Content uploaded by Nicoletta Corrocher
Author content
All content in this area was uploaded by Nicoletta Corrocher
Content may be subject to copyright.
Introduction
The notion of a ‘digital economy’ or ‘digitalization’
belongs to those concepts that are too wide to be rep-
resented and well described by a synthetic and precise
de nition. The emergence of technological platforms of
information and communications technology (ICT) is
determining signi cant and unprecedented changes in
many aspects of our social and economic life. However,
being directly involved in this process of change makes
it hard for researchers to understand, analyse and mea-
sure the phenomenon and also because the technolog-
ical dynamics are far from being consolidated.
The generality and vagueness of the notions that have
been employed in the last decade for describing the
phenomena related to the emergence of an ‘informa-
tion society’ (Lyon, 1988; Dordick and Wang, 1993;
Webster, 1995; Hill, 1999) have fostered a ‘legendary’
perception of such phenomena. At the same time, they
have been responsible for a widespread diffusion of the
associated concepts both in the media and in the com-
mon culture. However, from an empirical point of view,
the absence of a clear de nition hinders the process of
measurement (Sneath and Sