Out the Transition toward Information Societies” ............ 8
Figure 2 Social network of the spread of diseases and online communities .................... 20
Figure 3 Hypothetical example of homogeneous mixing Overview: “Mapping Out the
Transition toward Information Societies” ......................................................................... 22
Figure 4 Social network schematization of different perspectives on the digital divide .. 25
Figure 5 Diffusion of most common ICT with individuals. Capacity to communicate
through fixed line, mobile telephony and Internet in optimally compressed kilobits per
second per capita ............................................................................................................... 29
Figure 6 Availability of Email in local governments. Country’s access (ICT Development
access subindex) and income for 2007 (n=147) ............................................................... 33
Figure 7 Computer and Internet access in schools (2004/5). Mobile penetration in Brazil
(2005). Multivariate analysis of household Internet access for the individual ................. 37
Figure 8 ICT access at different locations. sophistication of Internet usage .................... 44
Figure 9 ICT policy coherence versus ICT infrastructure and access index, 2007. ICT
policy coherence versus online presence of e-government index, 2007 ........................... 60
Figure 10 Monthly household spending on communication, as % of total spending (top)
and as absolute values in current US$ (bottom) ............................................................... 82
Figure 11 Fighting longstanding discrimination with digital means .............................. 140
Figure 12 Schematization of the three basic information processes ............................... 155
Figure 13 Capacity to communicate through fixed line, mobile telephony and Internet 163
Figure 14 Capacity to communicate according to technology........................................ 165
Figure 15 Capacity to transmit information through radio and TV (terrestrial, satellite,
cable) ............................................................................................................................... 167
Figure 16 Capacity to store information in hard disks of PCs and laptops..................... 168
Figure 17 Capacity to compute information with PCs, notebooks, and mobile phones . 170
Figure 18: The three basic information operations and its most prominent technologies.
......................................................................................................................................... 180
MAPPING OUT THE TRANSITION 6
Figure 19 World’s technological installed capacity to store information, in optimally
compressed Megabytes (MB) per year, for 1986, 1993, 2000 and 2007, semi-log plot. 184
Figure 20 World’s technological effective capacity to broadcast information, in optimally
compressed Megabytes (MB) per year, for 1986, 1993, 2000 and 2007, semi-log ........ 187
Figure 21 World’s technological effective capacity to telecommunicate information, in
optimally compressed Megabytes (MB) per year, for 1986, 1993, 2000 and 2007, semi-
logarithmic plot ............................................................................................................... 188
Figure 22 World’s technological installed capacity to compute information on general-
purpose computers, in millions instructions per second (MIPS), distribution for 1986,
1993, 2000 and 2007, semi-logarithmic plot .................................................................. 191
Figure 23 Annual growth of installed general-purpose computational capacity as
percentage of all previous computations since 1977 (yeart / Σ[1977, yeart-1]). ............ 192
Figure 24 eLAC Policy Priority Delphi from eLAC2007 to eLAC2010........................ 212
Figure 25 Distribution of 1,274 online contributions (Delphi Round one, two and four)
according to education, professional affiliation, gender and geographic representation of
participants ...................................................................................................................... 214
Figure 26 Polarization Index for Subregions and Sectors of professional affiliation of
participants ...................................................................................................................... 220
Figure 27 Goals of eLAC2010 in comparison with goals of eLAC2007 ....................... 226
Figure 28 Goals of eLAC2010 in comparison to goals proposed by Policy Delphi results
......................................................................................................................................... 230

MAPPING OUT THE TRANSITION 7
Abstract
This research thesis sheds lights on different aspects of the transition toward
information societies. It consists of a collection of interrelated studies that analyze in
more rigorous terms three main and complementary aspects of the transition (see Figure
below). After and introductory CHAPTER ONE, the consecutive CHAPTER TWO of
this thesis looks at the social nature of the current transition toward the information
society, which is characterized by a diffusion process that is known as digital divide. This
chapter focuses on the socio-demographic characteristics of the transition, and
characterizes its bottlenecks, such as the cost-income relation of ICT and users, as well as
its opportunities, such as the opportunity to fight long-standing gender inequalities.
CHAPTER THREE focuses not only on equality, but also on growth of the world’s
information and communication capacity in absolute terms. The chapter consists of two
sections that quantify the magnitude and growth of information in the information
society, measured directly in bits and bytes. This provides insights into the speed and
general pattern of the transition from analog to digital information processing in society.
Both chapters combined provide complementary insights into what have been
traditionally the two main pillars of socio-economic development: equity and growth. In
this case the focus is set on the equality and growth of technologically mediated
information. Various particularities of the transition become evident, such as the
exponential rates of change of the transition, the all-pervasiveness of ICT in the social
realm, and the unequal diffusion process. The final CHAPTER FOUR studies a concrete
example of successful policy making in the digital age that takes these particularities
into consideration. The case study focuses on a foresight Delphi exercise aimed at
MAPPING OUT THE TRANSITION 8
identifying future policy priorities that offered input into the inter-governmental
negotiation of an Action Plan in Latin America. It is believed to have been the most
extensive online participatory policy-making foresight exercise in the history of
intergovernmental processes in the developing world. The process of policy-making in
this international multi-stakeholder Delphi embraces the particular characteristics of the
transition toward Information Societies by design.

Figure 1 Overview: “Mapping Out the Transition toward Information Societies”

The Chapters consists of a collection of complementary studies, which use a diverse
array of methodologies and data sources to map out diverse aspects of the transition
toward this new form of socio-economic organization. The three main Chapters consist of
6 articles that have been produced during the time of my doctoral program at USC’s
MAPPING OUT THE TRANSITION 9
Annenberg School of Communication (since August 2008). Chapter Two consists of
three articles (resulting in three complementary sections), Chapter Three consists of two
articles, and the final Chapter Four of one article. These articles are by now all published
in recognized peer-reviewed Journals, all of which are indexed in the Thomson Reuters
Social Science Index. Some of these Journals are leading in their fields (such as
Technological Forecasting and Social Change, the world’s leading journal in foresight
studies, with a 5-year Thomson Reuters Journal Citation Impact factor of 2.2; or World
Development, the world’s leading journal in international development studies, with a 5-
year impact factor of 2.5), while other Journals are among the most recognized outlets for
multidisciplinary science research in general (such as Science, with a 5-year impact
factor of 32). Each of the articles has passed the peer-review of at least 2 specialized
reviewers (most often 3 or 4), plus one editor, resulting in an estimated number of over
20 specialized and independent reviewers for this collection of articles. The articles, in
their order of inclusion in this thesis, include:

Hilbert, M. (2011). The end justifies the definition: The manifold outlooks on the
digital divide and their practical usefulness for policy-making. Telecommunications
Policy, 35(8), 715-736. doi:10.1016/j.telpol.2011.06.012

Hilbert, M. (2010). When is Cheap, Cheap Enough to Bridge the Digital Divide?
Modeling Income Related Structural Challenges of Technology Diffusion in Latin
America. World Development, 38(5), 756-770. doi:10.1016/j.worlddev.2009.11.019

MAPPING OUT THE TRANSITION 10
Hilbert, M. (2011). Digital gender divide or technologically empowered women in
developing countries? A typical case of lies, damned lies, and statistics. Women’s Studies
International Forum, 34(6), 479-489. doi:10.1016/j.wsif.2011.07.001

Hilbert, M., López, P., & Vásquez, C. (2010). Information Societies or “ICT
Equipment Societies?” Measuring the Digital Information-Processing Capacity of a
Society in Bits and Bytes. The Information Society, 26(3), 157-178.
doi:10.1080/01972241003712199

Hilbert, M., & López, P. (2011). The World’s Technological Capacity to Store,
Communicate, and Compute Information. Science, 332(6025), 60 -65.
doi:10.1126/science.1200970

Hilbert, M., Miles, I., & Othmer, J. (2009). Foresight tools for participative policy-
making in inter-governmental processes in developing countries: Lessons learned from
the eLAC Policy Priorities Delphi. Technological Forecasting and Social Change, 76(7),
880-896. doi:10.1016/j.techfore.2009.01.001

MAPPING OUT THE TRANSITION 11
Chapter One: Introduction

More than a quarter of a century ago, the USC Annenberg scholar Beniger (1986)
already enlisted dozens of works that took a outlook on the social transformations
provoked by the massive introduction of modern information and communication
technologies (ICT). Since these early days, the resulting form of social organization has
been given many names, including the “Computerized Society” (Martin and Norman,
1970), “Information Revolution” (Lamberton, 1974), “Electronics Revolution” (Evans,
1977), the “Information Economy” (Porat, 1977), the “Microelectronics Revolution”
(Forester, 1980), “Information Technology Revolution” (Forester, 1985), “Network
Society” (Castells, 2009), “age of Information and Communication Technology”
(Freeman and Louça, 2002), “the fifths Kondratieff” (Perez, 1983; 2004), “Information
Age” (Jorgenson, 2005; Castells, 2009; Brynjolfsson and Saunders, 2010), and
“Information Society” (Masuda, 1980; Martin and Butler, 1981, Miles, 1988, Webster,
2006; Mansell, 2009). This last term has stuck with many and even started to dominate
the global political agenda. Between 2003 and 2005 the highest possible political level of
the world gathered to discuss the social, political, economic and cultural implications of
this revolution during the “World Summit on the Information Society”.
1

1
A World Summit is a gathering of all acting Head of States or government. The World Summit on the
Information Society (WSIS) was held in two phases. The first phase took place in Geneva hosted by the
Government of Switzerland from 10 to 12 December 2003, and the second phase took place in Tunis
hosted by the Government of Tunisia, from 16 to 18 November 2005: http://www.itu.int/wsis
MAPPING OUT THE TRANSITION 12
Theoretical Background: Societies in Transition
All of these complementary descriptions of the ongoing social transformations find
their collective theoretical foundation in the Schumpeterian notion of socio-economic
evolution (Schumpeter, 1939; Perez, 2004; Freeman and Louça, 2002, Castells, 2009),
which holds that human progress “goes on in units separated from each other by
neighborhoods of equilibrium. Each of those units, in turn, consists of two distinct
phases, during the first of which the system, under the impulse of entrepreneurial activity,
draws away from an equilibrium position, and during the second of which it draws
toward another equilibrium position” (Schumpeter, 1939: p. 142). Schumpeter points out
that these phases seem especially related to different technological systems and their
diffusion and absorption by the social and economic system, which is reflected in their
usage levels, prices, employment, and economic output. Some of the fluctuations are as
short as a few months, while other span decades. Schumpeter explains that “innovations,
their immediate and ulterior effects and the response to them by the system, are the
common ‘cause’ of them all, although different types of innovations and different kinds
of effects may play different roles in each… If waves of innovations of shorter span play
around a wave of a similar character but of longer span, the sequence of the phases of the
latter will so determine the conditions under which the former rise and break as to make a
higher unit out of them, even if the innovations which create them are entirely
independent of the innovations which carry the longer wave”.
This results in an interpretation of development in terms of so-called “long waves”
that arise as a consequence of innovations on different levels of human activity, such as
previously proposed by the Russian economist Nikolai Kondratieff (1925) in his work on
MAPPING OUT THE TRANSITION 13
“The Major Economic Cycles”. This incessant process of “creative destruction”
modernizes the modus operandi of society as a whole, including its economic, social,
cultural and political organization.
As already noted by Schumpeter, the motor of this incessant force of creative
destruction is technological change (Perez, 1983). The literature distinguishes between
five recent so-called “long waves” of technological change (Freeman and Louça, 2002).
The key enabling technology of the first Industrial Revolution (1770-1850) was based on
water-powered mechanization (based on classical mechanics), the second long wave
(1950-1895) was enabled by steam-powered technology (thermodynamics), the third
(1895-1940) was characterized by the electrification of social and productive
organization (electromagnetism), and the fourth by motorization and the automated
mobilization of society (1940-1970) (mechanical and chemical engineering). The most
recent long wave that transforms the social fabric in based on the digitization of
information and communication processes in social systems with the help of so-called
information and communication technologies (ICT) (engineered on basis of insights
gained from information theory and computer science). Each one of those long waves (or
Kondratieff) has been characterized by a sustained period of social modernization, most
notably by sustained periods of increasing economic productivity.
2

While empirical evidence is much weaker for previous periods, the logic of
classifying phases of human development by its dominating technological tool is not
new, and has been borrowed by social scientists from historians. The archaeological

2
The reason why most theories on social evolution focus on economics instead of focusing on the
modernization of cultural or political processes might simply be due to the fact that the respective
performance indicator are much more accessible in the economic realm (i.e. monetary, productive output,
etc).
MAPPING OUT THE TRANSITION 14
period of early human history is commonly subdivided into the sequence of stone-age
(2,000,000 b.c. – 3,300 b.c., duration 1,996,700 years), bronze-age (3,300 b.c. – 1,200
b.c., duration 2,100 years) and iron-age (1,200 b.c. – 44 b.c., duration 1,156 years).
According to Perez’s seminal 1983 article, each of the resulting “quantum jump(s)
in productivity can be seen as a technological revolution, which is made possible by the
appearance in the general cost structure of a particular input that we could call the 'key
factor', fulfilling the following conditions: (1) clearly perceived low-and descending-
relative cost; (2) unlimited supply for all practical purposes; (3) potential all-
pervasiveness; (4) a capacity to reduce the costs of capital, labour and products as well as
to change them qualitatively.”
Digital Information and Communication Technologies (ICT) fulfill those
requirements: the cost-performance relationship of computers, storage and
communication devices has seen respective compound annual growth rates of 60-80 %
during the period from 1980-2005 (Hilbert and Cairo, 2009); their practically unlimited
supply has led to a technological diffusion process that is unprecedented in human history
(ITU, 2010); their nature as a general purpose technology touches all aspects of human
development (Hilbert and Peres, 2010); and ICT have proven to be the driver of sustained
productivity increases during recent decades (Jorgenson, 2005; Cimoli, Hoffman and
Mulder, 2010). As such, ICT are the enabling technology of the most recent and current
socio-economic transformations, and therefore represent the core of profound and
ongoing social transformations.

MAPPING OUT THE TRANSITION 15
Chapter Two: The Social Nature of the Transition

The transition toward information societies is not instantaneous, but is characterized
by an unequal diffusion process of technology adoption. This diffusion process is
generally known as the digital divide: the divide between those that are already part of
the new form of social organization, and those still marginalized. In order to get a better
understanding of this transition, I carried out three studies that concentrate on the socio-
economic and demographic characteristics of social adoption of ICT. In the first section
of this Chapter I created a conceptual framework to classify the diverse definitions of the
digital divide, and showed how complementary outlooks on this multidimensional
challenge are not necessarily harmful, but can rather be useful for impact oriented policy-
making. The section study focuses on the economic bottlenecks of technology diffusion,
in particular on the cost-income relation. I calculated how cheap ICT would have to be in
order to be affordable for everybody in developing countries. The final study of this
Chapter focuses on a widely debated demographic aspect of the divide: gender
differences in ICT access and usage. Using an extensive database from Latin America
and Africa, I provide empirical evidence that women are more enthusiastic ICT users
than men, but since women are discriminated in terms of income, education, and
employment, they end up with fewer possibilities to embrace their natural affinity with
digital communication technologies.

MAPPING OUT THE TRANSITION 16
The End Justifies the Definition: the Manifold Outlooks on the
Digital Divide and their Practical Usefulness for Policy-
Making
3
,
4

Based on the theory of the diffusion of innovations through social networks, the
article discusses the main approaches researchers have taken to conceptualize the digital
divide. The result is a common framework that addresses the questions of who (e.g.
divide between individuals, countries, etc.), with which kinds of characteristics (e.g.
income, geography, age, etc.), connects how (mere access or effective adoption), to what
(e.g. phones, Internet, digital-TV, etc.). Different constellations in these four variables
lead to a combinatorial array of choices to define the digital divide. This vast collection
of theoretically justifiable definitions is contrasted with the question of how the digital
divide is defined in practice by policy makers. The cases of the United States, South
Korea, and Chile are used to show that many diverse actors with dissimilar goals are
involved in confronting the digital divide. Each of them takes a different outlook on the
challenge. This leads to the question if this heterogeneity is harmful and if countries that
count with a coherent national strategy and common outlook on digital development do

3
This article is published as: Hilbert, M. (2011). The end justifies the definition: The manifold outlooks on
the digital divide and their practical usefulness for policy-making. Telecommunications Policy, 35(8), 715-
736. http://dx.doi.org/10.1016/j.telpol.2011.06.012
4
Acknowledgements: A considerable part of the figures and insights of this article have been produced by
the team of the Information Society Program of the United Nations Economic Commission for Latin
America and the Caribbean (UN ECLAC,http://www.eclac.org/SocInfo), which the author had the pleasure
to coordinate between 2000-2008, including Doris Olaya, Valeria Jordan, Massiel Guerra, Cesar
Cristancho, and Priscila Lopez. The author is also indebted with the blind peer reviewers for their
demanding, insisting, and very productive comments, with the participants of the 38th Research
Conference on Communication, Information, and Internet Policy (TPRC 2010), and with his students at the
Annenberg School of Communication, University of Southern California (USC).
MAPPING OUT THE TRANSITION 17
better than others. It is shown that the effect of a coherent vision is secondary to tailor-
made sector specific efforts. On the contrary, a one-size-fits-all outlook on a multifaceted
challenge might rather be harmful. This leads to the conclusion that it is neither
theoretically feasible, nor empirically justifiable to aim for one single definition of the
digital divide. The digital divide is best defined in terms of a desired impact. Since those
are diverse, so are the definitions of the challenge. The best that can be done is to come
up with a comprehensive theoretical framework that allows for the systematic
classification of different definitions, such as the one presented in this article.

Introduction
The term “digital divide” has defied a consented definition since its conception in
the early 1990s (e.g. NTIA, 1995). This has led to confusion and frustration among
researchers and policy makers. In an effort to clarify and separate distinct definitions, this
article returns to the most widely accepted theoretical basis for the digital divide: the
study of diffusion of innovations. Based on this theory the article starts by reviewing
existing literature and identify four broad classes of variables that have been used to
define the digital divide. Differences in definitions arise because scholars distinguish
between (1) the kinds of Information and Communication Technology (ICT) in question;
(2) the choice of subject; (3) diverse attributes of the chosen subjects; and (4) levels of
adoption, going from plain access to effective usage with real impact. This results in a
four dimensional matrix and a vast combinatorial array of different combinations that can
be used to define the digital divide. Is it possible to identify one overall definition that is
particularly useful?
MAPPING OUT THE TRANSITION 18
To answer this question, the methodological discussion is complemented with an
empirical analysis of policy relevance and impact. The article discusses how the diverse
definitions of the digital divide affect the understanding of who is in charge of fighting
the divide. The other way around, the article also stresses how the perspectives of the
actors influence the definition of the challenge. An analysis of the public ICT expenditure
budgets from the United States, South Korea and Chile shows that there are many diverse
authorities involved in confronting the digital divide. The evidence suggests that in
practice, policy makers have a much more heterogeneous outlook on the digital divide
than the infrastructure and access oriented definitions that were traditionally assumed in a
large part of the respective literature during the 1990s and 2000s. It is shown how those
diverse definitions of the digital divide are useful and often even necessary to achieve
sector specific development ends.
From a policy perspective, such diversity can easily be confused with immaturity of
the response to the challenge. In an effort to streamline the heterogeneous outlooks many
countries have started to create a coherent and consensus-oriented policy strategy on the
national level. Specific examples and empirical evidence from Latin America is analyzed
to show that a common outlook is very useful when it comes to the creation of synergies
among the different agents involved in the challenge, but that the measurable impact of
such common outlook is only secondary to sector-specific policies. In order to achieve
real-world impact, it seems more important to count with a tailor-made solution for a
particular problem, than with inter-sectorial coherence and analytical elegance in
definitions. The challenge is multi-dimensional and complex, and so are is solutions.
MAPPING OUT THE TRANSITION 19
In conclusion, from an analytical perspective, the literature has identified a large
variety of justifiable definitions of the digital divide. From a practical perspective, a large
variety of diverse policy makers aim to exploit ICT to achieve very different ends. There
is no evidence to suggest that the introduction of a common outlook leads to significant
positive impacts. Combining these analytical and practical conclusions leads to the same
consequence: the desired impact of ICT is the conditioning variable of any useful
working definition of the digital divide, and different ends justify different definitions.
Notwithstanding this defensible heterogeneity, there are important synergies and
complementarities that can be obtained by clearly keeping track of the manifold
definitions of the digital divide and of the agents that execute them. The common
framework presented in this article provides a tool for doing so.

Theoretical background: Diffusion of innovations
The study of the diffusion of innovation provides the general theoretical framework
to categorize the different approaches researchers have taken to analyze the digital divide.
The dynamic is well-understood by social scientists and related studies have their roots in
the 19
th
century (e.g. Frobenius, 1897; Tarde, 1903).
5
In 1962, Everett Rogers formulated
a coherent theory in his seminal work The Diffusion of Innovations (Rogers, 2003).
Rogers (2003, p.5) defines diffusion as "the process by which an innovation is
communicated through certain channels over time among the members of a social
system”. The logic behind this approach is today known as social network analysis and

5
The first innovation that was rigorously studied was the diffusion of hybrid seed corn among farmers
(Ryan & Gross, 1943).
MAPPING OUT THE TRANSITION 20
analyzes social systems in terms of nodes (or vertices) and edges (or ties) (e.g. Scott,
2000; Strogatz, 2001; Albert & Barabasi, 2002; Newman, 2003, 2010). Social networks
are usually depicted with graph-based structures and studied with the analytical tools of
graph theory and matrix algebra. The Figure shows two typical social networks.
Conceptually, the diffusion of ICT is not very different from any other kind of
diffusion through social networks, such as the prominent example of the diffusion of
contagious diseases, like the spread of the HIV epidemic or an airborne infectious disease
(Valente, 2010). The diffusion of both a contagion and an innovation through human
networks is influenced by the nature of the ties among agents (the network structure) and
by the characteristics of each agent (the personal adoption threshold).

Figure 2 Social network of the spread of diseases and online communities

Source: Krebs, 2003, 2005.

With the framework of a social network in mind, it is straightforward to model the
basic logic of the characteristic S-shaped diffusion curve that gives rise to the digital
divide. Figure assumes a social group of 100 people, whereas some technological

MAPPING OUT THE TRANSITION 21
innovation was adopted by 2 initial innovators. These 2 innovators interact randomly
with the 98 who have not yet adopted and they persuade (“infect”) them at a constant rate
of 1% (assuming homogeneous mixing, without any particular network structure). This
leads to 1.96 (say 2) new adopters during the next time period (2 x 98 x 0.01) (see
Figure). The resulting 3.96 early adopters of the innovation (2 + 1.96, say 4) again
interact randomly with 1% of the rest, leading to 3.80 new adopters, and so forth. The
lower curve of Figure shows the resulting number of new adopters, which first increases,
and then naturally decreases, because with increasing diffusion, there are less and less
potential new adopters available. In the example of Figure, the maximum amount of new
adopters per period is 24.94 (say 25), and happens in the sixth of the ten time intervals,
which represents the inflection point of the diffusion process. The upper curve in Figure
is the respective integral and yields the well-known S-shaped pattern, which Rogers
subdivided into five categories: “innovators, early adopters, early majority, late majority
and laggards” (Rogers, 2003, p.281). The growth in adoption occurs gradually at first and
then accelerates toward the middle of the diffusion process, in order to naturally taper off
as the number of non-adopter vanishes (Valente, 1995). Several mathematical models
have been developed to evaluate the rate and character of these kinds of diffusion curves
(Mahajan & Peterson, 1985).

MAPPING OUT THE TRANSITION 22
Figure 3 Hypothetical example of homogeneous mixing Overview: “Mapping Out
the Transition toward Information Societies”

Cumul.
Adopters
Rate of
adoption
Non-
adopters
New
Adopters
2.00 x 0.01 x 98.00 = 1.96
=> 3.96 x 0.01 x 96.04 = 3.80
=> 7.76 x 0.01 x 92.24 = 7.16
=> 14.92 x 0.01 x 85.08 = 12.70
=> 27.62 x 0.01 x 72.38 = 19.99
=> 47.61 x 0.01 x 52.39 = 24.94
=> 72.55 x 0.01 x 27.45 = 19.91
=> 92.47 x 0.01 x 7.53 = 6.97
=>
99.43
x
0.01
x
0.57

= 0.56
=> 100.00 x 0.01 x 0.00 = 0.00

Source: based on Rogers, 2003, and Valente, 2010.

This hypothetical example of homogeneous mixing is the simplest model. Besides
its simplicity, it has shown to represent the process of diffusion quite accurately.
However, in reality the characteristic S-shaped diffusion curves come in different shapes
and sizes. This is because real-world diffusion is not following a random process of
homogeneous contagion (such as assumed in Figures), but is influenced by the particular
structure of the social network and by the characteristics of its nodes (such as shown in
Figures). Some networks are more centralized than others, and others are characterized
by clusters and cliques. Besides, innovators can often be found at the periphery of social
networks, which means that they have few ties (Figure). As a result, the process tends to
start off even slower. Once opinion leaders in the center of the network adopt the
innovation (those with many ties), the novelty usually spreads quickly. Network
thresholds (Valente, 1995) and the notorious tipping points (Gladwell, 2002) play a
crucial role here. The specific attributes of the nodes act as can shape the diffusion curve.

-
10.00
20.00
30.00
40.00
50.00
60.00
70.00
80.00
90.00
100.00
1 2 3 4 5 6 7 8 9 10
Cumulative Adopters
New Adopters
Innovators Early Early Late Laggards
Adopters Majority Majority
MAPPING OUT THE TRANSITION 23
Income- and educational levels of individuals often act as adoption thresholds. The nature
of the ties also influences diffusion. Some actors might be connected by strong ties (e.g.
family, friends or formal work relations), while others only relate to each other through
informal weak ties or through some form of media (Granovetter, 1973). In short, the
structure and nature of the network (i.e. the ties and its nodes) influence the diffusion
process (Valente, 1995, 2010; Newman, 2010). This gives a particular and distinct shape
to each individual diffusion curve (e.g. Bass, 1969; Andrés, Cuberes, Diouf, &
Serebrisky, 2010).
6

Independent of the kind of network, the diffusion through a social network is never
immediate. While the innovation spreads through the network and the diffusion curve
unfolds, some are included and others excluded from the benefits of the new innovation.
The result is an unavoidable divide. This divide is inevitable. It is the inescapable result
of the fact that it takes a certain amount of time for innovations to spread through social
networks with particular shapes and characteristics. During the past century hundreds of
innovation divides have been identified in a myriad of studies on the diffusion of
innovations (Rogers, 2003). The diffusion of ICT, and its ensuing digital divide, has been
given special importance, given the outstanding socio-economic significance of this
powerful general purpose technology (Bell, 1976; Porat, 1977; Forester, 1985; Miles,
1988; Freeman & Louçã, 2002; Guerrieri & Padoan, 2007; Mansell, 2009; Castells,
2009).

6
According to the particularities of the diffusion process, some researchers also suggest that a modified
classification of adopters’ categories fits particular curves better than the five categories suggested by
Rogers (e.g. Kauffman & Techatassanasoontorn, 2009).
MAPPING OUT THE TRANSITION 24

How to define the digital divide analytically?
The theory of diffusion of innovations provides an adequate framework to classify
the diverse methodological approaches that have been taken to study the digital divide.
The Figure below shows a social network that tracks the diffusion of ICT. The “haves”
are filled nodes, while the “have-nots” are empty. The difference between the “haves”
and “have-nots” is called the “digital divide”.
Figure differentiates between four perspectives on the digital divide. Two of them
are concerned with the type of node: what does a node represent? Which are the attributes
that are considered for each node? In short, what constitutes a node? The other two
concern the diffusion of innovation: what kind of innovation diffuses through the
network? Is it sufficient to have access to the technology, or is it necessary to effectively
adopt the technology (e.g. requiring actual usage with measurable impact)? In short,
when to color a node?

MAPPING OUT THE TRANSITION 25

Figure 4 Social network schematization of different perspectives on the digital
divide

Source: Author’s own elaboration.

The following section classifies the literature on the digital divide according to the
conceptual schematization of Figure. It is shown that these differences in definitions and
focus not only lead to contradictory conclusions (one can show that the digital divide is
closing and widening at the same time, depending on the chosen definition), but also that
they have far-reaching consequences for policies aimed at confronting the digital divide.

Definition of nodes:
what is a node?
(2) Who is the subject
(country, organization,
individual, etc)?
(3) Which attributes
matter?
(income, education,
type of ownership,
geography, size, etc)
Definition of divide:
when is a node at which
side of the divide?
(1) What kind of
technology
(phones, Internet,
broadband, storage
devices, combination
of all, etc)?
(4) How to connect?
(access, usage,
effective adoption)?
Digital divide between “haves” &“have-nots”
MAPPING OUT THE TRANSITION 26
Type of technology
The key variable of interest in studies on the digital divide refers to the technology
in question. In social network graphs, nodes that have already adopted this technology
would typically be marked with some specific trait, such as a distinct color (Figure). This
allows to observe how the innovation spreads through the network (much like a disease
follows a pattern of contagion) (see Figures). There is a large variety of Information and
Communication Technologies (ICT) that might be of interest. Conceptually, ICT can be
divided into three broad groups: technologies that transmit and communicate information
(the movement of information through space); technologies that store information (the
movement of information through time); and technologies that compute information (the
transformation of information) (Hilbert & Cairo, 2009; Hilbert and Lopez, 2011a). Most
current studies focus on technologies that communicate, such as telephones and Internet
subscriptions. Figure shows the technologies that are most commonly studied. Depending
on the choice of the analyst, diverse studies reach different conclusions. For example, the
digital mobile phone divide is rapidly closing (Wareham, Levy & Shi, 2004; Barrantes &
Galperin, 2008; Castells, Fernandez-Ardevol, Qiu, & Sey, 2009), while the digital
broadband divide is quickly widening (e.g. Dutton, Gillett, McKnight, & Peltu, 2004;
Cohen, 2008, Guerra & Jordan, 2010). In terms of the previously presented theory, this
means that the mobile phone has already passed the inflection point of the diffusion curve
(in the concave part of the S-curve), while the diffusion of broadband technology is still
in the first (convex) period of the curve. It is inevitable that incessant technological
process will continuously reintroduce new inequalities that are caused by new
MAPPING OUT THE TRANSITION 27
technologies. Each new technology diffuses through the social network once again,
creating a new divide every time.
Some studies also merge these technologies into so-called indices, such as ITU’s
(2009) ICT Development subindex for access and infrastructure. This index takes
indicators such as fixed and mobile telephony, international Internet bandwidth,
proportion of households with a computer and Internet access, assigns each of them some
particular weight, and creates an average score. This approach implies that the digital
divide is not considered as being closed if a user counts with one specific technology, but
rather with a mix of technologies. The main problem with these indices is that it is at the
discretion of the researcher which weights to assign to which technology. Some studies
use experts opinions, others statistical methods (Hanafizadeha, Saghaeia & Hanafizadeh,
2009). Minges (2005) has evaluated twelve of those indices
7
and reconfirmed the
predictable conclusion that the weight of each ingredient predetermines the resulting
average score to a large extent. This leads to the well-known problem of subjectivity in
the creation of any kind of index and therefore does not solve the problem of clearly
stating which technology is relevant for closing the divide. It rather passes this
responsibility on to the methodological level.
Another, maybe more justifiable way of considering a combination of different
technologies into a single indicator is to measure them in terms of their performance,
measured in [MB], [MIPS] or [kbps] (Hilbert, López & Vasquez, 2010; Hilbert & López,

7
These include the twelve most widespread indices on a global level: Composite index of technological
capabilities across countries (ArCo); Digital Access Index (DAI); Digital Opportunity Index (DOI);
Economist Intelligence Unit (EIU) e-readiness; Index of Knowledge Societies (IKS); Knowledge Economy
Index (KEI); Network Readiness Index (NRI); Orbicom Digital Divide Index; Technology Achievement
Index (TAI); UNCTAD Index of ICT Diffusion; UN PAN E-Readiness Index; World Bank ICT Index.
MAPPING OUT THE TRANSITION 28
2011a). This implies not only to count the number of devices, but to multiply them with
their informational performance (Hilbert & López, 2011c). The resulting sum provides
insight into the technological capacity to process information. Figure combines the
communication capacities of analog and digital fixed and mobile voice telephony, and
mobile data and fixed Internet services, and shows the capacity to communicate of an
average inhabitant of the world’s most industrialized countries (member of the
Organisation for Economic Co-operation and Development, OECD), and the average
inhabitant of the (developing) rest of the world (based on Hilbert & López, 2011b).
Figure shows an increasing digital divide: inhabitants of the industrialized world increase
their informational capacity faster than their counterparts in developing countries. While
in 2002, every inhabitant of the OECD had 8 times more bandwidth available than its
non-OECD peer (79 kbps per capita versus 10 kbps per capita), the broadband revolution
increased this divide to a factor of 15 by 2007 (1,868 kbps per capita versus 126 kbps per
capita). This conclusion is different from the general conclusion that is reached when
merely counting the number of technological devices. While the divide in terms of
devices is closing around the world, the technological performance of those devices
results in an increasing divide (Hilbert, et al., 2010).
This introduces a new aspect: not all technological innovations are equal. Some
have more capacity than others. “Have” or “have-nots” is not a binary yes-no decision,
but consists of a gradient with different intensities: “have how much”? One can mark
these differences in performance in network graphs, for example by assigning different
shades of colors to the nodes, or by adjusting their sizes (compare upper-right node in
Figure).
MAPPING OUT THE TRANSITION 29
Figure 5 Diffusion of most common ICT with individuals. Capacity to communicate
through fixed line, mobile telephony and Internet in optimally compressed kilobits
per second per capita

Source: ITU, 2009; Hilbert & Lopez, 2011b.

0
10
20
30
40
50
60
70
1998 2000 2002 2004 2006 2008
Fixed Telephone Lines
Mobile Telephone Lines
Internet Users
per 100 inhabitant
0
200
400
600
800
1,000
1,200
1,400
1,600
1,800
2,000
1997 1999 2001 2003 2005 2007
kbps per capita
OECD
Rest of world

2002:

2007:

MAPPING OUT THE TRANSITION 30
This last form of looking at the digital divide leads to an important question: when
is the digital divide closed? How much bits does a person have to communicate to be a
member of the information society? When is the technology actually diffused? In the
analytical terms proposed in Figure, the question becomes when to color a node? These
kinds of questions cannot be answered in a technocratic manner, but require normative
decisions. It leads all the way back to the fundamental judgment of what is seen as
necessary and sufficient for development (Sen, 2000). Adam Smith (Smith, 1776) had a
very strong opinion on this issue and stated (almost 250 years ago): “By necessaries I
understand not only the commodities which are indispensably necessary for the support
of life, but whatever the customs of the country renders it indecent for creditable people,
even the lowest order to be without. A linen shirt, for example, is strictly speaking, not a
necessary of life. The Greeks and Romans lived, I suppose, very comfortably though they
had no linen. But in the present times, through the greater part of Europe, a creditable
day-labourer would be ashamed to appear in public without a linen shirt, the want of
which would be supposed to denote that disgraceful degree of poverty which, it is
presumed, nobody can well fall into without extreme bad conduct. Custom, in the same
manner, has rendered leather shoes a necessary of life in England. The poorest creditable
person of either sex would be ashamed to appear in public without them” (Vol.2, book5,
p. V.2.148). The question is if, and if yes, then which kind of ICT connectivity represents
the linen shirt or leather shoes of the 21
st
century? Do the customs of modern form of
social organization render it “indecent for creditable people, even the lowest order to be
without” a mobile phone or a broadband connection of a given bandwidth? Which level
of bandwidth? Measuring the digital divide means defining and then tracking the
MAPPING OUT THE TRANSITION 31
diffusion of the necessary and sufficient. This is a normative decision and part of the
broader process of political will-formation in a society.

Policy implications of the choice of technology
As seen, different definitions of the technology in question can lead to different
conclusions about the digital divide. Depending on the chosen technology, the divide can
at the same time be closing and widening. The same differences in definition also affect
the question of who is in charge to confront the divide. In many countries, different
technologies are regulated by different authorities. If the digital divide is defined in terms
of phones and Internet, telecommunications authorities should be in charge of the
challenge. If the digital divide is defined in terms of a broader group of digital
technologies, such as digital TV, storage devices and general computer equipment, then
broadcast associations, equipment producers, and industry authorities have to be involved
as well. The choice of technology influences who is in charge to bridge it.

Choice of subject
Another difference in studies about the digital divide refers to the subject of
interest. This refers to the decision of what the nodes of the network represent:
individuals, organizations, communities, societies, countries, or world regions. Figure
focused on individuals. In this case each node of the social network represents a person.
But on a higher level of abstraction, each node can also be a group of individuals, such as
organizations, enterprises, schools, hospitals or municipalities, etc. For example, Figure
shows the diffusion of Email among local governments of several Latin American
MAPPING OUT THE TRANSITION 32
countries at two distinct points in time (2004 and 2007). In this case, the choice of
technology is email, and the nodes of the social network represent municipalities. The
figure shows that some countries, like Chile, succeeded early on connecting the large
majority of their municipalities. The diffusion curve in Chile had already reached the
upper end of the S-shaped diffusion curve: saturation. The figure finds the situation in
other countries to be at the very steep middle part of the S-shaped curve. In only three
years, the availability of Email in local governments of El Salvador jumped from mere 10
% to 63 %. Countries like Nicaragua and Honduras were lagging behind and were still
struggling with reaching the critical mass of the diffusion process.
Figure looks at the digital divide from an even higher level of abstraction, whereas
whole countries are the subjects of interest. The choice of technology is an aggregated
index, which consists of a mix of different access technologies (ITU, 2009). The spread
along the y-axis shows that some countries count with much more access to ICT than
others. The divide among countries is often called the international digital divide (e.g.
Corrocher & Ordanini, 2002; ITU, 2009).

MAPPING OUT THE TRANSITION 33
Figure 6 Availability of Email in local governments. Country’s access (ICT
Development access subindex) and income for 2007 (n=147)

Source: OSILAC, 2007. ITU, 2009. Note: ITU’s ICT Development Index (IDI)
subindex for ICT infrastructure and access is a weighted average of fixed telephone lines
per 100 inhabitants, mobile cellular telephone subscriptions per 100 inhabitants,
international Internet bandwidth (bit/s) per Internet user, proportion of households with a
computer, proportion of households with Internet access at home.

0%
20%
40%
60%
80%
100%
Chile Costa Rica El
Salvador
Guatemala Nicaragua Honduras
2004 2007
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00
0 20,000 40,000 60,000 80,000 100,000 120,000
Luxembur
Norwa
y
Sweden
China
Greece
Spai
n
US
Ireland
Brazi
l
Nigeri
a
GDP per capita
ICT Development access subindex
MAPPING OUT THE TRANSITION 34
Policy implications of the choice of subject
The choice of the subject also influences policy responsibility. The digital divide
can be defined to exist between countries, regional, organizations or individuals.
Respectively, there are global, regional, national and local authorities that take actions at
these different levels. In general, the digital realm does not recognize geographic borders.
The policy response has therefore been leveraged at various levels simultaneously, which
is reminiscent of the Russian matryoshka dolls, one inside another. The result is a global
strategy, which was defined at the World Summit on the Information Society (WSIS)
8
;
several regional action plans, such as in Europe
9
and Latin America
10
; national policy
strategies (e.g. ECLAC, DIRSI, & UNDP, 2008; Guerra & Jordan, 2010), and local
strategies
11
. Every organization, hospital, school or family might also grapple with its
particular digital divide and a respective strategy to accelerate the internal diffusion
process. All of these levels, from the global big picture to local improvisation are
important to assure the success of ICT policy (Heeks, 2002). Unfortunately, the policy
responsibility among those agents is often ill-defined and it is not rare that such entangled
strategies end up in misunderstandings and conflicts.

8
The World Summit on the Information Society (WSIS) was held in two phases. The first phase took place
in Geneva from 10 to 12 December 2003, and the second phase took place in Tunis, from 16 to 18
November 2005: http://www.itu.int/wsis . It produced two political declarations and two action plans that
point towards the year 2015.
9
For the history and background of the three consecutive European Action Plans, eEurope2002,
eEurope2005 and i2010, see http://ec.europa.eu/information_society/eEurope/2002/index_en.htm
10
For the history and background of the two consecutive Latin American and Caribbean Action Plans,
eLAC2005 and eLAC2010 see: http://www.cepal.org/SocInfo/eLAC/default.asp?idioma=IN ;
http://en.wikipedia.org/wiki/eLAC
11
For example, Iberomunicipios (http://www.iberomunicipios.org) is a network of hundreds of
municipalities and local e-government initiatives throughout Latin America and Europe.
MAPPING OUT THE TRANSITION 35
Attributes of nodes and ties
Figure does not only show the nodes and their level of connectedness to ICT, but
also another attribute of the nodes: income per capita. This leads to another distinction.
The main attribute of interest is ICT connectivity, but nodes can have more than one
attribute. Individuals, for example, can be distinguished by income, educational level,
geographic location, age and gender, and their maternal language, among others (Parker,
2000; Katz & Rice, 2002; Rice & Katz, 2003; Roycroft & Anantho, 2003; Flamm &
Chaudhuri, 2007). Traditionally, income and geographic location (i.e. urban- rural divide)
are the two most frequently used attributes to describe the divide among individuals.
Organizations can be characterized by their type of ownership, size, profitability, sector,
geography, maturity and organizational culture (Taylor & Murphy, 2004; UNCTAD,
2009); and entire societies, countries or world regions are often classified by their level of
development, wealth, size, geography and ethnicity, among others (Corrocher &
Ordanini, 2002; ITU, 2009; Billon, Marco & Lera-Lopez, 2009). In social networks, the
attributes of the nodes are habitually represented by a particular combination of size,
shade, color and shape of the nodes. Figure uses triangles and circles to represent two
distinct characteristics of each node, as well as coloring for our main attribute of ICT
connectivity. Some nodes have all three of them, others only one. Figure suggests that
those nodes with both triangles and circles are more likely to be on the “have-side” of the
divide (filled) than those with only one of them.
Figure shows the digital divide between public and private schools in Argentina and
Peru. Here the nodes are schools, the chosen technologies are computers and the Internet,
and the additional attribute is related to the type of ownership of the educational
MAPPING OUT THE TRANSITION 36
establishment. In reality, each node has an uncountable number of attributes (at the end,
each node is unique in some detailed way). It is the decision of the analyst to emphasize
some of them and to silence others, which inevitable moves some aspects of the divide
into the spotlight at the expense of others. Figure suggests that the distinction between
private and public schools is an attribute that seems to play an important role in
understanding the threshold of adoption during the diffusion process of computers and
the Internet through the social network of schools. In both countries, Argentina and Peru,
private schools are much more connected than public schools.
Figure shows the diffusion of mobile telephony in Brazil according to two different
attributes: income and education of individuals. It can be seen that both attributes have
independent effects: at the same level of income, access grows with increasing education,
and independent of education, access grows with more income. Since the diffusion and
adoption of ICT is a complex phenomenon that is influenced by multiple attributes, it
makes often sense to track several of them and to analyze their combined effects.
Figure presents a multivariate discriminative analysis of ten attributes of
individuals, testing for household Internet access.
12
It shows that education and income
are the most significant indicators (see also Chaudhuri, Flamm & Horrigan, 2005). Age
turns out to be the third most important determinant of Internet access. It turns out that
other attributes, such as urban/rural, gender or ethnicity rather seem to be a mere
consequence of these three previous ones: women and ethnic minorities have less income
and less education and this is the reason why they lack Internet access. Carefully

12
Multicollinearity can usually be expected in these kinds of exercises (for example between income and
education) and has to be tested. In this case, the tests indicated a low level of multicollinearity. Also, all test
turn out to be highly significant (Hilbert & Peres, 2010).
MAPPING OUT THE TRANSITION 37
controlled studies have shown that being a woman by itself (with the same level of
income and education) rather turns out to be positively correlated with the use of ICT
(Hilbert, 2010a). Given that ICT diffusion is determined by so many different attributes,
it is important to be careful with these confounding relations when analyzing the different
attributes of the digital divide.

Figure 7 Computer and Internet access in schools (2004/5). Mobile penetration in
Brazil (2005). Multivariate analysis of household Internet access for the individual

Source: OSILAC, 2007; Hilbert & Peres, 2010. Note (12): Canonical correlation
coefficients are normalized between 0 (no correlation) and 1 (total dependence).

Standardized Canonical Discriminant Function Coefficients

Chile
2006
Mexico
2007
Paraguay
2007
Nicaragu
a 2006
Uruguay
2007
Brazil
2007
Education of person .591 .690 .716 .802 .464 .416
Income per decile (p.c. of household) .551 .469 .634 .475 .755 .753
Household size (single/pair vs family) .412 .209* .245 .056* .404 .345
Age .329 .348 .425 .252 .094 .131
Enrollment in school/education .180 .247 .310 .056 .122 .115
Job category .018 .107 .107 .021 .050 .113
Color TV in household n.a. .034^ .095^ .233 .028 .060
Geographical region (urban/rural) .189 .017 .122 .002 -.038 -.073
Gender .042 .037 .220 .039 -.038 -.023
Indigenous ethnicity n.a. n.a. n.a. n.a. n.a. .008
Strength of overall correlation
Wilks lambda .854 .792 .896 .983 .696 .682
Canonical correlation coefficient .382 .456 .322 .132 .522 .564

0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
Argentina Peru Argentina Peru
Schools with computer Schools with Internet
Public Schools
Private Schools
MAPPING OUT THE TRANSITION 38
The income dimension of the divide has probably been the one attribute that has
received most attention as a potential bottleneck: if some “node” is below a certain level
of income, it can potentially not been reached by the innovation that spreads throughout
the network. Since ICT has a certain cost, the income attribute represents an absolute
impediment that allows to predict to which nodes the innovation can (eventually) spread.
Therefore, several studies claim that affordability is the key attribute of interest to track
and bridge the digital divide (e.g. Barrantes & Galperin, 2008; Hilbert, 2010b; Beilock &
Dimitrova, 2003). It has been shown that in Latin America the threshold is roughly
around the “magical number” of US$ 10 per person per month, or US$ 120 per year
(Hilbert, 2010b, p. 761). This is how much ICT people seem to strive for and therefore
how much ICT everybody would like to have as necessary and sufficient.
Notwithstanding, this desire is not in agreement with what people actually have: around
40 % of the world population lives with less than US$ 2 per day, and around 20% on less
than US$ 1 per day, or less than US$ 365 per year. It can hardly be expected that the poor
spend one third of their income on ICT (120/365 = 1/3). Normally people spend less than
3 % of their income on communication (Hilbert, 2010b). This implies that 40 % of
human kind (who live with less than US$ 2 per day) count with less than US$ 1.80 per
month to spend on these technologies (30.5 days/month * 2 US$/day * 0.03), or less than
one fifths of the magical number of US$10 per month. This is the economic reality of the
poor. Nodes with this income level can only be reached by those innovations that are
available at this price level. While some suggest that relatively cheap mobile phones
(Wareham, et al., 2004) and public Internet access (Simpson, Daws & Pini, 2004) are
MAPPING OUT THE TRANSITION 39
adequate solutions to reach those income groups, even those solutions provide very
limited access if one counts with less than US$ 1.80 per month
13
(e.g. Hilbert, 2010b).
As already mentioned, there are also different kinds of ties among the agents of the
social network. The diffusion of innovations is characterized by the attributes of the
nodes, as well as by the attributes of the ties. If individuals are bound by a contract with
the same company, or if they are bound by the same law or regulation, the diffusion
pattern might be more dependent on their peers than individuals or countries that merely
have informal and sporadic ties with each other. In Figure, differences in the ties are
represented with different kinds of lines (thicker and dashed, etc). The hypothetical
schematization suggests that those nodes that have already bridged the divide have
stronger ties among them.
Unfortunately, the effect of the nature of the linkages and the resulting structure of
the network is often neglected when studying the diffusion of innovation. Statisticians are
used to collect statistics about the attributes of the nodes, and not about the nature of the
relations between the nodes.
14
It can be expected that the effect of the kind of ties and the
resulting network structure is quite large, since the effect of the attributes of the nodes
alone usually only explains about half of the diffusion process (see for example the
canonical correlation coefficients in Figure: with a correlation coefficient that is
normalized between 0 (no correlation) and 1 (total dependence), it reaches around 0.5). It
can be expected that the other half of the story can be explained when considering the

13
With an average mobile phone minute price of US$ 0.05 per minute in 2009 (ITU, 2009), one can obtain
around 1 minute of mobile phone traffic per day with US$ 1.80 per month(1.8/0.05 = 36).
14
In contrary to traditional statistical software programs (like SPSS and SAS), software programs for social
network analysis work (like Pajek or UCINET) use two different databases: one to register the attributes of
agents (like in traditional statistics) and another one to register the type of relations between those agents.
MAPPING OUT THE TRANSITION 40
relations between nodes (e.g. Bass, 1969). Much more research and adequate statistics are
required to better understand this issue.

Policy implications of the chosen attributes
The selection of the most important attributes is subject to much policy debate,
since it often directly influences the nature of any policy. For example, focusing on
individuals, the traditional focus of the digital divide set on the divide between urban or
rural areas. The reason is historical and has its origins in the times when access to fixed-
line telephony was determined by urban-rural infrastructure deployment constraints.
Nowadays, it seems that other variables, like income, are much more important (Navas-
Sabater, Dymond, & Juntunen, 2002). Figure suggests that the educational level of the
individual is seems to be very important as well. Therefore, many countries start to
involve the education authority into the policy strategy. Others argue that language
barriers are important (Roycroft & Anantho, 2003), which implies the involvement of
cultural and linguistic authorities. Another much-debated question is if there is a digital
gender divide, or if the lagging ICT usage by women is merely a reflection of the
unfavorable conditions of women in terms of income, education and working conditions
(Rice & Katz, 2003; Hilbert, 2010a). Which divide to fight: urban-rural, rich-poor, men-
women? These kinds of definitions of the digital divide often directly influence the nature
of policy interventions and decide which public and private authorities are involved in
fighting the potential inequality.
Making things more complex, this definition can of course be combined with the
previous distinctions between different kinds of subjects and various types of technology.
MAPPING OUT THE TRANSITION 41
For example, taking schools as subjects and broadband connectivity as the technology, a
program might provide connectivity subsidies when they are private or public, rural or
urban, large or small. A rural hospital might be subject to natural discrimination when
employing mobile devices to facilitate their interaction with patients. Different types of
technology can be combined with a diverse choice of subjects with particular attributes.
This creates an increasingly complex matrix that can be used to define the digital divide.

Level of digital adoption
Last but not least, the digital divide can be defined in terms far beyond sole access
to ICT. Usually, words like connectivity or adoption are used to refer to the diffusion
process. But what does it actually mean to be connected or have adopted?
Rogers (2003) originally distinguished among five stages of adoption: (i) initial
exposure to an innovation; (ii) persuasion and the development of positive or negative
attitude; (iii) decision to access or reject the innovation (this is the stage which is often
measured in contemporary ICT statistics); (iv) implementation and actual usage; and (v)
confirmation of its utility to continue and improve. This last step implies that the user is
not only using the innovation effectively, but has started to internalize its benefits and
mold it according to particular needs.
15

Statistical practitioners have simplified these five steps of adoption and mainly
distinguish between ICT access and usage (OECD, 2002). The first step, access, refers to
Rogers’ stages (i) to (iii) and is already ambiguous. The previous discussion about

15
This then often results in feedback that goes back to shape the very nature of the technology (for
example by users demanding a particular kind of technology from manufacturing companies; von Hippel,
2005).
MAPPING OUT THE TRANSITION 42
different technologies has already touched upon the question when a person can be
considered to be connected: which and how much of which technology is necessary to
reach the necessary and sufficient level of connectivity? (see Figure). Figure shows
another perspective on this question. There are different kinds of access within the same
technology, for example individual or shared access. Comparing several countries of
Latin America with the average of the 27 countries of the European Union, it can be seen
that patterns of access are quite different in the developing and the developed world. The
vast majority of Internet users in countries like Peru, Ecuador, Mexico and El Salvador
access the internet through public and shared access facilities, such as cybercafés,
community centers or ICT equipped libraries. These are quasi not existent in Europe
(Figure). Given that 40 % of the world population counts with less than US$ 1.80 per
month to spend on ICT (see discussion above), collective access seems to be the only
economically viable solution to bring them some kind of access to the digital realm (also
Simpson et al., 2004). Is sporadic public access enough to be considered as being on the
“have” side of digital connectivity? Dominating statistics, such as the Internet user
statistics from ITU (2009), consider only household Internet access, not potential access
through public access centers. Therefore, most studies that analyze the digital divide in
terms of Internet access miss the hidden alternative of public access.
The step from access to usage refers to Rogers’ stages (iv) and (v) and is also not
free from ambiguities. Figure compares the usage pattern of some Latin American
countries with those of Europe. It shows that more sophisticated services, such as e-
government, e-banking and e-commerce, were much more common in developed regions,
whereas the Internet was mainly used for simple communication in developing countries.
MAPPING OUT THE TRANSITION 43
One of the main benefits of digital conduct is the reduction of transaction costs.
Transaction costs can be largely reduced by online transactions such as those involved in
banking, e-government or e-business. While a financial transaction over the counter at a
branch of a bank costs on average over US$1, an online bank transaction costs less than
US$0.01 (Lustsik, 2004). Mere online communication might also contribute to the
reduction of transaction cost (by lowering search costs, etc.), but communication alone
does not reap the entire potential benefits. It is therefore not merely the use of ICT, but
the effective adoption of ICT.
The analysis of the digital divide at different levels of adoption can lead to
contradictory results. For example, analysts who measure international access levels to
ICT devices have long claimed that the access divide among countries is diminishing,
since the number of devices reaches a certain level of saturation in developed countries
and developing countries are quickly catching up (e.g. Compaine, 2001; Andrés et al.,
2010). At the same time, however, patterns of effective adoption, which depend on skills
and socio-cultural reorganization, show largely diverging trajectories and suggest a
widening digital divide (van Dijk, 2005).

MAPPING OUT THE TRANSITION 44
Figure 8 ICT access at different locations. sophistication of Internet usage

Source: OSILAC, in Hilbert & Peres, 2010.

The steps from access to usage to effective adoption turn out to be crucial and are
often not automatic (Katz & Rice, 2002). It has been shown that first use and
intensification of use represent independent choices (Battisti & Stoneman, 2003;

- -
4
1
8
5
2
9
74
70
44 44
41
36
29
21
6 6
29
8
10
32
14
23 22
13 15
26
26
17
35
33
26
46
43
19
24
36
27
21
31
44
51
81
-
10
20
30
40
50
60
70
80
90
100
PER
2007
ECU
2008
MEX
2007
SLV
2007
DOM
2005
PAN
2007
CHL
2006
BRA
2005
EU27
2007
Community Internet access facility
Commercial Internet access facility
Place of education
At place of work
At home
82
70
67
61
54
53
19
18
87
3
31
13
12
4
2
0 0
47
5
22
9
14
2
8
1 1
44
2
16
8 9
7
5
1 2
50
-
10
20
30
40
50
60
70
80
90
100
PER
2007
BRA
2005
CHL
2006
DOM
2005
MEX
2007
ECU
2008
PAN
2007
SLV
2007
EU27
2007
Communication
Interaction with public authorities
Banking
Purchaising or ordering
MAPPING OUT THE TRANSITION 45
Hollenstein & Woerter, 2008; Battisti, Hollenstein, Stoneman & Woerter, 2007). The
mere usage of ICT already requires skills, capabilities and involves adjustments in
attitudes (Mossberger, Tolbert, & Stansbury, 2003). The step from usage to effective
adoption entails the effective integration of technology into the daily lives of individuals,
communities, institutions, and societies (Warschauer, 2004). This implies cultural
transformations that modernize the way of doing things. It often requires a change in the
most basic modus operandi of daily routines, as well as changes in the setting of priorities
for long-established procedures. Brynjolfsson and Hitt (1995) talk about the necessity to
invest into so-called intangible assets that complement the deployment of infrastructure,
like the costs of implementing a new business process, acquiring a more highly skilled
staff, or undergoing a major organizational transformation, etc.
Scholars of innovation theory underline that the diffusion of ICT does not occur in a
vacuum and—as with any other general purpose technology—the relationship between
ICT and the complementary surrounding economy becomes essential to advance from
mere access to real impact (Guerrieri & Padoan, 2007). Carlota Perez (Perez, 2004)
points to three different requirements for the successful adoption of ICT: (a) the
development of surrounding services (required infrastructure, specialized suppliers,
distributors, maintenance services, etc.); (b) the cultural adaptation to the logic of the
interconnected technologies involved (among engineers, managers, sales and service
people, consumers, etc.); (c) the setting up of the institutional facilitators (rules and
regulations, specialized training and education, etc.). As long as these complementarities
are lacking, one might achieve universal access to some kind of technological
MAPPING OUT THE TRANSITION 46
infrastructure without achieving the desired positive impact for socio-economic
development.
In short, a broader definition of the digital divide calls for the broader approach to
digital development, which goes far beyond infrastructure deployment and includes the
creation of an enabling environment. In concrete this might include a focus on training
and capacity building, the creation of content and online presence, modernization of legal
frameworks and the creation of supporting industries.

Policy implications of the level of adoption
Is it enough to provide users with access to ICT, or is the divide still existing until
effective adoption leads to tangible impacts? In most countries, the telecommunication
regulator is in charge of confronting the digital divide (e.g. Guerra & Jordan, 2010). It is
also telling that at the global level, the International Telecommunications Union (ITU)
has led the World Summit on the Information Society on behalf of the United Nations
(WSIS).
8
These infrastructure authorities have the mandate to regulate the respective
infrastructure and its deployment, which is undoubtedly an indispensable first step.
Notwithstanding, scholars argue that it is not enough to define the digital divide in terms
of access to infrastructure (Mossberger, et.al., 2003; Warschauer, 2004; Battisti, et al.,
2007; Galperin, 2010), but to evaluate the divide in terms of the effective adoption of the
technologies and their impact. For example, one could call for the successful integration
of ICT into the sectors of education, health and public administration. To achieve this, it
is not sufficient to expand access in schools, hospitals and among government authorities.
E-education entails an adjustment of the curricula in educational establishments and
MAPPING OUT THE TRANSITION 47
therefore requires the educational authorities to be involved. E-health requires the
modernization of the health care sector by the digitization of medical records and
procedures, which demands that health and pharmaceutical authorities are present at the
table that defines an ICT strategy. It might even require changing health care legislation.
E-government implies the effective modernization of public administration, and therefore
calls for the leadership of the highest governmental level to introduce digital transparency
and efficiency in governmental processes of all levels. The digital revolution does not
stop here and continues to the realms of culture, business, family, youth, gender,
entertainment, democracy, transport, finance, sports, military defense and security,
among many others. The effective integration of ICT into the social organization of a
society requires the expertise and guidance of authorities that are concerned with issues
that are complementary to the deployment of technological infrastructure. If the divide is
defined in terms that go beyond access, it is indispensable to count with a much broader
group of expertise in the design and execution of respective policies.

Who, with which characteristics, connects how, to what?
Based on the long-established theory of the diffusion of innovation, it was
straightforward to distinguish among four broad classes of variables that have been used
to define the digital divide (Figure). These can be abridged in the question of: who, with
which characteristics, connects how, to what? All kinds of studies and approaches to the
digital divide can be classified into these four categories:
• WHO (choice of subject): individuals vs. organizations/communities, vs.
societies/countries/ world regions, etc.;
MAPPING OUT THE TRANSITION 48
• with WHICH characteristics (attributes of nodes and ties): income, education,
geography, age, gender, or type of ownership, size, profitability, sector, etc.;
• connects HOW (level digital sophistication): access vs. actual usage vs. effective
adoption;
• to WHAT (type of technology): phone, Internet, computer, digital TV, etc.
This results in a matrix with four distinct dimensions, whereas each dimension
consists of various variables. Each additional variable increases the combinatorial
complexity of this matrix exponentially. For example, counting with only 3 different
choices of subjects (individuals, organizations, or countries), each with 4 characteristics
(age, wealth, geography, sector), distinguishing between 3 levels of digital adoption
(access, actual usage and effective adoption), and 6 types of technologies (fixed phone,
mobile phone, computer, digital TV, general Internet, broadband with a certain speed),
already results in 3*4*3*6 = 216 different ways to define the digital divide. Each one of
them seems equally reasonable and depends on the objective pursued by the analyst.
Despite their heterogeneity, all of them are the result of a common generative
mechanism: diffusion through social networks. The existing diversity in the definitions of
the digital divide are simply the result of prioritizing some aspects of this general process,
while silencing others. Considering the vast combinatorial range of possibilities arising
from this matrix, it is not surprising that discrepancies among diverse methodological
approaches to the digital divide have often led to more confusion than common
understanding.
What determines the choice of a specific one of those possible combinations? Is
there an overall definition that transcends the differences? The preceding sections have
MAPPING OUT THE TRANSITION 49
put emphasis on the fact that the particular definition of the digital divide has far-reaching
implications for the decision on who is in charge to confront the challenge. This can also
be turned around: each policy authority in charge has a different outlook on the digital
divide. Given that the final impact and gain from ICT depends on the successful
integration into a particular environment, and given that all of these different thematic
fields have their particularities and characteristics, it seems very difficult to find a one-
size-fits all definition. Infrastructure authorities will naturally have different priorities
than education and health authorities, and the military or cultural communities have again
a different interpretation of what matters most. The nature of the digital divide is in the
eye of the beholder.
For example, telecommunications authorities traditionally emphasized the diffusion
of infrastructure to every home in their definitions of the digital divide, while others who
are concerned with social welfare and social equality might prefer a definition that
defines the divide in terms of a certain amount of kbps per capita as a socially accepted
minimum (this affects the choice of type of technology in the definition, compare
Figures). International authorities will naturally look at the divide between countries,
while local authorities will be concerned with the exclusion of specific parts of a
community (this affects the choice of subject in the definition, compare Figures).
Somebody who has the goal to modernize education with ICT will identify individuals
with different attributes as core subjects than somebody that focuses on improving
national security (this affects the choice of attributes in the definition, compare Figures).
And finally, somebody who wants to employ ICT to diffuse the work of national
museums has a different understanding of the necessary level of ICT adoption than
MAPPING OUT THE TRANSITION 50
somebody who wants to use ICT to effectively modernize the domestic health care
system (this affects the choice of level of digital adoption in the definition, compare
Figure). In short: the end determines the definition of choice.

Who defines the digital divide in practice?
With this in mind, the following section looks at empirical evidence on who is in
charge of fighting the digital divide in some selected countries. This will provide a better
understanding of the breadth of the existing perspectives in practice. The most
straightforward way to identify who is in charge on a national level is to see who counts
with how much resources to fight the digital divide. The identification of the funds that
each government authority has available to execute digital policies provides an idea how
governments perceive and define the digital divide in day to day policy making. If it
should turn out that most resources are spent by infrastructure authorities, it can be
concluded that –in practical terms—policy makers understand the digital divide in terms
of access to telecommunications. If other authorities receive even more resources (e.g. e-
government, education and health authorities), it can be concluded that the de facto
definition of the digital divide goes beyond access, etc.

Who manages the resources to fight the digital divide?
In the case of the United States, the Federal Communications Commission (FCC)
manages roughly US$ 8 billion annually to fight the digital divide in the country.
However, the newly appointed first Federal Chief Technology Officer of the United
States (CTO) estimates that the federal government spends up to US$ 70 billion (Chopra,
MAPPING OUT THE TRANSITION 51
2010). So even in times where the American Reinvestment and Recovery Act
appropriated an additional ad-hoc and onetime US$ 7.2 billion to expand digital
broadband access and adoption in communities across the country (NTIA, 2010), it
becomes clear that the bulk of the pie is by far still dispersed with authorities that do not
mainly focus on ICT itself, but try to make ICT work for the development of the country
from different perspectives. If one assumes that money talks louder than discourse, it
turns out that in reality the telecommunications authorities FCC and NTIA are not really
in charge of fighting the digital divide, independently of what any official mandate might
say. While the divide in the United States has traditionally been defined exclusively in
terms of access to infrastructure (NTIA, 1995), it turns out that the bulk of the respective
budget is spent elsewhere in the meantime.
16

The case of South Korea is also interesting in this regard. The country set up a
specialized Informatization Promotion Fund which invested a total of US$ 5.33 billion
between 1994 and 2003. Of that, 38 % was invested in ICT Research & Development, 20
% into informatization promotion, 18 % in ICT human resource development, 15 % in
broadband infrastructure and promotion, 7 % in ICT industries, and some 3 % in

16
In response to this fragmented challenge, President Obama has set up a coordination-trio, consisting of
three posts (Obama, 2009): the Federal Chief Information Officer was created by the e-government Act
from 2002 (Congress 107th, 2002) and is the administrator of the Office of Electronic Government, which
in turn is part of the Office of Management and Budget. It is “responsible for setting technology policy
across the government, and using technology to improve security, ensure transparency, and lower costs”
(Obama, 2009). This post has been complemented by the Chief Performance Officer, in 2009, which is also
part of the Office of Management and Budget and concentrates on general government reform. Finally, the
Chief Technology Officer was created in 2009 as a position within the Office of Science and Technology
Policy. It is its assigned task to “promote technological innovation to help achieve our most urgent
priorities – from creating jobs and reducing health care costs to keeping our nation secure” (Obama, 2009).
He is also tasked with increasing American's access to broadband. Even though none of these positions is
very high up in the hierarchy of the federal government, it is envisioned that their interventions gain
efficiency by working closely together. At the time of writing this article, the judgment is still out if such
trilogy of power is effective to coordinate the dispersed US$ 70 billion spent annually on ICT issues by the
federal government.
MAPPING OUT THE TRANSITION 52
standardization (Suh & Aubert, 2006). It can be seen that the ambitions of Korea have
been quite broadly defined: only 15 % of the total was dedicated to infrastructure
deployment, while the rest was dedicated to dimensions that complement the
infrastructural gap, such as the development of skills and the integration of ICT into
governmental processes (informatization).
Unfortunately, it was not possible to obtain more detailed information on the
budgetary distribution of the United States and South Korea or many other countries.
This is mainly because those inventories hardly exist. Not even the government often
knows how much it spends on ICT. One exception is the detailed ICT spending inventory
of the government of Chile. An analysis of this data will provide anecdotic evidence to
answer the question of who is in charge of the digital agenda in practice.

The Chilean ICT budget
The Chilean Telecommunications Development Fund (FDT) is seen as a best
practice in the developing world (Wellenius & Bank, 2002; Hawkins, 2005; Mena, 2006).
The fund was created in 1994 with an objective to improve payphone access in rural and
low-income urban areas with low teledensity. The fund offered subsidies to private
companies to provide payphone service. Subsidies were allocated through competitive
tenders and were taken from the national budget (Wellenius & Bank, 2002). Like all
funds of this interventionist nature in a privatized market it has been subject to much
debate and public and private conflicts of interest right from the days of its conception
(Rosenblut, 1998). The debate continues as the fund moves into the Internet age and
started to subsidize data services (Hawkins, 2005; Mena, 2006). The fund is managed by
MAPPING OUT THE TRANSITION 53
the national telecom regulator SUBTEL, a sub-secretariat of the Ministry of Transport
and Telecommunications.
In 2003 the Fund spent a total of US$ 4.86 million, allotted by two public contests
(SUBTEL, 2003). At the same time, the national government set up the first generation of
the Chilean Digital Agenda.
17
In the frame of this comprehensive strategy, the Ministry
of Finance carried out an inventory of public ICT spending, which covered 210 public
institutions from 22 budgetary rubrics (DIPRES, 2005). The initiative did not come
without protest, since it additionally complicated the already intricate process of national
budgeting. This might also be one of the reasons why detailed statistics on government
wide ICT spending are very scarce. For the same reason, the Chilean study excluded
entities that correspond directly to Congress, as well as establishments of higher
education (which make up a considerable part of public funding). It also does not include
the subsidies used in the Chilean Telecommunications Development Fund. The inventory
identified a total spending on ICT and related services of US$ 205 million in 2003. Table
1 enlists the 22 authorities that handle this budget and shows their fiscal power in the
field of digital development.
It turns out that the e-government projects of the Ministry of Finance occupied the
largest share of the pie (15.2%). The Ministries of Education and Defense almost spent
the same amount of resources on their ICT projects (14-15% each), while the US$ 22
million spent by the Ministry of Health account for 10.7% of the total. Together, these

17
Chile was one of the pioneers in national agenda setting for digital development in developing countries.
The first generation of the plan, between 2004-2006, was called Agenda Digital Chile, while the 2007-2012
plan is called Digital Strategy (http://www.estrategiadigital.gob.cl). It focuses on the modernization of the
State, ICT investments and the effective usage of ICT by the society at large.
MAPPING OUT THE TRANSITION 54
four authorities account for 55% of the national public spending on ICT. Note that none
of these authorities sees infrastructure deployment as their main task.
The lower part of Table 1 shows the ends toward which the government spending is
directed. 20.4% of the total goes toward salaries and specialized ICT staff. One could
argue that this amount does not really make part of any digital divide policy, since it is
spent on improving internal processes of the government, without being directly aimed at
the public. Without this amount, the total spending still amounts to US$ 163 million.
These resources are exclusive investments into the successful deployment of ICT and
related services for the benefit of the public, spent by the most diverse public authorities.
In the light of these kinds of resources, the much-cited US$ 4.86 million of the
Telecommunications Development Fund seem almost negligible: total government
spending consists of 34 times the resources of the specialized fund.
18
Alone the resources
the overall government spends on ICT investments and purchases (15.7% of the total, see
lower part of Table 1) sum up to almost 7 times the US$ 4.86 million of the
Telecommunications Development Fund ([205*0.157]/4.86). The amount of resources
the government spends yearly on general Development Projects that involve ICT (11.5%
of the total ICT budget, see lower part of Table 1) is almost five times as much as the
resources provided by the much-cited best-practice fund ([205*0.115]/4.86).
In short, the case of Chile suggests that in practice the digital divide is seen as a
challenge that goes far beyond mere infrastructure deployment. Only 3 % of the public
ICT budget is assigned to the national ICT access authority, while the bulk of the
available funds are dispersed among 210 institutions from 22 budgetary rubrics. The

18
163/4.86 = 33.6; or with a broader definition of ICT spending, including salaries: 205/4.86 = 42.
MAPPING OUT THE TRANSITION 55
Ministries of Finance, Education, Defense, Health, Labor and Social Security, and Justice
carry out the most important initiatives regarding digital development in the country, and
given the nature of their thematic priorities, they certainly count with different definitions
of the nature of the problem.

Table 1 Governmental spending on ICT in Chile, 2003, in percent of a total of US$
205 million

General
governm.,
security &
defense
Fiscal
functions
Regulatory
functions
Investment
functions
Social
functions
Ministry of Finance 6.2 0.0 9.0 0.0 0.0 15.2
Ministry of Education 11.2 3.0 0.2 0.4 0.0 14.9
Ministry of Defense 12.9 1.1 0.0 0.0 0.0 14.0
Ministry of Health 0.6 9.5 0.6 0.0 0.0 10.7
Ministry of Labor and Social Security 0.2 6.5 1.9 0.0 0.0 8.6
Ministry of Justice 5.3 0.7 1.3 0.1 0.0 7.3
Judicial Power 5.1 0.0 0.0 0.0 0.0 5.1
Ministry of Public 3.3 0.0 0.0 0.0 0.0 3.3
Ministry of Economy & Reconstruction 0.3 0.2 0.6 2.0 0.0 3.1
Ministry of Public Works 0.2 0.0 0.2 0.0 2.7 3.1
Ministry of Agriculture 0.2 0.0 0.0 2.6 0.0 2.8
Ministry of Interior 1.8 0.0 0.0 0.0 0.7 2.5
Ministry of Housing and Urban 1.4 0.0 0.0 0.0 0.5 2.0
Ministry of Planning and Cooperation 0.8 0.9 0.0 0.1 0.0 1.8
Ministry of General Secretary of Governm. 0.4 0.6 0.0 0.1 0.0 1.0
Ministry of General Secretary of President 0.2 0.0 0.0 0.7 0.0 0.9
General Accounting Office 0.0 0.0 0.9 0.0 0.0 0.9
Ministry of Exterior 0.4 0.0 0.0 0.4 0.0 0.8
Ministry of Mining 0.1 0.0 0.0 0.6 0.0 0.7
Ministry of Transport and Telecom 0.1 0.0 0.0 0.4 0.0 0.5
Presidency 0.4 0.0 0.0 0.0 0.0 0.4
Ministry of National Goods 0.0 0.0 0.0 0.4 0.0 0.4

51.0 22.6 14.7 7.7 3.9 100.0
Which are distributed toward the following ends:
Staff and salaries 8.9 5.2 1.5 0.8 4.0 20.4
Computer and telecom services/leasing 23.9 7.4 3.5 2.0 15.6 52.4
Investment and ICT purchases 8.0 2.1 1.6 1.1 2.9 15.7
Development projects involv. ICT 10.2 0.1 1.0 0.0 0.2 11.5
51.0 14.7 7.7 3.9 22.6 100.0
Source: DIPRES, 2005.

MAPPING OUT THE TRANSITION 56
It is to be expected that infrastructure and telecommunication authorities will
continue to play an important part in the challenge of narrowing the digital divide, but, as
suggested by the budgetary priorities from the United States, South Korea and Chile,
their role is in reality already much smaller than what is generally assumed. The funds
managed by the telecom authorities only represent a small fraction of the total
governmental ICT funds, which are distributed among the budget lines of diverse
agencies. In the case of Chile, authorities of the fields of finance, e-government,
education, health and social security spend much more on ICT policies than infrastructure
authorities that are exclusively concerned with the diffusion of technology. This is in line
with the previous argument that policies that aim at fighting the digital divide should aim
at the effective integration into a specific area of interest.

A fragmented policy response
The case of Chile shows that the realm of policy making counts with a very broad
and multifaceted interpretation of the challenge and opportunities posed by the digital
age. The heterogeneity in policy makers and their diverse tasks inevitably leads to
heterogeneity in the outlooks on the challenge. Different authorities have different
priorities in their interpretation of the digital divide. This incoherence can lead to double
efforts, waste of scarce resources, and even obvious contradictions.
For example, different government authorities have different ideas about the kind of
technologies that matters for the well-being of society, and reflect these priorities in
relevant policies. Baqir, Palvia, & Nemati (2009) report on inconsistent and contradictory
policies regarding sales tax and import duties on ICT services and equipment. As a result
MAPPING OUT THE TRANSITION 57
of the limited coordination between the diverse outlooks of different government
authorities, foreign investors and local manufactures were discouraged from committing
more of their resources.
A typical example of double efforts is the coordination of diverse public access
strategies (such as public access centers and libraries) with access at educational facilities
(computers in schools). Public ICT access centers target the larger public, while computer
labs in schools focus exclusively on school students. While it is natural that the latter use
their computer labs during morning hours, the general public usually visits public access
centers during the afternoon and evening. By allowing the public at large to use school
computer labs during times students are not at school, valuable synergies can be created.
This, however, requires a coordination of the diverse definitions of the challenge at hand.
Something similar accounts for the confrontation of the skill-gap, which is at the
core of the usage and impact dimensions of the digital divide. In the United States, the
national telecommunication authorities FCC (Federal Communications Commission) and
NTIA (National Telecommunications and Information Administration) have recently
started to support training courses focused, among other things, to support people finding
employment through online services. These are not directly coordinated with
longstanding similar efforts from employment and labor offices, social security and
industry authorities.
19
It can be expected that the funds of the latter largely surpass the
funds that the telecom sector dedicates to this end in an ad hoc effort. Double efforts do
not only waste resources, but are not sustainable, since the ad-hoc funds of the ICT-

19
This statement is based on comments made by Mignon Clyburn, Commissioner of Federal
Communications Commission, after a presentation at the 38th Research Conference on Communication,
Information and Internet Policy Friday, October 1, 2010 at the George Mason University School of Law,
Arlington.
MAPPING OUT THE TRANSITION 58
authority are only temporal in nature. It would be more sustainable to modernize
longstanding existing programs from non-ICT authorities, making digital development
part of their continuous mandate.

In search for evidence of impact of a common outlook
These examples point to the usefulness for one common and coherent strategy to
confront the multiple challenges of digital development. With this in mind, many
countries have started attempts to unify and streamline the different visions by bringing
ICT policy under one common and coherent umbrella with a shared outlook on the
challenge. These national ICT strategies consist of inter-ministerial and multi-sectorial
policy agendas, mainly led by the public sector, and aim at coordinating the diverse and
disconnected efforts carried out by different authorities (ECLAC, DIRSI, & UNDP,
2008; Guerra & Jordan, 2010). The goal of these coordination mechanisms is to create a
common outlook of the nature of the challenge among the different stakeholders, which is
typically written down in a public document (such as in Chile’s digital strategy,
Colombia’s connectivity agenda, or Mexico’s e-Mexico plan, among others). The natural
question is if the existence of such nationwide common strategies for digital development
leads to a detectable positive impact. Is a coherent outlook on the digital divide essential
for the advancement toward to digital age?
To answer this question, Figures compare the stage of development of such
strategies with measures of impact for several countries from Latin America and the
Caribbean. Both figures show three different stages of national strategy formulation on
the x-axis. Some countries, like Panama and Costa Rica, did not count with a coherent
MAPPING OUT THE TRANSITION 59
national strategy or dialogue on digital development in 2007. There was no coherence
among the different initiatives and projects, no common outlook or shared vision. A
second group of countries, like Brazil and Bolivia, found themselves in the phase of
formulating such national strategy; while a third group (the majority), including Mexico,
Chile and Jamaica, actively executed a coherent and nationwide strategy for digital
development. Figure measures these stages against a composite index that evaluates the
deployment of a quality infrastructure (ITU, 2009), and Fig 16 against a composite index
that measure the online presence of the national e-government (UN DESA, 2008). This
provides two complementary indicators for impact: one on the level of infrastructure, and
another one on the level of online content.

MAPPING OUT THE TRANSITION 60
Figure 9 ICT policy coherence versus ICT infrastructure and access index, 2007.
ICT policy coherence versus online presence of e-government index, 2007

Sources: OSILAC, 2009 and ITU, 2009. Note: ITU’s ICT Development Index (IDI)
subindex for ICT infrastructure and access is a weighted average of fixed telephone lines
per 100 inhabitants, mobile cellular telephone subscriptions per 100 inhabitants,
international Internet bandwidth (bit/s) per Internet user, proportion of households with a
computer, proportion of households with Internet access at home. OSILAC, 2009 and UN
DESA, 2008. Note: E-government online presence is a subindex of UN DESA’s World e-
Government Preparedness ranking and measures the online presence of national e-
government websites, including those of the ministries of health, education, welfare,
labor and finance of each State.

Chile
Mexico
Brazil
Argentina
Uruguay
Colombia
Venezuela
Peru
Jamaica Panama
Trinidad&
Tobago
Costa Rica
El Salvador
Dom.Rep.
Bolivia
Ecuador
Guatemala
Paraguay
Nicaragua
Honduras
2
2.5
3
3.5
4
4.5
5
5.5
6
ICT infrastructure and access index
initial stage:
no coherence
among actors
formulation stage:
mediate coherence
among actors
full execution:
strong coherence
among actors
older than 5 years
younger than 5 years
Chile
Mexico
Brazil
Argentina
Uruguay
Colombia
Venezuela
Peru
Jamaica
Barbados
Panama
Trinidad&
Tobago
Bahamas
Costa Rica
El Salvador
Dom.Rep.
Bolivia
Ecuador
Guatemala
Paraguay
Nicaragua
Honduras
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
e-government Web presence index
initial stage:
no coherence
among actors
formulation stage:
mediate coherence
among actors
full execution:
strong coherence
among actors
older than 5 years
younger than 5 years
MAPPING OUT THE TRANSITION 61
Both figures show a slight positive correlation between the existence of a national
policy strategy and digital development.
20
Figures also show that older strategies (older
than 5 years) seem to have more correlation with infrastructure deployment than with e-
government development. This makes sense, since it takes much longer to employ an
infrastructure than to set up transactional webpages. This suggests that the existence of a
national strategy does have a positive impact. However, in both cases this correlation is
not very strong. Some countries without any national vision for digital development, such
as Panama and Costa Rica, do better than countries that count with an established
strategy, like the Dominican Republic. Decisive differences can also be found between
areas of impact. Trinidad and Tobago counts with a quality ICT infrastructure, but is
struggling with their e-government initiative, while the opposite holds for Mexico. The
existence of a coherent policy vision can therefore not be the main cause for the observed
impact.
Digging deeper into the reasons for why some countries do well in some areas and
not in others, it shows that it is not the existence or non-existence of a national strategy
per se that explains success of failure in digital development, but rather sector specific
projects and tailor-made policies that address specific areas of interest (ECLAC, DIRSI,
& UNDP, 2008; Guerra & Jordan, 2010; Hilbert & Peres, 2010). For example, in the case
of Mexico, lack of competition has notoriously limited the deployment of
telecommunications infrastructure (Mariscal & Rivera, 2005), which led to a mediocre
performance in this sector (see Figure 15). Infrastructure deployment is a very concrete

20
Note that Figures represent correlations, without making any claim about causality. While the obvious
goal is to foster digital development with a national ICT for development strategy, it might well be that
higher impact in certain areas of digital development facilitate the existence of a national strategy.
MAPPING OUT THE TRANSITION 62
problem and requires a tailor-made response. At the same time, the Mexican e-
government authorities independently went ahead to set up one of the world’s leading e-
government service network (Luna-Reyes, Gil-Garcia & Cruz, 2007). The results are
clearly detectable in Figure. Something similar accounts for Bolivia. The country has
long struggled with the development of vibrant ICT infrastructure (ITU, 2001) (see
Figure), but has successfully set up relatively well-working e-government online content
(ADSIB, 2010) (see Figure). On the contrary, some small countries like Panama and
Jamaica counted with well-developed infrastructures in 2007, while they had not yet set
up a successful e-government strategy (see also Miranda, 2007; Lawton, 2010). The
provision of high quality content is a different challenge than the deployment of
infrastructure, and, to a certain extent, it is possible to advance in each area independently
of the other. Each implies a different outlook on digital development.
This shows that real-world impact does not primarily seem to depend on the
existence of one common outlook, but on how well a particular challenge is confronted
with a specific solution. It seems intuitive that it is much more important to confront
concrete challenges than to find methodological elegance and coherence. Second, the
different projects are certainly complementary. For example, a thriving e-government
will eventually depend on the existence of a quality infrastructure, and vice versa. These
kinds of complementarities are not to be underestimated (Hilbert & Peres, 2010) and they
can be harnessed by exploiting synergies as such the ones discussed in the previous
section. The slightly positive correlation between the existence of a coordination
mechanism and impact in Figures supports this intuitive idea. However, this effect seems
MAPPING OUT THE TRANSITION 63
to be merely secondary to the provision of concrete and tailor-made solutions for
particular challenges.

Conclusion: Impact over analytical coherence
This article started by deriving a conceptual framework that allows to differentiate
among the manifold definitions of the digital divide. This conceptual framework is based
on the network analysis approach to the diffusion of innovation and provides four
straightforward ways to categorize the existing literature on the topic. A systematic
literature review showed that there is a vast combinatorial array of different ways to
define the digital divide. These diverse definitions influence the choice of who is in
charge of confronting the digital divide. At the same time, the other way around, each
authority has a different outlook on the challenge. Empirical evidence shows that a large
and heterogeneous group of authorities with the most dissimilar thematic priorities is
invested in the challenge. This is good news, since it is generally accepted that real
impact and gains from ICT demand sector-specific expertise from the fields in which ICT
is employed. Given the diversity of the potential benefits and impacts of such a versatile
general-purpose technology as ICT, this finding argues in favor of a flexible definition of
the digital divide that considers specific ends with a final impact. It is unavoidable that
these different perspectives will lead to tailor-made and complementary definitions of the
divide (Vehovar, Sicherl, Husing & Dolnicar, 2006).
To strengthen this point, it was also shown that the existence of a unifying
institutional mechanism and common outlook on the digital divide do not necessarily lead
MAPPING OUT THE TRANSITION 64
to detectable impact. On the contrary, it is indeed conceivable that a very stringent one-
size-fits-all definition of the digital divide will be counterproductive. This is one of the
main critiques of general ICT development indexes, such as ITU’s ICT Development
Index (ITU, 2009) or the Network Readiness Index of the World Economic Forum
(INSEAD & WEF, 2009) (for an overview see Minges, 2005). If policy makers would
take such indexes seriously (and finance and economic authorities often do, because of
concerns related to foreign investments and national competitiveness), the content of any
policy agenda would have to follow the specifications of the components of the index.
For example, if computers in every household would receive a considerable weight in the
definition of the digital divide, the most impact effective policy of any country would be
to design projects and regulations that would assure to increase home computer
penetration. Other initiatives, which might be more valuable but not included in the
weighting (like mobile phone bandwidth, the establishment of software industries, or e-
government initiatives, etc.) would then inevitably suffer from such policy priorities. This
would of course put the cart before the horse. The ends should determine the means, not
the other way around. Since there are no common ends in the deployment of ICT, it is
counterproductive to pursue common means. There are only complementary definitions
of the digital that fall into common categories and pursue one multifaceted final goal:
achieving positive impacts from the deployment of ICT.
These insights lead to an emerging consensus among scholars. “The new consensus
recognizes that they key question is not how to connect people to a specific network
through a specific device, but how to extend the expected gains from new ICTs”
(Galperin, 2010; p. 55; see also Bar & Best, 2008; Khalil & Kenny, 2008; Heeks, 2009).
MAPPING OUT THE TRANSITION 65
The analytical focus shifts from the search of a definition by means of understanding the
diffusion process (inductive: from real-world observations to concepts), to the
identification of a desired impact, which then determines the adequate definition to solve
a particular problem (normatively deductive: from concept to desired real-world change).
Since the impacts of ICT are diverse, the definitions of the digital divide are as well.
Therefore, questions like “what is the best definition of the digital divide?” or “when is
the digital divide closed?” do not make sense by themselves, but have to be formulated
on basis of a conditioning variable:
Given the desired impact, who, with which characteristics, connects how to what?
Or, normatively speaking:
Given the desired impact, who, with which characteristics, should best be connected
how to what?
This leads to a relativistic and maybe unsatisfactory conclusion, which is
nonetheless very certain: there is no truth about what the digital divide is. It is subjective
and depends on what the aspired achievement. More formally speaking: the definition is
conditioned on the desired impact. The best that can be done is to come up with a single
and coherent framework based on a solid theory, which allows for the classification and
comparison of the different definitions, such as done in the first part of this article. The
challenge does not consist in reducing the heterogeneity in outlooks, but in better
understanding and keeping track of the communalities and differences among the
priorities of diverse actors and their definitions. In practical terms, a major part of this
task consists in institutionalizing a mechanism that takes inventory of the budget each
(public or private) authority allocates to ICT related policies and projects. This is a
MAPPING OUT THE TRANSITION 66
practical first step in the search for synergies among diverse outlooks. Unfortunately,
most countries do not count with any mechanism to track the complete amount of
resources that are dedicated to ICT policies and projects.
21
The combination of
conceptual clarity and relevant information among the diverse priorities eliminates
confusion and allows for the effective search for synergies among complementary
outlooks on a multifaceted challenge.

21
The effort of Chile in 2003 was a onetime effort, which was not continued.
MAPPING OUT THE TRANSITION 67
When is Cheap, Cheap Enough to Bridge the Digital Divide?
Modeling Income Related Structural Challenges of Technology
Diffusion in Latin America
22

The article presents a model that shows how income structures create
diffusion patterns of Information and Communication Technologies (ICT). The
model allows the creation of scenarios for potential cuts in access prices and/or
required subsidies for household spending in Mexico, Uruguay, Brazil and Costa
Rica. One analyzed scenario would require the reduction of ICT prices to as low
as 4% of the current price levels (to US$0.75 per month), or alternatively, a
subsidy as high as 6.2% of GDP (a figure equivalent to public spending on
education plus health). Neither existing technological solutions nor existing
financial mechanisms are sufficient to cope with this challenge. Alternatives are
discussed.

Introduction
The heated discussion about the digital revolution has already spanned two decades
and is far from cooling down. It has been widely reported that the increased
pervasiveness of digital Information and Communication Technologies (ICT) leads to
significant contributions to development, including economic growth (e.g. Cole, 1986;
Mody and Dahlman, 1992; Waverman, et.al., 2005; Indjikan and Siegel, 2005),

22
This article was published as: Hilbert, Martin (2010). When is Cheap, Cheap Enough to Bridge the
Digital Divide? Modeling Income Related Structural Challenges of Technology Diffusion in Latin
America. World Development, The Multi-Disciplinary International Journal Devoted to the Study and
Promotion of World Development, 38(5), 756-770. doi:10.1016/j.worlddev.2009.11.019
MAPPING OUT THE TRANSITION 68
democracy and transparency (e.g. Coleman, 2005; Banisar, 2006; Hilbert, 2009),
education (e.g. Marshal and Taylor, 2006) and cultural development (Castells, 2004),
among others. After a thorough and well-cited analysis, Castells (1998, p. 367) reaches
the conclusion that “the generation of wealth, the exercise of power, and the creation of
cultural codes came to depend on the technological capacity of societies and individuals,
with information technologies as the core of this capacity.”
The nature of the diffusion process of the related innovations resembles a well-
known S-curve from centre-periphery, whereby the centre can be characterized as being
more developed and the periphery as underdeveloped (e.g. Mahajan and Peterson, 1985;
Rogers, 2003). The unfolding of the curve inevitably creates a divide between those that
can first benefit from the innovation and those excluded. In the case of ICT diffusion
patterns the term “digital divide” has been coined to describe the fact that some already
use digital tools, while others are still deprived of access and the potential opportunities
that follow from it (e.g. NTIA, 1995, 1999; OECD, 2001; ITU, 2009).
Both arguments together describe a double-edged sword. Technology is not only a
parent of wealth and development (creates it), but also its child (stems from it). It has
therefore the potential to spawn either a virtuous or vicious circle between existing or
missing development and technology. The urgency to work on this challenge arises from
the rapidly closing window of opportunity related to the dynamic between development
and technological progress (e.g. Perez and Soete, 1988; Freeman and Louca, 2001). The
excluded could either be armed with new empowering tools despite their unfavorable
starting position or their relative situation could worsen while the benefits once more
enable the already well-off to make headway.
MAPPING OUT THE TRANSITION 69
This article focuses on the access dimension of the digital divide and analyzes the
economic feasibility of providing universal ICT services. It presents a model that
simulates the dynamic between access prices, eventual subsidies and ICT penetration
rates, and as such allows the exploration of feasible options. A quick thought experiment
sets the stage for the subsequent exercise: it has been claimed that a US$100 laptop
would enable us to provide “one laptop per child” throughout the developing world
(Negroponte, 2005). The World Bank reports that around 40% of the world population
lives with less than US$2 per day. Optimistically estimating that they spend 2.7% of their
income on communication (more on that in Figure), the poor would have US$20 per year
available to spend on ICT (2*365*0.027). That would be enough to buy one US$100
laptop each fifth year. While this result does not sound too shocking, it has to be
remembered that this person is not expected to spend another cent on any other form of
communication during these five years (no letter post, not one phone call, etc) and still
has not connected the laptop to any kind of network. Correspondingly, the scenario looks
worse for the 15% of the world population living with less than US$1 per day. One
alternative would be to call for public subsidies, providing free or at least cheap
connectivity for the poor. How much would this cost? The other alternative would be to
reduce prices even further. How cheap, is cheap enough to bridge the digital access
divide?
This article tackles both questions by following the outlined logic. It presents a
model to simulate various scenarios. The exercise is carried out for four representative
countries: Mexico, Uruguay, Brazil and Costa Rica. We start by finding a valid working
definition of the digital divide for our purposes. As a second step we empirically justify
MAPPING OUT THE TRANSITION 70
our income-based model by looking at the importance of income distribution as a
predictive variable for the digital divide. We then verify ICT spending per income level
in the third section. The available resources for ICT in each income quintile can be
understood as the potential market, which is then compared to the current cost of ICT
access for each income segment. As a next step, we can quantify the gap between access
prices, spending levels and actual ICT penetration rates per quintile, as done in the fourth
section. Subsequently, we apply the model to analyze three exemplary future scenarios.
The results show how much it would cost to bridge the digital divide or, alternatively,
when cheap is cheap enough for the market to close the divide. For example, one scenario
consists in bringing ICT access levels from below 20% for Internet and around 50% for
telephony up to electricity penetration rates, which are currently above 95% in Latin
America. In Mexico, for example, providing the economic possibility of ICT access to
the poorest 20% of society would require to reduce current access prices of estimated
US$244 per year down to US$35 per year (US$3 per month, or a price reduction down to
13%of current prices). In Brazil, the poorest 20% of the population counts with merely
US$9 per year to spend on ICT (US$0.75 per month), implying the necessity to reduce
current ICT prices down to 4% of its current level of US$220 per year. This is the
economic reality of these income segments and the magnitude of the challenge to provide
“one ICT access per child”. Alternatively, a (direct or indirect) subsidy could be given to
balance the purchasing power of these segments. In the case of Uruguay, this subsidy
would need to be as high as 6.2% of GDP, which is equivalent to Uruguay’s public
spending on education plus health. This shows that the governments of developing
countries (and their economies) currently do simply not posses the means to provide
MAPPING OUT THE TRANSITION 71
personalized access to all, even if they opted for the cheapest equipments. The article
ends by discussing the odds in the light of this reality and explores complementary ways
to bridge the digital divide. This includes the economic feasibility of sustainable public
access to ICT, which has proven to be a viable way to reduce access prices by sharing the
fixed costs of the technology. The income reality of the world’s poor argues for the need
of policy strategies that prepare for a long period in which public access will be the only
viable access solution for significant parts of the developing world. However, the
required financial mechanisms for such a period are currently not in place. The article
ends by discussing innovative proposals in this regard.

How to define the digital divide?
The term “digital divide” is that rare breed of a new concept that has the ability to
camouflage into the author’s intended meaning, often leaving confusion or tedious
semantic quarrels behind. It is usually defined as the divide between those included and
excluded from the digital age, leaving lots of room for interpretation.
In general, distinctions can be made with regards to the group of users (countries or
population segments), the kind of technology under consideration (mobile or fixed; voice
or data; communication or computing, etc) and the stage of adoption. The most
straightforward notions of the digital divide select a specific technological solution as a
representation of the bulk of digital technologies (such as telephones or Internet
subscription) and compare the amount of equipment or services between societies
(international digital divide) or within different social segments of one society (domestic
digital divide). Beyond this, different stages can be distinguished in the process of
MAPPING OUT THE TRANSITION 72
technology adaptation. In his traditional work on the “diffusion of innovations”, Rogers
(2003) has famously distinguished among five different stages of adoption. Statistical
practitioners interested in measuring the nature of the digital divide have merged these
five stages into three consecutive steps: ICT access, use and impact (OECD, 2002). Even
though there might be a positive relation between the amount of ICT equipment, its usage
and its impact, one of them does not automatically imply the next. The complexities in
the steps from access to productive usage and from usage to impact have been widely
discussed, such as in the productivity paradox of the 1980s or the new-economy hype of
the late 1990s.
As a result of these distinctions many different proposals have been made on how to
conceptualize the embracement of technology in each one the these stages and how they
are related and intertwined with existing socio-economic and geo-spatial inequalities (e.g.
Warschauer, 2003; Mossberger, et.al., 2003; van Dijk, 2006; Buys, et.al, 2009). The
determinants of the divide can be assessed in each stage of the adoption process and with
regards to all of the diverse existing technologies, or their combination. For example,
access might depend on financial limitations, which also manifest themselves in geo-
spatial requirements of infrastructure deployment, while productive usage will eventually
depend on how the installed equipment is embraced, drawing the focus of attention to
skills and motivations, among others. The final impact of the technology will ultimately
be influenced by the purposeful application of the installed equipment, often requiring the
readjustment of the general modus-operandi of the cultural and institutional setting,
which leads to a complex dynamic of social change. Depending on the definition and the
scope of the exercise, the results can be contradictory and either argue in favor of a
MAPPING OUT THE TRANSITION 73
rapidly closing digital divide (Compaine, 2001) or one that is still deepening (van Dijk,
2005).
The common feature of all studies is the inclusion of the access dimension of the
divide, which might not be sufficient, but is a necessary first step. Without neglecting that
the discussion of the digital divide can become much more complex, it will be enough for
the scope of this paper to focus only on this indispensable dimension.
Regarding the access dimension it is important to distinguish “universal access”
versus “universal service”. United Nations International Telecommunications Union
(ITU) defines that universal access implies that everyone in a population has access to
publicly available communication network facilities and services, typically provided
through such means as pay telephones, community telecentres and community Internet
access terminals; while universal service focuses on promoting or maintaining universal
connectivity of all households to public network facilities and services, and at affordable
prices (ITU, 2007). Over the late 1990s, it seemed that the focus should turn to universal
access, instead of pursuing the ambitious goal of universal service. However, the focus on
universal service was revived by a couple of occurrences, including the mobile phone
revolution, which had shown that it was possible to provide personalized services to most
of the world’s population (mobile phone penetration has reached 60 percent of the
world’s population in 2008), and by the discussion that has been unleashed by the “One
Laptop per Child” initiative (Negroponte, 2005).
In 2005, the MIT driven “One Laptop per Child” initiative announced the ambitious
goal to produce a “$100 laptop” that would revolutionize education by being within reach
of each child in the developing world. Due to various reasons, the plan did not
MAPPING OUT THE TRANSITION 74
accomplish its goal during the years to come (The Economist, 2008a). During 2008 the
initiative announced a temporary “Give One, Get One“ scheme in the developed world,
encouraging people to pay US$ 399 for two of the laptops, one of which would be given
to child in the developing world. The initiative set in motion a snowball effect that went
far beyond the potential benefits of a single quick-fix solution. The chipmaker Intel
swiftly announced an alliance with software giant Microsoft to commercialize their own
education-oriented laptop, called Classmate. The Classmate was first to push the price
under the US$ 300 mark and the “laptop war” has been raging ever since (The
Economist, 2008b). The Indian government surprised everyone with the announcement of
a “$10 laptop” in 2008, but quickly readjusted its target to the well-known US$100 mark
(Ribeiro 2008). At the same time private market analysts saw this goal at least three years
away, with an uncertain future (Gartner, 2008). Independent from the intricate details of
the ongoing projects, it had become clear that price reduction of digital access equipment
became a concern of mayor interest for policy makers and industry. The race is on and
universal service had once again become the benchmark for bridging the digital divide.
In this paper we will contribute to this and related discussions by evaluating basic
questions like if US$100 is actually low enough to provide “one laptop per child”? Is
universal service a feasible goal in the mid-term, maybe in combination with public
subsidies? Answering these specific questions does not provide coherent policy advice on
how to bridge the access divide, but it grounds the current discussion on the reality of
existing income structures in the developing world. The results provide normative
guidelines to direct policy options, much in line with Seneca’s logic that if one does not
MAPPING OUT THE TRANSITION 75
know to which port one is sailing, no wind is favorable
23
. It is beyond the scope of this
article to identify the right winds to guide us from normative objectives to concrete ICT
access solutions, as it is much further beyond our focus to explore how to get from access
to optimal usage and real impact.

How important is income distribution for the digital divide?
Models are idealizations of reality. Because reality is too complex, models work on
an elevated level of abstraction, selectively choosing to ignore parts of reality, while
focusing on what is seen as the variables with the largest explanatory power. Modeling
the access dimension of the digital divide is no exception and there are countless
variables that define the domestic dynamic. In this section we will explore if it is
justifiable to elaborate a model of the digital access divide that only focuses on income as
an explanatory variable.
It has been shown that the same long established determinants of socio-economic
inequality also define the digital divide, including income, education, skills, employment,
geography, age, gender and ethnicity, among others (e.g. Cullen, 2001; Hilbert and Katz,
2003). In Latin America, the region that persistently shows the world’s most skewed
income distribution (ECLAC, 2008), it is not surprising that income structures is seen as
the most influential attribute to explain access inequalities (Katz and Hilbert, 2003; Peres
and Hilbert, 2009). Brazil for example, the region’s largest country (with over 37% of the
region’s GDP and 34% of Latin America’s population), has one of the most unequal
income distributions. The richest 10% of society receive around half of the national

23
"Ignoranti, quem portum petat, nullus suus ventus est".
MAPPING OUT THE TRANSITION 76
income, while the second richest count with a 15% share, and the poorest decile has to
manage with less than 1% of the available resources. Uruguay, one of the smallest
countries of the region (.8% of the region’s GDP and .6% of its population), has the most
equal income distribution, with the richest 10% of the population receiving 37% of
income, the second decile 15% and the poorest decile 2%.
Let us take a closer look at the importance of income as a determinant of access in
Brazil and Uruguay. During recent years an effort has been made by national statistics
institutes to incorporate ICT indicators into household questionnaires throughout Latin
America (Olaya, 2008; Partnership, 2008). This now allows us to perform multivariate
analysis on the determinants of variables like Internet access. For our purposes,
discriminant analysis seems to be good choice among the various alternatives of
multivariate analysis. It tests for a linear combination of variables, which –when
appropriately weighted—will maximally discriminate between those who have access
and those who do not have access
24
.
Taking the 2007 household survey of Brazil (OSILAC, 2009) the following 10
potential variables have been chosen to explain household Internet access: per capita
household income per decile; personal level of education (without, primary, secondary,
postsecondary and tertiary); household size (single and double vs. larger families
25
); age
(in groups from 5-14, 15-24, etc); current school or education enrollment; job category

24
Discriminant analysis goes back to Fisher (1936) and can basically be understood as a multiple
regression where the criterion variable is nominal rather interval or ratio (in our case: access or not) and the
aim is to search for a combination of variables that maximally discriminates between those groups (e.g.
Williams and Monge, 2001).
25
This variable is supposed to test if people in single or double households tend to connect more to the
Internet to avoid social isolation.
MAPPING OUT THE TRANSITION 77
(in four broad groups
26
); geographical region (urban vs. rural); installation of color TV in
household
27
; gender (male-female); and last but not least, ethnicity (as auto-identification
of belonging to an indigenous group or not
28
). The results can be seen in Table, second
column.
Tests for multicollinearity indicated a low level of multicollinearity
29
. The overall
test was significant (Wilks λ = .682, Chi-square = 3.44E7, df = 10, Canonical
correlation
30
= .564, p <. 001), but not very strong. The identified discriminant function
accounted for 32% of the variance in Internet access. This is an acceptable degree of
correlation, but shows that there must be other variables to explain who has and who does
not have access to the Internet at home. Reclassification of people based on the newly
identified function was quite successful: based on the identified function 78.5% of
Brazilians were correctly reclassified into the original categories of having or not having
home access. The most interesting result for our purposes is the importance of the income
variable. The weights of the standardized canonical discriminant function (see Table) tell
us how the ten original variables combine to make a new one that maximally
discriminates between Brazilians who have and those who do not have household Internet

26
1
st
group (primary): agriculture/ hunting/ forestry/ fishing/ mining/ quarrying/ construction; 2
nd
group
(secondary): manufacturing/ electricity/ gas/ water; 3
rd
group (services1): wholesale/ retail/ repair/
household/ community/ hotel/ restaurants; 4
th
group (services2): Financial/ real estate/ business/ public
administration/ defense/ education/ health/ social work/ extraterritorial organizations.
27
This variable is supposed to test for the degree of previous technological adoption and a general interest
in ICT.
28
This variable is supposed to test for eventual language or cultural barriers.
29
Tolerance in the order of variables enlistment in text: .634; .617; .904; .759; .871; .662; .878; .722; .951;
.999. Similar values result in the subsequent tests of Brazil 2005 and 2002 and Uruguay 2005 and 2007.
30
The canonical correlation is the correlation between the new canonical variables formed by applying the
weights from the discriminant function to the 10 predictors, and the grouping variable of either having or
not having household Internet access.
MAPPING OUT THE TRANSITION 78
access. We can interpret the standardized discriminant function coefficients as a measure
of the relative importance of each of the original predictors. Income level (.753) is by far
the strongest predictor for Internet access, followed by the level of education, household
size and age.

Table 2 Results of discriminant analysis of household Internet access for the
individual

Source: own elaboration; based on OSILAC, 2009

One might wonder if this result depends on a specific stage of the diffusion process
and if these results change over time? The same exercise for Brazil 2005 and Brazil 2002
(see Table) shows that these results are pretty stable over time. In addition, discriminant
analysis has been carried out for Uruguay 2005 and 2007 to verify that the identified
tendency also holds for smaller countries with a more balanced income distribution (see
Table).
Standardized Canonical Discriminant Function Coefficients
Brazil 2007 Brazil 2005 Brazil 2002
Uruguay
2005
Uruguay
2007
Income per decile (p.c. of household) .753 .743 .704 .799 .755
Education of person .416 .487 .556 .423 .464
Household size (single/pair vs family) .345 .305 .249 .431 .404
Age .131 .167 .207 .084 .094
Enrollment in school/education .115 .118 .117 .114 .122
Job category .113 .077 .061 .026 .050
Color TV in household .060 .117 .164 .038 .028
Geographical region (urban/rural) -.073 -.007 .048 n.a -.038
Gender -.023 -.023 -.021 -.027 -.038
Indigenous ethnicity .008 -.007 -.004 n.a n.a.
Strength of overall correlation
Wilks lamba .682 .712 .751 .708 .696
Canonical correlation .564 .536 .499 .540 .522
Reclassification success 78.5% 81.2% 80.6% 79.4% 79.2%
MAPPING OUT THE TRANSITION 79
Of course these results cannot automatically be generalized to other countries. They
have been chosen on basis of available statistics and the exemplary nature of two
dissimilar countries. Table reveals that the level of education also turns out to be an
important variable and it remains to be seen what role it –and other variables—play in
different countries and with regard to other grouping variables (such as Internet usage
instead of plain access). Notwithstanding, these indicative results provide evidence that
ICT access is significantly influenced by income distribution. These results give us
enough justification to opt for a simplified model that simulates the complex dynamic of
the closing digital divide exclusively in terms of income distribution. Such reduction of
reality will surely not be perfect, because, as we have seen, there are other important
explanatory variables. However, models are always an abstraction of reality and the
presented results show that income is the single variable with the greatest predictive
potential.

How much resources are available for ICT?
Official consumer and household spending questionnaires are scarce and valuable
treasures for research in the developing world. Mexico, as a member of the statistics-
savvy OECD, stands out as an exception. It has been possible to collect and harmonize
some specific years for other countries as well (see Figure). The first thing that meats the
eye when looking at Figure is that spending unequivocally differs among income
segments (here presented in quintiles; I: lowest income 20% of population; V: highest
income 20%).
MAPPING OUT THE TRANSITION 80
The top of Figure shows a contradictory story with regard to the elasticity of
household spending on communications. Traditionally consumption goods are either
defined as ‘necessity goods’ or ‘luxury goods’, going back to Engel’s (1857) pioneering
work. Food is the prototype of a necessity good. Everybody needs to spend a comparable
amount on it, implicating that the poor will end up spending a larger part of their total
income on it. A luxury good, on the contrary, is characterized by elastic spending
patterns, implying that the rich will spend significantly more on it than the poor, who do
not depend on this good to satisfy their most basic needs. The top of Figure shows that
communication seems to be a luxury good for Peru and Colombia, which is in agreement
with more specific research about this question in these countries (Gamboa, 2007;
Aguero, 2008), while it seems to have some characteristics of a necessity good for the
other countries (especially in the richest segments), which is also in agreement of
previous research on those countries (Ureta, 2005). Is communication a luxury or a
necessity?
The solution of this riddle can be found at the bottom graph of Figure. It shows that
Peru and Colombia (with respective Gross-National-Income per capita of US$3,990 and
US$4,660 in 2008) spend noticeably less in absolute value on communications than
Costa Rica (GNI p.c.: US$6,060), Brazil (GNI p.c.: US$7,350), Uruguay (GNI p.c.:
US$8,260) and Mexico (GNI p.c.: US$9,980). ICT are tradable goods and therefore are
not much subject to price variations among different countries. This is different from
non-tradable services, such as teaching for example. A teacher in Peru might cost
considerably less than a teacher in Mexico. Not so for ICT. On the contrary, it can often
be seen that ICT is cheaper in rich countries (which often produce them and have larger
MAPPING OUT THE TRANSITION 81
and frequently more competitive markets), than in poor countries (which have small
markets and import the equipment). As a result, ICT spending is tied to absolute, not
relative levels of spending. It seems that a certain saturation threshold has only been
reached when household expenditure rises up to US$30-40 per month, making that
roughly US$10 per person (see income segment IV of Costa Rica, Brazil and Mexico in
both graphs). Before reaching this level, consumer continue to keep spending whatever
they have available. In other words, the Figures suggest that communication spending
below US$10 per person per month seems to be a necessity (for these years and these
countries), while everything above appears to be an additional luxury.

MAPPING OUT THE TRANSITION 82
Figure 10 Monthly household spending on communication, as % of total spending
(top) and as absolute values in current US$ (bottom)

Source: own elaboration; based on INEGI (2000-2009), Encuesta Nacional de
Ingresos y Gastos de los Hogares.2000-2005 en Mexico; INE (2006), Encuesta National
de Gastos e Ingresos de los Hogares 2005-2006 en Uruguay; IBGE (2003), Pesquisa de
Orçamentos Familiares 2002-2003 Brasil; INEC (2006), Encuesta Nacional de Ingresos y
Gastos de los Hogares 2004 en Costa Rica; DANE (2004), Encuesta de Calidad de Vida
del 2003 en Colombia; INEI (2004-2007), Encuesta Nacional de Hogares en el Peru.
MAPPING OUT THE TRANSITION 83
How much does ICT access currently cost?
The spending figures give us a rough idea about the size of the potential market.
Having quantified the demand side of ICT, we now need some figures from the supply
side, in other words the price of ICT. This has to start with a definition of what ICT
actually are.
The technological system that is generally referred to as ICT merges several
previously separated evolutionary paths into one technological system. Traditional
broadcasting technologies, such as radio and television, communication systems, such as
the telephone, as well as storage devices, such as music players, and of course computing
equipment make part of the dynamic that is often referred to as ICT-convergence. The
current introduction of digital television is an epitome for how a traditional unidirectional
information-disseminating technology is merging with bi-directional communication
systems, while increasing its data processing capacities by adding different hardware and
software solutions. Another example is third and forth generation mobile phones (3G and
4G), in which the telephony communication system is upgraded to include broadband
Internet services and software applications. Eventually, digital TV and mobile Internet
will also merge into a single technological system, becoming part of the network of
networks (the ‘Inter-net’), leading to all sorts of fixed or mobile solutions, for individual
or collective use, with a specific resolution, bandwidth and storage capacity, blurring the
traditional borderlines between broadcast, telecommunication and computation. The
downside of this high-speed technological progress is that every new technological
innovation reopens the digital abyss, constantly resulting in new inequalities between
those who possess more and better information and communication processing capacities
MAPPING OUT THE TRANSITION 84
and those with lower ones. The digital divide becomes a moving target, and on top of
that, one that moves at the rhythm of a technology that moves faster than anything we
have seen in history so far (e.g. Hilbert and Cairo, 2009).
In this sense, much more pronounced than with other challenges of technology
penetration, the digital divide is not only wide or narrow, but is also deep or shallow.
Actually, as the extraordinarily fast innovation cycle persistently emphasizes qualitative
differences, the digital divide will never be “closed” in a uniform sense. Some people will
always have more information processing capacity than others. However, we are able to
build a bridge over even the deepest abyss. “Bridging” the digital divide means therefore
that every member of the so-called Information Society has continuous access to
sufficient resources to be able to maintain minimum connectivity with its peers. The
result is the challenge of constantly defining the terms “sufficient” and “minimum”. Such
reasoning has led to the concept of “digital poverty,” with segments of society being
below the dynamically evolving digital poverty line, and others above (Galperin and
Mariscal, Eds, 2007).
Considering the dynamics of multiple access solutions and their short innovation
cycles, it seems advisable not limit our analysis of the digital divide to a specific
technology. Rather, we should consider connectivity as access to a combination of
technologies, an ‘ICT access packet’, that allows people to communicate and to process
information through digital means, leaving the technological implementation to
individual choices and later policy considerations. Consequently, bridging the digital
divide would mean that a minimum ICT access packet would be in reach for everybody,
still conceding that some will always continue to have more connectivity than others.
MAPPING OUT THE TRANSITION 85
Keeping in mind that our goal is to estimate the amount of resources dedicated to
ICT, independently of future technological implementation, we nevertheless have to start
with what is currently seen as the standard ‘ICT access packet’ to get a feeling for what is
right now being spent on it. This can be done by reviewing current penetration rates of
different ICT and include the most popular solutions into our packet. The most popular
ICT access solutions in Latin America are fixed-line telephony (70% of all households in
2007), mobile phones (70% of all individuals), personal computers
31
(45% of all
households) and Internet (17% of all households) (OSILAC, 2007). We will stick to those
four main technologies. We suppose that all members of the household share the fixed-
line, as well as the computer and the Internet access. This inevitably favors the low-
income segments, as low-income families tend to be larger in general (on average the
household size of the poorest quintile in Latin America is around 4.5 members, while a
household in the richest quintile counts with 2.5 members) (CEPALSTAT, 2009).
Table shows annual ICT access prices for Mexico 2002, Mexico 2007, Uruguay
2005, Brazil 2003, and Costa Rica 2004, as national average and per income quintile.
These numbers could easily have been obtained for more countries, but only these cases
also count with the necessary household spending surveys and the related penetration
rates (which we will need in the subsequent steps). Having said this, the final selection of
these cases was determined by the availability of all necessary variables in the respective
household questionnaires (OSILAC, ICT statistical information system, 2009), together
with the availability of household spending statistics. Luckily these countries are
relatively representative for the diversity of the region, being one large and one small

31
The term personal computer is used to refer to all kind of household computers, including laptops,
notebooks and Mac.
MAPPING OUT THE TRANSITION 86
country from South- and Meso-America in each case. Income levels of these countries
vary from Costa Rica (GNI: US$6,060), Brazil (US$7,350), Uruguay (US$8,260) to
Mexico (US$9,980) and are therefore almost exactly span the upper half of the region’s
income diversity. Unfortunately it is typical that poorer countries do not count with
detailed statistical information to include them in these kinds of exercises.
The underlying methodology of Table follows generally accepted assumptions. For
the calculation of the price of fixed line telephony we stick to the methodology proposed
by the United Nations International Telecommunications Union (ITU, 2009) and the
World Bank’s “price basket for residential fixed line”, which capture the average
monthly cost of a basic local fixed residential telephone service (in our case divided by
the average household size per income quintile), plus 30 three-minute local calls per
month to the same (fixed) network (15 peak ad 15 off-peak calls). For mobile telephony
we include 30 one-minute outgoing calls per month (15 on-net and 15 off-net, that is to a
network of the same and a different operator). ITU reports mobile tariffs for prepaid cell-
phones. This is justified not only because the overwhelming majority of mobile phone
users in the developing world uses prepaid phones instead of fixed contracts (in Mexico
95% of all mobile phone users; in Uruguay 85%), but because they are often the only
payment method available to low-income users who might not have a regular income and
will thus not qualify for a postpaid subscription based service. For Internet services it was
decided to opt for the price of a 256 kbit/s connection for the household. This is the
lowest non-dial-up connectivity usually available. It was decided not to opt for dial-up
because this would have required an estimation of usage minutes and because dial-up
option have become much more expensive per person than low-cost dedicated services,
MAPPING OUT THE TRANSITION 87
such as through a telephone line with DSL
32
. The price of this connection is divided by
household size per income quintile. Last but not least, a USD 500 computer is chosen
which would be in use for three years (which is seen as the average industry standard for
computer obsolescence). This refers to some of the cheapest industry offers in the
region
33
(Laplane, et.al. 2007). It is assumed that the entire household also shares the
computer.

32
For example, in Mexico 2002, only two minutes and 45 seconds daily dial-up Internet for one person cost
as much as unlimited 256kbps broadband shared with the average size of household members.
33
Industry and government collaboration, such as “mi primer PC” in Chile, offer financing plans for PCs
for around USD 500. Normal industry prices are slightly above this number.
MAPPING OUT THE TRANSITION 88
Table 3 Annual ICT prices per inhabitant in US$, national average and per income
quintile

MEXICO 2002 National I II III IV V
Fixed-line telephony 103.8 93.3 99.4 102.6 109.2 120.0
Mobile phone 99.1 99.1 99.1 99.1 99.1 99.1
Internet 155.0 121.6 140.9 151.2 172.2 206.6
Personal Computer 41.7 32.7 37.9 40.7 46.3 55.6
TOTAL PACKET yearly 399.5 346.6 377.2 393.5 426.7 481.2

MEXICO 2007 National I II III IV V
Fixed-line telephony 94.0 83.1 89.7 94.0 100.8 112.4
Mobile phone 57.3 57.3 57.3 57.3 57.3 57.3
Internet 111.5 84.8 100.9 111.5 128.4 157.0
Personal Computer 43.9 33.3 39.7 43.9 50.5 61.7
TOTAL PACKET yearly 306.7 258.5 287.6 306.7 337.1 388.4

URUGUAY 2005 National I II III IV V
Fixed-line telephony 69.4 57.9 67.2 73.4 80.4 84.9
Mobile phone 133.7 133.7 133.7 133.7 133.7 133.7
Internet 208.8 142.4 195.8 232.0 272.4 298.3
Personal Computer 55.6 37.9 52.1 61.7 72.5 79.4
TOTAL PACKET yearly 467.5 371.8 448.7 500.9 559.0 596.3

BRAZIL 2003 National I II III IV V
Fixed-line telephony 58.6 50.2 54.7 62.1 63.4 69.7
Mobile phone 85.7 85.7 85.7 85.7 85.7 85.7
Internet 102.2 79.5 91.7 111.8 115.4 132.5
Personal Computer 47.6 37.0 42.7 52.1 53.8 61.7
TOTAL PACKET yearly 294.1 252.5 274.9 311.7 318.3 349.6

COSTA RICA 2004 National I II III IV V
Fixed-line telephony 20.4 19.3 19.3 19.3 20.7 23.6
Mobile phone 21.8 21.8 21.8 21.8 21.8 21.8
Internet 150.2 139.0 139.0 139.0 154.4 185.3
Personal Computer 45.0 41.7 41.7 41.7 46.3 55.6
TOTAL PACKET yearly 237.4 221.7 221.7 221.7 243.2 286.2

Source: own elaboration; based on ITU, World Telecom. Database, 2009

These numbers therefore represent our ‘ICT access packet’, adjusted to family size
per income segment. Realistically looking at the chosen services, this packet is rather at
the very low end, meaning that it represents a minimum packet for somebody who would
MAPPING OUT THE TRANSITION 89
like to claim to be a full-fledged member on an “Information Society”. Of course, an
individual could make much more than one 3 minutes fixed-line call per day, or could
easily spend more than one daily minute on the cell-phone. We will look at these possible
variations in the following section.

How does income distribution create ICT diffusion patterns?
One way to check the validity of the estimations for our ‘ICT access packet’ is to
multiply the prices with the penetration rates of those technologies, which should give us
the amount of the respective ICT spending. If it turns out that this multiplication results
1:1 with the identified spending figures, it has been proven that our price estimations are
accurate. If the result of the multiplication turns out to be below actual spending, the
conclusion would be that individuals actually spend more on ICT than defined by our
minimum ‘ICT price packet’. For example, broadband connectivity could be increased
far beyond 256 kbps (1000-4000 kbps connections are commercially available in these
countries), while standard personal computers can easily cost up to US$1500-3000. If the
multiplication of price levels with penetration rates would turn out to be above actual
spending levels, it could be concluded that our packet is overestimated. In other words, in
these cases users spend less than 3 minutes per day on the fixed-line or less than one
minute per day on their mobile phone (for example by having a non-charged prepaid
phone). The existence of prepaid mobile phones that are only used sporadically has been
reported frequently in the developing world (e.g. Galperin and Mariscal, 2007). Another
explanation for the overestimation of our packet might be that computers are not
MAPPING OUT THE TRANSITION 90
modernized every three years. It is not unusual to find 10-15 year old PCs and recycled
computers throughout the developing world (e.g. REALC, 2009).
ICT penetration rates per income quintile can be seen in Table. Once again, the
chosen countries include Mexico 2002, Mexico 2007
34
, Uruguay 2005, Brazil 2003 and
Costa Rica 2004.

34
Given that no spending figures exist for Mexico 2007 (but given that this is the last year for which we
count with the harmonized household penetration levels), the Mexico 2008 spending levels have been
applied to model Mexico 2007. This is justifiable because Graph 1 confirms that spending levels in Mexico
stay pretty constant.
MAPPING OUT THE TRANSITION 91
Table 4 ICT penetration rates per inhabitant; Affordable portion of ‘ICT access
packet’; national average and per income quintile

MEXICO 2002 National I II III IV V
Fixed-line telephony 0.45 0.24 0.33 0.47 0.58 0.76
Mobile phone 0.25 0.18 0.24 0.29 0.31 0.27
Internet 0.07 0.01 0.02 0.04 0.12 0.22
Personal Computer 0.16 0.04 0.07 0.12 0.24 0.43
How much of the ‘ICT access packet’ can be bought:
[indiv.spending] / ∑[penetration rate * price] 1.03 0.80 0.87 0.88 0.95 1.64

MEXICO 2007 National I II III IV V
Fixed-line telephony 0.55 0.41 0.46 0.55 0.66 0.81
Mobile phone 0.60 0.34 0.54 0.67 0.72 0.91
Internet 0.12 0.03 0.05 0.09 0.13 0.44
Personal Computer 0.25 0.10 0.16 0.21 0.28 0.63
How much of the ‘ICT access packet’ can be bought:
[indiv.spending] / ∑[penetration rate * price] 0.95 0.59 0.71 0.86 1.11 1.49

URUGUAY 2005 National I II III IV V
Fixed-line telephony 0.73 0.48 0.75 0.85 0.91 0.98
Mobile phone 0.33 0.32 0.33 0.33 0.34 0.36
Internet 0.14 0.01 0.06 0.15 0.26 0.51
Personal Computer 0.24 0.05 0.16 0.29 0.42 0.62
How much of the ‘ICT access packet’ can be bought:
[indiv.spending] / ∑[penetration rate * price] 0.76 0.39 0.69 0.77 0.88 1.07

BRAZIL 2003 National I II III IV V
Fixed-line telephony 0.50 0.16 0.37 0.55 0.73 0.89
Mobile phone 0.39 0.13 0.27 0.38 0.54 0.79
Internet 0.11 0.00 0.01 0.04 0.14 0.46
Personal Computer 0.15 0.01 0.03 0.08 0.21 0.55
How much of the ‘ICT access packet’ can be bought:
[indiv.spending] / ∑[penetration rate * price] 0.84 0.47 0.71 0.92 0.98 1.13

COSTA RICA 2004 National I II III IV V
Fixed-line telephony 0.66 0.42 0.56 0.66 0.75 0.86
Mobile phone 0.45 0.17 0.28 0.39 0.55 0.80
Internet 0.21 0.01 0.04 0.09 0.21 0.57
Personal Computer 0.25 0.06 0.10 0.16 0.34 0.57
How much of the ‘ICT access packet’ can be bought:
[indiv.spending] / ∑[penetration rate * price] 1.34 1.08 1.42 1.50 1.28 1.41

Source: own elaboration; based on OSILAC, 2009.

MAPPING OUT THE TRANSITION 92
Table also shows the result of the ratio between the average US$ spending per
household member and the sum of the individual price of our ‘ICT access packet’
multiplied with the respective penetration rate (comparing real and expected spending).
The national figures show that our price estimations have been quite reasonable. In the
case of Mexico 2002, our estimations have only been 3% off the actual spending levels
for the national average (1.03). However, taking a closer look at how this is distributed
among the different income-quintiles, it can be seen that the richest 20% of the Mexican
population seems to spend more than what would be expected with our ‘ICT access
packet’ (i.e. 64% more; ratio [real spending]/[expected spending to purchase ‘ICT access
packet] = 1.64). This is not surprising, considering the large variety of expensive high-
tech equipment and services available (including 3G and 4G mobile phones and
expensive computer systems). Rich Mexicans buy much more ICT than are included in
our packet. At the same time, multiplying penetration rates with the respective costs of
the components of our packet shows that the average member of Mexico’s 2002 poorest
quintile can only buy 80% of our ‘ICT access packet’. In agreement with what we have
anticipated, we can conclude that those people make less intense use of ICT than what we
expect with our minimum ‘ICT access packet’. Those who have mobile phones do not
use it every day for one minute and/or those who have a computer at home can be
expected to have obsolete equipment, among other reasons.
The evolution of the digital divide in Mexico between 2002 and 2007 shows us that
this tendency has become more pronounced in the poor segments over time (the ratio of
[real spending]/[expected spending to purchase ‘ICT access packet] in the poorest
quintile falls from 0.80 to 0.59, implying that the average member of the poorest quintile
MAPPING OUT THE TRANSITION 93
can only by 59% of our ‘ICT access packet’ in Mexico 2007). While more of the poor are
connected, the poor spend on average less on digital communication. In other words,
while penetration rates have expanded (compare Table), it seems that this has, at least
partially, been on the expense of the quality of the services purchased by the user in the
low-income segment, while high-income segments proceed to higher quality services. As
a consequence, the qualitative dimension of the digital divide widens.
Comparing Tables, the data show nicely how income structures contribute to create
ICT diffusion patterns. For example, looking at the other quintiles of Mexico between
2002 and 2007 shows why connectivity has almost exclusively increased in the high-
income quintile. As shown by Table, Internet access in quintile I, II, III and IV has only
increased marginally (by 2, 3, 5 and 1 percentage points), while the high-income segment
increased from 22% to 44% Internet penetration during this period. This is because
Internet access (requiring a computer and connectivity) represents the largest part of ICT
spending (50% of our ‘ICT access packet’, see Table) and those segments only have part
of the required spending power. Mobile phone, on the contrary, has increased
substantially also in the poorer quintiles, which is not surprising, considering that our
mobile phone sub-packet has fallen from US$99 to US$57 (see Table, making that
around 20% of our packet). It is therefore within the possibilities of most income groups
(we will present the yearly personal spending per quintile in US$ in Table 4).
A special case is Costa Rica. Table shows that all segments actually do spend more
than what we would expect with our ‘ICT access packet’, and that despite the fact that it
is the poorest country of our sample. Costa Rica counts with a public telecommunications
provider (GRUPO ICE: soluciones de electricidad e Infocomunicaciones para la
MAPPING OUT THE TRANSITION 94
inversión en Costa Rica), a state owned monopoly created in 1963. It has survived
various attempts of privatization and, as can be verified in Table, is providing
exceptionally low fixed and mobile tariffs (Internet tariffs have also become among the
lowest in the region by 2008; see ITU, 2009). This leads to the paradoxical fact that the
poorest quintile of the (low-income) Costa Rican society (who spend 8% more than the
cost of our ‘ICT access packet’) can buy more access than the richest quintile in
(relatively wealthy) Uruguay (who only spend 7% more than our packet). Uruguay has a
history of inadequate international connectivity and therefore traditionally high costs for
Internet services, especially around 2005 (as can be verified in Table).
This closes our descriptive analysis. While our data do not give us a perfect
description of reality–which is not surprising remembering all the previously analyzed
factors that we have decided to leave out of our modeling effort—we have seen how
income patterns relate to ICT diffusion patterns and that we can roughly model the digital
access divide in this way.

What does our model tell us about future scenarios?
In this section we will apply the previously used logic to create three future
scenarios of the domestic digital access divide in the analyzed countries. The two basic
alternatives to bridge the income dimension of the digital access divide consist of either
reducing technology prices to the level of available income (supply side policy), or
increasing the purchasing power of the poor (demand subsidy, for example by public
subsidies, international development aid or private philanthropic donations). We can
analyze both alternatives with our model by changing the penetration levels to the desired
MAPPING OUT THE TRANSITION 95
level, and by multiplying these desired levels of connectivity with futuristic prices for our
‘ICT access packet’. Scenario 1 focuses exclusively on reducing prices and explores how
much we would have to cut the price of our ‘ICT access packet’ to reach a desired level
of penetration. Scenario 2 supposes stable prices and estimates the amount of resources
(as % of GDP) that would be needed in order to connect the desired percentage of
population. Scenario 3 is more realistic and combines both of these approaches.
First we have to choose a diffusion level that we can reasonable consider as our
“universal service” benchmark. We have to determine at which point the divide can be
considered as being bridged. It would be slightly naïve to suppose 100%, as in reality
there are always segments of society that fall between the cracks, especially in the
developing world. A more realistic approach would be to strive for lifting ICT
connectivity levels to the current levels of electricity services, supposing that, at least in
Latin America, electricity is generally accepted to be universally available (Mexico:
95.7%; Uruguay: 98.5%; Brazil: 96.6%; Costa Rica: 99.1%). Table shows electricity
penetrations per income quintile. The difference between current levels of ICT
penetrations and current levels of electricity penetration constitute our digital divide.
Scenario 1 in Table shows the levels of required price cuts to reach the electricity
penetration level. Scenario 2 in Table shows the required subsidies (as % of GDP), if we
would wish to enable these unconnected income segments to purchase our ‘ICT access
packet’ at the current price level from Table. Those figures are shown as national
aggregate and per income segment, whereby the poorest quintile has been divided into
deciles. Table also shows the personal spending levels in US$ per income segments.
MAPPING OUT THE TRANSITION 96
Based on the data of Brazil 2003, for example, prices would have to be cut down to
3% of 2003 prices (to less than US$7 per year) to connect the unconnected of the poorest
decile (Scenario 1), or alternatively 6.46% of GDP would have to be spent to subsidize
missing connectivity at current prices (Scenario 2). As also shown by Table, this would
be equal to Brazil’s total public spending in health (4.8% of GDP), plus national spending
on Research and Development (R&D, 1.0% of GDP), plus public spending on tertiary
education (0.8%). In the case of Uruguay 2005, ICT prices would have to be cut down to
6% of the 2005 prices (Scenario 1), because the poor have around US$20 per year to
spend on ICT, making that around 40 cents per week in one of Latin America’s richest
countries. Alternatively (Scenario 2), the required amount of resources to bridge the
digital divide (6.23% of GDP) would equal the amount of Uruguay’s public spending in
health (3.6% of GDP) and education (2.6% of GDP).
When evaluating the results of the model, we have to remember one complementary
factors that dynamically influences the evolution of the income dimension of the digital
access divide: growing spending levels. Spending levels would increase if households
start to spend a larger amount of their spending on communication. However, the time
series of Mexico 2002-2008 in Figure seems to suggest that relative portions of spending
on communication seem to fluctuate back and forth, but without a clear tendency toward
increased proportions
35
. Another catalyst of increased spending levels would be
economic growth. If the absolute income of households increases, so does spending on
communications. Reflecting on this possibility we have to remember that, first, household
spending does usually not grow as fast as economic growth would suggest, and second,

35
National household spending levels on communication as % of total household spending in Mexico:
2002: 3.01%; 2004: 2.89%; 2005: 3.20%; 2006: 3.08; 2008: 3.15%.
MAPPING OUT THE TRANSITION 97
traditionally only a very small portion of this trickles down to the lowest income
segments. The Mexican economy has grown at an average GDP growth rate of 6.46% per
year between 2002 and 2007 (CEPAL, 2009). This had some noticeable effects on the
communications spending of high-income segments. Table shows that the highest income
quintile has increased spending on communication from US$306 in 2002 to US$374 in
2007. This is lower than what economic growth would suggest (306*[1.0646]
5
=
US$418), but it is still an important increase. However, the trickle down effect of this
economic growth to the poorest 20% of Mexican society, has only led to an increase from
US$34 in 2002 to US$35 in 2007 of personal spending per year (far from what should be
expected considering economic growth: 34*[1.0646]
5
= US$47). Income levels in poor
segments grow very slow, which is the main causes for the persistence of worldwide
poverty and the reason why 22% of Latin America still lives with less than US$2 per day.
As a conclusion of this observation, we can place some (uncertain) hopes on economic
growth. On the bright side, we can hope that economic growth (which would partially be
also based on ICT-enabled productivity gains) would lift higher income segments to the
required levels of spending. On the down side, history has taught us that that spending
levels at the so-called “bottom of the pyramid” can be expected to move very slowly
beyond the presented levels.
This observation allows us to fine-tune our interpretation of Table. Expecting
economic growth to lift the richest 60% of Brazilian society up to the required spending
levels, “only” 2.84% of GDP (0.68%+0.70%+1.46%) would have to be spent to subsidize
the remaining 40% (see Scenario 2 in Table). This is more than the country receives
annually in Foreign Direct Investments (FDI). Expecting economic growth to lift the
MAPPING OUT THE TRANSITION 98
richest 60% of Uruguayans up to the required spending levels, “only” 1.15% of GDP
(0.59%+0.56%) would have to be spent to subsidize the remaining 40%. This is more
than twice as much as Uruguay’s public spending on tertiary education.

MAPPING OUT THE TRANSITION 99
Table 5 Required price reductions and subsidies for scenarios 1, 2 and 3; personal
communication spending; comparison to other national expenditures; weekly
minutes in public access center; all as national average and per income quintile

MEXICO 2002

National Decile 1 Decile 2
Quintile
II
Quintile
III
Quintile
IV
Quintile
V
Electricity penetration

95.7% 93.9% 94.7% 96.7% 96.8% 98.4% 98.6%
Scenario
1:
Required price cuts to reach electricity
levels (% of current price)
8% 12% 15% 20% 28% 64%
Personal spending per year in US$ $118 $ 34 $54 $78 $119 $306

MEXICO 2007

National Decile 1 Decile 2
Quintile
II
Quintile
III
Quintile
IV
Quintile
V
Scenario
1:
Required price cuts to reach electricity
levels (% of current price)
13% 15% 21% 31% 46% 98%
Scenario
2:
Required subsidies to reach
electricity levels (% of
GDP2007)
1.86% 0.20% 0.18% 0.41% 0.39% 0.40% 0.27%
Personal spending per year in US$ $143 $35 $59 $93 $154 $374
Scenario
3:
Reduced price packet $145 $118 $134 $145 $162 $190
Percentage of society that can
afford reduced price packet
46.7% 29.8% 44.1% 64.5% 95.2%
everybo
dy
Remaining required subsidy (%
of GDP)
0.42% 0.16% 0.15% 0.10% 0.01%
no
subsidy
required
Other expenditures as % of GDP Mexico (2002-2005):

Public
health:
3%
Public
educ:
5.4%
Tertiary
educ:
0.9%
Research
&
Develp:
0.4%
Official
develop
assist.:
0.0002%
Foreign
direct
invest.
inflow:
2.4%
Weekly minutes in public access center

83 20 34 54 89

URUGUAY 2005 National Decile 1 Decile 2
Quintile
II
Quintile
III
Quintile
IV
Quintile
V
Electricity penetration 98.5% 96.4% 97.6% 98.5% 99.2% 99.6% 99.7%
Scenario
1:
Required price cuts to reach electricity
levels (% of current price)
6% 10% 18% 25% 35% 60%
Scenario
2:
Required subsidies to reach
electricity levels (% of
GDP2005)
6.23% 0.59% 0.56% 1.31% 1.36% 1.35% 1.06%
Personal spending per year in US$ $155 $29 $79 $122 $193 $355
Scenario
3:
Reduced price packet $242 $176 $229 $265 $306 $332
Percentage of society that can
afford reduced price packet
32.0% 16.4% 34.5% 45.9% 63.0%
everybo
dy
Remaining required subsidy (%
of GDP)
2.18% 0.57% 0.59% 0.57% 0.45%
no
subsidy
required
Other expenditures as % of GDP Uruguay (2002-
2005):

Public
health:
3.6%
Public
educ:
2.6%
Tertiary
educ:
0.5%
Research
&
Develp:
0.3%
Official
develop
assist.:
0.1%
Foreign
direct
invest.
inflow:
4.2%
Weekly minutes in public access center 90 17 46 70 111

MAPPING OUT THE TRANSITION 100
BRAZIL 2003 National Decile 1 Decile 2
Quintile
II
Quintile
III
Quintile
IV
Quintile
V
Electricity penetration 96.6% 87.8% 94.6% 97.2% 98.4% 99.1% 99.8%
Scenario
1:
Required price cuts to reach electricity
levels (% of current price)
3% 5% 12% 22% 37% 73%
Scenario
2:
Required subsidies to reach
electricity levels (% of
GDP2003)
6.46% 0.68% 0.70% 1.46% 1.52% 1.28% 0.81%
Personal spending per year in US$ $96 $9 $32 $69 $117 $254
Scenario
3:
Reduced price packet $136 $113 $125 $145 $149 $166
Percentage of society that can
afford reduced price packet
32.0% 8.1% 25.7% 47.2% 79.0%
everybo
dy
Remaining required subsidy (%
of GDP)
1.89% 0.61% 0.59% 0.49% 0.20%
no
subsidy
required
Other expenditures as % of GDP Brazil (2002-2005):

Public
health:
4.8%
Public
educ:
4.4%
Tertiary
educ:
0.8%
Research
&
Develp:
1.0%
Official
develop
assist.:
(.)%
Foreign
direct
invest.
inflow:
1.9%
Weekly minutes in public access center 56 5 19 40 68

COSTA RICA 2004 National

Quintile I
Quintile
II
Quintile
III
Quintile
IV
Quintile
V
Electricity penetration 99.1% 97.3% 98.8% 99.6% 99.8% 99.9%
Scenario
1:
Required price cuts to reach electricity
levels (% of current price)
8% 18% 28% 40% 87%
Scenario
2:
Required subsidies to reach
electricity levels (% of
GDP2004)
3.74% 0.87% 0.84% 0.80% 0.74% 0.49%
Personal spending per year in US$ $92 $17 $39 $61 $97 $248
Scenario
2008:
2008 price packet $135 $123 $126 $126 $142 $165
Percentage of society that can
afford 2008 price packet (equal
to required price cuts of 2008
price to reach everybody)
32.2% 13.9% 30.6% 48.8% 67.9%
everybo
dy
Required subsidy to reach
electricity levels (% of
GDP2008)
0.85% 0.29% 0.25% 0.18% 0.13%
no
subsidy
required
Other expenditures as % of GDP Costa Rica (2002-
2005):

Public
health:
5.1%
Public
educ:
4.9%
Tertiary
educ:
n.a.
Research
&
Develp:
0.4%
Official
develop
assist.:
0.1%
Foreign
direct
invest.
inflow:
4.3%
Weekly minutes in public access center 53 10 22 35 56

Source: own elaboration; based on OSILAC, 2009; CEPALSTAT, 2009; UNDP,
2009; respective national household spending reports.

At this point it becomes clear that neither real price reduction of ICT nor demand
subsidy can be a solution by itself. The challenge has to be faces with a sophisticated
MAPPING OUT THE TRANSITION 101
combination of different options. We start with the assumption that the market still has
potential to reduce ICT prices, which will enable more people to afford connectivity. As
long as this price reduction does not go as far as identified in Scenario 1, we will still
require a reduced subsidy. Much in the same line of thought, a recent study by Latin
American telecom regulators (Regulatel, 2007) discusses the “market efficiency gap”
versus the “access gap.” The first concept denotes the difference between the current
level of service penetration and the level achievable in a hypothetically well-functioning
competitive market under a stable regulatory environment. The “access gap” takes note
of those segments of society that continue to be unable to afford access even under best-
case market conditions.
So let us become a little futuristic and explore Scenario 3. Let us start with the
assumption that eventually it would be possible to create the US$100 laptop (or any other
comparable device at that price, such as an evolved multimedia mobile or digital TV
device). Let us furthermore suppose that this connectivity would be enough to provide
voice services (such as Voice-over-IP) and that through some futuristic 4G wireless
connectivity, the device could be moved around as a mobile phone, constantly connecting
to the closest base-station to provide Internet service to each individual, while the price of
this Internet connectivity would still be shared by all members of the household (a kind
of joint family plan for mobile Internet services). This would reduce prices to the plain
price of equipment (US$100 per person for three years equipment lifetime, making it
US$33 per year) and a household-shared Internet connection (based on current prices),
eliminating fixed and mobile phone costs. Let this be our ‘reduced price ICT access
packet’.
MAPPING OUT THE TRANSITION 102
Table shows the result of this Scenario 3. In Mexico 2007, 47% of society could
buy such “reduced price access packet”, which is a clear improvement from the 12% of
society that has access to the Internet in 2007 (compare to Table). However, in order to
connect the ones that still remain unconnected, still 0.42% of GDP would be required as
subsidies. This is equal to the amount Mexico spent on Research and Development
(R&D) and half of Mexico’s public spending on tertiary education (see Table). In
Uruguay, Internet connectivity could be increased from 14% (see Table) to 32% of
society (Scenario 3 Table), but the amount required to bridge the digital divide for the
remaining unconnected (2.18% of GDP) would be 22 times the total amount of
international Official Development Assistance (ODA) the country receives per year (see
Table).
As already indicated, these scenarios depend on the future development of prices
and spending levels. While we do not have more recent statistics on spending, we do
have the price figures for 2008 (ITU, 2009). Uruguay and Costa Rica have been able to
reduce the price of their ‘ICT access packet’ from US$ 468 in 2005 to US$275
36
in 2008
and from US$ 237 in 2004 to US$135.3 in 2008 respectively (in Uruguay mainly due to
reductions in Internet and mobile prices, and in Costa Rica due to lower mobile prices).
The 2008 prices in Uruguay come close to the envisioned prices of our ‘reduced price
packet’ (US$242). If we could now also suppose that Uruguay’s economic growth at least
benefits the three higher income quintiles so that market forces would eventually enable
them to connect, we could conclude that in these cases the required subsidy would be

36
Uruguay’s 2008 ‘ICT access packet’ is the sum of
[US$77.4fixed]+[US$39.6mobile]+[US$100.7Internet]+[57.5PC].
MAPPING OUT THE TRANSITION 103
limited to the two lower income quintiles (which reduces subsidies to a little more than
0.57%+0.59% = 1.16% of GDP, see Scenario 3 in Table).
In the case of the public telecom operator in Costa Rica, prices in 2008 turn out to
be even lower than our ‘reduced price’ Scenario 3 (which would have turned out to be
US$184). Due to this impressive price cut in Internet and telephone tariffs, Table does
not show a Scenario 3 for Costa Rica, but a Scenario2008 (see Table). It takes 2008
prices and 2004 spending levels (unfortunately the last available spending levels). This
price reduction lowers required subsidies effectively to 0.85% of GDP, equal to “only”
twice the spending in R&D (see Scenario2008 in Table). In case that Costa Rica’s
economic growth had positively affected communication-spending levels between 2004-
2008, these required subsidies would accordingly be reduced.
This encouraging outlook, however, cannot be transferred to Mexico and Brazil. In
both countries the price of our ‘ICT access packet’ has risen in recent years, from US$
307 in 2007 to US$341 in 2008 in Mexico (mainly due to increases in Internet and fixed-
line prices), and from US$294 (2003) to US$ 462 in Brazil (due to the same reasons)
37
.
These unfortunate tendencies might be explained with two already observed facts: first,
high income segments have more spending power than required to purchase our
minimum ‘ICT access packet’ (see Table); and second, market expansion in high income
segments is still ongoing and has not even reached saturation for the ‘minimum access
packet’ (see Table). Therefore, the natural business strategy for private enterprises
consists in creaming off existing spending power in high-income segments. Proof of this

37
Mexico’s 2008 ‘ICT access packet’ is the sum of
[US$137.1fixed]+[US$39.6mobile]+[US$116.8Internet]+[47.4PC]; and Brazil’s
[US$146.0fixed]+[US$99.0mobile]+[US$172.1Internet]+[45.0PC].
MAPPING OUT THE TRANSITION 104
is the fact that increases in penetration rates can mainly be attributed to high-income
segments (compare Mexico2002 and Mexico2007 in Table). Combine this tendency with
the already mentioned fact that ongoing economic growth tends to favor the high-income
segments, and the result is a clear economic incentive for private enterprises to focus
their business model on the spending power of high-income segments. Considering that
in all Latin American countries, the richest 20% receive more than 50% of the national
income, it becomes beneficial to adjust commercial prices to the constantly increasing
spending power in this high-income segment. This also includes the constant introduction
of new high-quality ICT (broadband, 4G, etc), which can be expected to constitute a large
portion of the revenues of private enterprises. A consequence is increasing, not
decreasing ICT prices.
These mixed results show that the heterogeneity in access prices among countries
prohibits reaching a general conclusion. However, even in the most favorable cases, the
bridge over the digital access divide for the poorest of the region would still be a heavy
financial burden. It is actually quite utopian to think that the region’s finance ministers
could be convinced to take out even 0.5% of GDP from the general public budget to
reduce the digital divide in households. And even if developing countries would have the
resources to do so, one would have to ask if this step would be justifiable, or if this
money would rather be spend on other urgent issues, such as modernizing hospitals,
schools and municipalities, which can of course also be done by introducing ICT in these
entities. Such a policy would indirectly benefit citizens, who would remain unconnected
at their homes, and as such, would remain victims of the digital divide.
MAPPING OUT THE TRANSITION 105
Table also shows that the required funds are currently not available with the
international development community in forms of Official Development Assistance
(ODA). The required amount of yearly subsidies for the reduced price Scenario 3 in
Mexican 2007 only, would be almost twice as much (US$4.3 billon) as the annual net
budget of the United Nations (US$2.5 billon in 2009). In other words, even if market
forces would drastically reduce ICT prices, the domestic and international financial
mechanisms would stop far short from enabling universal service.
Before going on to drawing conclusions, we should once again point to the
limitations of our modeling effort. Table shows us that the ‘ICT access packet’ in
Mexico’s highest income segment has decreased from US$481 in 2002 to US$388 in
2007 (see Table) and Table confirms that spending levels have increased from US$306 to
US$374, thanks to economic growth. As a result, in theory, almost every member of the
highest income segment in Mexico should have had enough resources to buy our ‘ICT
access packet’ in 2007. Table confirms that Internet penetration has indeed risen in this
segment during these five years from 43% to 63% and mobile phone penetration from
27% to 91%. However, still not everybody is connected in 2007, despite the economic
capacity to do so. This is proof of the fact that our model is simplified and ignores other
important explanatory variables, such as discussed in the second section of this article. As
previously shown, the correlation between access and income is high, but not perfectly 1-
to-1.

MAPPING OUT THE TRANSITION 106
What are the odds for the digitally excluded?
The underlying question of this paper was as simple as it is direct: How much
would it cost and how far would ICT prices need to be reduced to bridge the domestic
digital divide in the Latin American context? The answer has been modeled on basis of
household spending statistics and adjustable ICT penetration rates and access prices. The
results turn out to pose an extremely challenging task, constrained by the reality of Latin
American income levels. Any realistic bridge over the digital divide will need to be
constructed as a combination of various solutions, including price reduction and
sophistically distributed subsidies in close public-private cooperation, and hopefully
alleviated by substantial (ICT-enabled) economic growth that trickles down to the poorest
segments. We will discuss some examples of the available alternatives in this concluding
section.
Given the fast and unforeseeable technological progress of ICT, it does not seem
advisable to limit price reduction strategies to a specific technological solution. New
technology is being developed at this moment and this innate uncertainty has to be taken
as an opportunity, not a threat. It is definitely easier to change technology than it is to
change the present reality of income levels. The development of cheap (or free) source
software, as well as of cheap hardware equipment makes part of this approach. The US$
100 laptop set off a debate that moves into the right direction. Another area with great
potential is the introduction of digital TV. Converter “set-up boxes” can upgrade
analogue television sets with digital interactivity and have been manufactured for less
than US$40. However, the eventual provisions of e-government, e-health and e-business
services through a converted analogue TV set still remain to be seen. Besides equipment
MAPPING OUT THE TRANSITION 107
and software, traffic prices also need to be reduced. The use of unlicensed radio
frequency spectrum, in combination with cheap broadband wireless technology, such as
the popular Wi-Fi, is part of this challenge.
Regarding public subsidies, our model has clearly shown that income levels of the
developing world are simply too low to strive for “universal service” in the short-term.
Household connectivity is a viable option for the high-income segments, but low-income
segments will have to be satisfied with a voice- and short-message based mobile
telephony. While mobile phones are an important first step, they do not convert the poor
into full-fledged members a true Information Society. In order to provide e-government
and other e-services to the poor, the provision of “universal access” seems much more
realistic than the ambitious goal of “universal service”. The economic model behind this
strategy is an old one. It is the same that gave birth to Thomas Jefferson’s ideal of giving
access to books to all people by sharing their fixed price through the establishment of
public libraries. Fitness clubs and the public transportation system follow the same line of
reasoning. The benefit lies in sharing the fixed cost, while covering the variable cost.
Even more than 125 years after the commercialization of the automobile, not everybody
in the developing world posses a car. But most people have nevertheless access to
automated mobility, thanks to decades of massive public and private investments in
public transportation. Special machines have been built for collective access to
transportation—such as busses, metros, auto rickshaws, “guaguas” and “tuk-tuks”—and
micro entrepreneurs provide transportation services and offer to share their motorcycles,
often resulting in breathtaking acrobatic acts. The logic of collectively sharing the access
price to technology becomes obvious when looking at the reality of transportation
MAPPING OUT THE TRANSITION 108
systems in developing countries even 125 years after the “mobility-revolution”. Of
course, this still leaves a qualitative dimension of the “mobility divide” (some have more
mobility than others, and the well-off have better cars and even go by helicopter), but the
“minimum access divide” to mobility has been bridged.
The same reasoning is behind the massive sprouting of the so-called “info-centros”,
or public ICT access centers. More than 140,000 of these public access points have been
identified throughout Latin America in 2006 already (Maeso and Hilbert, 2006). The only
viable solution for an important part of the region’s society might be to buy some minutes
in a public Internet access point. Table shows that the poorest 20% of the analyzed Latin
American societies have between US$ 0.18 – 0.67 per week to spend on ICT
(Mexico2007: US$ 0.67 per week; Uruguay2005: US$ 0.55; Brazil2003: US$ 0.18; Costa
Rica2004: US$ 0.33), while 40% of society has less than US$ 1.50 per week. Supposing
an average cost of US$2 per hour for Internet access at a commercial Internet café
38
, the
poorest 20% of Costa Rica2004 could buy some 10 minutes of access per week.
Generally speaking, Table shows that the poorest half can afford around half an hour of
Internet access at a public access place at a price of US$2 per hour. This is not a lot, but
could enable them to take care of an urgent transaction with the government or make a
reservation at a far away hospital, for example. One additional benefit of public access
places is the frequent updating of equipment and service (usually broadband connectivity
is available and the equipment is being maintained by the owner), while there is always
help around to assist the user to overcome skill limitations.

38
Prices for commercial cyber-cafes in 2008/9 range between US$1 per hour (Brazil and Uruguay) up to
US$3 per hour in Mexico.
MAPPING OUT THE TRANSITION 109
While shared access seems to be a viable solution to “bridge” the divide, the
financial sustainability of the applied business models is a mayor concern (e.g. Celedon
and Razeto, 2009). The results of the analysis presented here show a reality in which it
does not seem likely that the poor will gain sufficient purchasing power to attain
personalized access in the short term. The logical conclusion is to prepare for a long
period in which public access is the only viable access solution to assure quality and up-
to-date access for these income segments. An apparatus would be comparable to the
institutional structure of today’s public transportation system, which might of course
consist of public and private components (similar to public transportation).
Notwithstanding, any sustainable solution of such institution requires a reliable stream of
resources. Currently, few financial mechanisms are in place to support such an
institutional structure of public access to information.
Throughout Latin America, public Universal Access Funds have started to support
those shared access initiatives. Most Latin American countries maintain such Universal
Access Funds, which are alimented by an earmarked tax on the telecommunications
industry (and are therefore highly polemic, given their interventionist nature). These
funds have their historical roots in the days of public pay phones and usually charge
telecom operators around 1% (up to 5%) of their revenues, in order to finance
connectivity in underserved populations. Until 2006, these funds have collected US$2.7
billion throughout Latin America (Regulatel, 2007). The Brazilian FUST (Fundo de
Universalização dos Serviços de Telecomunicação), has collected a yearly average of
US$ 354 million between 2001 and 2006. This constitutes 3.4% (or
1
/
30
) of the subsidies
that would be required to provide individual universal service in the case of our ‘reduced
MAPPING OUT THE TRANSITION 110
price access packet’ (see Table). In other words, if prices could be reduced to the aspired
level of our reduced price scenario and if the fund would be modernized to actually
subsidize public ICT access initiatives, then the Brazilian Universal Access Fund could
provide one shared connectivity packet per 30 unconnected inhabitants. This does not
lead to “universal service”, but to “universal access” and 30 people per station is not
overly crowded for a public access cyber café. This scenario enables to end our analysis
on a “glass-half-full” outlook. It seems that the innovative combination of public and
private efforts to reduce prices, together with public and private efforts to provide shared
access provide at least one viable path to follow.
It is important to point out that not all countries around the world have such funds,
and many of the existing funds face fierce opposition from groups that defend free-
market mechanisms. Others have suggested exploring the possibility to extend this model
to the international level and hold the globally connected responsible to connect the
global poor. This fits the global nature of digital networks and it would also enable to
expand the logic of digital solidarity from telecom, to hardware and software services. As
already mentioned, most ICT are tradable goods. In the established Universal Access
Funds, however, only the non-tradable telecom services are subject to contributions.
National authorities only tax non-tradable national telecommunications service
companies, and do not place a tax on imported hardware or software products. This does
not only result in indirect subsidies from telecom companies to hardware and software
MAPPING OUT THE TRANSITION 111
companies
39
, but is also difficult to justify ideologically, because “ICT access” requires
hardware and software just as much as telecom services.
During the 2003-3005 World Summit on the Information Society, Heads of States
and governments have proposed the creation of a public-private “Digital Solidarity
Fund”, “as an innovative financial mechanism of a voluntary nature” (WSIS, 2005). The
initial idea of some participants was to establish global contribution system, similar to the
national Universal Access Funds in Latin America, which will then be alimented by the
global ICT industry. A contribution of 1% of the revenues of the world’s ten largest non-
telecom ICT enterprises
40
(Fortune, 2008) would have provided US$ 4.6 billion to
provide connectivity to the world’s poor in 2008. Used for public access or for R&D to
develop cheap equipment suitable for the poor, such a constant source of yearly resources
would surely have a significant and sustainable impact. Notwithstanding the potential of
the idea, the voluntary nature of the Digital Solidarity Fund has raised less than US$10
million during the entire period of its first five years of existence (until 2009), with a
substantial part of the donations coming from developing country governments in a good-
will effort to show their sympathy for the idea. This reality shows the “glass-half-empty”
side of the current challenge: the low degree of commitment to the overcome of the
digital abyss. Many more innovative and creative ideas—and their practical

39
By forcing telecom operators to expand their networks to underserved areas, new markets are
consequently opened for hardware and software producer in the developing world. Telecom operators have
to chip in to make this step possible, while hardware and software producers are under no regulation to
provide cost-effective solutions to marginalized populations. Additionally, if public access centers are
financed by the universal service funds, these centers need to buy hardware and software equipment, which
in some cases is financed with the resources that are collected from the telecom operators through the fund.
40
In 2008 this included: Hewlett-Packard, IBM, Dell, Microsoft, Intel, Cisco Systems, Apple, Oracle,
Xerox and Google.
MAPPING OUT THE TRANSITION 112
implementation—will have to be explored to find sustainable solutions for the digital
excluded.
Let us relativize these findings with a final word of caution on the limitations of the
presented approach. It has to be remembered that the presented numbers will inevitably
change over time as new household spending surveys become available. What will not
change as easily, however, is the general logic of combining the structural characteristic
of the highly skewed income distribution in low-income countries with tradable ICT
equipment, whose prices are internationally defined. While the numbers have to be
checked in the future, it can be expected that the digital divide will persist over the
coming years. Besides new statistical input, the current model can and should also be
refined (which will happen at the cost of increasing complexity). Two of the most
straightforward improvements would include estimations on future spending levels
(considering estimations for economic growth and its trickle down effect to different
income levels) and the inclusion of explanatory variables beyond income distribution
(education levels seems especially promising).
Summing up, by putting numbers and quantities to the omnipresent rhetoric about
the digital divide, this paper has shown that, once again, long-standing structural
characteristics of the developing world could be about to deepen the vicious circle
between inequality and technology diffusion. The numbers have shown that the challenge
of breaking this circle is a formidable one. The presented model allows for a
quantification of the challenge and the identification of normative goals to break it. The
development of such normative models is an important first step, but not sufficient. The
exploration of practical policy tools requires complementary further research in the light
MAPPING OUT THE TRANSITION 113
of the presented findings. This research will have to be realistic and convincing enough to
motivate national and international public and private sectors to take up those concepts
and implement sustainable solutions that enable the world’s unconnected to become full-
fledged members of a truly global Information Society.

MAPPING OUT THE TRANSITION 114
Digital Gender Divide or Technologically Empowered Women
in Developing Countries? A Typical Case of Lies, Damned
Lies, and Statistics
41,42

The discussion about women’s access to and use of digital Information and
Communication Technologies (ICT) in developing countries has been inconclusive so
far. Some claim that women are rather technophobic and that men are much better users
of digital tools, while others argue that women enthusiastically embrace digital
communication. This article puts this question to an empirical test. We analyze data sets
from 12 Latin American and 13 African countries from 2005-08. This is believed to be
the most extensive empirical study in this field so far. The results are surprisingly
consistent and revealing: the reason why fewer women access and use ICT is a direct
result of their unfavorable conditions with respect to employment, education and income.
When controlling for these variables, women turn out to be more active users of digital
tools than men. This turns the alleged digital gender divide into an opportunity: given
women’s affinity for ICT, and given that digital technologies are tools that can improve
living conditions, ICT represent a concrete and tangible opportunity to tackle
longstanding challenges of gender inequalities in developing countries, including access
to employment, income, education and health services.

41
This article was published as: Hilbert, M. (2001). Digital gender divide or technologically empowered
women in developing countries? A typical case of lies, damned lies, and statistics. Women’s Studies
International Forum, 34(6), 479-489. http://dx.doi.org/10.1016/j.wsif.2011.07.001
42
The author would like to thank the support of Canada’s International Development Research Centre
(IDRC), which has been the driving force behind the creation of important statistics throughout the
developing world for decades. Without the long-term vision, dedication and trust of its professionals, like
Ben Petrazzini, this, and many other studies of this kind, would not exist
MAPPING OUT THE TRANSITION 115
During the second half of the last century, human kind has turned to the “massive
task of making our bewildering store of knowledge more accessible” (Bush, 1945). The
result has brought on irrevocable social, productive, political and cultural
transformations, which are based on a global communication infrastructure that includes
innovations like the Internet, mobile telephony and social networking applications in all
shapes and sizes. During the beginning of this new century, society at large is starting to
embrace these new tools, changing forever the way we communicate, coordinate our
activities and organize social interactions (Bell, 1973; Perez, 1983; Webster, 1995;
Negroponte, 1995; Castells, 1996; Freeman and Louça, 2001). At the core is the question
of access to digital networks, and, in particular, who gets empowered and who is
informationally marginalized by use of these new tools.
As a contribution to this ongoing discussion, this article analyzes the differences
between men’s and women’s access to and use of Information and Communication
Technology (ICT) in developing countries. We start with a literature review that shows
that some see digital technologies as practical and tangible tools for women to overcome
longstanding inequalities. ICT can help women to gain employment (for example through
telework or newly created information jobs), obtain cost-effective health services and
education (such as through online courses or software-based literacy programs) and to
increase their income (such as through e-business channels and online transactions). In
contrast to this glass-half-full outlook stands the pervasive and persistent
counterargument that women are at a natural disadvantage to benefit from the digital
revolution because they are less tech savvy, and more technophobic, and because the
technology is not built for their needs and intuition. If this were the case, the increasing
MAPPING OUT THE TRANSITION 116
socio-economic importance of ICT would add a new dimension to the already existing
vicious circle between discrimination and women’s backwardness, which can be
expected to be particularly severe in developing countries, where four out of five women
live worldwide. Unfortunately very few of the related studies control for potentially
confounding variables.
43
We know that the lack of employment, income and education
affect ICT usage negatively (e.g. NTIA, 1999; Cullen, 2001; Warschauer, 2003;
Mossberger, et.at., 2003; OSILAC, 2007). We also know that women are discriminated
against in many aspects of social life, including employment, income and education.
Given these potential confounders, it is not clear if being a woman per se has a negative,
neutral or positive effect on ICT usage. In the first case, the digital revolution would pose
a severe threat to women. In the latter case, the increasing socio-economic importance of
ICT would pose a unique opportunity: the new tools would be a perfect tool to fight
existing inequalities between men and women.
To this point, lack of adequate statistical data had prevented us from testing this
question empirically. Arguments were often based on anecdotal evidence from case
studies or uncontrolled correlations, which sometimes lead to contradictory results. In
recent years, statistical institutes and academic research centers in the developing world
have made a significant effort to collect adequate statistical information. For this study
we employ 25 datasets from 12 Latin American and 13 African countries (total of
1,176,816 observations), which allows us to execute a series of uncontrolled and
controlled empirical tests that will provide further insight into this unresolved question.

43
A notable exception for the case of the United States is Rice and Katz (2003), which, after control, do not
detect any significant digital divide between men and women. The author is not aware of any controlled
studies for the digital gender divide on the international level.
MAPPING OUT THE TRANSITION 117
What is the digital divide?
The unfolding of the digital revolution is happening at unprecedented speed (for
ICT penetration rates during the past 15 years, see ITU, 2010). However fast, it is not
immediate and the related diffusion process follows the form of a well-known S-shaped
curve, which distinguishes between early adopters and latecomers (Rogers, 2003). While
this process unfolds, a new form of inequality is added to all the existing forms of
discrimination: an inequality in the power to communicate and to process information
digitally. The term “digital divide” has been coined to refer to this concept (e.g. NTIA,
1995, 1999; OECD, 2001).
Studies on the digital divide differ in their focus and methodological approach.
Despite their differences, all of them answer (part of) the following questions: who
(individuals vs. organizations/communities, vs. societies/countries/ world regions, etc.),
with which attributes (income, education, geography, age, gender, or type of ownership,
size, profitability, sector, etc.), connect how (pain access vs. usage vs. real impact), to
what kind of technology (phone, Internet, computer, digital TV, etc.) (see Hilbert, 2011).
In this article we test for one specific attribute of ICT users: their gender.
The main focus of this article is set on analyzing access to ICT in Latin America
and Africa, while we also sneak an exploratory peak into how men and women use the
Internet in Latin America. This is important because literature has shown that access and
usage foster the well-being in multiple aspects of life (e.g. Castells, 1996; Webster, 1995;
Waverman, et.al. 2005; Cimoli, et.al, 2010; Hilbert and Peres, 2010). Previous research
has shown that ICT adoption patterns are characterized by the same long established
determinants of inequality as other aspects of social life, such as those related to income,
MAPPING OUT THE TRANSITION 118
education, skills, employment, geography, age and ethnicity, and gender, among others
(e.g. Cullen, 2001; Compaine, 2001; OECD, 2002; Warschauer, 2003; Mossberger, et.at.,
2003; van Dijk, 2005; OSILAC, 2007; Hilbert, 2010).

What do we know about the digital gender divide?
Let us begin with clarifying that most literature in this field refer to ICT access and
usage patterns among biologically identifiable men and women (sex), not the self-
identified gender identity of an individual, such as understood in the field of gender
studies. While it would be very interesting to explore the relationship between the digital
divide and gender identity, the paucity of data on the last variable forces us to follow
most existing research and equate gender with sex in this article.
During the 1990s, researchers were quick to observe that women tend to be
latecomers to the digital age (e.g. Dholakia, 1994; NTIA, 1999). As a consequence, the
new technology was popularly portrayed as a male domain (Badagliacco, 1990). Bimber
(2000: 2) concluded that the gap in ICT usage between women and men “is the product
of both socioeconomic differences and some combination of underlying, gender-specific
effects”. Researchers claimed that those gender-specific differences had their origins in
the fact that women underestimated their actual usage skills, which lead to lower self-
efficacy to use ICT (Busch, 1995; Joiner, et.al, 1996; Hargittai and Shafer, 2006), as well
as in their general attitudes toward computers (Shashaani, 1994). It was concluded that
“men are more interested in technology than women, and they are also more tech savvy”
(Fallows, 2005: 5). In short, women were seen as being more likely to be technophobic
and were ascribed a certain computer anxiety. This type of reasoning is in line with a
MAPPING OUT THE TRANSITION 119
longstanding argument that technology is gendered (Lohan and Faulkner, 2004; Puente,
2008). ICT are seen as yet another “toy for the boys” (Faulkner, 2001).
As more statistics became available and Internet and mobile telephony penetration
rates began to rise, women started to catch up in many developed countries (Rice and
Katz, 2003). In the United States, most new users were women around the year 2000
(Cummings and Krout, 2002). Gender differences remained, but were smaller (Leggon,
2006) and mainly concentrated on marginalized groups, such as ethnic minorities
(Tolbert, et.al., 2007). However, once online, women remained less frequent and less
intense users of the Internet (Ono and Zavodny, 2003; Wasserman and Richmond-
Abbott, 2005). The focus of attention started to shift towards differences in how men and
women use ICT (Bonfadelli, 2002). For example, it was found that girls use the Internet
for instant messaging and chat-rooms, whereas boys downloaded games and music,
engaged in online trading, and created Web pages (Lenhart, Rainie, & Lewis, 2001;
Roberts and Foehr, 2004). Fallows (2005: 1) summarized a survey in the United States
with the conclusion: “men like the internet for the experiences it offers, while women like
it for the human connections it promotes”. As already mentioned, we will look at both
aspects in our subsequent analysis: access and usage, with a focus on the first one.
Statistical data from the USC led World Internet Project (2009) reconfirm these
findings. In Canada, 79% of men and 75% of women were online in 2007. This
difference grows to 56% to 46% for citizens of 60 years and older. The study also
confirms differences in usage. In 2004, Canadian men spent on average more time online
than women (14.3 to 12.0 hours per week). This difference increased from 2.3 to 3.5
hours in 2007 (18.8 hours to 15.3 hours). As the main reasons for non-usage, Australian
MAPPING OUT THE TRANSITION 120
women state lack of interest (35%), not having a computer or Internet connection (26%)
or lack of skills (16%). The percentage of men to women who use the Internet is reported
for the following developed countries: Australia: 74% to 71%. Czech Republic: 55% to
46%; Hungary: 45% to 39%; Israel: 71% to 64%; New Zealand: 78% to 77%; Singapore:
69% to 54%; United Kingdom: 68% to 65%. The two exceptions to this trend seem to be
Sweden (with 75% of men online and 78% of women) and the United States (71% to
73%). However, even in these countries, men are more frequent and more intense users.
In 2008, men from the U.S. are more likely than women to surf the web “at least daily”
(54 to 41 percent) and men spend 1.5 hours more than women at their monitors reading.
In short, differences have become smaller in developed countries, but still remain,
especially in usage.

What about women and ICT in developing countries?
Due to the paucity of adequate statistics about the world’s poor, technology-related
research and respective policy-advice is often exclusively focusing on the roughly 20 %
of the world population living in the most industrialized countries
44
, while the remaining
80 % of the global population is frequently ignored or inappropriately subsumed under
these findings. This is delicate, because living conditions, opportunities and threats differ
decisively in developed and developing countries. The vast majority of women live in
developing countries and they often suffer even more gender related discrimination than
their counterparts in developed countries. At the same time, if ICT were to hold a

44
In 2006, the countries member countries of the Organisation for Economic Co-operation and
Development (OECD), which represents the world’s “industrialized countries”, was home to 1,184 million
inhabitants, within a world population of 6,555 million (18 %).
MAPPING OUT THE TRANSITION 121
promise to empower women, than this promise is much larger in the developing world,
given that the lower starting point provides for greater potential gains.

ICT: a threat for women
Similar to the above-cited data from developed countries (World Internet Project,
2009), existing data from developing countries show that women are less likely than men
to use ICT. This leads related research to the conclusion that a digital gender divide
clearly exists and is a severe threat to women: “In many countries such gaps become
dramatic, putting women at a significant disadvantage” (Hafkin and Huyer, 2007: 33).
Similar to findings in developed countries, this divide applies to access and to the
frequency and intensity of usage (Park, 2009). Looking for reasons, researchers normally
fall back on anecdotal case studies and local evidence, which found that women face
barriers that include lack of access and training, and that they were confronted with
software and hardware applications that did not reflect their female interests and needs
(Arun and Arun, 2002; Ng and Mitter, 2005; Best and Maier, 2007). In this sense, the
same technophobic arguments that had been raised in the developed world during the
1990s, have been transferred to women in the developing world in recent years. It is
argued that women have a negative attitude toward ICT (Varank, 2007) and that the
introduction of technologies has often implicitly been designed to meet the needs of men,
not of women (Basu, 2000; Hafkin, 2000).

MAPPING OUT THE TRANSITION 122
ICT: an opportunity for women
In contrary to these findings, some case studies and anecdotal evidence show that
ICT can and are empowering women in developing countries. For example, ICT provide
women entrepreneurs with access to worldwide e-business channels, which and can be
operated 24 hours a day from home in real-time (Heeks, et.al. 2004; Schaefer Davis,
2007; Brodman and Berazneva, 2007). Ng and Mitter (2005) look beyond ICT’s
contributions to economic well-being, and show how ICT are used by women for the
purposes of community building and political organization. ICT enable meaningful
participation and make female voices heard, as proven by the role of digital networks in
feminist movements (Harcourt, 1999). Others have argued that ICT have the potential to
completely redefine traditional gender roles, especially for women who have limited
skills or who lack the resources to invest in higher education (Kelkar and Nathan, 2002).
In short, ICT can be “powerful tools for women to overcome discrimination, achieve full
equality, well-being and participation in the decisions that determine their lives and the
future of their communities. […] ICT [...] opens up a direct window for women to the
outside world. Information flows to them without distortion or any form of censoring,
and they have access to the same information as their counterparts” (Sharma 2003: 1).
However, this potential to empower women in the developing world depends on access to
and actual usage of these technologies, which is a necessary first step (see e.g. Scott,
2001).

MAPPING OUT THE TRANSITION 123
How can misleading statistics mask the reality about the gender divide?
We have seen that the literature is inconclusive. We do not know if ICT are a severe
threat or an opportunity for women. What could be the reason for this apparent
contradiction?
There is a subtle message that can often be read between the lines of research
related to the digital gender divide. For example, Sharma (2003) points out that “women
have less online access than men, for all the usual gender-related reasons—time, money,
control, learning opportunities, other commitments, prioritising others’ needs”. Arguing
that longstanding gender-related inequalities are the reason for less usage is very different
from arguing that women are naturally technophobic. It has widely been measured that
women around the world are discriminated in fields like employment, income and
education (see e.g. Anand and Sen, 1995). It is therefore not clear if these existing
inequalities lead to the fact that women make less usage of ICT or, if being a woman per
se has a negative effect on ICT usage. This problem is well known in statistics and is
treated under the topic of so-called “confounding variables” (e.g. Freedman, et.al., 2007).
Often the confounder is easy to spot. For example, if somebody would realize that
children’s ICT usage is positively correlated to the size of their shoes, most people would
become suspicious and reason that age, and therefore literacy skills, might confound this
relation. There is no reason to believe that the shoe size of children with the same level of
schooling would make any significant difference. Often it is not as easy. But the cure
remains the same: as soon as there is a suspicion of confounding variables, it is wise to
control for them and to compare subjects on the same level of such variables. If the result
MAPPING OUT THE TRANSITION 124
still makes a difference, it is more probable that the original variable has explanatory
power. If not, the confounder made the difference.
45

What does the data say?
The most frequently analyzed statistics so far have been collected by
telecommunication administrative authorities and have been harmonized by the United
Nations Telecommunications Union (e.g. ITU, 2010). Traditionally, these administrative
registers collect the national aggregates of the numbers of subscriptions, connections and
devices and therefore do not allow detailed cross-tabulations with user attributes (like
gender, income, employment, education, etc). Those are provided by the household
surveys that we will use.
Our databases are both products of the initial seed funding of Canada’s
International Development Research Centre (IDRC). In Latin America, IDRC has
cooperated since 2002 with the United Nations Economic Commission for Latin America
and the Caribbean (ECLAC) to operate OSILAC (Observatory for the Information
Society in Latin America and the Caribbean)
46
. During the last decade, OSILAC has
successfully worked with National Statistics Offices all over the region to include ICT
indicators in existing household surveys. Given the large samples of official household
surveys, this data is very robust (all used sample sizes are between 21,000 and

45
This does not change the fact that children with larger shoes will have better ICT usage scores. Same
accounts for the case of ICT and women: might be that women use ICT less than men, but the question is
why: because they are women, or because of some other reason that come with being a woman?
46
OSILAC: http://www.eclac.org/SocInfo/OSILAC/
MAPPING OUT THE TRANSITION 125
408,000)
47
. Parts of these databases are publicly available (OSILAC, 2009). In Africa,
IDRC is cooperating with the Research ICT Africa Network
48
, which has conducted their
own household and individual user surveys of ICT access and usage between 2007 and
2008 (sample sizes between 819 and 2,355).
49
Despite the smaller sample size, this is
nonetheless an important effort, as Africa is normally considered a black hole for
technology related statistics (for overview of these surveys see Gillwald and Stork, 2008).

Controlling correlations in Latin America
Let us start with a series of simple correlations between gender and both, Internet
usage and mobile phone usage. It is important to underline that we use the question of
active usage by a specific person as an indicator for access, not the plain existence of
equipment in a household. We use the Pearson correlation coefficient to measure the
degree of association between two correlation coefficients (e.g. Williams and Monge,
2001; Freedman, et.al., 2007), in our case between being a woman and using the Internet.
We code in a way that a negative correlations (r < 0) means that women use less Internet

47
Sample sizes: Brazil 2005: 408,148; Chile 2006: 268,873; Costa Rica 2005: 43,682; Ecuador 2006:
55,666; El Salvador 2006: 68,312; Honduras 2007: 100,028; Mexico 2007: 21,292; Nicaragua 2006:
40,190; Panama 2007: 48,295; Paraguay 2007: 21,053; Dominican Republic 2005: 20,610; Uruguay 2006:
64,164.
48
Research ICT Africa: http://www.researchICTafrica.net/
49
Sample sizes: Benin: 1101; Botswana: 818; Cote d’Ivoire: 1112; Ethiopia: 2355; Ghana: 1092; Kenya:
1461; Mozambique: 1131; Namibia: 885; Rwanda: 1078; Senegal: 1081; South Africa: 1771; Tanzania:
1490; Uganda: 1127.
MAPPING OUT THE TRANSITION 126
than men, while r < 0 implies that Internet usage is positively correlated with being a
woman.
50

A first look at the upper two rows of Table reveals that most of the correlations
between ICT usage and being a woman turn out to be negative. In agreement with
previous findings, the overall data show that women are less likely than men to use the
Internet or a mobile phone. In Brazil, for example, the region’s largest country with over
a third of the region’s GDP and population, being a women is negatively correlated with
using the Internet (with a correlation coefficient of r = -0.022), and with using a mobile
phone (with r = -0.029).
51

In the following rows of Table show two kinds of the inequalities between men and
women in percentage points. It is shown that there are real inequalities between men and
women regarding their working status (being employed or self-employed) and their
current attendance at an educational institution. Continuing with the example of Brazil,
92.2% of all men are actively working, compared to only 83.7% of women, and 31.6% of
all men currently attend an educational establishments, compared to 30.8% of all women.
These differences do not seem to be very large, but let us see what happens when we
control for them. This can be done with a partial correlation, which measures the degree
of association between two variables when the effects of a third variable are removed (see

50
Pearson’s r is normalized between +1 and -1, where +1 is a perfect positive association and -1 is a perfect
negative association. A correlation near zero indicates that there is no relationship between the two
variables.
51
A second look at the data reveals that the identified correlations are not very strong (even though they are
all statistically significant, weighted and stratified samples with p < .001). Squaring Pearson’s correlation
coefficient tells us how much of the variation in ICT usage can be explained by variation in being male or
female. In all cases, less than 0.5% of the variation in ICT usage can be explained by gender (for example,
in Chile r
2
= -0.047
2
= 0.2%). There must be much more powerful explanatory variables that determine ICT
usage than gender.
MAPPING OUT THE TRANSITION 127
e.g. Williams and Monge, 2001). To be more precise: what is the relationship between
being a women and ICT usage when the effect of work or current schooling is removed?
Table shows that in the controlled environment, being a woman is positively
correlated with using the Internet (for Brazil r = +0.056) and with using a mobile phone
(for Brazil r = +0.033). While this correlation is still very low, it is striking that this turn-
around effect is consistent throughout almost all analyzed countries, which represent a
very heterogeneous group of socio-demographic and cultural societies. There are some
countries in which women are more active Internet users to begin with (Panama,
Honduras, Nicaragua) or more active mobile phone users (Dominican Republic, Panama,
Nicaragua), which of course naturally argues in favor of women being more active ICT
users to begin with. There are also cases in which women are not discriminated in the
fields of employment status (Nicaragua) or current attendance at an educational
establishment (Panama and Ecuador). This does not affect the logic of our result. The
overwhelming majority of the cases show that, when controlling for working and
educational enrollment conditions, women make more use of digital ICT than men. The
only exception in the 20 changes in tendency that can be observed in Table is mobile
phone usage in Ecuador: the correlation coefficient becomes weaker in the controlled
test, but continues to stay negative (r = -0.037). This reminds us of the fact that social
science is not an exact science.

MAPPING OUT THE TRANSITION 128
Table 6 Correlations and controlled correlations of gender with ICT usage in Latin
America; working and studying populations by men and women.

Chile 2006
Brazil 2005
Uruguay
2006
Mexico 2007
Paraguay
2007
El Salvador
2006
Costa Rica
2005
Dominican
Rep. 2005
Panama 2007
Honduras
2007
Nicaragua
2006
Ecuador 2006
Correlation
coefficient, r
Internet use with
being a woman
-.047 -.022 -.024 -.045 -
.002/
-.023 -.032 -.033 .020 .008 .004 -.026
Mobile use with
being a woman
-.004 -.029 n.a. n.a. n.a. n.a. -.044 .009 .011 -.029 .013 -.070

Real world
inequalities
(in %)
Actively
working
% of
Men
98.0 92.2 98.7 95.1 88.0 89.3 98.3 93.7 95.8 87.3 88.3 86.5
% of
Wome
n
98.0 83.7 96.3 90.4 74.0 79.3 96.5 78.0 85.0 81.3 91.7 64.8
Attending
educational
establishment
% of
Men
31.1 31.6 29.6 33.0 44.2 34.5 35.0 34.2 28.6 31.4 34.7 10.8
% of
Wome
n
28.2 30.8 27.6 29.3 44.1 30.3 34.2 34.3 29.9 31.2 33.1 11.3

Correlation
coefficient, r:
being a
woman,
controlled for
working and
assisting educ.
establishment
Internet use .050 .056 .047 .048 .088 .030 .047 .082 .148 .093 .066 .007
Mobile use .039 .033 n.a. n.a. n.a. n.a. .019 .094 .139 .069 .085 -.037

Source: own elaboration, based on OSILAC, 2009.

Let us dig deeper into this question and open up this statistical black box to see
what actually accounts for these results. The first two rows of Table show the actual
percentages that lead to the previous finding that in most countries more men than
women use the Internet. Continuing with our example of Brazil, 22.0% of men use the
Internet compared to 20.2% of women. In mobile phone usage the divide is at 38.5% to
35.4%. In agreement with the results from Table, the notable exceptions for Internet
usage are Panama, Honduras and Nicaragua, and for cell phone users, Dominican
MAPPING OUT THE TRANSITION 129
Republic, Panama and Nicaragua. The following rows show what happens if we put men
and women on “equal footing” regarding their working condition. We only consider men
and women who are either employed or self-employed, neglecting those who are
unemployed, retired or stay at home without salary. Based on this condition, it turns out
that in all countries more women than men use ICT actively, again with the sole
exception of mobile phone usage in Ecuador. In Brazil, only 22.8% of all working men
use the Internet, while 28.5% of all working women are online. Only 47.0% of all
Brazilian working men use a mobile phone, while 50.6% of all working women
telecommunicate on the go. The same general change in direction accounts for ICT when
controlled for current attendance at an educational establishment. Once in school, women
turn out to be more active users of digital opportunities (35.6% to 36.2% for Internet use
in Brazil; 32.5% to 39.9% for mobile usage). The exceptions to the general rule for
Internet usage are again Ecuador, as well as Costa Rica and Dominican Republic.
Generally speaking, the differences are much more pronounced for mobile phone usage
than for Internet use. Once set on equal footing in terms of employment and education,
women seem to embrace mobile voice communication quite a bit more enthusiastically
than men.

MAPPING OUT THE TRANSITION 130
Table 7 Percentage of man/women that use the Internet and own a mobile phone in
Latin America; place of Internet usage, Internet use frequency.

Chile 2006
Brazil 2005
Uruguay
2006
Mexico 2007
Paraguay
2007
El Salvador
2006
Costa Rica
2005
Dominican
Rep. 2005
Panama 2007
Honduras
2007
Nicaragua
2006
Ecuador 2006
Overall
ICT
inequalities
Internet
use
Men 39.6 22.0 30.5 24.3 11.3 5.6 23.4 17.1 22.8 9.7 11.6 7.9
Women 35.1 20.2 28.3 20.6 11.1 4.6 20.8 14.7 24.5 10.2 11.9 6.6
Mobile use Men 54.2 38.2 n.a. n.a. n.a. n.a 35.0 56.6 45.0 26.1 40.9 41.6
Women 53.8 35.4 n.a. n.a. n.a. n.a 30.9 57.4 46.1 23.6 42.1 34.8

Men and
women
actively
working
Internet
use
Men 31.0 22.8 30.9 21.0 10.6 4.6 24.0 17.0 21.1 8.9 10.7 7.3-
Women 36.6 28.5 37.3 25.8 15.1 5.7 30.4 24.7 37.2 15.4 15.1 8.2-
Mobile
phone
Men 68.9 47.0 n.a. n.a n.a. n.a. 44.2 62.5 53.7 41.3 49.2 49.0
Women 73.1 50.6 n.a. n.a n.a. n.a. 46.8 71.9 68.3 47.6 56.4 45.2
Men and
women
attending
educational
establishment
Internet
use
Men 70.2 35.6 49.8 39.2 19.9 11.5 35.8 32.6 42.4 16.9 19.1 26.8
Women 70.3 36.2 53.1 41.2 23.8 12.6 35.5 29.6 48.5 18.8 22.0 26.1
Mobile use Men 39.1 32.5 n.a. n.a. n.a. n.a. 29.4 60.7 37.4 13.6 37.7 58.9
Women 44.2 39.9 n.a. n.a. n.a. n.a. 33.6 66.8 49.7 17.0 41.8 60.5

Place of Internet usage, given that
the person is actively working
At home
Men 14.3 11.7 13.6 7.9 3.3 1.4 7.9 3.8 6.7 1.9 0.7 n.a.
Women 16.4 14.2 16.4 10.2 4.4 1.5 9.4 5.5 11.5 2.9 0.8 n.a.
At work
Men 14.5 14.8 15.4 9.8 4.6 2.0 12.3 7.8 9.8 3.8 4.3 n.a.
Women 18.1 17.8 18.3 12.9 5.1 2.8 15.5 13.2 19.9 6.3 6.6 n.a.
Communal
public
access
Men 0.3 1.6 0.5 0.8 n.a. 0.1 0.1 1.7 0.7 0.01 0.0 n.a.
Women 0.3 2.3 0.8 0.6 n.a. 0.01 0.1 1.2 1.3 0.01 0.1 n.a.
Commerci
al public
access
Men 6.9 4.2 12.3 7.0 3.5 0.9 9.3 7.1 6.9 5.7 5.8 n.a.
Women 7.6 4.3 14.9 6.9 5.8 1.0 12.1 7.5 10.3 10.1 7.7 n.a.
Other
person’s
home
Men n.a. 6.2 2.6 0.2 0.2 0.01 1.1 4.9 1.1 n.a. 0.2 n.a.
Women n.a. 7.4 2.8 0.4 0.5 0.01 1.2 5.7 1.5 n.a. 0.1 n.a.

Source: own elaboration, based on OSILAC, 2009.

These results seem to indicate that women are the more enthusiastic ICT users.
However, one could argue that this tendency originates in the fact that women are more
likely to be forced to use computers at work for unsophisticated and repetitive secretarial
tasks (e.g. Kaplan, 1994). In this case, force, not enthusiasm would be the reason for our
results. Saying it very bluntly, the argument would be that men at construction sites do
MAPPING OUT THE TRANSITION 131
not need Internet access, while female secretaries are forced by their employers to
execute trivial typing jobs and routine office activities, such as banal word processing and
spreadsheet work. This general tendency could also affect ICT-enthusiasm in school,
since girls and boys often already anticipate their future job. While this sounds like a
possible hypothesis, this argument cannot explain the detected differences in mobile
phone usage. Besides, as shown by the lower rows in Table, working women do not only
access the Internet when forced to do so by their employers in their working
environment, but women are also more active online users at home, at public access
centers or commercial cyber cafes, and even at other people’s homes. Continuing with
our example of Brazil, Table confirms that more working women use the Internet at their
job than men (14.8% to 17.8%), but at the same time more women also go online at home
(11.7% to 14.2%) at a communal access center (1.6% to 2.3%), a commercial public
access center (4.2% to 4.3%) or at the home of family and friends (6.2% to 7.4%). While
ICT access at work might still have a catalyzing role, it can be seen that working women
also make use of their digital skill outside the working environment. Rather than being
forced to ICT usage against their will by an external force, it seems that women naturally
enjoy the use of digital communication wherever they get the opportunity to do so.

Women and ICT in Africa
Let us now compare uncontrolled and controlled usage rates in Africa (see Table).
Sample sizes are much smaller in these surveys and ICT usage rates in Africa are lower,
making it more difficult to detect differences. While all results of the weighted samples
turn out to be significant (weighted stratified samples with p < .01), they are less robust
MAPPING OUT THE TRANSITION 132
than the ones from Latin America. Having said this, the general tendencies are the same.
In agreement with the traditional findings of literature, the overall correlation between
gender and ICT usage shows that in 11 of the 13 countries, a larger percentage of men
use the Internet than women (with the exception of Rwanda and Tanzania, in which
women already represent the larger share). In Kenya, for example, one of the larger and
technologically most advanced African countries of our sample, 21.1% of all men have
been online in 2007/8, while only 11.5% of all women use the Internet. 56.0% of all men
use a mobile phone, versus only 46.9% of all women.
Notwithstanding, the following rows of Table show that, in general, African women
are also less literate
52
(in Kenya 77.2% of men to 68.0% of women), and that fewer
women are actively working or studying (employed, self-employed or full-time student)
(81.4% of Kenyan men to 49.9% of women). Women also have less income (29.8% of all
Kenyan men belong to the top 25% income group of the country, while only 16.6% of all
women do).
On the basis of these characteristics, a new group was created. We will refer to it as
“women on equal footing”, simply for the sake of giving it a name. In this group we only
consider men and women who are literate, are actively working or studying and who
belong to the top 25% income group
53
. Controlling for these three inequalities, we can
see that the gender divide disappears in most African countries for women “on equal
footing”. In the case of Kenya, the divide in Internet usage is erased at 29.7% for both

52
Literacy was defined by including all respondents that claimed to be able to read the newspaper easily
and to write a letter easily.
53
In the case of Africa it is necessary to focus on this high-income group of the top-25%, since income
levels in general are relatively low (in absolute terms) and ICT are tradable goods with prices levels that are
only accessible to segments that reach a certain absolute level of income (see Hilbert, 2010).
MAPPING OUT THE TRANSITION 133
men and women, while women on equal footing turn out to be more active mobile phone
users (90.0% to 92.7%). When placed on equal footing, the ratio of women versus men
turns around for Internet usage in four of the 13 analyzed countries (Namibia, Ethiopia,
Mozambique, Senegal). For another six countries, men continue to use the Internet more,
but the relative difference diminishes in all cases (South Africa, Benin, Botswana, Ghana,
Uganda, Cote d’Ivoire). For example, in South Africa, in the uncontrolled environment,
the share of men online is almost twice as large (20.2% of men to 11.3% of women),
while it shrinks to a difference of merely five percent for men and women on equal
footing (39.9/37.7 = 1.05).
This observed change in tendency is again much more pronounced for mobile
phone usage. In nine of the 13 countries, these controls turn the inequality around. With
the exception of Senegal and Tanzania, women on equal footing tend to embrace mobile
telephony more than men.

MAPPING OUT THE TRANSITION 134
Table 8 Percentage of man/women that use the Internet and own a mobile phone;
literacy, working and income inequalities; in Africa 2007/08.

Kenya
Namibia
Ethiopia
Rwanda
Mozambique
Senegal
Tanzania
South Africa
Benin
Botswana
Ghana
Uganda
Cote d’Ivoire
Overall ICT
inequalities
Internet
use
Men 21.1 11.2 0.9 1.8 1.1 14.4 1.9 20.2 11.9 8.1 7.9 10.1 3.7
Women 11.5 7.2 0.4 2.1 0.9 6.7 2.3 11.3 5.3 4.0 3.2 4.0 1.1
Mobile
use
Men 56.0 53.3 3.7 11.8 21.9 55.1 26.2 56.3 37.9 42.5 60.7 59.4 26.3
Women 46.9 45.4 2.5 7.5 32.4 26.2 17.6 64.9 20.5 37.9 57.2 58.9 12.2

Real world inequalities
Literate Men 77.2 56.2 33.5 47.1 38.9 33.4 72.5 75.9 42.0 45.1 49.1 77.5 49.9
Women 68.0 58.5 26.5 40.0 23.5 19.9 67.7 74.4 21.2 38.0 39.8 72.5 25.2
Actively
working/
student
Men 81.4 58.3 93.0 79.4 86.6 84.4 77.5 66.4 94.0 86.7 89.0 64.8 87.5
Women 49.8 42.5 32.3 62.8 35.8 53.6 53.9 38.5 49.7 51.6 80.7 43.4 47.0
Top 25%
income
Men 29.8 34.2 50.1 30.5 27.2 44.4 39.2 37.0 38.3 32.6 31.3 34.5 32.2
Women 16.6 17.9 10.7 21.0 15.2 9.3 20.9 17.4 14.5 9.0 19.0 21.1 10.6
Equal
footing
Men 25.3 21.4 13.9 16.5 13.8 18.1 25.1 28.5 18.1 14.8 15.0 31.1 21.7
Women 13.6 12.2 4.4 8.0 2.7 3.0 10.5 12.8 3.8 3.9 8.1 17.8 6.3

E
Man and
women on equal
footing Internet
use
Men 29.7 26.9 3.8 6.2 2.8 31.0 4.0 39.9 27.6 23.9 26.7 23.9 13.5
Women 29.7 37.8 6.3 7.6 14.2 37.4 12.6 37.9 26.9 17.5 11.7 17.5 7.4
Mobile
phone
Men 90.0 90.3 18.7 39.9 57.7 91.1 56.4 89.9 86.9 84.4 82.3 84.4 62.6
Women 92.7 93.1 34.3 43.4 92.5 87.9 47.8 94.8 95.9 94.9 94.7 94.9 71.4

Source: own elaboration, based on Research ICT Africa, 2008.

How do men and women use the Internet
As seen during the literature review, studies from developed countries reported that
men and women use the Internet for different ends, which can lead to diverse definition
of the digital divide. Let us now take a look at the kind of online services used in Latin
America. This is of particular interest because we have already seen that women welcome
the use of digital tools; therefore, the kinds of services they use might give us hints about
possible digital opportunities for women (Table). The first row confirms the previously
MAPPING OUT THE TRANSITION 135
mentioned finding from developed countries that men seem to be more frequent online
users than women. This accounts for Chile, Uruguay, Costa Rica, Dominican Republic
and Nicaragua, while in Mexico and Honduras more women tend to be online every day.
When evaluating these statistics, we have to remember that usage frequency does not tell
us anything about the length of each session. Longer sessions could by far offset lower
frequency. Unfortunately the available statistics do not give us insight into the overall
intensity of usage.
When asked about the kinds of services used online, men reveal that they are much
more enthusiastic about using the Internet for entertainment reasons than women. When
it comes to using digital channels for education and training, the data is clear that women
tend to make much better use of the existing opportunities than men. This is especially
encouraging when considering the previously presented results of female disadvantages
in terms of literacy and educational attendance throughout the developing world (see
Tables ). It shows that women already started to make use of the digital opportunities to
fight those existing inequalities.
Table also shows that women still do not yet fully exploit many of the other
opportunities the digital world provides for them. Women are less enthusiastic about
applications of e-business and e-government. The use of e-business and online banking
channels could provide women with important steps to improve their financial
independence, while e-government services facilitate necessary, but often burdensome
interactions with public authorities. The use of the Internet for plain communication
purposes provides a mixed picture, as do the statistics on health services. Women from
Mexico and Dominican Republic are already using online networks to improve the health
MAPPING OUT THE TRANSITION 136
conditions for themselves and those close to them. Overall, there still seems to be a large
potential to take advantage (or maybe create) adequate online content to improve living
conditions for women in Latin America.

Table 9 Frequency of Internet usage; online service used by men and women in
Latin America.

Chile 2006
Brazil 2005
Uruguay
2006
Mexico 2007
Paraguay
2007
El Salvador
2006
Costa Rica
2005
Dominican
Rep. 2005
Panama 2007
Honduras
2007
Nicaragua
2006
Internet users that use daily Men 39.5 n.a. 25.2 34.3 n.a. n.a. 35.8 36.7 n.a. 33.8 35.6
Women 34.3 n.a. 22.6 34.5 n.a. n.a. 33.1 30.1 n.a. 34.2 29.0

Given that the person uses the Internet
Entertainment
Men 54.7 74.1 49.5 19.9 11.0 5.7 51.5 60.7 4.9 41.7 61.8
Women 50.5 67.2 34.5 14.2 4.9 1.9 43.2 51.1 1.9 32.8 56.1
Education and
training
Men 12.1 68.4 41.4 41.5 39.7 53.7 58.5 67.6 1.3 60.9 58.7
Women 12.4 75.0 46.5 44.9 49.3 65.0 66.8 72.5 1.8 63.0 62.3
Buying and
contracting
Men 7.3 16.5 5.8 7.2 2.3 3.7 9.9 10.7 1.4 5.0 3.5
Women 5.5 10.8 2.6 3.3 1.2 2.5 5.7 5.7 1.2 3.3 2.2
Online banking
Men 7.1 21.7 4.6 2.4 n.a. 2.2 21.8 14.8 0.8 n.a. 5.3
Women 5.6 16.4 3.0 1.2 n.a. 2.2 17.3 11.4 0.9 n.a. 4.3
Government
interaction
Men 9.9 29.4 n.a. 3.4 n.a. 0.7 n.a. 13.2 0.5 n.a. n.a.
Women 8.9 25.5 n.a. 2.3 n.a. 0.01 n.a. 9.2 0.2 n.a. n.a.
Communication
Men 58.8 68.8 79.0 48.4 51.8 18.1 73.5 63.1 17.9 69.6 77.8
Women 60.2 68.5 81.1 49.5 55.7 13.9 74.3 55.0 18.4 71.5 77.6
Health
Men n.a. n.a. n.a. 6.3 1.8 1.8 n.a. 18.7 n.a. n.a. n.a.
Women n.a. n.a. n.a. 8.9 0.9 1.5 n.a. 25.1 n.a. n.a. n.a.

Source: own elaboration, based on OSILAC, 2009.

A word of caution on the presented statistics
The humorist Mark Twain (1835-1910) has popularized the wisdom that “there are
three kinds of lies: lies, damned lies, and statistics”. This does not only apply to the
statistical practices of not controlling for confounding variables, such as criticized in this
MAPPING OUT THE TRANSITION 137
article, but a word of caution is also in order when interpreting the statistics in the
presented Tables.
Given the large sample sizes of the weighted household samples, all results turn out
to be statistically significant. However, all results are based on stratified samples,
meaning that the survey organizers took the sample according to their knowledge about
how the population is distributed in a particular country. Once collected, stratified
samples are weighted according to the proportions of the actual society. The answers of
one survey correspondent from a more common socio-demographic group might be
multiplied with a factor much larger than a person from a minority. This weighting turns
a small sample into the representative of a large population. Nevertheless, it also affects
significance tests. The theory of significance test is based on random sampling, not on
stratified samples that are subsequently expanded. If the weighted number of cases
exceeds the sample size, tests of significance tend to be inflated, which is our case.
Therefore, even though all our results are statistically significant, meaning that it is very
probable that the observed differences between men and women are real and not just due
to chance, these tests are inflated.
As a consequence, results that are very close (such as 49.5% to 50.5% or a
correlation of 0.008), have to be taken with a large grain of salt, as pure luck of sample
drawing might play us a trick here. Given the much smaller sample size in Africa than in
Latin America, the Latin American results are more reliable and stable than the African
surveys (in the Latin American samples each observation was weighted with hundreds of
people on average, while in Africa, factors of thousands were applied). Having said this,
differences for one country on the decimal level surely would not make a strong case by
MAPPING OUT THE TRANSITION 138
itself. However, the consistency of our results across a large number of very
heterogeneous societies makes a relatively strong case: even though some results are
close and might be influenced by chance, they tend to show the same direction in general.
In other words, the presented results should be interpreted as a mutually confirming
whole, while specific results for particular countries might be subject to small variations.
Particularly close results from one specific country should not be used as a standalone
argument and might require more detailed sampling and further analysis.

A small change in mindset can sometimes make a large difference
We have analyzed a very heterogeneous group of 25 countries, representing
different levels of development, geography, culture and social structure. According to the
United Nations Human Development Index (UNDP, 2008), Chile is the most developed
country of our sample, reaching a rank 40 of the 179 countries included in the 2008
Index. Mozambique is the least developed with a rank of 175. Independent from these
differences, our results have been surprisingly consistent: ICT per se does not have
anything on them that might keep women and girls from using it in developing countries.
In fact, when controlled for existing inequalities, it shows that women embrace digital
technology more enthusiastically than men. One might be tempted to speculate that
women are simply better communicators and that therefore the use of these technologies
seems more intuitive for women than for men. Unfortunately, the presented data do not
tell us why women use ICT more than men; they just tell us that this is the case.
Notwithstanding, women continue to be discriminated in many other aspects of
social life, including employment, literacy and income. These inequalities also throw
MAPPING OUT THE TRANSITION 139
their shadows on ICT usage. More specifically, being a woman is positively correlated
with ICT usage, and negatively correlated with employment, income and education (see
Tables). Uncontrolled correlations mix both effects, resulting in the fact that
underemployed, underpaid and undereducated women use ICT less than men. Traditional
discrimination in the fields of employment, income and education turn the positive
correlation between women and ICT into a negative one. At the same time, as shown
during the literature review, ICT have the potential to provide access to employment,
education and income. Therefore, ICT provide women with a bootstrapping opportunity
to pull themselves out of these unfavorable starting conditions. In other words, if woman
are provided with ICT, digital tools represent an opportunity for women to fight
longstanding inequalities.
The resulting logic is schematized in Figure. Traditionally, longstanding
inequalities prevent women from accessing ICT, leading to a vicious circle between
digital exclusion, unemployment, low income and lacking education. However, once
having access to ICT, this vicious circle can be turned into a virtuous circle, whereas the
identified positive attitudes of women toward ICT enable them to circumvent and fight
existing inequalities.

MAPPING OUT THE TRANSITION 140
Figure 11 Fighting longstanding discrimination with digital means

Source: author’s own elaboration.

This finding is by no means the end, but leads to the question of how to provide
more women with access to digital opportunities. For example, Table indicates that
communal or commercial public access centers might be a viable option (see also Maeso
and Hilbert, 2006). Others have pointed to the need of regulations and incentives to
facilitate the actual usage of applications that would favor women, such as legislation to
promote telework (see e.g. Boiarov, 2008). Besides, the development of adequate content
becomes a major concern, especially in key areas such as education (see e.g. RELPE,
2008).
Summing up, the empirical evidence in this article argues for a re-thinking about
women and ICT usage. This rethinking should also affect policy making, which is
unfortunately still influenced by the superficial and unsustainable argument that women
are technophobic. For example, in the final declarations of the United Nations World
negative
overall
correlation
ICT
WOMEN
who have
less less less
employment income education
ICT
WOMEN
who have
less less less
employment income education
WOMEN
employment income education
ICT
positive
positive
negative negative negative
positive
MAPPING OUT THE TRANSITION 141
Summit on the Information Society (2003-2005), heads of States and governments have
recognized “that a gender divide exists as part of the digital divide in society” (WSIS,
2005) and declared a need for “enhancing communication and media literacy for women
with a view to building the capacity of girls and women to understand and to develop ICT
content” (WSIS, 2003). These statements seem to be based on the idea that women are
less digitally capable. Based on the results here presented, this is not at all the case and a
change in mindset seems appropriate. These policy statements should rather be
reformulated to something along the following lines: “a digital gender divide exists only
as a direct reflection of existing gender-related inequalities and policy actions should
make use of the natural communication skills and media capacities of women and their
proven embrace of the new digital opportunities to overcome longstanding gender
inequalities”. Such re-thinking is necessary to create policies and projects that truly allow
girls and women to become equal members of an information society, digital society,
network society, knowledge society, or simply equal members of society, independent
from the forename it may be given.

MAPPING OUT THE TRANSITION 142
Chapter Three: Magnitude and Growth of the Transition

Studies about digital social transformations, such as the ones from the previous
Chapter, traditionally use proxies to measure the advancement of the ongoing digitization
of information and communication processes in society, such as a headcount of the
installed technological devices, or the amount of financial investments into these
technologies. As with all proxies, the conclusions that can be drawn from these indicators
are flawed, and, as it turns out, in the case of the transition toward information societies,
these flaws can be severe. This Chapter presents a methodology to quantify the amount of
information and communication in bits, therefore directly measuring the magnitude and
growth of information in information societies. I applied this methodology by measuring
the world’s technological capacity to store, communicate, and compute information. This
line of inquiry started with an application of this idea to a selected group of countries,
with a selected list of technologies, studying their diverging informational capacities. A
follow-up study was much more comprehensive and complete, and covered more than 60
technologies, for more than 20 years for the entire world.

MAPPING OUT THE TRANSITION 143
Information Societies or “ICT Equipment Societies”?
Measuring the Digital Information Processing Capacity of a
Society in Bits and Bytes
54,55

The digital divide is conventionally measured in terms of ICT equipment diffusion,
which comes down to counting the number of computers or phones, among others. This
article fine-tunes these approximations by estimating the amount of digital information
that is stored, communicated and computed by these devices. The installed stock of ICT
equipment in the consumer segment is multiplied with its corresponding technological
performance, resulting in the “installed technological capacity” for storage (in bits),
bandwidth (in bits per second) and computational power (in computations per second).
This leads to new insights. Despite the rapidly decreasing digital equipment divide, there
is an increasing gap in terms of information processing capacity. It is shown that in 1996
the average inhabitant of the industrialized countries of the OECD had a capacity of 49
kibps more than its counterpart from Latin America and the Caribbean. Ten years later,
this gap widened to 577 kibps per inhabitant. This innovative approach towards the
quantification of the digital divide leads to numerous new challenges for the research
agenda.

54
This article is published as Hilbert, M., López, P., & Vasquez, C. (2010). Information Societies or “ICT
equipment societies”? Measuring the digital information processing capacity of a society in bits and bytes.
The Information Society, 26(3). Retrieved from
http://www.tandfonline.com/doi/abs/10.1080/01972241003712199
55
The authors would like to thank the support of the @LIS project (Alliance for the Information Society) of
the European Commission and UN-ECLAC, as well as the encouraging comments from Manuel Castells,
François Bar and two supportive blind reviewers.
MAPPING OUT THE TRANSITION 144
The far-reaching and profound impact of the digitization of information and
communication processes has long been noted (e.g. Wiener, 1948; Machlup, 1962; Bell,
1973). It is widely recognized that the advancement of digital information and
communication technologies (ICT) has led to a new mode of development (e.g. Perez,
1983; Freeman and Louça, 2001). With the arrival of digital systems, the storage,
communication and computation of information became the omnipresent core of social
and political activity, and of economic and cultural production (e.g. Webster, 1995;
Castells, 1996). This has put the question of how to track and measure the diffusion and
eventual impacts of these new technologies on the centre stage.
This article is a contribution to this discussion. We propose to improve the measure
of traditional ICT access indicators by adjusting existing ICT equipment statistics with
the corresponding quality of their performance. The stock of available technologies is
multiplied with their corresponding performance measures. The result is three new
aggregate indicators which represent the “installed information processing capacity”: (1)
how much information can be stored (in bits), (2) communicated (in bits per second), and
(3) computed (in computations per second). This improvement contributes not only to the
sustainability of the traditional ICT indicators (new ICT equipments emerge faster than
indicators ever can), but it also consolidates the array of currently available indicators,
merging them into three straightforward measures.

Multiple dimensions of technology diffusion
As with previous innovations, the nature of the ICT diffusion process is
characterized by a well-known S-curve from centre-periphery, wherein the centre can be
MAPPING OUT THE TRANSITION 145
depicted as being more developed and the periphery as underdeveloped (Rogers, 1962).
As a result, technological revolutions create a divide between those who benefit from it
first and those reached by it later on. In the case of ICT diffusion patterns, the term
“digital divide” was coined to describe the reality where some can access and use digital
tools, while others are excluded from the ensuing opportunities (NTIA, 1995-2000; ITU,
1999; UNDP, 2001 ITU, 2009).
The increasing importance of ICT in the socio-economic development has lead to a
broad variety of proposals on how to adequately measure this process of diffusion, and
thereby, how to conceptualize the digital divide. The most straightforward measures
focus on a specific technological solution as a proxy of the vast bulk of digital
technologies (such as Internet access or telephones) and compare the amount of
equipment or services in different societies (international digital divide) or within
different social segments of one society (domestic digital divide). More complex
measures distinguish between three consecutive steps in the adoption of the technology:
ICT access, use and impact (OECD, 2002). Even though there might be a positive
relation between the amount of ICT equipment, its usage and its impact, one of them does
not automatically imply the next. The determinants of the divide can be assessed in each
stage of the adoption process and with regards to all of the diverse existing technologies,
or their combination.
For example, on the access level it has been shown that the same long established
determinants of socio-economic inequality also define the digital divide, including
income, education, geography, age, gender, and ethnicity, among others (e.g. Cullen,
2001; Norris, 2001; Hilbert and Katz, 2003). Moving on to the usage stage of technology
MAPPING OUT THE TRANSITION 146
adoption, the importance of computer skills and motivations has been emphasized (e.g.
van Dijk and Hacker, 2003; Mossberger, et.al., 2003; Shelly, et.al., 2004). The final
impact of the technology will ultimately be influenced by the purposeful application of
the installed equipment, often requiring the readjustment of the general modus-operandi
of the cultural and institutional setting, which leads to a complex dynamic of social
change (e.g. Warschauer, 2003; van Dijk, 2006). Depending on the definition and the
scope of the exercise, the results can be contradictory. Most typically, research that
focuses on the access dimension (diffusion of technological equipment) argues in favour
of a rapidly closing digital divide (e.g. Compaine, 2001; Howard, et.al, 2009), while
research focusing on skill-related usage and impact indicators claims that the divide is
still deepening (e.g. van Dijk, 2005; James, 2008).
In an attempt to create a coherent picture, various compound measures have been
created, so-called e-Readiness indexes, such as the ICT Development Index (ITU, 2009).
They integrate a number of variables into a single index (access indicators and others,
such as skills). The weight of each component of the index, as well as the chosen
statistics, differs among indices (see Barzilai-Nahon, 2006; Vehovar, 2006; Hanafizadeh,
2009). Minges (2005), who has personally designed some of these indices at the leading
United Nations agency ITU, has evaluated twelve of them
56
and reconfirmed the
predictable conclusion that—besides problems of transparency, data reliability and
subjectivity—the weight of each ingredient predetermines the result to a large extent.

56
These include the twelve most widespread indices on a global level: Composite index of technological
capabilities across countries (ArCo); Digital Access Index (DAI); Digital Opportunity Index (DOI);
Economist Intelligence Unit (EIU) e-readiness; Index of Knowledge Societies (IKS); Knowledge Economy
Index (KEI); Network Readiness Index (NRI); Orbicom Digital Divide Index; Technology Achievement
Index (TAI); UNCTAD Index of ICT Diffusion; UN PAN E-Readiness Index; World Bank ICT Index.
MAPPING OUT THE TRANSITION 147
This leads to the well-known problem of subjectivity in the creation of any kind of index
and therefore does not solve the problem of adequately measuring the divide, but rather
passes the buck on to the methodological level.
In short, the digital divide is one of the rare types of concept that flexibly adapts to
the meaning that the analyst decides to give it. This can lead to much confusion, or, at
least, to tedious semantic quarrels. Despite all differences, there is one feature that all of
these studies and indexes have in common: the inclusion of the access dimension, such as
the diffusion of telephony, computers and Internet, among others (mostly those
harmonized by ITU, 2007). Access might not be sufficient, but it is a necessary first step.
Without neglecting that the discussion of the digital divide can become much more
complex, we will focus on improving the measurement of this indispensable dimension.
At a later stage, the proposed measurement of ICT access could easily be integrated into
more complex modular methodologies and indexes that—additionally to access
measurements— might also include computer skills and cultural considerations, among
others. In the meantime, we will limit our focus to the improvement of ICT access
measures.
The article starts by reviewing the traditional measure of ICT access, which is
usually done by counting the number of existing devices. We then propose an analytical
framework for tackling the task of measuring the installed information processing
capacity of a society, defined as the capacity to store, communicate and compute
information with digital tools. This new framework is applied to one concrete example.
We decided to compare the private consumer segment of the industrialized OECD
(Organisation for Economic Co-operation and Development) with the one in Latin
MAPPING OUT THE TRANSITION 148
America and the Caribbean (LAC), as representatives for developed and developing
countries on both sides of the international digital divide. While the scope of this article
only allows for one concrete example, it is important to underline that the chosen
example is just one case out of many that could have been chosen. It represents the
international digital divide (neglecting domestic differences among population segments),
and—in agreement with the conventional thinking on the digital divide—the analysis is
restricted to the private consumer segment (this is mainly due to the lack of coherent
statistics beyond households at the time of writing). The selection of this particular
example should not prevent future studies from applying the general framework of this
article to analyse the domestic digital divide and to assess the “installed information
processing capacity” of enterprises, public or private organizations or government
agencies. The final section takes up the underlying methodological discussion, which is
again independent from the concrete example that has been discussed before. The
resulting differences between the traditional approach and the proposed approach are
discussed, as well as the limitations and remaining challenges on the research agenda.

The closing digital equipment divide
After analyzing patterns of ICT equipment diffusion, some policy-related reports
come to the delicate conclusion that the access dimension of the digital divide is closing
rapidly and that underserved sections of the population are in an unprecedented process
of catching-up (e.g. Compaine, 2001; ITU, 2006; UNCTAD, 2006; WEF-INSEAD, 2006;
ITU and UNCTAD, 2007; Howard, et.al, 2009). In particular, it is argued that the divide
diminishes rapidly as the markets in the developed countries get increasingly saturated.
MAPPING OUT THE TRANSITION 149
Table shows that ICT equipment penetration rates in the 30 industrialized countries of the
OECD (1 184 million inhabitants in 2006
57
) are relatively high. The numbers in Table
also show that growth rates have been much higher in the 37 developing countries of
Latin America and the Caribbean (LAC) (456 million inhabitants
58
). In accordance with
these indicators, the theory of the diffusion of innovation, and the previously cited
research, it can be expected that the typical S-shaped diffusion curve is starting to
diminish in the more developed countries, while LAC seem to be on the upward slope of
the S-curve. There seems to be an upper limit on the amount of equipment an individual
possesses, even if one person can posses several devices of the same sort. The table
shows that while in 1996 OECD countries had 8.1 times more mobile phones per hundred
inhabitants than LAC countries, in 2006 the gap was reduced to a multiplication factor of
1.6. With regard to Internet users, the catching-up has even been more impressive,
reducing the ratio between both groups of countries from 18.5 to 3.0 in ten years. Thus,
analyses on the basis of these indicators seem to suggest convergence with a rapidly
disappearing inequality in access to digital information.

57
Australia, Austria, Belgium, Canada, Czech Republic, Denmark, Finland, France, Germany, Greece,
Hungary (starting 1996), Iceland, Ireland, Italy, Japan, Korea (Rep.) (1996), Luxemburg, Mexico (joined
OECD in 1994, and is therefore considered a OECD member for the ten year time frame considered in the
graphs, and not as Latin America), Netherlands, New Zealand, Norway, Poland (1996), Portugal, Slovak
Republic (2000), Spain, Sweden, Switzerland, Turkey, United Kingdom, United States.
58
Antigua and Barbuda, Argentina, Aruba, Bahamas, Barbados, Belize, Bolivia, Brazil, Chile, Colombia,
Costa Rica, Cuba, Dominica, Dominican Rep., Ecuador, El Salvador, French Guiana, Grenada, Guatemala,
Guadeloupe, Guyana, Haiti, Honduras, Jamaica, Martinique, Neth. Antilles, Nicaragua, Panama, Paraguay,
Peru, Saint Kitts and Nevis, Saint Lucia, St. Vincent and the Grenadines, Suriname, Trinidad y Tobago,
Uruguay, Venezuela.
MAPPING OUT THE TRANSITION 150
Table 10 ICT equipment diffusion per 100 inhabitants in OECD and Latin America
and the Caribbean, 1996-2006

Source: ITU, World Telecommunications Database, 2007.

A resulting, but premature, policy conclusion of this analysis could be that public
policies, such as market regulation and public access incentive programs, would be less
and less necessary to close the access dimension of the gap. This seems to be emphasized
by the success story of mobile telephony, which is the consumer technology with the
fastest technological diffusion record in history. Competitive markets seem to take
telecommunications networks and related hardware and software solutions to everybody
around the globe, such as regularly pointed out by industry representatives (GSM
Association, 2006; Frost and Sullivan, 2006).
Such conclusions are based on the simple accounting of equipment to assess the
situation of access to the digital realm. One of the main limitations of the traditional
equipment analysis is that technological progress is not considered. There are qualitative
Technology per 100 inhabitants 1996 2006
Fixed phones
OECD 46.5 46.8
LAC 9.8 17.2
ratio OECD/LAC 4.7 2.7
Mobile phones
OECD 11.0 86.6
LAC 1.4 54.7
ratio OECD/LAC 8.1 1.6
Personal Computers
OECD 18.5 56.6
LAC 3.0 16.7
ratio OECD/LAC 6.2 3.4
Internet users
OECD 3.7 23.4
LAC 0.2 7.8
ratio OECD/LAC 18.5 3.0
Broad band subscribers (2000-
2006)
OECD 3.0 16.8
LAC 0.1 2.3
ratio OECD/LAC 30.0 7.3
MAPPING OUT THE TRANSITION 151
differences in access. These differences vary with the calendar year under consideration
and also the user segment. The importance is easy to see. For example, Internet users
with a 56 kbps modem connection are not able to access the multimedia content
broadband users are benefiting from. However, in a simplistic count, both would be
considered as generic “Internet user” (see Table). The same applies to other ICT and also
holds for difference within one society. One hard disk from 1995 is not equal one hard
disk from 2005. Older equipment is much less powerful. Besides technological progress
in time, there are also differences in performance of different technological gear in the
same year
59
. While most mobile phones that are bought by the poor enable short-
message-services (SMS) through a 14 kbps data communication, third and fourth
generation mobile phones provide wealthy members of the Information Society with
mobile videoconferencing capabilities of several hundred kbps. Even if both, rich and
poor, would have the same quantity of equipment (by equipment headcount), their real
“access to digital information” might be very unequal. The currently available statistics
(such as shown in Table) do not show this difference.
This problem is recognized by recent literature, for example through the emphasis
in broadband connectivity (for example NTIA, 2002-2004). The current solution is to
simply add additional indicators (such as broadband), which cannot easily be compared
with the previous indicator of dial-up Internet. The resulting grab bag of indicators can be
expected to become more confusing as ICT-convergence continues. The ongoing
substitution between various services renders many traditional indicators quickly
obsolete. Voice services can be transmitted with Voice-over-Internet-Protocol (VoIP)

59
The relevant statistics for this consideration are often more difficult to obtain than performance
adjustment according to year-related technological progress.
MAPPING OUT THE TRANSITION 152
over the Internet and WebPages can be accessed through mobile phones. Actually, the
traditional separation into the aforementioned technologies according to their hardware
and functionality (and not according to capacity) is not really helpful for understanding
and coping with the ongoing dynamics.
In order to obtain a deeper insight into current developments, we measure the total
sum of technological information processing capacity of diverse technological solutions.
There are two major benefits of considering the bits and bytes that can be processed by
the different solutions. One, it enables the consideration of technological progress in the
performance of the different generations of equipment. It therefore recognizes the digital
divide as a constantly moving target. Two, it permits harmonization of substituting
technologies on a common unit of measurement (if two service substitute each other, as
per definition of substitution, they provide the same performance measure, for example
bits per second, which are jointly aggregated to sum up to the installed capacity).
Three steps are necessary. One, ICT systems need to be classified according to their
informational functionality. The following section identifies three distinct groups of basic
information operations (communication, storage, computation). Two, adequate
measurement units for each of the three identified technological subsystems need to be
developed, which are discussed in the subsequent section. Three, the evolving
performance of each technology needs to be estimated for various years and respectively
multiplied with the available technological equipment. This will result in the installed
information processing capacity of a society.

MAPPING OUT THE TRANSITION 153
The three subsystems of information processing
ICT systems do not represent a single technology, but are the result of a
combination of symbiotic technological trajectories that converge into one larger
technological system. As already mentioned, some of them might be potential substitutes
(e.g. fixed and mobile voice-communication) and others serve different ends (e.g. hard
disks and telephones). To avoid such confusion let us return to a basic definition of what
is technology. Technology has been defined as patterns of solutions that are based on
selected principles derived from the natural sciences and are applied to confront a
specific question or promise (Dosi, 1988). Following this definition, ICT answers three
different questions: (1) how to store information in some deposit for later usage; (2) how
to convert and compute some kind of information in a meaningful manner into another
kind of information; and (3) how to transmit and communicate information from one
place to another. In order to adequately reflect existing technologies, we further subdivide
this last function and differentiate between “transmission”, which we define as being
unidirectional (only down-link, such as broadcast), and “communication”, which we
define to be bidirectional (up-link and down-link, such as telecommunication)
60
.
The scope of the technological system that is often loosely referred to as digital
technologies is defined by the use of the “bit”. It is based on the idea of representing and
manipulating information through its most basic code, the binary digit
61
. The binary
codification and processing of information has not only improved and amplified the

60
For conceptual reasons the separation is justified, because communication is expected to have different
socio-economic potential than mere transmission. For practical reasons this separation is necessary
because, in terms of bit-rates, broadcasting technologies transmit a much larger amount of data.
61
A common unit of storage is the byte, equal to 8 bits, that is, eight consecutive yes-no decisions resulting
in a decision tree with 2
8
= 256 possible combinations.
MAPPING OUT THE TRANSITION 154
performance of each technological subsystem, it also meant that for the first time all three
of them began to function according to a common logic: binary logic. This led to the
integration of the three different technological subsystems into one system, a process
colloquially referred to as ICT-convergence.
While the “frictionless” interconnection of storage, communication and computing
devices has manifold advantages and leads to increased complementarities among the
different tasks, the introduction of the bit has not changed the fact that each of the three
operations has a distinct end. Figure depicts the basic schematization on the basis of
which the following analysis is structured. ICT dynamics are the result of the interplay of
all three technological subsystems, unified by the paradigm of the bit, which defines the
scope of the technological system. Above and beyond the three technological information
operations, the human brain is our indispensable recipient of, and contributor to, this
dynamic process. We define the “information processing capacity” of an individual or a
society as a construct comprised of these three informational operations.

MAPPING OUT THE TRANSITION 155
Figure 12 Schematization of the three basic information processes

Source: own elaboration.

All three subsystems have experienced extraordinary growth rates in their
performance development during recent decades (see Table). The Table shows, for
example, that it would be deceiving to compare a hard disk from 1980 with another one
from 2005. Actually, one hard disk in 2005 would be equal to 792 hard disks from 1995
and 750 000 from 1980. Sustained annual growth rates of 56-76% over 25 years are
outstanding, which can be seen when compared to more common socio-economic rates of
change (annual economic growth rates are traditionally between 3-4%). In a field of such
0110001010011100010100010101
1010110111100101101000110101
0101010110101100010100111000
1010001010110101101111001011
0100011010101010101101001100
0101001110001010001010110101
1011110010110100011010101010
1011010011000101001110001010
0010101101011011110010110100
0110101010101011010110001010
0111000101000101011010110111
1001011010001101010101010110
1001100010100111000101000101
0110101101111001011010001101
0101010101101011000101001110
0010100010101101011011110010
1101000110101010101011010110
0010100111000101000101011010
1101111001011010001101010101
0101101001100010100111000101
Storage
Capture & interoperation
Transmission /
Communication
Computation
Biologic
processing
01100010100111000
10100010101101011
01111001011010001
10101010101011010
interface
interface
interface
interface
MAPPING OUT THE TRANSITION 156
rapid change, it is essential to consider this constantly moving performance frontier in the
measurement of the dynamics
62
.

Table 11 Price decline and performance increase in the technological frontier of all
three ICT subsystems, 1980-2005

Basic function/
representative of the
technological frontier
(US$ 2006)
1980 1995 2005
Compound
annual growth
rate between
1980-2005
Transmission
telecommunication
(kilobits / sec / US$)
0.0007
(Modem Apple II)
0.06
(US Robotics v.34
modems)
48
(WiMax)
56%
Storage (MB / US$) 0.0032
(hard disk 5MD
HD)
3.03
(hard disk
MC191AV)
2400
(hard disk 320GB,
7,200 rpm, 8MB)
72%
Computation (millions of
computations / sec / US$)
7 x 10
3

(IBM4341)
1 x 10
8

(Dell Dimension
XPS P133c)
1 x 10
10

(Precision
Workstation690)
76%
Source: own elaboration.

The amount of digital information
The amount of digital information that can be processed by the available
technologies is calculated by multiplying the amount of equipment with its respective
performance, i.e. with the amount of bits that each equipment can store, the amount of
kbps it can communicate, and the amount of MCps is can compute. This approach is
inspired by two groundbreaking studies done by the School of Information at the
University of California, Berkeley in 2000 and 2003, that gauge the quantity of
information that exists worldwide (Lyman, Varian and Swearingen, 2003). Working with
proxies and assumption is unavoidable when working in this new field. Therefore we

62
A second look on Table 2 reveals that the technological frontier in each subsystem has advanced at a
different pace. It is interesting to note that the advancement of telecommunications, which is often
celebrated as the epitome of the networked revolution, shows the slowest technological progress, when
measured in terms of its price/performance relation.
MAPPING OUT THE TRANSITION 157
have taken great care to be transparent with our estimations, enabling replication and
improvements in the future. Our methodological details are presented in the Appendix.
The decision of the unit of measurement of information transmission and storage is
straightforward: the BI-nary digi-T. The encoded bits represent information, which can
reduce uncertainty with regard to a specified probability space
63
. From an engineering
perspective, transmission and storage are conceptually similar: one transports information
through space (bits per second through a transmission channel) and the other one
transports information through time (bits on a storage device). Storage will be measured
in bits and transmission in bits per second, or, to be more precise, in kbits and kibps
(kibibit per second, equal to [2^10]=1024; while a kilobit per second remains be equal to
[10^3]=1000).
64

Computers also function according to binary logic (manipulating bits through
Boolean logic gates). Unfortunately for us, the amounts of bits [1s and 0s] that are
manipulated per second do not provide any interesting performance indicators. A
computer with the universal design of a Turing machine consists of different information
operations, such as reading from and writing on different storage devices and the speed of
computation depends on the chosen architecture of the system. The diverse functionality
of computers leads to a large variety of quantitative approaches to performance

63
The binary code is a kind of alphabet with only two letters that can represent all other kinds of alphabets.
In the best of cases (in which a bit represents information entropy), one bit of information can reduce
uncertainty by half (Shannon, 1948). Every bit represents information and has the potential to reduce
uncertainty. In this technical definition of information, uncertainty and information are seen as opposites
and the reduction of a possibility space by half is the most efficient way to communicate information.
64
To solve the longstanding ambiguity regarding the units of kilobit in storage (one kilo being traditionally
equal to 1024 bits) and communication technologies (one kilo being traditionally equal to 1000 bits), the
latter can be measured in kibibit per second [Kibps] and mebibit per second [Mibps], which correspond to
1024 bits and 1024
2
bits per second, respectively.
MAPPING OUT THE TRANSITION 158
measurement (Hennessy and Patterson, 2007). For pragmatic reasons, we refer to the
historic data produced by Nordhaus (2006), which are mainly based on MacCallum
(2003). The resulting index is called CPS (computations per second) and is oriented by
the instructions per second a computer executes. In agreement with industry standards, it
is calibrated on the computer Digital Equipment Corporation VAX 11/780 from the year
1978 and correlated to the millions of instructions per second that a computer can execute
(MIPS) (see Appendix). The VAX 11/780 is considered to perform exactly 1 MIPS,
which, as a rough guide, is 150 million times as powerful as manual computations
65
.
It is important to remember that the number of bits does not consider the meaning
or value of the information content. As the father of information theory, Claude Shannon
(1948), points out: “Frequently the messages have meaning […but…] these semantic
aspects of communication are irrelevant to the engineering problem” (p. 379). From an
engineering perspective, a bit only gives a measure of how much uncertainty can
potentially be reduced with regard to a known possibility space (such as the selection of a
letter from an alphabet to construct words or the selection of a color to fill an image). It
does not reveal anything about the ‘meaning’ or ‘value’ of the information in the message
(in the sense that some words might be more important to the receiver than others).
Currently, there is no universally accepted scientific measure to classify the ‘meaningful
value’ of information. This is not tragic for our purposes, as we estimate the installed
information processing capacity to transmit and store information, independently of a
specific purpose. Going one step further, some might want to trust in the common
assumption that individual users are rational and self-interested actors, and would assume

65
Nordhaus (2006) defines that manual computation would imply that “you can add two five-digit numbers
in 7 seconds and multiply two five-digit numbers in 80 seconds” (p. 11).
MAPPING OUT THE TRANSITION 159
that they would utilize the provided technologies for ends that are useful and meaningful
to their specific ends. This additional step, however, is independent from our basic
exercise to estimate the available installed capacity in bits and bytes. Our estimations do
not differentiate among the meanings of information contents (by the way, the same also
holds for estimations that are based on equipment headcounts).
Having defined the measurement units, the two required statistics are the quantity of
ICT equipment and their respective performance. The first statistic is mainly extracted
from ITU’s World ICT indicators Database (2007), which is the world’s most complete
historical administrative registry for ICT. It receives its inputs from national
telecommunications and industry authorities. We estimated missing years and
complemented these data with information from mainly private sector sources, including
the assessments of the distribution of the various generations of the particular
technologies, such as the distinction between the share of existing mobile standards (such
as analogue, GSM, GPRS, CDMA2000, etc), the different television standards, and the
distribution of various existing hard disks according to their diameter, among others (see
Appendix). The historical performance of the various technologies has been gathered by
industry and academic sources, such as detailed in the Appendix.
The fact that we use national statistics as a basis for our calculations conceals the
fact that the digital information infrastructure is global in nature. If a user from one
country uses a hard disk in another country over an Internet connection, this international
outsourcing of informational capacity cannot be covered by our estimations. This lack of
coverage is not too damaging in the case of our specific example that estimates the
installed information processing capacity of the consumer segment. The amount of
MAPPING OUT THE TRANSITION 160
international infrastructure sharing, such as cloud and grid computing, is minimum in the
consumer segment. This would change, however, when applying the presented logic to
the broader economy, including businesses, universities and research center. Super-
computing facilities are often shared on the international level.
For reasons of simplicity and missing statistics, estimations focus on the installed
capacity, not on its real usage. In other words, it is assumed that the installed technology
would be running 24 hours for 365 days a year. As another general rule we decided that
estimations adopt an “optimistic bias” in favor of developing countries. This means that
in case of missing statistical information, it was assumed that the newly introduced
equipment performs at the technological frontier. This surely leads to an overestimation
of the installed capacity in all countries, as consumers might purchase older technology
from earlier years. Assuming that the technology consumed in developed countries is
generally closer to the technological frontier, this bias overestimates the installed capacity
in developing countries and is therefore “optimistic” from a development perspective.
The result is summarized in Table. It is to be understood as an optimistic estimation
of the worldwide installed capacity to communicate, transmit
60
, store and compute
information through digital systems. It shows that the personal capacity to compute and
store information has clearly experienced the largest progress. This is in agreement with
what we have already observed in Table, and is rather surprising
62
, as the advancement of
telecommunications is often celebrated as the epitome of the network revolution. The
plain numbers of Table question this generally accepted notion. This conceptual
concentration on communications – instead of computation or storage – is not only
prevalent in academic writing, but also in the area of policy making. In most countries,
MAPPING OUT THE TRANSITION 161
for example, the telecommunications authority is in charge of shaping the road toward the
Information Society, and private and public authorities of computer engineering often do
not even participate in policy agenda setting (for Latin America and the Caribbean see for
example Guerra, et.al., 2008). The United Nations World Summit on the Information
Society (2003-2005; e.g. Klein, 2004), as another example, has been organized by the
International Telecommunications Union, and its audience and the discussed topics have
been largely determined by this bias. Looking at Table, one can no longer say that
technological progress in communication technologies is the defining characteristic of the
Information Society. The table instead suggests that the storage of information in vast
memories and its meaningful computation are the principal characteristics of the
Information Society.
An interesting insight can be garnered by comparing the capacity of the different
subsystems. For example, the table allows for the following thought experiment: if
communication channels were running at full capacity and if every kind of communicated
information was original and saved as soon as it was received, then every user could have
filled the available per-capita storage capacity in roughly two weeks in 2006
66
. It shows,
however, that under the same assumptions
67
in 1990, the available storage capacity would
have been filled up completely in less than one and a half hours
68
. This shows how the

66
(299 951 493 kbits/inhabitant storage) divided by (224 kibps/inhabitant communication) = 1 339 069
seconds, which is 15.5 days.
67
In reality, not all information is, of course, original, and neither is all received information saved on a
hard disk right away. Therefore, the period between required erasures of memory are actually expected to
be much longer.
68
(56 438 kbits/inhabitant storage) divided by (12 kibps/inhabitant communication) = 4 703 seconds, which
is 1.31 hours.
MAPPING OUT THE TRANSITION 162
estimated global capacity to store information has increased much more remarkably than
the global capacity to communicate.

Table 12 Worldwide installed capacity to compute, communicate and store digital
information

1980 1990 2000 2006
Compound annual
growth rate
between 1980 and
2006
Communication (telephony and
Internet)
Kibps/inhabitant
9 12 34 224 13.2%
Transmission (radio and TV)
Kibps/inhabitant
2 653 4 403 7 230 8 143 4.4%
Computation (computers and
mobile devices)
MCps/inhabitant
0.0020 0.0958 63.15 957.74 65.4%
Storage (hard disks)
Kbits/inhabitant
9 475 56 438 14 501 988 299 951 493 49.0%

Source: own elaboration, based on various sources, see specifications in Appendix.

The digital divide as a moving target
The result of comparing the countries of the OECD with the countries of Latin
America and the Caribbean is shown in Figures. Figure represents the digital divide of
the capacity to communicate and exchange information through ICT, considering fixed
lines and mobile telephony, as well as Internet (including broadband). It shows that in
1996 the average inhabitant of the OECD counted with a capacity of 49 kibps (equal to
49*1024 bits per second, see Appendix) more than its counterpart from LAC (62 kibps
versus 13 kibps). Ten years later, this gap widened to 577 kibps (756 kibps as OECD
average versus 179 kibps as LAC average). It is important to point out that this
MAPPING OUT THE TRANSITION 163
development also represents a slight reduction of the digital divide in relative terms,
given that the ratio between OECD/LAC lowered from 4.7 to 4.3 (reduction to 91% of
original). However, this relative reduction is significantly smaller than the ratios
presented in Table (which show ratio reductions between 16-57%). Furthermore, in
contrast to the signs of saturation of the advanced OECD countries in ICT equipment
diffusion (see Table), Figure does not show any significant signs of saturation. The
amount of information that is communicated by the average member of the developed
region of OECD continues to grow explosively.

Figure 13 Capacity to communicate through fixed line, mobile telephony and
Internet

Source: own elaboration, based on various sources, see specifications in Appendix.

0
100
200
300
400
500
600
700
800
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
Kibps/capita .
OECD LAC
MAPPING OUT THE TRANSITION 164
Figure takes a closer look at the reasons for this result. The bulk of installed
communication capacity of a country is explained by fixed broadband Internet
connections, especially DSL, cable modem and fixed-wireless, such as WiFi
(representing 73% of the installed communication capacity in OECD and 61% in LAC in
2006). The capacity of the broadband Internet has surpassed the installed fixed line
capacity (including fixed-line telephony or alternatively dial-up Internet) in the OECD in
2000 and in LAC in 2003. An interesting insight points to the importance of mobile
telephony. In terms of equipment diffusion, the number of mobile phones equaled the
number of fixed lines in 2001 in the OECD and in 2002 in LAC (ITU, 2007). However,
this does not directly lead to a conclusion about the installed communication capacity
through fixed or mobile networks. One needs to consider that 2G mobile communications
(such as GSM and cdmaOne) provide a bandwidth of roughly 14 kibps, which only
allows very limited data services, such as SMS messaging. A fixed line opens up a
communication channel of up to 125 kibps, which can for example be used for dial-up
Internet. Therefore, in terms of communication capacity, a 2G mobile phone channel is
only a partial substitute for a fixed line in terms of data transmission rates. On the other
hand, 2.5G or 3G mobile communications allow up to 350 kibps (such as WCDMA). As
a result, the bulked communication capacity of mobile technology only surpassed fixed
line communication with the introduction of advanced mobile data services, such as
EDGE and CDMA2000. The breakeven point between fixed line and mobile
communication was delayed for two years in both regions (2003 in OECD and 2004 in
LAC). On the one hand, these cost-effective solutions also lead to the fact that mobile
communication is increasingly becoming important in developing countries: in 2006,
MAPPING OUT THE TRANSITION 165
mobile channels represented 28% of the communication capacity in LAC and only 19%
in the OECD (partly due to lack of fixed lines in developing countries). On the other
hand, while the number of mobile phones has started to slow down in the OECD during
recent years (with 86.6% of the population having a mobile phone in 2006), the amount
of information communicated through mobile networks in the OECD does not show any
sign of deceleration. The introduction of multimedia 3G and 4G communication
continues to push the capacity to communicate on the go, even though the number of
devices might not grow as fast anymore. These findings demonstrate that the analysis of
communication capacity can lead to different results and insights than the analysis of the
quantity of equipment.

Figure 14 Capacity to communicate according to technology

Source: own elaboration, based on various sources, see specifications in Appendix.
0
50
100
150
200
250
300
350
400
450
500
550
600
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
Kibps/capita .
Fixed Lines OECD Fixed Lines LAC
Mobile Communications OECD Mobile Communications LAC
Fixed Broadband Internet OECD Fixed Broadband Internet LAC
Fixed Broadband Internet
surpasses fixed line
Mobile communications
surpasses fixed line
MAPPING OUT THE TRANSITION 166

Figure shows the capacity to transmit and disseminate information through one-way
broadcasting channels, such as TV and radio. In 1996, the OECD had 8800 kibps more
than LAC (14 200 kibps versus 5 400 kibps). In 2006, this gap widened to 10900 kibps
(every OECD inhabitant on average had 17 800 kibps, versus 6900 kibps for every LAC
inhabitant). The massive diffusion of satellite and cable-TV in developed countries is
contributing to this widening of the gap. Actually, the data reveal that in 2006 around
62% of the OECD’s broadcast capacity was installed in high-quality cable and satellite
technology, while in LAC 71% was still transmitted through unreliable analogue
terrestrial TV systems. While the data show a relatively stable OECD/LAC ratio in
relative terms (around 2.6 throughout the decade), the absolute numbers disclose that the
6900 kibps/capita broadcast capacity of LAC in 2006 corresponds to the installed OECD
capacity of the year 1973. In other words, in terms of installed broadcast capacity per-
capita, LAC is 33 years behind the OECD. It can be expected that the introduction of
digital TV will very soon introduce a new dynamic in both regions.

MAPPING OUT THE TRANSITION 167
Figure 15 Capacity to transmit information through radio and TV (terrestrial,
satellite, cable)

Source: own elaboration, based on various sources, see specifications in Appendix.

A similar situation accounts for the storage of information in computer hard disk
drives (Figure). In 1996, an inhabitant of the OECD had on average 3 780 000 kilobits
more storage capacity in hard drives of PCs and laptops than its LAC counterparts (4 552
000 vs. 772 000). Ten years later, the advantage of the OECD increased to almost 750
000 000 kilobits per capita (1 090 000 000 vs. 341 160 000).

0
2,000
4,000
6,000
8,000
10,000
12,000
14,000
16,000
18,000
20,000
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005
Kibps/capita .
OECD LAC
6900 kibps: gap of 33 years
MAPPING OUT THE TRANSITION 168
Figure 16 Capacity to store information in hard disks of PCs and laptops

Source: own elaboration, based on various sources, see specifications in Appendix.

We have repeated this – and other – exercises with different methodological
assumptions. For example, in agreement with Jorgenson and Vu (2005) we have
estimated an economic utility lifetime of seven years for a computer and its hard disk
(this estimate is based on economic depreciation rates) (see Appendix). Changing this
assumption to five or three years (which might be close to the actual usage period of
usage, not its complete economic depreciation), the results do not significantly affect the
ratio between both regions
69
. It does, however, affect absolute storage capacity in both

69
The gap increases roughly 200-fold regardless of the seven years assumption
([750000000]/[3780000]=198); five years assumption ([878000000]/[4345500]=202) or three years
assumption of computer life-time ([1091000000]/[5560000]=196). We can conclude that changing the
lifetime of hard disk drives does not significantly affect the final results in terms of the ratio between the
differences of the storage capacity in OECD and Latin America, as those methodological changes affect
both regions in a similar manner.
0
200.000.000
400.000.000
600.000.000
800.000.000
1.000.000.000
1.200.000.000
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
kb/capita
OECD LAC
MAPPING OUT THE TRANSITION 169
regions. In the case of reducing lifetime, the installed equipment is updated to the
technological frontier more frequently, increasing the final storage capacity in 15-17 or
45-47 per cent with the five and three years assumptions, respectively. Methodological
considerations surely can make a difference in the absolute numbers, but our tests and re-
tests have shown that they do not change the general tendency and therefore the validity
of the arguments that are presented here.
Figure shows the capacity to compute information. For computers (PCs and
notebooks) we apply the same performance indicators to both regions. We include
computers (including Mac and PC), laptops and mobile phones (which have started to
possess considerable computing power). Regarding mobile phones, we use the available
statistics of 2G, 2.5G and 3G communication services to estimate the computational
power of mobile devices. As a result, we can see that in 1996 the OECD counted with 19
million computations per second per capita more than LAC, while in 2006 this gap
widened to 2 520 MCPS/capita
70
. It is interesting to observe the increasing importance of
computing capacity of mobile phones, which rely on a small processor. We estimate that
the individual processing power of a computer or notebook in 2006 is 22 times larger
than the computational power of a multi-service mobile phone (see Appendix). As a
result, in 2006, mobile devices represent 3.5% of the installed computational power in the
OECD and in LAC 2.3%. It is expected that the rapid diffusion of multimedia phones

70
Retesting these results with varying utility lifetime of computers, we observe something similar as
already observed with storage. The total installed computing capacity increases in both regions between 10-
20 when changing lifetime from seven to five years, and 20-40% with three years. Notwithstanding, the
ratio between both region does not change too much: with the seven years supposition the gap increases
over 130-fold [2520]/[19]=132.6, with five years [2920]/[22]=132.7 and with three years, it reduces to 115-
fold [3456]/[30]=115.2.
MAPPING OUT THE TRANSITION 170
will decisively increase the computational importance of mobile devices in the short-term
future.

Figure 17 Capacity to compute information with PCs, notebooks, and mobile phones

Source: own elaboration, based on various sources, see specifications in Appendix.

Summing up, measuring the digital divide in terms of information processing
capacity leads to different conclusions than comparing diffusion of ICT equipment.
Contrary to the superficial conclusion of a rapidly closing digital divide in terms of plain
access to the technology, the change in perspective presented here shows that the digital
divide is a moving target. Increasing saturation of ICT equipments diffusion in developed
markets does not imply a stagnation of increases in information processing capacity, due
to the incessant creative destruction of technological innovation. The number of devices a
0
500
1.000
1.500
2.000
2.500
3.000
3.500
4.000
1996 1998 2000 2002 2004 2006
MCPS/capita
OECD LAC
MAPPING OUT THE TRANSITION 171
person can possess might be limited, but this does not give us insight into how much
information a person can process with them.

Limitations and corresponding research challenges
This article proposes to measure digital development in terms of information
processing capacity, not in terms of the mere number of installed equipment. The
headcount of devices has long served as a rough proxy for the development of the
Information Society. While it is a fact of socio-economic research that measurement
efforts have to work with proxies most of the time, we need to take care that the usage of
proxies does not disguise the nature of the analyzed phenomena, which can result in
misleading policy conclusions. The simple exercise presented here shows that a
refinement of indicators can tell a quite different story about the same observed
phenomena. There is a difference in measuring the quantity of equipment in a society, we
might call the result “ICT-equipment-societies”, and measuring the amount of
information that a society processes: “Information Societies”.
With presented exercise as a starting point, we can see that a number of areas open
up for research. The first set of issues focus on the limitations of the presented exercise
and on potential refinements.
Access and real usage: In our estimations, we have assumed that ICT run 24 hours
for 365 days a year. An important refinement would be to estimate the “actual usage” of
these technologies in hours, not simply the installed and potentially usable capacity.
Available statistics are the limiting factor, but can be found in local or national samples,
such as time-budget studies.
MAPPING OUT THE TRANSITION 172
Analogue ICT: In accordance with the conventional definition of the digital divide,
we only consider digital ICT as access tools. In the Information Society, however,
analogue technologies, such as books, newspapers, radio and analogue TV, VHS and
music cassettes, among others, also play an important role. Presentation of all of these
technologies in approximated bit rates would allow for the first time to quantitatively
compare the capacity of “analogue” and “digital” solutions. Is most of the world’s
information already in digital format? If yes, when did it happen? What is the current
ratio? Right now, nobody knows the answer to these questions. Nevertheless, this
translation from analogue to digital is not straightforward. Analogue technologies do not
work with bits and any translation would require a set of reasonable assumptions.
Aggregate measures disguise their underlying distribution: One of the main benefits
of the approach presented in this article is that it reduces an array of traditional indicators
to only three indicators, expressed in bits, bits per second and computations per second.
In the presented exercise we have not focused on the nature of the distribution that leads
to these aggregate values. For example, four 14 kbps 2G mobile phones reach the same
amount of kbps as one 56 kbps modem connection (4*14 = 56). However, in the former
case, lower capacity is distributed among four tools (and most probably four distinct
users), and in the latter case it is concentrated in one tool. This leads to questions of
equality: Is there a trade-off between having few with much resources (capacity) and
having many with little? The total amount of bits per inhabitant does not tell us anything
about its concentration. The inclusion of this aspect could lead to interesting insights into
the effects of the concentration of information processing capacities. The installed
information processing capacity of one particular Information Society might be built on
MAPPING OUT THE TRANSITION 173
high-quality broadband connections for a few, while another Information Society with the
same aggregate information processing capacity (ceteris paribus), might be constructed
on the basis of low-quality mobile phones for everybody. What difference does it make?
This leads to the analysis of the domestic digital divide. As already mentioned in the
introduction, even though the presented example focuses on the international scenario, its
underlying logic can easily be applied to the domestic setting.
ICT Functionality: A similar logic that looks for patterns disguised by aggregate
statistics applies to the consideration of other determinants of functionality. Mobile
solutions are different from fixed solutions, and storage devices come with all kinds of
different storage latencies and throughputs (reading and writing speeds). A multi-
dimensional definition of ICT functionality would certainly make any analysis more
complex, but could enable deeper insights.
Type of content: The present analysis does not differentiate between the type of
content, such as voice, text, images, videos, etc. The main restriction to this refinement is
the availability of statistics about the nature of digital content. Consideration of the type
of content would not only allow for an analysis of its relevance, but also estimation of the
ultimate information entropy of installed systems (in Shannon’s sense), since
compression algorithms heavily depend on the type of content.
The second set of issues is conceptual and policy oriented in nature.
Moving target: If the Information Society is defined by its capacity to work with
information, the digital access divide becomes an extremely rapidly moving target. From
this perspective, it becomes clear that increasing pace of technological change will make
it impossible to “close” the digital access divide in a uniform sense. Some will always
MAPPING OUT THE TRANSITION 174
have more access than others and the approach presented in this paper points out to these
qualitative differences. While qualitative difference will remain, the divide could be
bridged nevertheless. This implies that every member of an Information Society could
have sufficient resources to continuously maintain minimum connectivity to the public on
the basis of equal entitlement. This is a constant challenge, and much depends on how the
terms “sufficient” and “minimum” are defined.
Sustainable policies: Given that the digital access divide is a constantly moving
target that widens informational abyss inside and between societies with every digital
innovation, and given the importance of digital ICT for today’s socio-economic
organization, appropriate policies will not cease to be part of the policy agenda. During
the past decades, the private sector has led the deployment of infrastructure, in most cases
under vigilant observation by regulators, such as the FCC, the European Commission and
other national authorities in countries all over the world. The regulation of ICT
infrastructure has become a complex subject by itself and even the fiercest market
competition is often closely regulated. This task will continue as technological progress
continues. The data presented in this article shows no sign of an innovation downturn.
Users continue to strive for more and more information processing power all over the
world. One of the resulting research questions is how to design policies that consider the
fast innovation cycles, but are independently of a specific – and rapidly outdated –
technological solution.
Is there an end? Even though there might be a limit to the number of devices a
person can possess, is not evident that there is a limit to the number of bits an individual
or a society can process. The human brain seems to have an upper limit for conscious
MAPPING OUT THE TRANSITION 175
information processing, but when is it reached? And even when it is reached, the theory
of biological evolution suggests that human intelligence is a flexible and expandable
variable. While our grandparents could hardly imagine the amount of information we
consume today on a daily basis, there seems to be no reason why our grandchildren
would not shake their heads in amazement when looking back at our informational snail-
systems.
From a methodological perspective, the conclusion is that it is necessary to go
beyond simplistic approximations of ICT equipment penetration rates. This will also
deepen our comprehension of the Information Society. This article presented an
alternative perspective that will perhaps contribute to the elaboration of new approaches
for this challenging undertaking.

MAPPING OUT THE TRANSITION 176

The world’s Technological Capacity to Store, Communicate
and Compute Information
71,72

We estimate the world’s technological capacity to store, communicate, and compute
information, tracking 60 analog and digital technologies during the period from 1986 to
2007. In 2007, humankind was able to store 295 x 10
18
optimally compressed bytes,
communicated almost 2 x 10
21
bytes, and carry out 6.4 x 10
18
instructions per second on
general-purpose computers. Our sample of application-specific embedded computing
power grew at a compound annual growth rate of 83 % over two decades and general-
purpose computing capacity at 58 % annually. The world’s capacity for bidirectional
telecommunication grew at 28 % per year, closely followed by the increase in globally
stored information (23 %). Humankind’s capacity for unidirectional information diffusion
through broadcasting channels has experienced comparatively modest annual growth (6
%). Telecommunication has been dominated by digital technologies since 1990 (99.9 %
in digital format in 2007) and the majority of our technological memory has been in
digital format since the early 2000s (94% digital in 2007).

71
This article has been published as Hilbert, M., & López, P. (2011). The World’s Technological Capacity
to Store, Communicate, and Compute Information. Science, 332(6025), 60 –65.
doi:10.1126/science.1200970
72
We would like to thank the Information Society Program of United Nations ECLAC (in Chile) for its
support, Tom Coughlin, John McCallum, Don Franz, Miguel Gonzalez, Cristian Vasquez, Len Adleman,
Manuel Castells and the statisticians from UPU (Universal Post Union) and ITU (International
Telecommunications Union), as well as numerous colleagues who motivated us by doubting the feasibility
of this undertaking.
MAPPING OUT THE TRANSITION 177
Leading social scientists have recognized that we are living through an age in which
“the generation of wealth, the exercise of power, and the creation of cultural codes came
to depend on the technological capacity of societies and individuals, with information
technologies as the core of this capacity” (Castells, 1998; 367). Despite this insight, most
evaluations of society’s technological capacity to handle information are based on either
qualitative assessments or indirect approximations, such as the stock of installed devices
or the economic value of related products and services.

Previous work
Some pioneering studies have taken a more direct approach to quantify the amount
of information that society processes with its information and communication
technologies (ICT). Following pioneering work in Japan (Ito, 1981), Pool (1983)
estimated the growth trends of the “amount of words” transmitted by 17 major
communications media in the United States from 1960 to 1977. This study was the first to
show empirically the declining relevance of print media with respect to electronic media.
In 1997, Lesk (1997) asked “how much information is there in the world?” and presented
a brief outline on how to go about estimating the global information storage capacity. A
group of researchers at the University of California, at Berkeley, took up the
measurement challenge between 2000 and 2003 (Lyman and Varian, 2003). Their focus
on “uniquely created” information resulted in the conclusion that “most of the total
volume of new information flows is derived from the volume of voice telephone traffic,
most of which is unique content” (97 %); as broadcasted television and most information
storage mainly consists of duplicate information, these omnipresent categories
MAPPING OUT THE TRANSITION 178
contributed relatively little. A storage company hired a private sector research firm
(International Data Corporation, IDC) to estimate the global hardware capacity of digital
ICT for the years 2007-2008 (Gantz, 2008). For digital storage, IDC estimates that in
2007 “all the empty or usable space on hard drives, tapes, CDs, DVDs, and memory
(volatile and nonvolatile) in the market equaled 264 exabytes” (Gantz, 2008; 3). During
2008, an industry and university collaboration explicitly focused on information
consumption (Bohn and Short, 2009), measured in hardware capacity, words, and hours.
The results are highly reliant on media time-budget studies, which estimate how many
hours people interact with a media device. The result obtained with this methodology was
that computer games, TV and movies represent 99.2 % of the total amount of data
“consumed”.

Scope of our exercise
To reconcile these different results, we focus on the world’s technological capacity
to handle information. We do not account for uniqueness of information, since it is very
difficult to differentiate between truly new and merely recombined, duplicate
information. Instead we assume that all information has some relevance for some
individual. Aside from the traditional focus on the transmission through space
(communication) and time (storage), we also consider the computation of information.
We define storage as the maintenance of information over a considerable amount of time
for explicit later retrieval and estimate the installed (available) capacity. We do not
consider volatile storage in the respective inventory (such as RAM), since the ultimate
end of volatile memory is computation, not storage per se. Communication is defined as
MAPPING OUT THE TRANSITION 179
the amount of information that is effectively received or sent by the user, while being
transmitted over a considerable distance (outside the local area). This includes those
transmissions whose main purpose consists in the overcoming of distances, not the local
sharing of information (such as the distribution of copies at a meeting, or communication
through private local area networks). We take inventory of the effective communication
capacity (the actual amount of bits transmitted). We define computation as the
meaningful transformation of information and estimate the installed (available) capacity.
More precisely, as shown in Figure, we distinguish between: (a) storage of
information in bits; (b1) unidirectional diffusion through broadcasting in bits per second;
(b2) bidirectional telecommunication in bits per second; (c1) computation of information
by general purpose computers in instructions per second (or MIPS); and (c2) we estimate
the computational capacity of a selected sample of application-specific devices (MIPS).
While previous studies tracked some two or three dozen categories of ICT over three
consecutive years at most, our study encompasses worldwide estimates for 60 categories
(21 analog and 39 digital) and spans over two decades (1986-2007).

MAPPING OUT THE TRANSITION 180
Figure 18: The three basic information operations and its most prominent
technologies.

We obtain the technological capacity by basically multiplying the number of
installed technological devices with their respective performances. All estimates are
yearly averages, but we adjust for the fact that the installed technological stock of a given
year is the result of an accumulation process of previous years, whereas each year’s
technologies contribute with different performance rates. We used 1,120 sources and
explain our assumptions in detail in Supporting Online Material (López and Hilbert,
2012). The statistics we rely on include databases from international organizations,
MAPPING OUT THE TRANSITION 181
historical inventories from individuals for commercial or academic purposes, publicly
available statistics from private research firms, as well as a myriad of sales and product
specifications from equipment producers. We filled in occasional blanks with either
linear or exponential interpolations, depending on the nature of the process in question.
Frequently we compared diverse sources for the same phenomena and strove for
reasonable middle grounds in case of contradictions. In cases where specific country data
were not available, we aimed for a globally balanced outlook by creating at least two
international profiles, one for the “developed” member countries of the Organisation for
Economic Co-operation and Development (OECD), and another one for the rest of the
world.

Information, not hardware with redundant data
While the estimation of the global hardware capacity for information storage and
communication is of interest for the ICT industry, we are more interested in the amount
of information that is handled by this hardware. Therefore, we convert the data contained
in storage and communication hardware capacity into informational bits by normalizing
on compression rates. This addresses the fact that information sources have different
degrees of redundancy. The redundancy (or predictability) of the source is primarily
determined by the content in question, such as text, images, audio or video (Shannon,
1948; Cover and Thomas, 2006). Considering the kind of content, we measure
information as if all redundancy were removed with the most efficient compression
algorithms available in 2007 (we call this level of compression “optimally compressed”).
Shannon (1948) showed that the uttermost compression of information approximates the
MAPPING OUT THE TRANSITION 182
entropy of the source, which unambiguously quantifies the amount of information
contained in the message. In an information theoretic sense, information is defined as the
opposite of uncertainty. Shannon defined one bit as the amount of information that
reduces uncertainty by half (regarding a given probability space, such as letters from an
alphabet or pixels from a color scale). This definition is independent of the specific task
or content. For example, after normalization on optimally compressed bits we can say
things like “a 6 square-cm newspaper image is worth a 1000 words”, because both
require the same average number of binary yes/no decisions to resolve the same amount
of uncertainty.
Normalization on compression rates is essential for comparing the informational
performance of analog and digital technologies. It is also indispensable for obtaining
meaningful time series of digital technologies, since more efficient compression
algorithms enable us to handle more information with the same amount of hardware. For
example, we estimate that a hard disk with a hardware performance of 1 MB for video
storage was holding the equivalent of 1 optimally compressed MB in 2007 (“optimally
compressed” with MPEG-4), but only 0.45 optimally compressed MB in 2000
(compressed with MPEG-1), 0.33 in 1993 (compressed with cinepack) and merely 0.017
optimally compressed MB in 1986 (supposing that no compression algorithms were
used). Given that statistics on the most commonly used compression algorithms are
scarce, we limit our estimations of information storage and communication to the years
1986, 1993, 2000 and 2007 (see López and Hilbert, 2012).
Conventionally bits are abbreviated with a small “b” (such as in kilobits per second:
kbps) and bytes (equal to 8 bits) with a capital “B” (such as in Megabyte: MB). Standard
MAPPING OUT THE TRANSITION 183
decimal prefixes are used: kilo (10
3
), mega (10
6
), giga (10
9
), tera (10
12
), peta (10
15
), exa
(10
18
), zetta (10
21
).

Storage
We estimate how much information could possibly have been stored by the 12 most
widely used families of analog storage technologies and the 13 most prominent families
of digital memory, from paper-based advertisement to the memory chips installed on a
credit card (Figure). The total amount of information grew from 2.6 optimally
compressed exabytes in 1986, to 15.8 in 1993, over 54.5 in 2000, to 295 optimally
compressed exabytes in 2007. This is equivalent to less than one 730 MB CD-ROM per
person in 1986 (539 MB per person), roughly 4 CD-ROM per person of 1993, 12 in the
year 2000 and almost 61 CD-ROM per person in 2007. Piling up the imagined 404 billion
CD-ROM from 2007 would create a stack from the earth to the moon and a quarter of
this distance beyond (with 1.2 mm thickness per CD).
Our estimate is significantly larger than the previously cited hardware estimate from
IDC for the same year (Gantz, 2008) (IDC estimates 264 exabytes of digital hardware,
not normalized for compression, while we count 276 optimally compressed exabytes on
digital devices, which occupy 363 exabytes of digital hardware). While our study is more
comprehensive, we are not in a position to fully analyze all differences, since IDC’s
methodological assumptions and statistics are based on inaccessible and proprietary
company sources.
Before the digital revolution, the amount of stored information was dominated by
the bits stored in analog videotapes, such as VHS cassettes (Figure). In 1986, vinyl Long-
MAPPING OUT THE TRANSITION 184
Play records still made up a significant part (14 %), as did analog audio cassettes (12 %)
and photography (5 % and 8 %). It was not until the year 2000 that digital storage made a
significant contribution to our technological memory, contributing 25 % of the total in
2000. Hard disks make up the lion share of storage in 2007 (summing up to 52 %), while
optical storage contributes more than a quarter (28 %) and digital tape some 11 %. Paper-
based storage solutions capture a decreasing share of the total (0.33 % in 1986 and 0.007
% in 2007), even though their capacity was steadily increasing in absolute terms (from
8.7 to 19.4 optimally compressed petabytes).

Figure 19 World’s technological installed capacity to store information, in optimally
compressed Megabytes (MB) per year, for 1986, 1993, 2000 and 2007, semi-log plot.

1.E+12
1.E+13
1.E+14
1.E+15
1986 1993 2000 2007
A
n
a
l
o
g
D
i
g
i
t
a
l
(MB)
ChipCard Floppy disks
Camera/camcorder internal Videogames others
Mobile phones & PDA Memory Cards
Portable Media Player Other hard-disks (portable)
CDs and MiniDiscs Server & Mainframe hard-disk
Digital tape DVD and Blu-Ray
PC hard-disk
Books Other paper and print
Newsprint TV movie film
X-Rays TV episodes film
Vinyl LP Cine movie film
Photo negative Audio cassette
Photo print Video analog
1%
71%
10%
1%
8%
5%
6%
2%
6%
8%
11%
21%
42%
2%
4%
2%
86%
2%
1%
1%
14%
5%
12%
8%
58%
MAPPING OUT THE TRANSITION 185
Communication
We divide the world’s technological communication capacity into two broad
groups: one includes technological systems that provide only unidirectional downstream
capacity to diffuse information (referred to as broadcasting), and one provides
bidirectional upstream and downstream channels (telecommunication). The ongoing
technological convergence between broadcasting and telecommunication is blurring this
distinction, as exemplified by the case of digital TV, which we count as broadcasting,
even though it incorporates a small, but existent upstream channel (e.g. video-on-
demand).
The inventories of Figures account for only those bits that are actually
communicated. In the case of telecommunication, the sum of the effective usages of all
users is quite similar to the total installed capacity (any difference represents an over- or
future investment). This is because most backbone networks are shared and only used
sporadically by an individual user. If all users demanded their promised bandwidth
simultaneously, the network would collapse. This is not the case for individual broadcast
subscribers, who could continuously receive incoming information. In order to
meaningfully compare the carrying capacities of each, we apply effective consumption
rates to the installed capacity of broadcasting (calling it the effective capacity). This
reduces the installed capacity by a stable factor (by 9 in 1986; 9.1 in 1993; 8.7 in 2000;
and 8.4 in 2007), implying an average individual broadcast consumption of roughly 2
hours and 45 minutes per 24 hours. It does not significantly change the relative
distribution of the diverse technologies (Figure \).
MAPPING OUT THE TRANSITION 186
Figure displays the capacity of 6 analog and 5 digital broadcast technologies,
including newspapers and personal navigation devices (GPS). In 1986, the world’s
technological receivers picked up around 432 exabytes of optimally compressed
information, 715 in 1993, 1.2 optimally compressed zettabytes in 2000 and 1.9 in 2007.
Cable and satellite TV steadily gained importance but analog over-the-air terrestrial
television still dominates the evolutionary trajectory. Digital satellite television leads the
pack into the digital age, receiving 50 % of all digital broadcast signals in 2007. Only a
quarter of all broadcasting information was in digital format in 2007. The share of radio
declined gradually from 7.2 % in 1986 to 2.2 % in 2007.

MAPPING OUT THE TRANSITION 187
Figure 20 World’s technological effective capacity to broadcast information, in
optimally compressed Megabytes (MB) per year, for 1986, 1993, 2000 and 2007,
semi-log

Figure presents effective capacity of the 3 most common bidirectional analog
telecommunication technologies and their 4 most prominent digital heirs. The 281
petabytes of optimally compressed information from 1986 were overwhelmingly
dominated by fixed line telephony, while postal letters contributed with a mere 0.34 %.
1993 was characterized by the digitization of the fixed phone network (471 optimally
compressed petabytes). We estimate the year 1990 to be the turning point from analog to
digital supremacy. The Internet revolution began shortly after the year 2000. In only 7
years, the introduction of broadband Internet effectively multiplied the world’s
telecommunication capacity by a factor of 29, from 2.2 optimally compressed exabytes in
1.0E+14
1.0E+15
1.0E+16
1986 1993 2000 2007
(MB)
A
n
a
l
o
g
D
i
g
i
t
a
l
TV-Terrestrial analog
TV-Cable analog
TV-Satellite analog
Radio analog
Paper Newspapers
Paper advertisement
TV-Satellite digital
TV-Cable digital
TV-Terrestrial digital
Radio digital
GPS
79%
13%
1%
7%
73%
19%
3%
5%
50%
21%
3%
2%
12%
8%
4%
61%
25%
4%
3%
5%
2%
MAPPING OUT THE TRANSITION 188
2000, to 65 in 2007. The most widespread telecommunication technology was the mobile
phone, with 3.4 billion devices in 2007 (versus 1.2 billion fixed line phones and 0.6
billion Internet subscriptions). Nevertheless, the fixed-line phone is still the solution of
choice for voice communication (1.5 % of the total), while the mobile phone network is
increasingly dominated by data traffic in 2007 (1.1 % for mobile data versus 0.8 % for
mobile voice).

Figure 21 World’s technological effective capacity to telecommunicate information,
in optimally compressed Megabytes (MB) per year, for 1986, 1993, 2000 and 2007,
semi-logarithmic plot

Compared with broadcasting, telecommunications still makes a quite modest, but rapidly
growing part of the global communications landscape (3.3 % of their sum in 2007, up
from 0.07 % in 1986). While there are only 8 % more broadcast devices in the world than
1.0E+11
1.0E+12
1.0E+13
1.0E+14
1986 1993 2000 2007
(MB)
Analog
Digital
Fixed (voice) phone analog
Postal letters
Mobile (voice) phone analog
Internet fixed
Fixed (voice) phone digital
Mobile (data) phone digital
Mobile (voice) phone digital
2%
1%
51%
42%
4%
31%
1%
67%
80%
0.34%
20%
97%
1%
1% 1%
MAPPING OUT THE TRANSITION 189
telecommunication equipment (6.66 billion vs. 6.15 billion in 2007), the average
broadcasting device communicates 27 times more information per day than the average
telecommunications gadget. This result might be unexpected at first sight, especially
considering the omnipresence of the Internet, but can be understood when considering
that an average Internet subscription effectively uses its full bandwidth for only around 9
minutes per day (during an average 1 hour and 36 minutes daily session).

Computation
From a theoretical standpoint, a “computation” is nothing else than the repeated
transmission of information through space (communication) and time (storage), guided
by some algorithmic procedure (Turing, 1937). The problem is that the applied
algorithmic procedure influences the overall performance of a computer, both in terms of
hardware design and in terms of the contributions of software. As a result, the theoretical,
methodological, and statistical bases for our estimates for computation are less solid than
the ones for storage and communication. In contrast to Shannon’s bit (1948), there is no
generally accepted theory that provides us with an ultimate performance measure for
computers. There are several ways to measure computational hardware performance,
such as FLOPS and SPECs. Our hardware performance variable of choice is MIPS
(Million or Mega Instructions Per Second), which was imposed upon us by the reality of
available statistics. Regarding the contributions of software, it would theoretically be
possible to normalize the resulting hardware capacity for algorithmic efficiency (such as
measured by O-notation). This would recognize the constant progress of algorithms,
which continuously make more efficient use of existing hardware. However, the
MAPPING OUT THE TRANSITION 190
weighted contribution of each algorithm would require statistics on respective execution
intensities of diverse algorithms on different computational devices. We are not aware of
such statistics. As a result of these limitations, our estimates refer to the installed
hardware capacity of computers.
We distinguish between two broad groups of computers. The first group includes all
computers whose functionality is directly guided by their human users. We call this group
“general-purpose computers” and include 6 technological families (Figure). The second
group carries out automated computations that are incidental to the primary task, such as
in electronic appliances or visual interfaces. While the user may have a range of
predefined choices regarding their functionality, the user cannot change the automated
logic of these embedded systems. We call this group “application-specific computers”.
While general-purpose computers are also equipped with application-specific parts
(mobile phones come with digital signal processors, and PCs contain microcontroller
units, etc.), we only include the capacity of humanly guidable microprocessors in the
respective inventory. The calculator laid the cornerstone for modern microprocessors and
was still the dominant way to compute information in 1986 (41 % of 300 general-purpose
tera-IPS). The landscape changed quickly during the early 1990s, as personal computers
and servers and mainframe computers pushed the evolutionary trajectory to 4.4 peta-IPS.
The personal computer extended its dominance during the year 2000 (289 peta-IPS), to
be rivaled by videogame consoles and increasingly relevant mobile phones by 2007 (6.4
exa-IPS). Videogame consoles contributed 25 % of the total in 2007. Nowadays, clusters
of videogame consoles are occasionally used as supercomputer substitutes for scientific
purposes and other data intensive computational tasks.
MAPPING OUT THE TRANSITION 191

Figure 22 World’s technological installed capacity to compute information on
general-purpose computers, in millions instructions per second (MIPS), distribution
for 1986, 1993, 2000 and 2007, semi-logarithmic plot

The relatively small role of supercomputers (less than 0.5 % throughout) and
professional servers and mainframes might come as a surprise. It can partially be
explained by the fact that the inventory of Figure presents the installed capacity,
independent of effective usage rates. We also carried out some estimations based on the
effective gross usage of the computers, which considers the time users interact with
computers (not the net computational time). As a result we get between 5.8 % and 9.1 %
of the installed capacity. The share of servers and mainframes grows to 89 % in 1986 and
11 % in 2007, and supercomputers contribute 4 % to the effective capacity in 2007.
The data also allows us to look at respective growth rates. Until the early 1990s, the
annual growth rate was quite stable, at roughly 40 % (Figure). The 1990s show
1.0E+08
1.0E+09
1.0E+10
1.0E+11
1.0E+12
1.0E+13
1986 1989 1992 1995 1998 2001 2004 2007
(MIPS)
Personal computers Videogame consoles Mobile phones/ PDA
Servers and mainframe Supercomputers Pocket calculators
66%
25%
6%
3% 0.3%
2007
86%
5%
3%
6%
2000
33%
9% 17%
41%
1986
64%
6%
1%
23%
6%
1993
MAPPING OUT THE TRANSITION 192
outstanding growth, reaching a peak of 88 % in 1998. Since then, the technological
progress has slowed. In recent times, every new year allows humankind to carry out
roughly 60 % of the computations that could have possibly been executed by all existing
general-purpose computers before that year.

Figure 23 Annual growth of installed general-purpose computational capacity as
percentage of all previous computations since 1977 (yeart / Σ[1977, yeart-1]).

Our inventory of application-specific computations is the least complete one. The
entire group of application-specific computers is very large and diverse (for example,
dice cups and roulette wheels are application-specific analog random number generators)
and it is often not straightforward to translate their performance into MIPS. The main
goal of our inventory of this group was to show that the computational hardware capacity
of application-specific computers is larger than the computational capacity of general-
purpose computers. To achieve this we focused on a sample that includes three prominent
groups: digital signal processors (DSP), which translate between analog and digital
signals (including CD, DVD and PVR devices, cameras and camcorders, modems and
0%
20%
40%
60%
80%
100%
1983 1986 1989 1992 1995 1998 2001 2004 2007
MAPPING OUT THE TRANSITION 193
setup boxes, GPS, portable media, printer and fax, radio, fixed and mobile phones);
microcontrollers (MCU) (which regulate electronics and appliances); and graphic
processing units (GPU) (an increasingly powerful microprocessor for visual displays).
While microcontrollers dominated our sample of application-specific computing support
in 1986 (90 % of the 433 application-specific tera-IPS from our sample), graphic
processing units clearly made up the lion share in 2007 (97 % of 189 exa-IPS).

Comparisons and growth rates
The world’s technological capacity to compute information has by far experienced
the highest growth (Table). The per capita capacity of our sample of application-specific
machine mediators grew with a compound annual growth rate of 83 % between 1986 and
2007 and humanly guided general-purpose computers with 58 % per year. The world’s
technological capacity to telecommunicate only grew half as fast (CAGR of 28 %). This
might seem a little surprising, as the advancement of telecommunications, and especially
the Internet, is often celebrated as the epitome of the digital revolution. The results from
Table challenge this idea and move human kinds’ ability to compute information into the
spotlight. The storage of information in vast technological memories has experienced a
growth rate almost similar to telecommunication (CAGR of 23 % per capita over two
decades). The lower growth rate results from the relatively high base level provided by
prevalent analog storage devices. The main characteristic of the storage trajectory is the
digitalization of previously analog information (from 0.8 % digital in 1986, to 94 % in
2007). The global capacity to broadcast information has experienced the least progress, at
6 % CAGR per capita. Broadcasting is also the only information operation that is still
MAPPING OUT THE TRANSITION 194
dominated by analog ICT. As a result, the capacity to store information has grown at a
much faster rate than the combined growth rate of tele- and broadcast communication. In
1986 it would have been possible to fill the global storage capacity with help of all
effectively used communication technologies in roughly 2.2 days (539/241.16). In 1993 it
would have taken almost 8 days, in the year 2000 roughly 2.5 weeks, and in 2007 almost
8 weeks.
The presented compound annual growth rates represent the temporal average of
periods with differential intensities of technological change. While general-purpose
computation had its peak growth around the turn of the millennia (Figure), storage
capacity had slowed down around the year 2000, just to restart accelerated growth in
recent years (CAGR of 27 % for 1986-1993, 18 % for 1993-2000 and 26 % for 2000-
2007; Table). The introduction of broadband has led to a continuous acceleration of the
telecommunication landscape (CAGR of 6 % for 1986-1993, 23 % for 1993-2000 and 60
% for 2000-2007; Table), and broadcasting is subject to a relatively stable rate of change
(CAGRs of 5.7 %, 5.6 % and 6.1 % for 1986-1993, 1993-2000 and 2000-2007; Table).
MAPPING OUT THE TRANSITION 195
Table 13Evolution of the world’s capacity to store, communicate and compute
information, absolute per capita, compound annual growth rate (CAGR), and
percentage in digital format.

1986 1993 2000 2007
CARG
86-07
Storage
MB optimal compression per
capita (installed capacity)
539 2,866 8,988 44,716 23%
% digital 0.8% 3% 25% 94%

Broadcast
MB optimal compression per
capita per day (effective
capacity)
241 356 520 784 6%
% digital 0.0% 0.0% 7.3% 25%

Telecom
MB optimal compression per
capita per day (effective
capacity)
0.16 0.23 1.01 27 28%
% digital 19.8% 68.5% 97.7% 99.9%

General-purpose
computation
MIPS per capita
(installed capacity)
0.06 0.8 48 968 58%
Sample of application-
specific computation
MIPS per capita
(installed capacity)
0.09 3.3 239 28,620 83%

The growth rates also allow us to formulate some kinds of Moore’s laws for the
technological information processing capacity of humankind. Machines’ application-
specific capacity to compute information per capita has roughly doubled every 14 months
over the past decades in our sample, while the per capita capacity of the world’s general-
purpose computers has doubled every 18 months. The global telecommunication capacity
per capita doubled every 2 years and 10 months, while the world’s storage capacity per
capita required roughly 3 years and 4 months to increase twofold. Per capita broadcast
information has doubled roughly every 12.3 years. Of course, such averages disguise the
varying nature of technological innovation avenues.

Perspectives
To put our findings in perspective, the 6.4*10^18 instructions per second that
human kind can carry out on its general-purpose computers in 2007 are in the same
MAPPING OUT THE TRANSITION 196
ballpark area as the maximum number of nerve impulses executed by one human brain
per second (10^17).
73
The 2.4*10^21 bits stored by humanity in all of its technological
devices in 2007 is approaching order of magnitude of the roughly 10^23 bits stored in all
the DNA molecules of a human adult
74
, but it is still minuscule compared to the 10^90
bits stored in the observable universe (Lloyd, 2002). However, in contrast to natural
information processing, the world’s technological information processing capacities are
quickly growing at clearly exponential rates.

73
Assuming 100 billion neurons * 1,000 connections per neuron * max 1,000 nerve impulses per second
74
Considering a quaternary DNA alphabet, in which each base pair can store 4 bits * 3 billion DNA base
pairs per human cell * 60 trillion cells per adult human
MAPPING OUT THE TRANSITION 197
Chapter Four: Policy Actions for the Transition

This last Chapter draws from the lessons learned about the particularities of the
transition from the previous Chapters and analyzes how they were integrated into an
international policy making process in Latin America. The article analyzes the experience
of a foresight Delphi study that I had previously designed and carried out with policy-
makers in Latin America. The Delphi aimed at creating a future policy-agenda in Latin
America, and led to several interesting lessons about policy-making in the field of digital
development.

MAPPING OUT THE TRANSITION 198
Foresight Tools for Participative Policy-Making in Inter-
Governmental Processes in Developing Countries: Lessons
Learned from the eLAC Policy Priorities Delphi
75

The paper shows how international foresight exercises, through online and offline
tools, can make policy-making in developing countries more participatory, fostering
transparency and accountability of public decision-making. A five-round Delphi exercise
(with 1,454 contributions), based on the priorities of the 2005-2007 Latin American and
Caribbean Action Plan for the Information Society (eLAC2007), was implemented. This
exercise aimed at identifying future priorities that offered input into the inter-
governmental negotiation of a 2008-2010 Action Plan (eLAC2010). It is believed to be
the most extensive online participatory policy-making foresight exercise in the history of
intergovernmental processes in the developing world to date. In addition to the specific
policy guidance provided, the major lessons learned include (1) the potential of Policy
Delphi methods to introduce transparency and accountability into public decision-
making, especially in developing countries; (2) the utility of foresight exercises to foster
multi-agency networking in the development community; (3) the usefulness of
embedding foresight exercises into established mechanisms of representative democracy
and international multilateralism, such as the United Nations; (4) the potential of online
tools to facilitate participation in resource-scarce developing countries; and (5) the

75
This article is published as: Hilbert, M., Miles, I., & Othmer, J. (2009). Foresight tools for participative
policy-making in inter-governmental processes in developing countries: Lessons learned from the eLAC
Policy Priorities Delphi. Technological Forecasting and Social Change, 76(7), 880–896.
doi:10.1016/j.techfore.2009.01.001
MAPPING OUT THE TRANSITION 199
resource-efficiency stemming from the scale of international foresight exercises, and
therefore its adequacy for resource-scarce regions. Two different types of practical
implications have been observed. One is the governments’ acknowledgement of the value
of collective intelligence from civil society, academic and private sector participants of
the Delphi and the ensuing appreciation of participative policy-making. The other is the
demonstration of the role that can be played by the United Nations (and potentially by
other inter-governmental agencies) in international participatory policy-making in the
digital age, especially if they modernize the way they assist member countries in
developing public policy agendas.
MAPPING OUT THE TRANSITION 200
Over the last decades, much has been written about the structural changes in
societies and economies associated with the advent of modern Information and
Communication Technologies (ICT). Change continues at a rapid pace. We are
continuing to see the emergence of technologies like the Internet and mobile phones with
applicability to practically all kinds of human endeavours, some of them displaying
unprecedented speed of diffusion (with the Internet having reached almost every fifth
inhabitant of the world, and mobile telephony almost every second, in less than two
decades). The digital paradigm is characterised by fast innovation cycles and accelerating
technological progress. These factors have led to a high level of uncertainty concerning
the options for, and implications of, this technological change. At a global level, the
problem of the digital divide and prospects of digital opportunities for development have
been underlined at the highest possible political levels, during the two phases of the
United Nations’ World Summit on the Information Society
76
(WSIS) in Geneva in 2003
and Tunis in 2005. This globally approved policy agenda spans a variety of subjects and
sets goals to be worked on by the international community between 2005 and 2015.
Latin American and Caribbean (LAC) countries have responded to this global
challenge by identifying the most urgent and important short-term policy goals for the
region. The result was a selection of thirty areas of interest and seventy concrete goals to
be implemented during 2005-2007, through a plan dubbed eLAC2007
77
. This Regional

76
See: http://www.itu.int/wsis. Heads of State and government discussed the implications of the digital
revolution and approved two ambitious agendas (Geneva Plan of Action, 2003; and the Tunis Agenda for
the Information Society, 2005). It is part of a global political undertaking known as the Millennium
Development Goals (see http://www.un.org/millenniumgoals/), which recognizes the role of ICT in
enhancing development and focuses on partnerships with the private sector to "ensure that the benefits of
new technologies, especially information and communication technologies ... are available to all”.
77
For further details on the eLAC process see: http://www.eclac.org/SocInfo/eLAC
MAPPING OUT THE TRANSITION 201
Action Plan for the Information Society was approved at the Regional Preparatory
Ministerial Conference of Latin America and the Caribbean for the second phase of the
World Summit on the Information Society, in Rio de Janeiro from 8-10 June 2005
78
, and
was seen as a first partial step towards the goals for 2015. The plan’s purpose is to
mediate between the ambitious goals of the global agenda, and the local demands of
individual countries of the region, by identifying common regional priorities. The plan’s
nature is a short-term. Accelerating technological progress, proliferating applications, and
the related uncertainty in this field of development, have forced policy-makers to opt for
a short-term approach of no more than two or three years, allowing for continuous
revision and adjustment to constantly changing challenges
79
. The logic applied here calls
for a series of consecutive short-term Action Plans in order to implement the long-term
vision until 2015.
During the execution of eLAC2007, significant advancements have been observed
in the development of Information Societies in Latin America and the Caribbean
(OSILAC, 2007), while at the same time an increased level of policy activity could be
evidenced. As a result, countries and international organizations evaluated the plan as a
success (ECALC Resolution, 2006) and decided to start the discussion about future
priorities concerning the effective usage of ICT to tackle pending challenges in the LAC
development agenda. The promises of effective ICT usage are multifaceted. These

78
See the Conference portal: http://www.riocmsi.gov.br
79
The uncertainty associated with technological progress implies that even if there was an “optimum path
towards an information society”, in the sense of an ideal recipe for policy-making, the rapid pace of ICT
development makes it unlikely that any such ideal could be grasped -the world would have moved on
before it could be established. This leads to the necessity of adaptive short-term policy-planning – which
needs to be informed by a view of long-term technology development possibilities..
MAPPING OUT THE TRANSITION 202
promises include economic growth and productivity, social inclusion, the modernization
of public administration, education and health sectors, security and disaster management,
cultural development, and many potentials. The Information Society Programme of the
United Nations’ Economic Commission for Latin America and the Caribbean (UN-
ECLAC), acted as the technical secretariat for the Regional Action Plan eLAC2007. In
response to the task of elaborating a new regional agenda for the year 2010, the
Programme elaborated the “eLAC Policy Priorities Delphi”.

The exercise aimed to
identify public policy priorities and strategic policy alternatives regarding the use of ICT
for development in the LAC region, for the period between 2008 and 2010. It did so by
systematically collecting, and analyzing information so as to provide results that can help
to improve the quality of policy choices made in a public policy agenda. The Delphi
policy process was conducted between April 2006 and February 2008, with the report of
the Delphi exercise being used as an input for the new 2008-2010 Regional Action Plan
eLAC2010. (A short description of the inter-governmentally approved eLAC2010 Action
Plan can be found in the left-hand columns of the Annex to this paper, which also
provides a general overview of the different topics and thematic areas on the agenda an
Action Plan for the development of Information Societies).
Below we review this participatory exercise and highlight the lessons learned. It
starts with a summary of particularities to consider when working on regional-level
agenda-building on such a dynamic and cross-cutting topic as ICT in developing regions.
It then presents the Delphi policy process, before drawing conclusions as to the conduct
of foresight and participatory policy-making exercises, in developing countries and more
generally.
MAPPING OUT THE TRANSITION 203

Regional agenda-building for digital development in developing
countries
A public policy agenda may be defined as “the list of subjects or problems to which
governmental officials, and people outside of government circles who are closely
associated with those officials, are paying some serious attention at any given time.”
[Kingdon, 1995; 3]. The agenda-setting process then, involves determining which
approaches to understanding and tackling these “subjects or problems” are liable to be
most effective ones. Policymakers require good advice as to these approaches, and for the
strategic formulation of goals and of plans to achieve these goals. This may result in a
shortlist of proposals. “Having a viable alternative available for adoption facilitates the
placement of a subject high on a governmental agenda and dramatically increases the
high placement of a subject on a governmental agenda” [Kingdon, 1995; 144]. The
ensuing question for policy-making in the field of ICT in Latin America and the
Caribbean is therefore who is in the best position to formulate this “good advice” and the
required “shortlist of policy proposals” on a regional level? This section discusses how
the consideration of this question resulted in the decision to opt for a Policy Priority
Delphi as the method of choice to determine the 2008-2010 Regional Information Society
Action Plan.
At the two extremes, the choice is to opt either for a technocratic or a more
democratic and participative way of securing such advice. A technocratic approach will
involve relying on a group of experts who are members of a highly skilled and
legitimized elite group (“technocrats” such as public government officials and those
MAPPING OUT THE TRANSITION 204
knowledgeable and well-connected people known in the U.K. as “the great and the
good”). More participative approaches, in contrast, will aim to draw on a wider range of
inputs as to the nature of problems and possible solutions. This dichotomy suggests that
adequate knowledge to determine policy options may be found either within a small and
centralized subgroup of society, or dispersed more widely as decentralized intelligence
among an unlimited number of people. This duality has been at the heart of the
governance question ever since institutional mechanisms were established to govern the
fate of societies through political organization, including Athens’s Polis, monarchic
forms of governance, or different kinds of democratic governance in the contemporary
world.
From the point of view of the prevailing political system and legitimization of
public decision making, representative democracy is the current choice of governance
throughout the LAC region. In practice, this represents an intermediate solution between
rather “technocratic” and more “democratic/participatory” approaches to deciding who
formulates “good policy advice”. Elected and legitimized by popular vote, politicians and
the public officials who serve them have the task of arriving at policy options that can
rise to the challenges confronting their countries. In the words of James Madison, one of
the founding fathers of representative democracy (Madison, 1787; 6): “the delegation of
the government ... to a smaller number of citizens elected by the rest ... [aims] ... to
refine and enlarge the public views by passing them through the medium of a chosen body
of citizens whose wisdom may best discern the true interest of their country”. This logic of
governance seems to have produced acceptable results throughout recent centuries. The
task of elaborating a regional public policy agenda for ICT development in a developing
MAPPING OUT THE TRANSITION 205
region, however, can challenge important aspects of Madison’s logic. Various
particularities associated with ICT and international development suggest that it is
beneficial to expand the circle of participation and to enrich the established mechanisms
of representative democracy with the opinions of a decentralized group.
First of all, ICT is a pervasive technology with multisectoral impacts and
implications. Common core technologies associated with digital information processing
are extremely widely adopted. But the pace of adoption, the development of applications,
the configuration of systems, and the modes of usage can evolve in dissimilar ways across
different branches of the economy and society (e.g. firms, hospitals, schools,
municipalities), and across organizations of different types (e.g. larger and smaller firms),
and in different locations (e.g. metropolitan and more remote areas). The notion of creating
Information Societies implies that the transformation affects every aspect of society. The
cross-cutting nature of the transformations affects the development of infrastructure and
science and technology, as well as the educational and health sectors, the economy and
entrepreneurship, community and local life, cultural heritage, legislation, the
management of disasters and national security, public administration, among many
others. Knowledge concerning these patterns of digital development is thus distributed in a
decentralized manner throughout society. Very few people having much overview of the
overall contours of this development, let alone possessing insight into the variety of
localized experiences and initiatives. But this also implies that the centralization of
activities with any specific actor is liable to run into difficulties. The scope and diversity
of the challenges make it improbable that a small group, in the public sector or elsewhere,
MAPPING OUT THE TRANSITION 206
can have developed a complete basis of information in all of these fields. This implies the
need for a decentralized structure in the development of a public policy ICT agenda.
Second, uncertainty is very high in a field of such dizzying technological progress.
ICT performance systems display exponential trajectories ever since they came into
existence. The past 35 years of ICT development have been extremely fast and diverse,
starting with the black and white Xerox text-processors developed by the Palo Alto
Research Centre, passing by the PC and the first graphical PC operating systems, over the
Internet and mobile phones, to wireless multimedia broadband computers through which
millions of virtual avatars spend considerable amounts of time in massively three-
dimensional multiplayer online environments.
The velocity of innovation, and the constantly changing supply-side and user-side
environments make it extremely difficult for the public sector to gather all the necessary
data to found its decisions on a solid information base – and to continue to do so on an
ongoing basis. This uncertainty with regard to technological progress is especially severe
for developing regions such as Latin America and the Caribbean, given that the great
majority of technological Research and Development is exogenous to the region. This
means that regional decision-makers are not even involved in the definition of choices that
shape the final technological choices. In turn this means that the expertise to provide
intelligence on future opportunities needs to be acquired from a wider audience, which
might have better insights, such as the academic and private sector.
Third, efforts to create an international public policy agenda come up against the
demographic and socio-economic heterogeneity of LAC countries. The region is host to
countries with a population of more than 100 million (Brazil, Mexico) - and to others
MAPPING OUT THE TRANSITION 207
with less than 50,000 (Saint Kitts and Nevis). In 2008, telephone penetration is over
100% is some countries (Argentina, Bahamas, Barbados, Jamaica, Trinidad and Tobago)
- but less than 40% in others (Bolivia, Cuba, Haiti, Nicaragua, Peru). There are LAC
countries in which half of the population is already connected to the Internet (Barbados,
Jamaica) - and others in which Internet penetration is less than 5% (Cuba, Honduras,
Nicaragua, Paraguay). This heterogeneity means that experiences will be very diverse.
The implication, then, is that a diverse group of people need to be involved in the search
for “good advice” relevant for the region as a whole.
These three sets of considerations informed the planning and design of the eLAC
Policy Priorities Delphi. This aimed to elicit decentralized intelligence that would be
useful for preparing adequate policy choices. In accordance with Turoff’s definition, a
Policy Delphi is “a tool for the analysis of Policy issues and not a mechanism for making
decisions” (Turoff, 1975). In this sense, the goal is to inform the traditionally closed
circle of decision-makers in public offices, aiming to enable them to access knowledge
dispersed throughout society through open-ended consultations with the larger
stakeholder community concerned with ICT for development. A similar widening of
perspectives to inform decisions around Science and Technology has underpinned the
Technology Foresight programmes that many countries instituted over the part 10-15
years.
But widening participation is not unproblematic. Should consultation be open to
everybody – after all, uninformed people may opt for unrealistic policy choices? What
about the danger that powerful private sector actors will use these new channels of
participation to manipulate policy choices in their favour? Both possibilities have to be
MAPPING OUT THE TRANSITION 208
taken seriously. The first echoes the longstanding concerns expressed by critics of direct
democratic systems, fearing that decisions will be made on the basis of fickle sentiment
rather than deliberation and analysis. The second danger is especially relevant in the field
of ICT, with the world’s two richest individuals in 2007 being from ICT industries -- one
of them having business priorities almost exclusively focused on Latin America’s
connectivity
80
.
Opening up the consultation to everybody carries the risk of uninformed decisions.
Students of public opinion frequently caution that ordinary citizens generally lack well-
developed attitudes or opinions on most public issues. Max Weber (1918) famously
argued that the mass public – across all social classes —thinks only as far as the day after
tomorrow, and is always susceptible to emotional and irrational influences.
Notwithstanding, handpicking experts carries the risk of including a limited selection of
technocrats.
81
For practical reasons, the eLAC Policy Priorities Delphi opted for a mix
between the criteria of self-selection and handpicking. Rounds one, two and four of the
exercise were carried out virtually, receiving 1274 contributions from an open-ended
group that was filtered by the criterion of self-selection – are you prepared to dedicate the
time to fill out the online questionnaire? (see Figure). Since it took between thirty and
forty-five minutes to complete each online questionnaire, and given the specificity of the
nature of the topic, the willingness to devote sufficient time to it was a decisive criterion.

80
Carlos Slim Helú is a Mexican business man who controls Teléfonos de México (Telmex), Telcel and
América Móvil. He was identified as the richest man on earth by Forbes Magazine in 2007, followed by
Bill Gates, co-founder and chairman of Microsoft.
81
Even in the pioneering European Foresight exercises of the mid-1990s, a common experience was that of
discovering that the implications of the technological developments that were being considered required
knowledge that went well beyond that possessed by those recruited to participate in Panels – knowledge of
such diverse issues as entrepreneurship and the problems of small firms, consumer behaviour and public
opinion, impacts of tax and pension schemes, ethical issues around new technologies, and much else.
MAPPING OUT THE TRANSITION 209
Rounds three and five of the exercise were carried out by face-to-face consultations, via
personal interviews of selected experts or face-to-face workshops with handpicked
invitees. 180 contributions were achieved by this process of handpicking regional opinion
leaders.
The threat of opinion manipulation by lobby groups is extremely difficult to control.
Promoting the consultation with a very heterogeneous group of participants, spanning
diverse industries and communities, is one way to reduce the potential influence of a
particular lobby group. However, this cannot assure the absence of vested interests. Efforts
to make the selection of Delphi participants more transparent, democratic, and
representative would have been extremely resource-intensive. This would have been
disproportionate for the intended purpose – since the eLAC Policy Priority Delphi is not a
final decision making tool itself. Rather, it serves as informal input for the established
system of representative democracy. Democratically elected representatives evaluate the
usefulness and validity of the Delphi results. The Policy Delphi is not intended to
undermine the legitimacy of representative democracy, but to enrich it with the opinions of
a broader group enlisted into direct participation.
Similarly, the eLAC Policy Priorities Delphi supports multilateral policymaking. It
feeds into established governmental processes carried forward in the form of the Ministerial
Conferences held in Rio de Janeiro (June, 2005) and El Salvador (February 2008), under
the auspices of the United Nations. This multilateral organization is the only inclusive
intergovernmental body in today’s world, and thus plays an important role in formulating
and diffusing ICT policy.
MAPPING OUT THE TRANSITION 210
In summary, this Delphi plays a role in constructing the policy agenda: with an
approach that is (1) based on open-ended consultations that exploits the decentralized
intelligence of the group “from the bottom up” (through the direct contributions to the
Policy Delphi from a specific group of stakeholders, such as civil society, private sector
foundations and academia), while (2) assuring the legitimacy of the process “from top-
down” (through democratically legitimized bodies and their selected technocrats, such as
national governments and their intergovernmental institutions, such as the United Nations
system).

The eLAC Policy Priority Delphi
The eLAC Policy Priorities Delphi was carried out between April 2006 and
September 2007 with the financial assistance of the European Commission’s @LIS
project
82
. Its design was inspired by the European Union’s policy priority foresight
experiences
83
. The natural starting point was the existing 2005-2007 Regional Plan of
Action, eLAC2007.

82
@LIS - Alliance for the Information Society – is a Programme of the European Commission aiming to
reinforce the partnership between the European Union and Latin America in the field of the Information
Society. Its objectives are to establish dialogue and cooperation on policy and regulatory frameworks in
key areas and to boost interconnections between research networks and communities in both regions
reducing the digital divide and integrating Latin America into a Global Information Society. After the
successful execution of the 2003-2007 total budget of €77.5m (€63.5m financed by the EC and the rest by
the partners of the programme) @LIS is going into its second phase for the 2008-2014 period. See
http://www.dft.gov.uk/pgr/scienceresearch/futures/secsceniss/wrdsenariotoolv2
83
The European Union has a strong track record of information society foresight exercises [20], in order to
provide regular insight and updated intelligence for the eEurope2002-eEurope2005-i2010 agendas. Some
recent ones include the Delphi-based EUFORIA scenarios [21], the survey-based STAR scenarios, the
projection-based SEAMATE scenarios, the workshop-based ISTAG scenarios, the panel-based FLOWS
scenarios and the [22] foresight exercise.
MAPPING OUT THE TRANSITION 211
The first Delphi round basically presented the thirty priority areas of this inter-
governmentally consented Action Plan for revision and comments in the light of
upcoming challenges. Four additional rounds of consultation followed, leading to a
revised priority agenda for the period 2008-2010. The entire process was presented in a
comprehensive report (eLAC Delphi Report, 2007). In February 2008 it served as the
main input for the inter-governmental negotiations that led up to the approval of the
2008-2010 Action Plan eLAC2010, during the Latin American and Caribbean Ministerial
Conference on the Information Society in El Salvador.
84

Overall, the Delphi process received 1,454 contributions from public, private and
academic sectors and civil society throughout five consecutive rounds (see Figure 1). The
distribution of contributions for each round can be seen in the Figure. Following the
general design of a Policy Delphi, it used the results of previous rounds as feedback
during subsequent rounds, in order to enable judgments to be reconsidered in the light of
opinions collected in those rounds and thus identify areas of emerging consensus and
potential differences of interests. The five rounds of the eLAC Policy Priorities Delphi
were implemented through three online questionnaires (receiving 1,274 contributions)
and two face-to-face consultations (180 contributions). The first two rounds were carried
out online and aimed at the reconsideration of the priority areas of the (outgoing) Action
Plan eLAC2007. The final three rounds were carried out through a mix of personal
interviews and online questionnaires and aimed at the elaboration of practical policy
options to work on the newly identified priority areas.

84
See the Conference portal: http://www.eLAC2007.org.sv
MAPPING OUT THE TRANSITION 212
Figure 24 eLAC Policy Priority Delphi from eLAC2007 to eLAC2010

MAPPING OUT THE TRANSITION 213
As the eLAC community had not established a network of active stakeholders
before embarking on this exercise, the design involved an open-ended opinion poll. The
invitations to participate in the three online questionnaires were sent to about 7,000
contacts from the region gathered by the ECLAC Information Society Programme, with a
request for further dissemination of the invitation. Thirteen regional institutions from the
public and private sector and civil society, joined in this multi-stakeholder effort -
disseminating the questionnaires over their e-mail networks, posting them on their
Websites, and including them in their newsletters and bulletin boards. As already noted,
participants were self-selected from this pool of contactees. This process attracted a
relatively educated group, in which the majority of the participants were individuals with
a master’s degree or doctorate (62%) (Figure 2). Figure 2 also demonstrates that, at least
in broad terms, the range of participants was highly representative geographically.
The distribution of participants in terms of regional and professional affiliation,
gender and education level proved to be very similar over the three online rounds of the
exercise; it may provide a rough idea of the structure of the general ICT-for-development
community in developing countries. The high degree of participation of private sector
informants (39%) shows the proactive interest of the industry. But it is also notable that
the online method admitted traditionally underrepresented voices, such as from the
Caribbean or Central America.

MAPPING OUT THE TRANSITION 214
Figure 25 Distribution of 1,274 online contributions (Delphi Round one, two and
four) according to education, professional affiliation, gender and geographic
representation of participants

Gender Subregion Participation Real population
Women 34% South America 69% 67%
Men 66% Meso-America 23% 26%
Caribbean 8% 7%
Note: The virtual consultation was carried out during Delphi rounds one, two and
four (see Figure 1). Duplicates have not been removed from the presented count of
contributions, given the limited possibility to do so (due to the anonymity of the open-
ended group of participants and the fact that formal registration was not sought, since
this might be a barrier to participation).

In agreement with common Policy Delphi practice, anonymity was assured during
the virtual rounds, resulting in a focus of individual opinions, and not in organizational
statements (these have been collected during the two face-to-face rounds). Anonymity
enables participants to avoid potential repercussions and embarrassment, including the
difficulty of publicly contradicting colleagues or superiors. The participants were invited
to provide an e-mail address, without the need for self-identification, in order to be able
to invite them to receive feedback and participate in consecutive rounds (for percentages
of returning participants, see note in Figure 1). Of the 1,274 contributions in the three
Master
and PhD,
62%
University,
36%
Non
university
, 2%
Private
sector,
39%
Public
sector,
25%
Aca-
demic
sector,
24%
Civil
society,
12%
MAPPING OUT THE TRANSITION 215
online rounds, 720 Email addresses have been obtained, thus laying the foundations for a
more solid eLAC multi-stakeholder community. It was decided not to use a mechanism
that would have allowed for anonymous tracking of participants (such as the option to
register with an anonymous username and password). The analytical drawback of this
choice is the inability to identify with certainty how many of the participants returned in
each round. Nevertheless, it was consciously decided against this alternative given the
implied risk of reducing the number of participants. Delphi exercises are not very
common in LAC and initially it was not even clear if a critical mass of participants could
be reached at all. The extra effort of registration and eventual doubts about the reasons
behind a need for registration (i.e. for a political susceptibility exercise like the eLAC
Policy Priority Delphi) might have lowered the participant turn out. The obtained Email
addresses provide some insight on returning participants. While it is important to
remember that the same person does not necessarily provide the same Email address in
each round, we estimate that between 10-15% of the second round had already
participated in the first round and between 25-35% of the last online questionnaire (fourth
round) had already participated in the second round. This leaves a rather small group of
people that participated consistently in all online rounds (less than 10% of the 501
contributors of the second round), but overall increasing amount of returning participants
is interpreted as a positive sign towards the establishment of a committed stakeholder
community (and a related participative culture) for these kinds of practices in LAC.
The two rounds of personal face-to-face consultations have also been carried out,
through a selection of 180 regional opinion leaders, mainly government representatives,
officers of international organizations, and researchers and experts from academia, the
MAPPING OUT THE TRANSITION 216
private sector and civil society. The names and affiliations of the participants of the face-
to-face rounds are listed in the final eLAC Policy Priority Delphi report and attest the
representativeness of this international multi-sector exercise.
Most Delphi studies ask for forecasts of when various developments are liable to
occur, or how far such developments will have progressed by a particular point in time.
The present study first used a Policy Delphi approach, focusing on the extent to which
various trends and actions might contribute to overall goals. Subsequently, a Goals
Delphi was used which examined what targets should be associated with particular goals,
or how far particular goals were prioritized. For example, the 2005-2007 Action Plan
eLAC2007 covered thirty thematic areas, related to issues like public ICT access centres,
computers in schools, connectivity of hospitals, ICT alphabetization and training of the
work force, digital management of disasters, regional backbone infrastructure, telework,
required legislative frameworks, e-government, among others (see the Annex for brief
characterization of the eLAC issues). The questions to be answered by the first two
rounds of the Policy Delphi focused the relevance of these, and other thematic areas, for
the period 2008-2010. Thematic areas were prioritized and ranked. Rounds three to five
focused on the formulation or reformulation of concrete goals to move forward in the
identified thematic areas. The consultations aimed at the determination of the aspired
level of connectivity, the intensity of training programs, the kind of work that would need
to be done on various legislative challenges, and the creation of regional working groups
to deepen comprehension about specific issues, among other policy goals.
A major characteristic of a Policy Delphi is that participants are presented with
various voting dimensions. Traditionally, the preferred voting dimensions include
MAPPING OUT THE TRANSITION 217
dimensions are desirability and feasibility, as are consistency with existing values and
anticipation of future constraints. The subjects addressed are complicated ones, but – like
more conventional Delphis - Policy Delphi questionnaires are necessarily limited in terms
of their scope and depth. The eLAC Policy Priority Delphi had the ambition to mobilize
the largest number of stakeholders possible, fulfilling an advocacy and network role,
which led to the decision to simplify voting dimensions by summarizing several of the
possible evaluation criteria in the question into one major variable. In this sense, the
eLAC Policy Priorities Delphi asked the participants to themselves synthesise the
possible evaluation criteria in the question into one major variable for each issue
addressed. They were asked to assess the impact of each thematic area for “the
development of Latin American and Caribbean Information Societies for the year 2010”.
Another characteristic of a Policy Delphi is that it aims at determining the areas of
disagreement among groups of participants. The delineation of differing views aims at
providing an opportunity for the recipient audience members to prepare their respective
cases adequately (in our case the Ministerial Conference that will finally approve the
eLAC2010 Action Plan). The questionnaires of the first two survey rounds asked
participants to differentiate between economic, social and political impacts for
development and to evaluate each one of them on a Likert scale from “negative impact”
to “positive impact”. This information then allows analyzing the degree of consensus and
disagreement between various groups, such as subregions (South America, Meso-
America, Caribbean) and participating sectors (Public Sector, Private Sector, Civil
Society, Academia).
MAPPING OUT THE TRANSITION 218
The choice of these questions allows to gather enough information to satisfy the
defined goals of a Policy Delphi, without demanding too much of the participants’ time.
It is important to point out that throughout all Delphi rounds, participants have always
been invited to write and hand in open comments and ideas. The attained summary
insights were complemented in rounds three and five of the Delphi by means of face-to-
face consultations and personal interviews. These provided the opportunity to dig a little
deeper into the different opinions, and to explore the underlying arguments and patterns
of reasoning associated with different topics, and with specific priorities and policy
actions.

The five consecutive Delphi rounds
The thematic areas addressed in the first Delphi round (during April and June
2006) were based on the thirty priority areas for development of the information society
in LAC countries that the countries of the region had identified in their 2005-2007
Regional Action Plan (eLAC2007). 155 participants took part, using the virtual eLAC
platform
85
to rank these thirty areas (see Annex) by social, economic and political impact
for development up to the year 2010, on a Likert scale from one (“negative impact”) to
five (“positive impact”). Since the dynamics of the ICT revolution are ongoing, they were
also invited to suggest new fields of interest to be considered. The analysis of the first
round shows that participants esteem the greatest impacts for social development from
thematic areas related to access to ICT (connectivity and equipment diffusion), the largest

85
The platform was set up at: http://www.eLAC2007.info and the underling software was provided by
Groupmind Solutions, based on the tool GroupMind Express.
MAPPING OUT THE TRANSITION 219
impacts for economiv development from areas related to capabilities (training and
education), and the greatest impacts for political development from policy areas related to
the coordination of activities across sectors.
In the second Delphi round, during October and December 2006, a revised ladder
of 47 priority issues was displayed, in order to construct a fresh impact ranking, including
the thematic areas of eLAC2007 (ranked according to the first round results), as well as
suggestions for new thematic areas, such as “e-democracy”, “electronic management for
agriculture and fishing”, “content for mobile phones”, “ICT connectivity for tourist
centres”, among others
86
. This time, 501 contributors from twenty-two LAC countries
answered the online questionnaire. Connectivity for schools and local governments, ICT
training for enterprises and the workforce, e-government and national Information
Society strategies and agendas were identified as the top priorities throughout the region.
In order to get a general idea of the magnitude of the differences in opinion among
subregions (South America, Meso-America
87
, Caribbean) and participating sectors
(Public Sector, Private Sector, Civil Society, Academia), a simple polarization index was
adopted from (Schneider, 1972). In his article, Schneider proposes to compute the
absolute value of differences of the average vote for each one-to-one combination of
subgroup pairs. Following this method, the average difference among each pair of groups
was computed for the votes on the 30 thematic areas of the first round, and the 47
thematic areas of the second round. The results can be seen in Figure 3.

86
For the ranking that resulted from the first round, see
http://www.cepal.org/socinfo/noticias/paginas/8/26998/RANKINGS_english.xls . For the ranking that
resulted from the second round, including all 47 thematic areas, see:
http://www.cepal.org/socinfo/noticias/paginas/8/26998/Ranking%20by%20area%20of%20impact.pdf
87
For the purpose of this study, Meso-America is defined as Central-America plus Mexico.
MAPPING OUT THE TRANSITION 220
Figure 26 Polarization Index for Subregions and Sectors of professional affiliation
of participants

First Round Subregions First Round Sectors

Second Round Subregions Second Round Sectors

The first conclusion from Figure 3 is that despite the notorious heterogeneity in the
region, the results shows a striking coincidence of interests and a surprisingly large and
stable consensus (on average the differences are less than 0.2 on a Likert scale from one
to five). This has been a positive surprise and shows the feasibility and readiness of the
LAC region to create and foster a regional vision and policy agenda. While it is the
declared goal of a Policy Delphi to identify areas of disagreement, the politically
MAPPING OUT THE TRANSITION 221
sensitive environment of the eLAC Policy Priority Delphi motivated the project team to
highlight and foster this important evidence of a concerted regional outlook on a common
challenge. The risk of political escalation is always latent in inter-governmental processes
and LAC governments do not have any obligation to approve a Regional Action Plan.
Care needed to be taken in order to foster a climate of cooperation and compromise,
which then enables the creation and commitment to an innovative policy tool such as
eLAC.
However, the graphs allow the identification of several trends in disagreements. As
expected, among subregions the difference in opinion of the Caribbean sticks out. In both
rounds, it becomes clear that the small island States of the Caribbean have somewhat
different priorities than South- and Meso-America. Regarding the disagreement among
the various groups of professional affiliation of participants, the first round shows a pretty
harmonious picture, while in the second round civil society takes up its traditional
advocacy role and seems to increase opinion polarization by returning to their habitual
role of opposition of the public and private sectors. It is also interesting to note that in the
first round, the consensus among professional sectors was larger than the consensus
among regions, while in the second round, more disagreement could be fond among
sectors than among subregions. We will refer to specific cases of this polarization index
(P.I., as the average of the Subregion and Sector values of the second round) throughout
the rest of the article.
Given the need for prioritization that underpins the existing eLAC2007, the top
thirty priority areas (as identified in the second survey round) were selected for the rest of
MAPPING OUT THE TRANSITION 222
the Delphi. Twenty-three of these areas of interest coincided with eLAC2007, while
seven new issues had entered the list of regional ICT priorities.
In terms of topics that were dropped, it is somewhat surprising that the participants
did not consider that thematic areas such as “free software and open source software”
(ranked 33 out of 47) and “Internet Governance” (ranked 37) would have a large impact
on LAC development up to the year 2010. Both issues took up a large share of the
discussion that took place during the World Summit on the Information Society (WSIS,
2003-2005) - it seems that the broader ICT-for-development community assigns a
different importance to these controversial issues than their political representatives.
While the polarization index shows a large degree on disagreement on the Internet
Governance issue (0.30 as the average of Subregion and Sector P.I. of second round), the
low evaluation of the impact of free and open source software seems not as controversial
(P.I. of 0.16). Another thematic area which did not make it into the top thirty priorities
was the “local supply of hardware-related goods and services” (ranked 42) – this is only a
concern of larger countries, such as Brazil, in which the thematic area reached rank 31
88
,
and generally there was a fairly strong consensus on this issue (P.I. of 0.09).
Among the newly identified areas of interest are: “electronic democracy” (rank 10),
the “inclusion of the gender perspective” (rank 16), “distance medicine” (rank 25),
“intellectual property and copyright” (rank 26) and “Voice-over-Internet-Protocol” (rank
29). This reorganization of priority areas at the time of the second Delphi round
suggested some reprioritization of main concerns since the outgoing Action Plan

88
For the individual country ranks see: Argentina
(http://www.cepal.org/socinfo/noticias/paginas/2/27002/Argentina.pdf), aNNd similar pages for Brazil
(.../Brasil.pdf), Chile (.../Chile.pdf), Colombia (.../Colombia.pdf), Mexico (.../Mexico.pdf), Peru
(.../Per%FA.pdf) and Venezuela (.../Venezuela.pdf)
MAPPING OUT THE TRANSITION 223
eLAC2007 had been approved. Later, this paper will consider how far the final version of
the inter-governmentally approved new Action Plan eLAC2010 adopted these newly
identified priorities.
The third Delphi round consisted of personal interviews with 116 experts from the
public and private sectors, academia, and civil society, from nineteen countries. The
interviews were intended to result in the formulation of concrete goals and activities to
implement the 30 priority areas that had been identified in the second survey round. The
input that was received by the project team during the interviews was immense, partially
consisting of very unique and creative ideas, and partly recurring to well-known policy
options.
In order to facilitate work on the thirty priority areas of a new Regional Action
Plan, the project team needed to filter this input. It followed two simple rules to select
100 goals from the material:
 Policy options and goals need to be quantifiable and measurable (results-
oriented); and/or
 Policy options and goals need to rely on existing international mechanisms, in the
sense that specific action-oriented international agencies or institutions have to be
identified that actively work on this challenge in the region (action-oriented).
The two rules aim at avoiding “utopian wishful thinking” at contributing to the
practicality of the plan’s implementation (and its monitoring and follow-up). They helped
to narrow down the scope of goals considerably, leaving the project team with 100
potential goals have that could facilitate advancements in the thirty identified priority
areas. Twenty-four of the 100 concrete goals were mainly results-oriented and therefore
MAPPING OUT THE TRANSITION 224
quantifiable, while the remaining seventy-six goals were mainly action-oriented and more
qualitative in nature. The project team was responsible to assure a harmonized and
understandable formulation of the selected 100 goals and for the respective translation in
English, Spanish and Portuguese.
The goals were submitted to the regional stakeholder community for ranking during
July and August 2007. In this fourth round of the exercise, 618 contributions were
received with a view to fine-tuning the contents of the goals. For the quantifiable results-
oriented goals, the participants were asked to identify a target number, expressed in
absolute and relative terms. For example: “Train [select from the following four choices:
100% / 75% / 50% / 25%] of teachers in the use of ICT or [select from the following five
choices: increase by ⅓ / increase by one half / increase by ⅔ / double / triple] the share
of so-trained teachers.” The identified target numbers are expressed as regional averages
for Latin America and the Caribbean as a whole. For the qualitative action-oriented goals,
participants were asked to evaluate each activity’s importance to the regional
development agenda up to 2010, on a Likert scale from one (not important at all) to five
(very important), and to identify agencies that are active in the field and could potentially
help to implement them.
The fifth and final round of the Delphi exercise was a face-to-face consultation
with the main public- and private-sector agencies and international NGOs. This inter-
institutional meeting took place on 12 September 2007 at ECLAC headquarters in
Santiago, Chile. Sixty-four experts from fifty institutions helped to refine the priority
agenda for 2010. Twenty of these institutions were intergovernmental organizations from
the region, eighteen were public sector institutions, ten were private sector foundations
MAPPING OUT THE TRANSITION 225
and entities, nine were civil society institutions and NGOs, and seven were academic
networks and institutions active in the region. The final proposal for a new Regional
Action Plan contained sixty-three concrete goals in six Chapters - in comparison with
eLAC2007, which included seventy goals in five Chapters (see Figure 1 for a
schematization of the five rounds and Annex for a comparison between eLAC2007 and
eLAC2010).

Evaluation of the effectiveness of the Delphi: the acceptance by the
inter-governmental process
The complete report, with its final proposals, was presented at the regional
consultations that took place in Buenos Aires on 4 and 5 October 2007. These
consultations served as input for the inter-governmental negotiations leading up to the
Ministerial Conference on the Information Society in Latin America and the Caribbean,
which was held in San Salvador on 6-8 February 2008. This inter-governmental
Conference approved the new Action Plan eLAC2010 – this is the substantive part of the
so-called “San Salvador Commitment”
89
. Based on the sixty-three goals of the Delphi
proposal, eLAC2010 contains eighty-three concrete goals following the same six Chapter
format proposed by the eLAC Policy Priority Delphi result. The following sections will
explore the changes, additions and droppings of the goals that were identified in the
Policy Delphi, as compared with the officially approved Action Plan. This will then allow
us to form an opinion about the effectiveness of the exercise, and therefore its success or
failure.

89
For a complete version of the San Salvador Commitment see: http://www.eLAC2007.org.sv
MAPPING OUT THE TRANSITION 226
The evolution of goals depicted in Figure 4 suggests that an interval of three years
can be a reasonable period to review a policy agenda in fields related to rapid
technological change. Only 20% of the goals in eLAC2010 are very similar to goals in
eLAC2007 (almost literally adopted, including minor changes without substantive
importance). Half of the goals have been adjusted to a changing environment. Around
30% of the goals of eLAC2010 are completely new on the agenda, with no equivalent in
the old Action Plan (see Annex). The figure also reveals that 60% of the goals of the
2008-2010 ICT policy agenda (fifty out of the eighty-three goals) were already found in
the proposal of the Policy Delphi results and were maintained by the policy makers
throughout the inter-governmental negotiations until their final appearance in the
officially approved version of eLAC2010
90
.

Figure 27 Goals of eLAC2010 in comparison with goals of eLAC2007

90
This is in spite of their being no direct recognition, in the final political document approved by the
Regional Ministerial Conference, of the input and contributions made by the regional stakeholder
community towards the elaboration of the goals through the Policy Delphi.
MAPPING OUT THE TRANSITION 227

Figure 5 takes a closer look at the relation between the eLAC Policy Priorities
Delphi results and the final political agenda eLAC2010 agreed on an inter-governmental
basis. It shows that thirty-eight of the eighty-three goals of eLAC2010 have been adopted
by the inter-governmental Conference almost literally from the Delphi results (46%).
Notably, the quantitative, measurable goals stand out among these. It seems that
governmental policy-makers respected the expertise and knowledge of the wider
community of Delphi participants in setting realistic quantitative benchmarks for
connectivity and other targets. Only thirteen goals (13%) proposed by the Policy Delphi
exercise were actually rejected by the governments (in the sense of their not appearing in
eLAC2010). Governmental representatives also brought their own current (and
sometimes very specific) concerns to the table and introduced sixteen new goals, such as
the creation of a regional education content market, promotion of IPv6, geo-referenced
information systems, technological waste and garbage, among others (see sixth column in
Annex). Even though most of the newly introduced goals (nine) did not count with a
concrete policy formulation in the final Delphi proposal, they are in agreement with the
thematic areas identified by the second Delphi round. Then again, the governmental
representatives resuscitated seventeen goals from the old eLAC2007 agenda (which
makes up 20% of the new eLAC2010). It seems that decision-makers were comfortable
with the language and formulations previously agreed around these earlier -debated goals,
and/or they did not want to tamper and meddle with the established political consensus.
The resuscitation of these goals makes the eLAC2010 agenda much closer to its
predecessor eLAC2007. The Delphi proposal was considerably more radical in its
MAPPING OUT THE TRANSITION 228
suggestions for change. But political processes seem to favour the safeguarding of an
established and balanced consensus, as opposed to endorsing major change – at least on
this occasion.
Two conclusions can be drawn from Figure 5. First, as a result of the political
desire to resuscitate old goals from eLAC2007 and at the same time to introduce new
goals, eLAC2010 is longer than its predecessor (eighty-three instead of seventy goals).
This follows the political logic of maintaining and adding to the original formulations,
without being willing to reduce, simplify and prioritize.
Second, a considerable number of goals that appear in eLAC2010 are in agreement
with the broad thematic areas that were identified in the second Delphi round, but that do
not appear as a concrete goal suggestion in the final Delphi proposal. (This applies to
nine of the goals that have been “newly invented” by policy makers, and fifteen of the
goals resuscitated from eLAC2007). This suggests that final rounds of the Delphi (rounds
three to five) had difficulty in translating the identified thematic areas into the desired
policy goals. For example, the second Delphi round identified the thematic area of
“Regional infrastructure and interconnection of networks among countries” as one of the
top priority areas (rank 11). But the policy goals elaborated during rounds three to five of
the Delphi did not foresee the necessity of promoting IPv6 (Internet Protocol version6)
91

as part of this challenge. Similarly, the thematic area of “Legislative Frameworks” was
prioritized during the second Delphi round (rank 21). But successive consultations
during Delphi rounds three to five did not result in the formulation of separate policy

91
Internet Protocol version 6 (IPv6) is designated as the successor of IPv4, the current version of the
Internet Protocol, for general use on the Internet. The main change brought by IPv6 is a much larger
address space that allows greater flexibility in assigning addresses.
MAPPING OUT THE TRANSITION 229
goals for “digital signature”, “electronic payment”, or “electronic contracting”, which
have been introduced by policy makers during the inter-governmental negotiations during
the Ministerial Conference in February 2008. In short, while the Delphi was able to
anticipate the majority of the broad interests of policy makers, it was not, then, able to
anticipate how these thematic areas would translate into concrete policy actions. There
appear to be limits to the scope of elaborating qualitative policy goal formulations in a
collective manner, at least when resources limited. The 180 personal interviews might not
have been sufficient (in number or range) and the methodology to collectively formulate
policy goals through digital media (such as intended during the fourth Delphi round),
might not have been sufficiently sophisticated to assure that the broad thematic areas of
interest would be translated into all relevant policy goals.

MAPPING OUT THE TRANSITION 230
Figure 28 Goals of eLAC2010 in comparison to goals proposed by Policy Delphi
results

Let us consider the nine goals that governments included in the eLAC2010 Action
Plan, though they had not been among the priorities identified by the Policy Delphi – and
also take a look at the thirteen policy options rejected by the public officials. This will
give us further insight on the dynamic between the Policy Delphi and its main
beneficiaries, the Ministerial Conference.
Among the nine goals that have been adopted by the Ministers against the opinion
of the Delphi participants, the politically relevant goals of “Internet Governance” (goal
72 of eLAC2010, with a very high disagreement of 0.30 P.I. average among subregions
and sectors in the second Delphi round), “software development” (goal 74 of eLAC2010,
middle P.I. of 0.16) and “hardware and industry development” (goals 50 and 53,
eLAC2010, high consensus with P.I. of 0.09) stand out. During recent years both of these
thematic areas have received much political visibility and attracted above-average
Very similar
38
With adjustments
12
coherent with
Delphi priorities
15
not identified as
Delphi priorities
2
coherent with
Delphi priorities
9
not identified as
Delphi priorities
7
Adopted from
Delphi
Resuscitated
from
eLAC2007
Newly
invented
Proposed by
Delphi, but
not adopted
13
MAPPING OUT THE TRANSITION 231
political attention in the region. To justify the inclusion of these goals despite the Delphi
proposal, governmental officials argued that the wider ICT-for-development community
might not fully recognize the importance of Internet Governance for development of
sustainable Information Societies, and likewise the strategic importance of ICT
production capacities
92
. Though the wider stakeholder group of Delphi participants
considers that these issues do not have a major impact on regional development in the
years to come, governmental representatives insisted on these issues for development in
the digital age.
With regard to the thirteen policy goals proposed by the eLAC Policy Priorities
Delphi that the Ministerial Conference rejected, no straightforward explanation was given
at the time, nor is it easy to find a common thread. But some points are notable:
 Some of the newly suggested priorities have been rejected by the inter-
governmental group. A policy action to foster “distance and telemedicine” (goal 25 of the
final Delphi proposal, and a slightly elevated P.I. of 0.23) was rejected, as were the
regulatory debates about “Voice-over-Internet-Protocol” (goal 18 of the final Delphi
proposal, with a rather low P.I. of 0.13) and “intellectual property” (goal 49 of the final
Delphi proposal, with a middle P.I. of 0.20). The latter two are critical issues of public
concern, but are also delicate issues on the international policy agenda, involving strong
industry interests. Even though the Polarization Index shows that there was no

92
During the negotiations, one governmental official used a metaphor to justify the inclusion of these goals.
According to some Latin American narratives, the indigenous inhabitants of the region had not been able to
recognize the ships of Christopher Columbus and other colonizers during their week-long anchoring in
front of the coasts of the Americas. They had never seen such a thing as a ship on the open ocean, and their
cognitive processes could not classify the phenomenon - leading to the well-known bloody result of the
unforeseen surprise visit. Internet Governance and the strategic importance of software (especially open
source) might be compared to such unknown and cognitively incomprehensible phenomena and –according
to the government official— it should not be a surprise that the larger public would not recognize their
importance.
MAPPING OUT THE TRANSITION 232
extraordinary disagreement among the Delphi community on these issues, decision
makers were not able to find common ground. It might be that the positive spirit of
consensus-building did not leave room to bring these discussions to an agreeable result
for the 2008-2010 policy debate.
 Five of the thirteen rejected goals concerned democratic and transparent
governance. One of these concerned the modernization of the justice system, including
the introduction of digital tools for transparency and judicial efficiency (goal 33 of the
final Delphi proposal, with an elevated P.I. of 0.25 among subregions and sectors); two
referred to the strengthening of democratic practices, including the usage of ICT in
parliaments and the approval of freedom of information legislation (goals 36 and 60 of
the final Delphi proposal, with an elevated P.I. of 0.24); and two goals concerned privacy
issues and the protection of personal data (goals 35 and 61 of the final Delphi proposal,
also with an elevated P.I. of 0.25). The Polarization Index shows clearly how
controversial these issues are. However, it is surprising that in a Regional Action Plan for
Information Society development, crucial topics associated with the strengthening of
democratic institutions and practices, the transparency of the judicial system, and the
protection of privacy rights, did not join the eighty-three priority issues. It is a cause for
concern that the LAC developing countries do not consider these issues in their current
policy agenda.

Conclusions and lessons learned
In retrospect, it is possible to identify a number of issues that could have been done
differently, and might have improved the effectiveness of the exercise. One of the mayor
MAPPING OUT THE TRANSITION 233
challenges was the formulation of concrete policy options to implement the identified
thematic priority areas. The received online comments turned out not to be very useful to
find the adequate wording of concrete policy actions, and more personal interviews or
workshops during the face-to-face meetings in round three and five would have required
more resources. It might have been a cost-effective intermediate solution to establish a
multi-sector editorial board to assist and supervise the project team’s work on synthesising
online comments and the results of face-to-face interviews. If resources are available, this
group could also serve as a focus group of regional opinion leaders. Sequential workshops
could provide regular input and guidance throughout the exercise. A very positive
externality of this approach is the creation of a group of regional agents of change. Future
exercises should consider the creation of such board to accompany the process.
Several design options have been influenced by trade-offs between the “theoretically
desirable” and what was thought to be “politically practicable”. One such choice of design
related to the right balance between the logic of a Policy Delphi to generate opposing views
and the need for a political consensus to go ahead with the inter-governmental negotiation
of a common LAC Action Plan. Given that the eLAC process has matured decisively and
has entered its second generation already, it might be possible and beneficial for future
exercises to deepen analysis and to focus with more detail on the disagreements of
participants. In the same line of reasoning it would surely be of analytical interest to
employ some kind of mechanism or software to register online users anonymously (such as
with a username and password). This would allow tracking the evolution of disagreements
over several rounds and the stability of the emerging consent or dissent. Contrary to all
initial concerns, the LAC stakeholder community reacted very positively to the chance of
MAPPING OUT THE TRANSITION 234
participation, which seems to suggest that the addition of some minor user hurdles that
enrich later analysis (such as registration or additional background questions) might not
necessarily lower participant turn out.
Besides lessons learned on design issues, the exercise discussed here also led to
several insights about the nature and potential of foresight exercises in developing
countries. The eLAC Policy Priorities Delphi brought a considerable amount of
transparency and accountability, by introducing public debate into the traditionally obscure
and somewhat arbitrary nature of inter-governmental agreements. Decision-makers found
themselves asked to justify publicly why they rejected and preferred certain thematic
priorities - though they did not always respond to this request. In this sense, the use of
foresight tools to enhance participative policy-making in inter-governmental processes is
not a quick fix or magic bullet for the longstanding challenges of more democratic and
transparent approaches to policy-making. It is a gradual innovation, which respects
established customs and procedures of inter-governmental decision-making. While the
overwhelming bulk of the results of the Policy Delphi results were accepted by the inter-
governmental power structures, the open-ended Delphi community did not replace
traditional decision-making mechanisms - nor could it remedy all of their defects. The
eLAC Policy Delphi supported public decision making by providing a more open and
transparent mechanism, but governments remained free to follow what they see as their
given mandates and to act as they see fit.
The Policy Delphi did highlight points of mismatch between governmental opinions
and the result of the open-ended multi-stakeholder consultations. For example, the newly
approved Regional Action Plan for the Information Society in Latin America and the
MAPPING OUT THE TRANSITION 235
Caribbean, eLAC2010, does not mention the issues (identified as crucial by the Delphi
participants) related to the strengthening of democratic institutions and practices, the
transparency and efficiency of the judicial system and the protection of privacy. This is a
significant mismatch, whose wider implications deserve more consideration than we can
give them now. This “democratic deficit” would have been much less visible if there had
been no Policy Delphi exercise to urge the inclusion of these issues in the public policy
agenda. The fact that the inter-governmental Ministerial Conference rejected five
democracy-related policy actions that have been proposed by the multi-stakeholder group
provides rather tangible evidence of political perspectives. Such tangible evidence would
be less apparent were all arguments and decisions taking place in the opaque channels of
inter-governmental negotiations – the common situation in such political decision-
making. Thus, Policy Delphi exercises can help to augment transparency and
accountability in public decision-making by simply reveal mismatches of opinion
between political leaders and stakeholders. This is of special importance in developing
countries, where institutional structures are frequently immature. In this sense, the
exercise can serve as a demonstration of a cost-effective way to foster transparent and
accountable public decision-making - and not only in developing countries.
Furthermore, the potential of participatory mechanisms was not just a theoretical
matter, but has already shown practical results. The open-ended consultations during the
five consecutive Delphi rounds strengthened a network of stakeholders and institutions
that are involved in complementary tasks related to the areas of interest of the Regional
Action Plan. As a result, the Annex of the eLAC2010 Action Plan lists eighty-eight
regional agencies that are active in the various challenges outlined by the regional
MAPPING OUT THE TRANSITION 236
strategy. This Annex can be seen as a first “who-is-who” and “who-does-what” in the
LAC ICT-for-development community. This incipient multi-agency networking is surely
one of the most valuable results of the exercise. Such approaches are likely to be
especially important in cross-cutting and multi-thematic areas as ICT-for-development. In
this sense - as is so often the case in Delphi exercises and in Foresight more generally -
the process itself turned out to be just as important, or maybe even more important, for
advancement on the ground, than the final product presented in the report.
One reason for the acceptance and success of the exercise lies in its positioning and
presentation. The Policy Delphi was not seen to be questioning the legitimacy of
representative democracy and established multilateralism. Rather, the process aimed at
enriching its functionality, at supporting established inter-governmental practices in the
framework of the United Nations system (in form of its Regional Commission UN-
ECLAC). This is the basic ambition of participatory policy-making in a representative
democracy, where the aim is more for incremental change than for a radical break with
traditional notions of democracy. Nevertheless, it does recognize that there are techniques
and technological opportunities to gradually modernize the relationship between State
and societal actors, and it has been shown that traditional policy makers are receptive to
such innovations.
These changes involve a growing prominence of such notions as stakeholder and of
shared responsibility. Civil society, private and academic sectors and governments are
viewed as different parts of society that affect and are affected by the public policy
making process, and can contribute their knowledge and take responsibility in a
collective way through public foresight consultations. The process relies on the
MAPPING OUT THE TRANSITION 237
involvement of publicly legitimized technocrats, but recognizes that they cannot possibly
possess all the information required to make sound policy choices in such a dynamic and
generic topic as ICT-for-development. It therefore mobilizes the collective intelligence of
an open-ended stakeholder group, while respecting the established legitimization of
democratically elected governments, following the basic definition of a Policy Delphi as
a “decision-analysis tool”, and not a decision making tool.
Additionally, the eLAC Policy Priorities Delphi is not only about ICT-for-
development, but it also exploits the ICTs` involved, drawing on the benefits of digital
communication. This helps to overcome geographical barriers and provides a 24/7
availability of the online platform. Cheap channels for participation are essential in Latin
America and the Caribbean, given the region’s large size and scarce resources. For
example, while a four and a half hours flight between Stockholm and Lisbon represents
one of the largest barriers to a face-to-face meeting in Europe (3,000 km), a flight from
Mexico City to Buenos Aires easily takes over eleven hours (7,500 km). Thus, every
personal meeting with regional scope in Latin America and the Caribbean requires at
least one day of travel each way for all participants to be able to come, whereas, in
Europe, most participants in European Union or OECD meetings can go to a meeting and
return the same day. Combined with the cost of travel and the much more limited
resources in developing countries – not to mention environmental and personal burdens
associated with long-distance travel - virtual channels prove to have great potential to
facilitate participatory policy-making. There is bound to be much more use of such
methods in coming years.
MAPPING OUT THE TRANSITION 238
Last but not least, the most important lesson learned might be the contribution of
international collaboration in this exercise. The elaboration of a Public Policy Action
Plan in the fast-changing field of innovation and technological change is an ongoing
challenge that can be confronted internationally. Technological progress is a moving
target, but many developing countries do not have sufficient resources to maintain
continuous foresight exercises. In this case, the support from Europe (in form of the funds
from the European Commission’s @LIS project and the conceptual collaboration
between UN-ECLAC and the much more experienced team at the University of
Manchester, UK), and the South-South deliberations within Latin America and the
Caribbean, has shown that international cooperation provides an adequate platform and
sufficient scale for developing countries to adjust their policy actions to permanent
technological change. Foresight exercises are resource intensive and if developing
countries want to prepare better for the future, international collaboration at the regional
level seems like the most feasible level to start with an institutionalization of such
exercises in the developing world.

MAPPING OUT THE TRANSITION 239

References
ADSIB. (2010). Agencia para el Desarrollo de la Información en Bolivia. Retrieved from
http://www.adsib.gob.bo
Aguero, A. (2008). Telecommunications expenditure in peruvian households. Research
Briefs, Ottawa: DIRSI (Regional Dialogue on the Information Society), IDRC. Retrieved
from http://www.dirsi.net/english/content/view/148/72/.
Albert, R., & Barabasi, A.-L. (2002). Statistical mechanics of complex networks.
Reviews of Modern Physics, 74(1), 47. doi:10.1103/RevModPhys.74.47
Anand, S., & Sen , A. (1995). Gender Inequality in Human Development. In Occasional
Paper 19 (Vol. 1995). Presented at the Gender inequality in human development: theories
and measurement, United Nations.
Andrés, L., Cuberes, D., Diouf, M., & Serebrisky, T. (2010). The diffusion of the
Internet: A cross-country analysis. Telecommunications Policy, 34(5-6), 323-340.
doi:10.1016/j.telpol.2010.01.003
Arun, S., & Arun, T. (2002). ICTs, gender and development: women in software
production in Kerala. Journal of International Development, 14(1), 39-50.
Badagliacco, J. M. (1990). Gender and Race Differences in Computing Attitudes and
Experience. Social Science Computer Review, 8(1), 42-63.
Banisar, D. (2006). Freedom of Information around the World: A Global Survey of
Access to Government Information Laws. Privacy International. Retrieved August 3,
2009, from http://www.freedominfo.org/documents/global_survey2006.pdf.
Baqir, N., Palvia, P., & Nemati, H. (2009). Evaluating government ICT policies: An
extended design-actuality gaps framework. Proceedings of Second Annual SIG GlobDev
Workshop. Phoenix, USA. Retrieved from
http://www.globdev.org/files/proceedings2009/21-FINAL_Baqir-
et_al_Evaluating_Government_ICT_2009.pdf
Bar, F., & Best, M. (2008). Announcing ITID 2.0. Information Technologies and
International Development, 4(4), iii-iv.
Barrantes, R., & Galperin, H. (2008). Can the poor afford mobile telephony? Evidence
from Latin America. Telecommunications Policy, 32(8), 521-530.
doi:10.1016/j.telpol.2008.06.002
MAPPING OUT THE TRANSITION 240
Barzilai-Nahon, Karine. 2006. Gaps and Bits: Conceptualizing Measurements for Digital
Divide/s. The Information Society 22 (5): 269-278.
Bass, F. M. (1969). A new product growth for model consumer durables. Management
Science, 15(5), 215-227. doi:10.1287/mnsc.15.5.215
Basu, M. (2000). Feminist Perspectives and Gender Link in Development: The Critical
Role of Women's Organisations. The Ralph Bunche Institute on the United Nations.
Retrieved August 17, 2009.
Battisti, G., & Stoneman, P. (2003). Inter- and intra-firm effects in the diffusion of new
process technology. Research Policy, 32(9), 1641-1655. doi:10.1016/S0048-
7333(03)00055-6
Battisti, G., Hollenstein, H., Stoneman, P., & Woerter, M. (2007). Inter- and intra-firm
diffusion of ICT in the United Kingdom (UK) and Switerland (CH): An internationally
comparative study on firm-level data. Economics of Innovation and New Technology,
16(8), 669-687. doi:10.1080/10438590600984026
Beilock, R., & Dimitrova, D. V. (2003). An exploratory model of inter-country Internet
diffusion. Telecommunications Policy, 27(3-4), 237-252. doi:10.1016/S0308-
5961(02)00100-3
Bell, D. (1976). The Coming Of Post-industrial Society. Basic Books.
Beniger, J. R. (1986). The control revolution: technological and economic origins of the
information society. Harvard University Press.
Best, M. L., & Maier, S. G. (2007). Gender, Culture and ICT Use in Rural South India.
Gender Technology and Development, 11(2), 137-155.
Billon, M., Marco, R., & Lera-Lopez, F. (2009). Disparities in ICT adoption: A
multidimensional approach to study the cross-country digital divide. Telecommunications
Policy, 33(10-11), 596-610. doi:10.1016/j.telpol.2009.08.006
Bimber, B. (2000). Measuring the gender gap on the internet. Social Science Quarterly,
81(3), 868–76.
Bohn, R., & Short, J. (2009). How Much Information? 2009 Report on American
Consumers. San Diego: Global Information Industry Center of University of California,
San Diego. Retrieved from http://hmi.ucsd.edu/howmuchinfo.php
Boiarov, S. (2008). Informe sobre legislación y normativa vinculada al teletrabajo en
América Latina y el Caribe. eLAC2007, goal 5. United Nations ECLAC, Information
Society Programme. Retrieved August 23, 2009, from
http://www.cepal.org/socinfo/noticias/noticias/2/32222/GdT_eLAC_meta_5.pdf.
MAPPING OUT THE TRANSITION 241
Bonfadelli, H. (2002). The Internet and Knowledge Gaps: A Theoretical and Empirical
Investigation. European Journal of Communication, 17(1), 65-84..
Brodman, J., & Berazneva, J. (2007). Transformina Opportunities for Women
Entrepreneurs. Information Technologies and International Development, Special issue
on Women's Empowerment and the Information Society, 4(2), 3-10.
Brynjolfsson, E., & Hitt, L. M. (1995). Beyond computation: Information technology,
organizational transformation and business performance. Journal of Economic
Perspectives, 14, 23-48.
Brynjolfsson, E., & Saunders, A. (2009). Wired for Innovation: How Information
Technology is Reshaping the Economy. USA: The MIT Press.
Busch, T. (1995). Gender differences in self-efficacy and attitudes toward computers.
Journal of Educational Computing Research, 12(2), 147-158.
Bush, V. (1945, July). As We May Think. The Atlantic Monthly, , 176(1). Retrieved
August 17, 2009, from http://www.theatlantic.com/doc/194507/bush.
Buys, P., Dasgupta, S., Thomas, T. S., & Wheeler, D. (2009). Determinants of a Digital
Divide in Sub-Saharan Africa: A Spatial Econometric Analysis of Cell Phone Coverage.
World Development, 37(9), 1494-1505.
Castells, M. (1996). The Rise of the Network Society. Blackwell Pub.
Castells, M. (1998). End of Millennium (Castells, Manuel. Information Age, V. 3,).
Blackwell Pub.
Castells, M. (2004). The Power of Identity: The Information Age: Economy, Society and
Culture, Volume II (The Information Age) 2nd Edition (2nd ed.). Wiley-Blackwell.
Castells, M. (2009). The Rise of the Network Society: The Information Age: Economy,
Society, and Culture Volume I (2nd ed.). Wiley-Blackwell.
Castells, M., Fernandez-Ardevol, M., Qiu, J. L., & Sey, A. (2009). Mobile
communication and society: A global perspective. The MIT Press.
CEPAL (Comisión Económica para America Latina y el Caribe). (2009). CEPALSTAT.
CEPALSTAT Estadisticas de America Latina y el Caribe. Retrieved August 7, 2009,
from http://websie.eclac.cl/sisgen/ConsultaIntegrada.asp.
Chaudhuri, A., Flamm, K. S., & Horrigan, J. (2005). An analysis of the determinants of
internet access. Telecommunications Policy, 29(9-10), 731-755.
doi:10.1016/j.telpol.2005.07.001
Chopra, A. (2010, June). Round 1 Post-Award Recipient Workshop | BroadbandUSA -
NTIA. Chief Technology Officer and Associate Director for Technology, Office of
MAPPING OUT THE TRANSITION 242
Science and Technology Policy presented at the Broadband Technology Opportunities
Program (BTOP). Retrieved from http://www2.ntia.doc.gov/JuneWorkshop
Cimoli, M., Hofman, A., & Mulder, N. (Eds.). (2010). The Impact of Information and
Communication Technologies in Latin America (In Association with the United Nations
Economic Commission for Latin America and the Caribbean.). Edward Elgar. Retrieved
from http://www.e-elgar.co.uk/Bookentry_Main.lasso?id=13996
Cohen, E. S. (2008). Broadband Internet: Access, regulation and policy. Nova Publishers,
New York.
Cole, S. (1986). The global impact of information technology. World Development,
14(10-11), 1277-1292.
Coleman, S. (2005). The Lonely Citizen: Indirect Representation in an Age of Networks.
Political Communication, 22, 197-214.
Compaine, B. M. (2001). The digital divide: facing a crisis or creating a myth? MIT
Press.
Congress 107th. (2002). E-Government Act of 2002; Public Law 107–347—Dec. 17
2002, 116 STAT. 2899. Retrieved from http://frwebgate.access.gpo.gov/cgi-
bin/getdoc.cgi?dbname=107_cong_public_laws&docid=f:publ347.107.pdf
Corrocher, N., & Ordanini, A. (2002). Measuring the digital divide: A framework for the
analysis of cross-country differences. Journal of Information Technology, 17(1), 9.
doi:10.1080/02683960210132061
Cover, T. M., & Thomas, J. A. (2006). Elements of Information Theory (2nd ed.).
Hoboken, NJ: Wiley-Interscience.
Cullen, R. (2001). Addressing the digital divide. Online Information Review, 25(5), 311 -
320.
Cummings, J. N., & Kraut, R. (2002). Domesticating Computers and the Internet. The
Information Society: An International Journal, 18(3), 221.
de Tarde, G. (1903). The laws of imitation. H. Holt and Company, New York.
Dholakia, R. R., Dholakia, N., & Pedersen, B. (1994). Putting a byte in the gender gap.
American Demographics, 16(12), 20.
DIPRES. (2005). Informe de cuantificacion del gasto en gobierno electronico, Ano 2003.
Direccion de Presupuesto, Ministerio de Hacienda, Gobierno de Chile. Retrieved from
http://www.dipres.cl/572/articles-21677_doc_pdf.pdf
Dosi, G. (1988). Sources, Procedures and Microeconomic Effects of Innovation. Journal
of Economic Literature 26: 1120-1171.
MAPPING OUT THE TRANSITION 243
Duff, A. S. (2000). Information society studies. Psychology Press.
Duff, A. S., Craig, D., & McNeill, D. A. (1996). A note on the origins of the “information
society.” Journal of Information Science, 22(2), 117 -122.
doi:10.1177/016555159602200204
Dutta, Soumitra, Augusto López-Claros, Irene Mia. 2006. The Global Information
Technology Report, 2005-2006: Leveraging ICT for Development (INSEAD and World
Economic Forum report). Basingstoke, UK: Palgrave Macmillan.
Dutton, W. H., Gillett, S. E., McKnight, L. W., & Peltu, M. (2004). Bridging broadband
Internet divides: reconfiguring access to enhance communicative power. Journal of
Information Technology, 19(1), 28-38. doi:10.1057/palgrave.jit.2000007
ECLAC (Economic Commission for Latin America and the Caribbean). (2008). Social
Panorama of Latin America 2008. Social Panorama of Latin America. Santiago Chile:
United Nations ECLAC. Retrieved August 4, 2009, from
http://www.eclac.org/publicaciones/default.asp?idioma=IN.
ECLAC Resolution (2006), Resolution 629 (XXXI), Report of the Thirty-First Session of
the Commission, Montevideo, 20-24 March 2006, [online]
http://www.cepal.org/pses31/default.asp?idioma=IN , 2006.
ECLAC, DIRSI, & UNDP. (2008). Digital Review 2007 of Latin America and the
Caribbean:Progress and Challenges Policies for Development with ICT-Abridged
document (Information Society Programme No. 37). Santiago: United Nations ECLAC.
Retrieved from http://www.cepal.org/SocInfo/publicaciones
eLAC Delphi Report (2007), eLAC Policies Priorities Delphi, Martin Hilbert and Julia
Othmer, September 2007, Information Society Programme, Economic Commission for
Latin America and the Caribbean, United Nations, ECLAC, [online]
http://www.eclac.org/cgi-
bin/getProd.asp?xml=/socinfo/noticias/noticias/5/29955/P29955.xml&xsl=/socinfo/tpl-
i/p1f.xsl&base=/socinfo/tpl-i/top-bottom.xslt , 2007.
Engel, E. (1857). Die Productions- und Consumptionsverhaeltnisse des Koenigsreichs
Sachsen. Zeitschrift des Statistischen Bureaus des Koniglich Sachsischen Ministeriums
des Inneren, 8, 9.
Evans, L. B. (1977). Impact of the Electronics Revolution on Industrial Process Control.
Science, 195(4283), 1146-1151. doi:10.1126/science.195.4283.1146
Fallows, D. (2005). How Women and Men Use the Internet | Pew Internet & American
Life Project. Pew Internet and American Life Project. Retrieved August 16, 2009, from
http://www.pewinternet.org/Reports/2005/How-Women-and-Men-Use-the-Internet.aspx.
Faulkner, W. (2001). The technology question in feminism : : A view from feminist
technology studies. Women’s Studies International Forum, 24(1), 79-95.
MAPPING OUT THE TRANSITION 244
Fisher, R. (1936). The use of multiple measurements in taxonomic problems. Annals of
Eugenics, 7(2), 188, 179.
Flamm, K., & Chaudhuri, A. (2007). An analysis of the determinants of broadband
access. Telecommunications Policy, 31(6-7), 312-326. doi:10.1016/j.telpol.2007.05.006
Forester, T. (1981). The Microelectronics Revolution. The MIT Press.
Forester, T. (1985). The Information Technology Revolution (1st ed.). Cambridge, MA:
The MIT Press.
Fortune. (2008). Fortune 500. Retrieved August 7, 2009, from
http://money.cnn.com/magazines/fortune/fortune500/2008/full_list/.
Freedman, D., Purves, R., & Pisani, R. (2007). Statistics, 4th Edition (4th ed.). W.W.
Norton & Co.
Freeman, C., & Louca, F. (2002). As Time Goes By: From the Industrial Revolutions to
the Information Revolution. USA, Oxford University Press.
Frobenius, L. (1897). Der westafrikanische Kulturkreis. Petermanns Mitteilungen, 43/44,
225-271.
Frost and Sullivan Research Group. (2006). Social Impact of Mobile Telephony in Latin
America (GSM Association in Latin America and AHCIET report). New York: Frost and
Sullivan Research Group.
Galperin, H. (2010). Goodbye digital divide, Hello digital confusion? A critical embrace
of the emerging ICT4D consensus. Information Technologies and International
Development, 6 Special Edition, 53-55.
Galperin, H., & Mariscal, J. (2007). Mobile Opportunities: Poverty and Mobile
Telephony in Latin America and the Caribbean. Ottawa: DIRSI (Regional Dialogue on
the Information Society), IDRC. Retrieved August 7, 2009, from
http://www.dirsi.net/english/content/view/139/69/.
Galperin, H., & Mariscal, J. (Eds.). (2007). Digital Poverty: Latin America and the
Caribbean perspectives. Ottawa: DIRSI (Regional Dialogue on the Information Society),
IDRC. Retrieved August 5, 2009, from http://www.dirsi.net/english/content/view/21/41/.
Gamboa, L. F. (2007). Patterns of access and analysis of spending on cellular mobile
telephone service in Colombia, 2001-2006. . Research Briefs, Ottawa: DIRSI (Regional
Dialogue on the Information Society), IDRC. Retrieved August 5, 2009, from
http://www.dirsi.net/english/content/view/148/72/.
Gantz, J., Chute, C., Manfrediz, A., Minton, S., Reinsel, D., Schlichting, W., &
Toncheva, A. (2008). The Diverse and Exploding Digital Universe: An Updated Forecast
of Worldwide Information Growth Through 2011. Framingham: IDC (International Data
MAPPING OUT THE TRANSITION 245
Corporation) sponsored by EMC. Retrieved from http://www.emc.com/leadership/digital-
universe/expanding-digital-universe.htm
Gartner. (2008, July 28). Gartner Says the $100 Laptop Is at Least Three Years Away.
Press Center. Retrieved August 4, 2009, from
http://www.gartner.com/it/page.jsp?id=733316.
Gillwald, A., & Stork, C. (2008). Towards Evidence Based ICT Policy and Regulation:
ICT access and usage in Africa (Policy Paper 2 No. Vol. 1). ResearchICTAfrica (RIA).
Retrieved from http://www.researchictafrica.net.
Gladwell, M. (2002). The tipping point: How little things can make a big difference.
Back Bay Books.
Granovetter, M. S. (1973). The strength of weak ties. American Journal of Sociology,
78(6), 1360. doi:10.1086/225469
GSM Association. 2006. Universal Access: How Mobile can Bring Communications to
all. Atlanta, GA: GSM Association.
Guerra, M., & Jordan, V. (2010). Public policies for the information society: a shared
vision? ( No. LC/W.314). Information Society Program (p. 28). United Nations ECLAC.
Retrieved from http://www.cepal.org/socinfo
Guerra, Massiel, Martin Hilbert, Valeria Jordán, and Christian Nicolai. (2008). Digital
Review 2007 of Latin America and the Caribbean: Progress and Challenges Policies for
Development with ICT (Abridged version). New York: United Nations. Available at:
http://www.unctad.org/sections/dite_dir/docs/dite_pcbb_stdev0126_en.pdf (accessed
January 25, 2010).
Guerrieri, P., & Padoan, P. C. (2007). Modelling ICT as a General Purpose Technology:
Evaluation Models and Tools for Assessment of Innovation and Sustainable
Development at the EU Level (Special Edition No. 35). Collegium (p. 196).
Hafkin, N. (2000). Convergence of Concepts: Gender and ICTs in Africa. In E.
Rathgeber & E. Ofwana Adera (Eds.), Gender and the information revolution in Africa
(pp. Chapter 1, document 4). Ottawa: IDRC. Retrieved August 17, 2009, from
http://www.idrc.ca/en/ev-32939-201-1-DO_TOPIC.html.
Hafkin, N., & Huyer, S. (2007). Women and Gender in ICT Statistics and Indicators for
Development. Information Technologies and International Development, Special issue on
Women's Empowerment and the Information Society, 4(2), 25-41.
Hanafizadeh, M. R., Saghaei, A., & Hanafizadeh, P. (2009). An index for cross-country
analysis of ICT infrastructure and access. Telecommunications Policy, 33(7), 385-405.
doi:10.1016/j.telpol.2009.03.008
MAPPING OUT THE TRANSITION 246
Hanafizadeh, Payam, Mohammad Reza Hanafizadeh, and Mohsen Khodabakhshi. 2009.
Extracting Core ICT Indicators Using Entropy Method. The Information Society 25 (4):
236 - 247.
Harcourt, W. (Ed.). (1999). Women@Internet (p. 240). Palgrave Macmillan.
Hargittai, E., & Shafer, S. (2006). Differences in Actual and Perceived Online Skills: The
Role of Gender. Social Science Quarterly, 87(2), 432-448.
Hawkins, E. T. (2005). Creating a national strategy for Internet development in Chile.
Telecommunications Policy, 29(5-6), 351-365. doi:10.1016/j.telpol.2004.12.005
Heeks, R. (2002). Information systems and developing countries: Failure, success, and
local improvisations. The Information Society: An International Journal, 18(2), 101.
doi:10.1080/01972240290075039
Heeks, R. (2009). The ICT4D 2.0 Manifesto: Where next for ICTs and international
development? (Development Informatics Group). Institute for Development Policy and
Management. Institute for Development Policy and Management: University of
Manchester. Retrieved from
http://www.sed.manchester.ac.uk/idpm/research/publications/wp/di/index.htm
Heeks, R., Arun, S., & Morgan, S. (2004). Researching ICT-based enterprise for women
in developing countries: a gender perspective. In Women's ICT-Based Enterprise for
Development project. Institute for Development Policy and Management (IDPM),
Manchester. Retrieved August 17, 2009, from
http://www.womenictenterprise.org/LivelihoodsResearch.doc.
Hennessy, John L., David A. Patterson, and Andrea C. Arpaci-Dusseau. 2007. Computer
Architecture: A Quantitative Approach (4th edition). Boston: Morgan Kaufmann.
Hilbert, M. & Katz, J. (2003). Building an Information Society: a Latin American and
Caribbean Perspective (p. 336). United Nations ECLAC. Retrieved April 11, 2009, from
http://www.cepal.org/cgi-
bin/getProd.asp?xml=/publicaciones/xml/2/11672/P11672.xml&xsl=/ddpe/tpl-
i/p9f.xsl&base=/socinfo/tpl/top-bottom.xslt.
Hilbert, M. (2009). The Maturing Concept of E-Democracy: From E-Voting and Online
Consultations to Democratic Value Out of Jumbled Online Chatter. Journal of
Information Technology & Politics, 6(2), 87-110. doi:10.1080/19331680802715242
Hilbert, M. (2010). When is Cheap, Cheap Enough to Bridge the Digital Divide?
Modeling Income Related Structural Challenges of Technology Diffusion in Latin
America. World Development, 38(5), 756-770. doi:10.1016/j.worlddev.2009.11.019
Hilbert, M. (2010a). Digital gender divide or technologically empowered women in
developing countries? A typical case of lies, damned lies, and statistics. Women’s
MAPPING OUT THE TRANSITION 247
Studies International Forum, forthcoming. Retrieved from
http://www.martinhilbert.net/DigitalGenderDivide.pdf
Hilbert, M. (2011). The end justifies the definition: The manifold outlooks on the digital
divide and their practical usefulness for policy-making. Telecommunications Policy,
35(8), 715-736. doi:10.1016/j.telpol.2011.06.012
Hilbert, M. (2011a). Toward a synthesis of cognitive biases: How noisy information
processing can bias human decision making. Psychological Bulletin;Psychological
Bulletin, No Pagination Specified. doi:10.1037/a0025940
Hilbert, M., & Cario, O. (Eds.). (2009). ¿Quo vadis Tecnologia de la información y de las
comunicaciones? Conceptos fundamentales, trayectorias tecnológicas y el Estado del
Arte de los sistemas digitales (p. 372). Colombia: Mayol Ediciones. Retrieved August 8,
2009, from http://www.mayolediciones.com/portal/detallelibro.php?idlibro=633.
Hilbert, M., & López, P. (2011). The World’s Technological Capacity to Store,
Communicate, and Compute Information. Science, 332(6025), 60 -65.
doi:10.1126/science.1200970
Hilbert, M., & López, P. (2011b). The growing information inequality: Concentration
and international distribution of the world’s technological capacity to communicate, store
and compute information. inedito. Retrieved from http://www.martinhilbert.net/
Hilbert, M., & López, P. (2011c). How to measure the world’s technological capacity to
communicate, store and compute information? International Journal of Communication,
forthcoming. Retrieved from http://www.martinhilbert.net/
Hilbert, M., & Peres, W. (Eds.). (2010). Information Societies in Latin America and the
Caribbean: Development of Technologies and Technologies for Development. ECLAC
institutional books. Santiago: United Nations Economic Commission for Latin America
and the Caribbean. Retrieved from http://www.cepal.org/SocInfo/
Hilbert, M., López, P., & Vasquez, C. (2010). Information societies or “ICT equipment
societies”? Measuring the digital information processing capacity of a society in bits and
bytes. The Information Society, 26(3). Retrieved from
http://www.indiana.edu/~tisj/inpress/3/ab-hilbert.html
Hilbert, M., Miles, I., & Othmer, J. (2009). Foresight tools for participative policy-
making in inter-governmental processes in developing countries: Lessons learned from
the eLAC Policy Priorities Delphi. Technological Forecasting and Social Change, 76(7),
880-896. doi:10.1016/j.techfore.2009.01.001
Hilbert, Martin and Jorge Katz. (2003). Building an Information Society: A Latin
American and Caribbean Perspective (LC/L.1845 edition). Santiago, Chile: Economic
Commission for Latin America and the Caribbean, United Nations.
MAPPING OUT THE TRANSITION 248
Hollenstein, H., & Woerter, M. (2008). Inter- and intra-firm diffusion of technology: The
example of E-commerce: An analysis based on Swiss firm-level data. Research Policy,
37(3), 545-564. doi:10.1016/j.respol.2007.12.006
Howard, Philip N., Ken Anderson, Laura Busch, and Dawn Nafus. 2009. Sizing Up
Information Societies: Toward a Better Metric for the Cultures of ICT - Adoption. The
Information Society 25(3): 208-219.
Indjikian, R., & Siegel, D. S. (2005). The Impact of Investment in IT on Economic
Performance: Implications for Developing Countries. World Development, 33(5), 681-
700.
INSEAD, & WEF (World Economic Forum). (2009). The global Information technology
report 2008-2009. Retrieved from http://www.weforum.org/reports
Ito, Y. (1981). The Johoka Shakai approach to the study of communication in Japan. In
C. Wilhoit & H. de Bock (Eds.), Mass Communication Review Yearbook (Vol. Volume
2, pp. 671–698). Beverly Hills, CA: Sage.
ITU (2001). The Internet in the Andes: Bolivia case Study. Michael Minges, Sonia Jorge,
and Ben Petrazzini. Geneva: International Telecommunication Union. Retrieved from
http://www.itu.int/ITU-D/ict/cs/bolivia/bolivia.html
ITU (2009). Measuring the Information Society - The ICT Development Index 2009.
Geneva: International Telecommunications Union. Retrieved from
http://www.itu.int/ITU-D/ict/publications/idi/2009/index.html
ITU (International Telecommunication Union) and UNCTAD (United Nations
Conference on Trade and Development). 2007. World Information Society Report 2007:
Beyond WSIS. Geneva: United Nations.
ITU (International Telecommunication Union). (2007). Trends in Telecommunication
Reform 2007: The Road to Next-Generation Networks (NGN). Geneva. Retrieved
August 4, 2009, from http://www.itu.int/publ/D-REG-TTR.9-2007.
ITU (International Telecommunication Union). (2009). Measuring the Information
Society - The ICT Development Index 2009. Geneva. Retrieved August 4, 2009, from
http://www.itu.int/ITU-D/ict/publications/idi/2009/index.html.
ITU (International Telecommunication Union). (2010). Measuring the Information
Society 2010 - The ICT Development Index 2010. Geneva. Retrieved August 4, 2010,
from http://www.itu.int/publ/D-IND-ICTOI-2010/en.
ITU (International Telecommunication Union). 2006. Measuring ICT for Social and
Economic Development. Geneva: ITU Secretariat, United Nations.
MAPPING OUT THE TRANSITION 249
ITU (International Telecommunications Union). (2010). Measuring the Information
Society 2010. Geneva: International Telecommunication Union, ITU-D. Retrieved from
http://www.itu.int/publ/D-IND-ICTOI-2010/en
ITU (International Telecommunications Union). 1999. Challenges to the Network:
Internet for Development. Geneva: ITU Secretariat, United Nations.
ITU (International Telecommunications Union). 2007. World Telecommunication/ICT
Indicators Database. Geneva: ITU Secretariat, United Nations.
ITU (International Telecommunications Union). 2009. Measuring the Information
Society - The ICT Development Index 2009. Geneva.
J. Schneider, “The Policy Delphi: A Regional Planning Application”, Technological
Forecasting and Social Change 3, 481-497, 1972.
James, Jeffrey. 2008. Digital Divide Complacency: Misconceptions and Dangers. The
Information Society 24 (1): 54-61.
Joiner, R., Messer, D., Littleton, K., & Light, P. (1996). Gender, computer experience
and computer-based problem solving. Computers & Education, 26(1-3), 179-187.
Jorgenson, D. W. (2005). Accounting for Growth in the Information Age. Handbook of
Economic Growth (Vol. 1, pp. 743-815). Elsevier. Retrieved from
http://ideas.repec.org/h/eee/grochp/1-10.html
Jorgenson, Dale W., and Khuong Vu. 2005. Information Technology and the World
Economy. Scandinavian Journal of Economics 107 (4): 631-650.
Kaplan, R. (1994). The gender gap at the PC keyboard. American Demographics, 16(1),
18.
Katz, J. E. & Rice, R. E. (2002). Social consequences of Internet use: access,
involvement, and interaction. MIT Press.
Katz, J., & Hilbert, M. (2003). Road maps towards an information society in Latin
America and the Caribbean. ECLAC books. Santiago Chile: United Nations ECLAC
(Economic Commission for Latin America and the Caribbean). Retrieved April 11, 2009,
from http://www.cepal.org/cgi-
bin/getProd.asp?xml=/publicaciones/xml/0/12900/P12900.xml&xsl=/ddpe/tpl-
i/p9f.xsl&base=/socinfo/tpl-i/top-bottom.xslt.
Kauffman, R. J., & Techatassanasoontorn, A. A. (2009). Understanding early diffusion of
digital wireless phones. Telecommunications Policy, 33(8), 432-450.
doi:10.1016/j.telpol.2009.03.006
Kelkar, G., & Nathan, D. (2002). Gender Relations and Technological Change in Asia.
Current Sociology, 50(3), 427-441.
MAPPING OUT THE TRANSITION 250
Khalil, M., & Kenny, C. (2008). The next decade of ICT development: Access,
applications, and the forces of convergence. Information Technologies and International
Development, 4(3), 1-6.
Kingdon, J. W. (1995), Agendas, Alternatives, and Public Policies, New York,
HarperCollins, second edition.
Klein, Hans. (2004). Understanding WSIS: An Institutional Analysis of the UN World
Summit on the Information Society. Information Technologies and International
Development 1 (3-4): 3-13.
Krebs, V. (2003). Spread of contagions via contact tracing. visualcomplexity.com.
Retrieved August 8, 2010, from http://www.visualcomplexity.com/vc/project.cfm?id=28
Krebs, V. (2005). Emergent online community. visualcomplexity.com. Retrieved August
8, 2010, from
http://www.visualcomplexity.com/vc/project_details.cfm?index=27&id=139&domain=S
ocial%20Networks
Kvasny, Lynette. (2003) Virtual Inequality: Beyond the Digital Divide, by Karen
Mossberger, Caroline J. Tolbert, and Mary Stansbury. Washington, DC: Georgetown
University Press, 2003. Xvi + 192 Pp. $19.95/£14.25 (Paper). ISBN: 0-87840-999-8. The
Information Society 20, no. 5 (2004): 409.
Lamberton, D. M. (1974). Information Revolution. Amer Academy of Political &
Laplane, M., Rodrigues, F., Guitierrez, F., & Rojas, F. (2007). Asimetrías de información
en el mercado de computadoras personales: los casos de financiación de PC para
consumidores de bajos ingresos. Santiago Chile: United Nations ECLAC, Information
Society Programme. Retrieved August 6, 2009, from http://www.cepal.org/cgi-
bin/getProd.asp?xml=/publicaciones/xml/8/28528/P28528.xml&xsl=/ddpe/tpl/p9f.xsl&ba
se=/socinfo/tpl/top-bottom.xslt.
Lawton, O. (2010). Monitoring Caribbean information societies (Information Society
Program No. LC/W.315). Santiago: United Nations ECLAC Observatory for the
Information Society in Latin America and the Caribbean. Retrieved from
http://www.eclac.org/publicaciones/xml/9/38899/W315.pdf
Leggon, C. (2006). Gender, race/ethnicity, and the digital divide. In M. F. Fox, D. G.
Johnson, & S. V. Rosser (Eds.), Women, gender, and technology (p. 204). University of
Illinois Press.
Lenhart, A., Lewis, O., & Rainie, L. (2001). Teenage Life Online . Pew Internet &
American Life Project. Retrieved August 16, 2009, from
http://www.pewinternet.org/Reports/2001/Teenage-Life-Online.aspx.
Lesk, M. (1997). How Much Information Is There In the World? lesk.com. Retrieved
from http://www.lesk.com/mlesk/ksg97/ksg.html
MAPPING OUT THE TRANSITION 251
Lloyd, S. (2002). Computational Capacity of the Universe. Physical Review Letters,
88(23), 237901. doi:10.1103/PhysRevLett.88.237901
Lohan, M., & Faulknet, W. (2004). Masculinities and Technologies: Some Introductory
Remarks. Men and Masculinities, 6(4), 319-329.
López, P., & Hilbert, M. (2012). Methodological and Statistical Background on The
World’s Technological Capacity to Store, Communicate, and Compute Information
(online document). Retrieved from http://www.martinhilbert.net/WorldInfoCapacity.html
Luna-Reyes, L., Gil-Garcia, R., & Betiny Cruz, C. (2007). E-Mexico: Collaborative
structures in mexican public administration. International Journal of Cases on Electronic
Commerce (IJCEC), 3(2), 54-70.
Lustsik, O. (2004). Can e-banking services be profitable? University of Tartu Faculty of
Economics and Business Administration. Retrieved from
http://www.mtk.ut.ee/doc/febawb30.pdf
Lyman, Peter, Hal R. Varian, Kirsten Swearingen, Peter Charles, Nathan Good, Laheem
Lamar Jordan, and Joyojeet Pal. 2003. How Much Information? 2003. Berkeley, CA:
School of Information Management and Systems, University of California at Berkeley.
Lyon, D. (1991). The Information Society: Issues and Illusions (1st ed.). Polity.
MacCallum, John. CPU Price Performance 1944-2003. Available at:
http://www.jcmit.com/cpu-performance.htm (accessed January 25, 2010).
Madison, J (1787), Federalist 10, The Utility of the Union as a Safeguard Against
Domestic Faction and Insurrection, Independent Journal, 22 November, Constitution
Society, Austin, Texas [online] http://www.constitution.org/fed/federa10.htm , 1787.
Maeso, O., & Hilbert, M. (2006). Centros de acceso público a las tecnologías de
información y comunicación en América Latina: características y desafíos. United
Nations, Santiago Chile: United Nations ECLAC, Information Society Programme.
Retrieved August 7, 2009, from http://www.cepal.org/SocInfo/publicaciones
Mahajan, V., & Peterson, D. R. A. (1985). Models for innovation diffusion. Sage
Publications, Inc., London.
Maier, S., & Nair-Reichert , U. (2007). Empowering Women Through ICT-Based
Business Initiatives: An Overview of Best Practices in E-Commerce/E-Retailing Projects.
Information Technologies and International Development, Special issue on Women's
Empowerment and the Information Society, 4(2), 43-60.
Mansell, R. (2009). The Information Society (1st ed.). Routledge, London.
MAPPING OUT THE TRANSITION 252
Mariscal, J., & Rivera, E. (2005). New trends in the Latin American telecommunications
market: Telefonica & Telmex. Telecommunications Policy, 29(9-10), 757-777.
doi:10.1016/j.telpol.2005.05.007
Marshall, S., & Taylor, W. (2006). Editorial: ICT for education and training.
International Journal of Education and Development using Information and
Communication Technology, 2(4), 3-5.
Martin, J., & Butler, D. (1981). Viewdata and the information society. Prentice-Hall.
Martin, J., & Norman, A. R. D. (1970). The computerized society: an appraisal of the
impact of computers on society over the next fifteen years. Prentice-Hall.
Masuda, Y. (1980). The Information Society as Post-Industrial Society. Transaction
Publishers.
Mena, L. (2006). Regulation and investment in telecommunications: Chile case study.
Regulation and Investment in Telecommunications. DIRSI (Regional Dialogue on the
Information Society). Retrieved from http://www.dirsi.net/en/node/322
Miles, I. (1988). Information technology and information society: options for the future.
PICT policy research papers; no. 2. London: Economic and Social Research Council
(ESRC), London.
Minges, M. (2005). Evaluation of e-readiness indices in Latin America and the
Caribbean. Santiago: United Nations ECLAC. Retrieved from
http://www.cepal.org/SocInfo/publicaciones
Miranda, C. (2007). Information society and public ICT policies in the Caribbean: a
review of advances and challenges, policy instruments and country experiences
(Information Society Programme No. 155). Santiago: United Nations ECLAC. Retrieved
from http://www.eclac.org/publicaciones/xml/2/32162/W155.pdf
Mody, A., & Dahlman, C. (1992). Performance and potential of information technology:
An international perspective. World Development, 20(12), 1703-1719.
Mossberger, K., Tolbert, C. J., & Stansbury, M. (2003). Virtual Inequality: Beyond the
Digital Divide (1st ed.). Georgetown University Press.
Mossberger, Karen, Caroline J. Tolbert, and Mary Stansbury. 2003. Virtual Inequality:
Beyond the Digital Divide. Washington, D.C.: Georgetown University Press.
Navas-Sabater, J., Dymond, A., & Juntunen, N. (2002). Telecommunications and
information services for the poor - toward a strategy for universal access (World Bank
Discussion Paper No. 432 No. WDP432) (p. 118). World Bank. Retrieved from
http://www-
wds.worldbank.org/external/default/main?pagePK=64193027&piPK=64187937&theSite
MAPPING OUT THE TRANSITION 253
PK=523679&menuPK=64187510&searchMenuPK=64187283&siteName=WDS&entityI
D=000094946_02041804225061
Negroponte, N. (1995). Being digital (p. 255). Vintage Books.
Negroponte, N. (2005). One laptop per child. The Economist, Science(Nov 18).
Retrieved August 2, 2009, from
http://www.economist.com/theworldin/displaystory.cfm?story_id=E1_VTPQGTJ.
Newman, M. (2010). Networks: An introduction. Oxford University Press, New York.
Newman, M. E. J. (2003). The structure and function of complex networks. SIAM
REVIEW, 45, 167--256.
Ng, C., & Mitter, M. S. (2005). Gender and the Digital Economy: Perspectives from the
Developing World. SAGE Publications.
Nordhaus, William D. 2006. An Economic History of Computing. New Haven, CT: Yale
University and the NBER. Available at:
http://www.econ.yale.edu/~nordhaus/homepage/computing_June2006.pdf (accessed
January 25, 2010).
Norris, Pippa. 2001. Digital Divide: Civic Engagement, Information Poverty, and the
Internet Worldwide. New York: Cambridge University Press.
NTIA (National Telecommunications and Information Administration). (1995). Falling
through the Net: A Survey of the Have Nots in Rural and Urban America. Retrieved
August 2, 2009, from http://www.ntia.doc.gov/ntiahome/fallingthru.html.
NTIA (National Telecommunications and Information Administration). (1999). Falling
through the Net: Defining the Digital Divide. A Report on the Telecommunications and
Information Technology Gap in America. Revised. U.S. Government Printing Office,
Superintendent of Documents, Mail Stop: SSOP, Washington, DC 20402-9328. Tel: 202-
482-7002. For full text: http://www.ntia.doc.gov. Retrieved August 2, 2009, from
http://www.ntia.doc.gov/ntiahome/fttn99/contents.html.
NTIA. (2010). BroadbandUSA. Washington: National Telecommunications and
Information Administration. Retrieved October 2, 2010, from
http://broadbandusa.sc.egov.usda.gov/
Obama, B. (2009). Weekly address: President Obama discusses efforts to reform
spending, government waste; Names Chief Performance Officer and Chief Technology
Officer. Washington: The White House, Office of the Press Secretary. Retrieved from
http://www.whitehouse.gov/the_press_office/Weekly-Address-President-Obama-
Discusses-Efforts-to-Reform-Spending/
MAPPING OUT THE TRANSITION 254
OECD (Organization for Economic Co-operation and Development). (2001).
Understanding the Digital Divide. Paris, France: OECD (Organization for Economic Co-
operation and Development).
OECD (Organization for Economic Co-operation and Development). (2002). Measuring
the Information Economy 2002 (p. 94). Paris, France: OECD (Organization for Economic
Co-operation and Development).
Olaya, D. (2008). Compendium of Practices on the implementation of ICT questions in
household and business surveys in Latin America and the Caribbean. United Nations,
Santiago Chile: United Nations ECLAC, Information Society Programme, OSILAC.
Retrieved August 5, 2009, from http://www.cepal.org/cgi-
bin/getProd.asp?xml=/publicaciones/xml/4/32164/P32164.xml&xsl=/ddpe/tpl-
i/p9f.xsl&base=/socinfo/tpl-i/top-bottom.xsl.
Ono, H., & Zavodny, M. (2003). Gender and the Internet. Social Science Quarterly,
84(1), 111-121.
OSILAC (Observatory for the Information Society in Latin America and the Caribbean).
(2007). Monitoring eLAC2007:Progress and current state of development of Latin
American and Caribbean information societies. Santiago Chile: United Nations ECLAC,
Information Society Programme, OSILAC. Retrieved August 6, 2009, from
http://www.eclac.org/socinfo/publicaciones/default.asp?idioma=IN.
OSILAC (Observatory for the Information Society in Latin America and the Caribbean).
(2009). ICT Statistical Information System - OSILAC. Santiago Chile: United Nations
ECLAC. Retrieved August 5, 2009, from
http://www.cepal.org/tic/flash/default.asp?idioma=IN.
OSILAC (Observatory for the Information Society in Latin America and the Caribbean).
(n.d.). ICT Statistical Information System - OSILAC. Santiago Chile: United Nations
ECLAC. Retrieved August 5, 2009, from
http://www.cepal.org/tic/flash/default.asp?idioma=IN.
Park, S. (2009). Concentration of internet usage and its relation to exposure to negative
content: Does the gender gap differ among adults and adolescents? Women's Studies
International Forum, 32(2), 98-107.
Parker, E. B. (2000). Closing the digital divide in rural America. Telecommunications
Policy, 24(4), 281-290. doi:10.1016/S0308-5961(00)00018-5
Partnership on Measuring ICT for Development. (2008). The Global Information Society:
a Statistical View. United Nations, Geneva: ITU, OECD, UNCTAD, UNESCO,
UNDESA, World Bank, ECA, ECLAC, ESCAP, ESCWA, EUROSTAT. Retrieved
August 5, 2009, from http://www.itu.int/ITU-D/ict/partnership/.
Peres, W., & Hilbert, M. (Eds.). (2009). La sociedad de la información en América Latina
y el Caribe Desarrollo de las tecnologías y tecnologías para el desarrollo. Institutional
MAPPING OUT THE TRANSITION 255
books United Nations ECLAC (CEPAL) (Vol. 98). Santiago Chile: United Nations
ECLAC. Retrieved August 4, 2009, from
http://www.eclac.org/socinfo/publicaciones/default.asp?idioma=IN.
Peres, W., & Hilbert, M. (Eds.). (2010). Information Societies in Latin America and the
Caribbean: Development of Technologies and Technologies for Development. ECLAC
institutional books. Santiago: United Nations Economic Commission for Latin America
and the Caribbean. Retrieved from http://www.cepal.org/SocInfo/
Perez, C. (1983). Structural change and assimilation of new technologies in the economic
and social systems. Futures, 15(4), 357-375.
Perez, C. (2004). Technological Revolutions, Paradigm Shifts and Socio-Institutional
Change. In E. Reinert (Ed.), Globalization, Economic Development and Inequality: An
alternative Perspective (pp. 217-242). Cheltenham: Edward Elgar. Retrieved from
http://www.carlotaperez.org/papers/basic-technologicalrevolutionsparadigm.htm
Perez, C., & Soete, L. (1988). Catching up in technology: entry barriers and windows of
opportunity. In Technical change and economic theory, ed. by Giovanni Dosi, Freeman,
Christopher; Nelson, Richard; Silverberg, Gerald and Soete, Luc (pp. 458-479). Frances
Pinter. Retrieved August 3, 2009, from http://arno.unimaas.nl/show.cgi?did=6335.
Perez, Carlota. 1983. Structural Change and Assimilation of New Technologies in the
Economic and Social Systems. Futures 15 (5): 357-375.
Pierce, J. R. (1980). An Introduction to Information Theory (2nd Revised ed.). Dover
Publications.
Pool, I. de S. (1983). Tracking the Flow of Information. Science, 221(4611), 609–613.
doi:10.1126/science.221.4611.609
Porat, M. U. (1977, May). The information economy: Definition and measurement.
Superintendent of Documents, U.S. Government Printing Office, Washington, D.C.
20402 (Stock No. 003-000-00512-7). Retrieved from
http://www.eric.ed.gov/ERICWebPortal/contentdelivery/servlet/ERICServlet?accno=ED
142205
Puente, S. N. (2008). From cyberfeminism to technofeminism: From an essentialist
perspective to social cyberfeminism in certain feminist practices in Spain. Women's
Studies International Forum, 31(6), 434-440.
REALC. (2009). Plataforma Reigonal sobre residuos electronicos de PC en
Latinoamerica y el Caribe 2007-2010. Retrieved August 6, 2009, from
http://www.rrrtic.net/.
Regulatel (Foro Latinoamericano de Entes Reguladores de Telecomunicaciones), World
Bank, & United Nations ECLAC. (2007). Nuevos Modelos para el Acceso Uiversal en
MAPPING OUT THE TRANSITION 256
America Latina. Regulatel. Retrieved August 7, 2009, from
http://www.regulatel.org/publica/publica.htm.
RELPE, (Red Latinoamericana Portales Educativos) (2008). Analisis comparativo
Portales Educativos Latinoamericanos | RELPE (p. 35). RELPE. Retrieved January 20,
2010, from http://www.relpe.org/relpe/documentos-
tecnicos/analisis_comparativo_portales.php.
Ribeiro, J. (2008, July 29). India's '$10 Laptop' to Cost US$100 After All . PCWorld,
IDG News Service. Retrieved August 4, 2009, from
http://www.pcworld.com/businesscenter/article/149131/indias_10_laptop_to_cost_us100
_after_all.html.
Rice, R. E., & Katz, J. E. (2003). Comparing internet and mobile phone usage: digital
divides of usage, adoption, and dropouts. Telecommunications Policy, 27(8-9), 597-623.
Rice, R. E., & Katz, J. E. (2003). Comparing internet and mobile phone usage: digital
divides of usage, adoption, and dropouts. Telecommunications Policy, 27(8-9), 597-623.
doi:10.1016/S0308-5961(03)00068-5
Roberts, D. F., Foehr, U. G., & Foundation, H. J. K. F. (2004). Kids and media in
America (p. 380). Cambridge University Press.
Rogers, E. M., (2003). Diffusion of Innovations, 5th Edition (5th ed.). Free Press.
Rogers, Everett M. 1962. Diffusion of Innovations. New York: Free Press.
Rosenblut, Jorge. (1998). Telecommunications in Chile: Success and post-deregulatory
challenges in a rapidly emerging economy. Journal of International Affairs, 51(2), 565.
Roycroft, T. R., & Anantho, S. (2003). Internet subscription in Africa: Policy for a dual
digital divide. Telecommunications Policy, 27(1-2), 61-74. doi:10.1016/S0308-
5961(02)00091-5
Ryan, B., & Gross, N. (1943). The diffusion of hybrid seed corn in two Iowa
communities. Rural Sociology, 8(1), 15-24.
Schaefer Davis, S. (2007). Empowering Women Weavers? The Internet in Rural
Morocco. Information Technologies and International Development, Special issue on
Women's Empowerment and the Information Society, 4(2), 17-23.
Schumpeter, J. (1939). Business Cycles: A Theoretical, Historical, And Statistical
Analysis of the Capitalist Process. New York: McGraw-Hill. Retrieved from
http://classiques.uqac.ca/classiques/Schumpeter_joseph/business_cycles/schumpeter_busi
ness_cycles.pdf
Scott, A. (2001). (In)Forming politics: Processes of feminist activism in the information
age. Women's Studies International Forum, 24(3-4), 409-421.
MAPPING OUT THE TRANSITION 257
Scott, P. J. P. (2000). Social network analysis: A handbook (2nd ed.). Sage Publications,
London.
Sen, A. (2000). Development as freedom. Anchor, New York.
Shannon, C. (1948). A Mathematical Theory of Communication. Bell System Technical
Journal, 27, 379–423, 623–656. doi:10.1145/584091.584093
Sharma, U. (2003). Women Empowerment Through Information Technology. Authors
Press.
Shashaani, L. (1994). Gender-differences in computer experience and its influence on
computer attitudes. Journal of Educational Computing Research, 11(4), 347-367.
Shelley, Mack, Lisa Thrane, Stuart Shulman, Evette Lang, Sally Beisser, Teressa Larson,
and James Mutiti (2004). Digital Citizenship: Parameters of the Digital Divide. Social
Science Computer Review 22(2): 256-269.
Simpson, L., Daws, L., & Pini, B. (2004). Public internet access revisited.
Telecommunications Policy, 28(3-4), 323-337. doi:10.1016/j.telpol.2003.10.001
Smith, A. (1776). Wealth of nations: An inquiry into the nature and causes of the wealth
of nations (online). Library of Economics and Liberty. Retrieved from
http://www.econlib.org/library/Smith/smWN.html
Strogatz, S. H. (2001). Exploring complex networks. Nature, 410(6825), 268-276.
doi:10.1038/35065725
SUBTEL. (2003). Fondo de desarrollo de las telecomunicaciones: Memoria 2003 (p. 43).
Subsecretaria de Telecomunicaciones, Ministerio de Transporte y telecomunicaciones,
Chile.
Suh, J., & Aubert, J.-E. (Eds.). (2006). Korea as a knowledge economy: Evolutionary
process and lessons learned. International Bank for Reconstruction and Development /
The World Bank. Retrieved from
http://info.worldbank.org.libproxy.usc.edu/etools/docs/library/235384/KoreaKE-
Overview.pdf
Taylor, M., & Murphy, A. (2004). SMEs and e-business. Journal of Small Business and
Enterprise Development, 11(3), 280-289. doi:10.1108/14626000410551546
The Economist. (2008a, January 4). One clunky laptop per child. The Economist, Jan
4(Tech.view). Retrieved August 2, 2009, from
http://www.economist.com/science/displaystory.cfm?story_id=E1_TDQSNPDQ.
The Economist. (2008b, January 8). The laptop wars. The Economist, Business.view(Jan
8). Retrieved August 2, 2009, from
http://www.economist.com/business/displaystory.cfm?story_id=E1_TDQRJGTG.
MAPPING OUT THE TRANSITION 258
Tolbert, C., Mossberger, K., King, B., & Miller, G. (2007). Are All American Women
Making Progress Online? African–Americans and Latinas. Information Technologies and
International Development, Special issue on Women's Empowerment and the
Information Society, 4(2), 61-88.
Turing, A. M. (1937). On Computable Numbers, with an Application to the
Entscheidungsproblem. Proceedings of the London Mathematical Society, s2-42(1), 230–
265. doi:10.1112/plms/s2-42.1.230
Turoff, M. (1975). The Policy Delphi, in H. A. Linstone, and M. Turoff eds., “The Delphi
Method”, , Reading MA: Addison-Wesley Publishing Co., Inc., [online]
http://www.is.njit.edu/pubs/delphibook .
UN DESA. (2008). Global e-government survey 2008: From e-government to connected
governance. New York: United Nations Department of Economic and Social Affairs
(DESA).
UNCTAD. (2009). Information economy report 2009, trends and outlook in turbulent
times ( No. UNCTAD/IER/2009) (p. 153). Geneva: United Nations Conference on Trade
and Development. Retrieved from
http://www.unctad.org/Templates/webflyer.asp?docid=11915&intItemID=2068&lang=1
UNDP (United Nations Development Programme). (2008). Human Development Report
2007/8. United Nations Development Programme (UNDP). Retrieved August 23, 2009,
from http://hdr.undp.org/en/reports/global/hdr2007-2008/.
UNDP (United Nations Development Programme). (2009). Human Development Report
2007/2008 - Indicators. HDI Indicators - 2007/2008 Report. Retrieved August 7, 2009,
from http://hdrstats.undp.org/en/indicators/.
UNDP (United Nations Development Programme). 2001. Human Development Report
2001: Making New Technologies Work for Human Development. New York: Oxford
University Press for UNDP.
United States Economics and Statistics Administration and United States National
Telecommunications and Information Administration. 2000. Falling through the Net:
Toward Digital Inclusion: A Report on Americans' Access to Technology Tools.
Washington, D.C.: United States Department of Commerce.
United States National Telecommunications and Information Administration. 1995.
Falling through the Net: A Survey of the have Nots in Rural and Urban America.
Washington, D.C.: United States Department of Commerce.
United States National Telecommunications and Information Administration. 1998.
Falling through the Net II: New Data on the Digital Divide. Washington, D.C.: United
States Department of Commerce.
MAPPING OUT THE TRANSITION 259
United States National Telecommunications and Information Administration. 1999.
Falling through the Net: Defining the Digital Divide: A Report on the
Telecommunications and Information Technology Gap in America. Washington, D.C.:
United States Department of Commerce.
Ureta, S. (2005). Variations on the expenditure in communications in developing
countries: A synthesis of the evidence from Albania, Mexico, Nepal and South Africa
(2000-2003) (p. 23). RegulateOnline.org: World Dialogue on Regulation for Network
Economies. Retrieved August 5, 2009, from
http://www.regulateonline.org/content/view/528/68/.
Valente, T. W. (1995). Network models of the diffusion of innovations. Hampton Press,
New York.
Valente, T. W. (2010). Social networks and health: Models, methods, and applications
(1st ed.). Oxford University Press, New York.
van Dijk, J. A. (2005). The Deepening Divide: Inequality in the Information Society (1st
ed.). Sage Publications, Inc.
van Dijk, J. A. (2006). Digital divide research, achievements and shortcomings. Poetics,
34(4-5), 221-235.
van Dijk, Jan and Kenneth Hacker. 2003. The Digital Divide as a Complex and Dynamic
Phenomenon. The Information Society 19 (4): 315-326.
Varank, I. (2007). Effectiveness of Quantitative Skills, Qualitative Skills, and Gender in
Determining Computer Skills and Attitudes: A Causal Analysis. The Clearing House,
81(2), 71-80.
Vehovar, V., Sicherl, P., Husing, T., & Dolnicar, V. (2006). Methodological challenges
of digital divide measurements. Information Society, 22(5), 279-290.
Vehovar, Vasja, Pavle Sicherl, Tobias Hüsing, and Vesna Dolnicar. 2006.
Methodological Challenges of Digital Divide Measurements. The Information Society 22
(5): 279-290.
von Hippel, E. (2005). Democratizing innovation. The MIT Press, Cambridge,
Massachusetts.
Wareham, J., Levy, A., & Shi, W. (2004). Wireless diffusion and mobile computing:
Implications for the digital divide. Telecommunications Policy, 28(5-6), 439-457.
doi:10.1016/j.telpol.2003.11.005
Warschauer, M. (2003). Technology and Social Inclusion: Rethinking the Digital Divide.
The MIT Press.
MAPPING OUT THE TRANSITION 260
Wasserman, I. M., & Richmond-Abbott, M. (2005). Gender and the Internet: Causes of
Variation in Access, Level, and Scope of Use. Social Science Quarterly, 86(1), 252-270.
Waverman, L., Meschi, M., & Fuss, M. (2005). The Impact of Telecoms on Economic
Growth in Developing Countries . ICT Regulation Toolkit. Retrieved August 2, 2009,
from http://www.ictregulationtoolkit.org/en/Publication.3532.html.
Waverman, L., Meschi, M., & Fuss, M. (2005). The Impact of Telecoms on Economic
Growth in Developing Countries | ICT Regulation Toolkit (ICT Regulation Toolkit).
InfoDev, ITU. Retrieved from
http://www.ictregulationtoolkit.org/en/Publication.3532.html
Weber, Max (1918). Parlament und Regierung im neugeordneten Deutschland, Zur
politischen Kritik des Beamtentums und Parteiwesens, in M. Weber, Gesammelte
Politische Schriften. Tübingen: J.C.B. Mohr, Paul Siebeck.
Webster, F. (1995). Theories of the information society (p. 304). Routledge.
Wellenius, B., & Bank, W. (2002). Closing the gap in access to rural communications:
Chile 1995-2002. World Bank Publications, Washington.
Wiener, Norbert. 1948. Cybernetics : Or Control and Communication in the Animal and
the Machine. Cambridge, MA: MIT Press.
Williams, F., & Monge, P. R. (2001). Reasoning With Statistics: How To Read
Quantitative Research (5th ed.). Wadsworth Publishing.
World Internet Project. (2009). World Internet international Report 2009. USC
Annenberg School Center for the Digital Future. Retrieved August 18, 2009, from
http://store.digitalcenter.org/world-internet-report.html.
WSIS (World Summit on the Information Society). (2003, December). WSIS: Plan of
Action. United Nations. Retrieved August 17, 2009, from
http://www.itu.int/wsis/docs/geneva/official/poa.html.
WSIS (World Summit on the Information Society). (2005, November). Tunis
Commitment. United Nations. Retrieved August 17, 2009, from
http://www.itu.int/wsis/docs2/tunis/off/7.html.

Abstract (if available)

Abstract This research thesis sheds lights on different aspects of the transition toward information societies. It consists of a collection of interrelated studies that analyze in more rigorous terms three main and complementary aspects of the transition (see Figure below). After and introductory CHAPTER ONE, the consecutive CHAPTER TWO of this thesis looks at the social nature of the current transition toward the information society, which is characterized by a diffusion process that is known as digital divide. This chapter focuses on the socio-demographic characteristics of the transition, and characterizes its bottlenecks, such as the cost-income relation of ICT and users, as well as its opportunities, such as the opportunity to fight long-standing gender inequalities. CHAPTER THREE focuses not only on equality, but also on growth of the world’s information and communication capacity in absolute terms. The chapter consists of two sections that quantify the magnitude and growth of information in the information society, measured directly in bits and bytes. This provides insights into the speed and general pattern of the transition from analog to digital information processing in society. Both chapters combined provide complementary insights into what have been traditionally the two main pillars of socio-economic development: equity and growth. In this case the focus is set on the equality and growth of technologically mediated information. Various particularities of the transition become evident, such as the exponential rates of change of the transition, the all-pervasiveness of ICT in the social realm, and the unequal diffusion process. The final CHAPTER FOUR studies a concrete example of successful policy making in the digital age that takes these particularities into consideration. The case study focuses on a foresight Delphi exercise aimed at identifying future policy priorities that offered input into the inter-governmental negotiation of an Action Plan in Latin America. It is believed to have been the most extensive online participatory policy-making foresight exercise in the history of intergovernmental processes in the developing world. The process of policy-making in this international multi-stakeholder Delphi embraces the particular characteristics of the transition toward Information Societies by design. ❧ Figure 1 Overview: “Mapping Out the Transition toward Information Societies” ❧ The Chapters consists of a collection of complementary studies, which use a diverse array of methodologies and data sources to map out diverse aspects of the transition toward this new form of socio-economic organization. The three main Chapters consist of 6 articles that have been produced during the time of my doctoral program at USC’s Annenberg School of Communication (since August 2008). Chapter Two consists of three articles (resulting in three complementary sections), Chapter Three consists of two articles, and the final Chapter Four of one article. These articles are by now all published in recognized peer-reviewed Journals, all of which are indexed in the Thomson Reuters Social Science Index. Some of these Journals are leading in their fields (such as Technological Forecasting and Social Change, the world’s leading journal in foresight studies, with a 5-year Thomson Reuters Journal Citation Impact factor of 2.2

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Mobile apps and services for development: What can we learn from the non-smartphone era in ICT4D?

PDF

Mobile communication and development: a study of mobile phone appropriation in Ghana

PDF

The coevolution of multimodal, multiplex, and multilevel organizational networks in development communities

PDF

The informationalization of race: communication technologies and genomics in the information age

PDF

Examining internal communication practices to improve teachers’ motivation to use ICTs in instruction: a case study of Peruvian secondary schools

PDF

Tokens, ledgers, and rails: the communication of money

PDF

The Silicon Valley startup ecosystem in the 21st century: entrepreneurial resilience and networked innovation

PDF

Domestic and international tensions in artificial intelligence policy

PDF

Essays examining nutrition behavior and policy in California

PDF

The impact of social capital: a case study on the role of social capital in the restoration and recovery of communities after disasters

PDF

The impact of state-society relations on the design and implementation of digital-communication networks: a Franco-American comparative perspective

PDF

Quango reforms and challenges in South Korea: social relations, informal networks, and hidden actions

PDF

The struggle for communication rights: Global civil society and Internet govenance

PDF

Connected: information and state power between the United States and South Africa

PDF

Media reinvented: the transformation of news in a networked society

PDF

The impact of engaged communication scholarship on the re-imagination of space and place in South L.A.

PDF

The evolution of scientific collaboration networks

PDF

The paradoxes of network neutralities

PDF

Technical base, interests, and power in the two-level game of international telecom standards setting: the political economy of China's initiatives

PDF

Civic games with 'local fit': embedding with real‐world neighborhoods and place‐based networks

Asset Metadata

Creator Hilbert, Martin (author)

Core Title Mapping out the transition toward information societies: social nature, growth, and policies

School Annenberg School for Communication

Degree Doctor of Philosophy

Degree Program Communication

Publication Date 11/29/2012

Defense Date 10/05/2012

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag digital,digital divide,ICT for development (ICT4D),information and communication technologies (ICT),information society,OAI-PMH Harvest,public policy

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Castells, Manuel (committee chair), Bar, François (committee member), Gross, Larry P. (committee member), Monge, Peter R. (committee member), Valente, Thomas W. (committee member)

Creator Email martinhilbert@gmail.com,mhilbert@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-122978

Unique identifier UC11291026

Identifier usctheses-c3-122978 (legacy record id)

Legacy Identifier etd-HilbertMar-1369.pdf

Dmrecord 122978

Document Type Dissertation

Rights Hilbert, Martin

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA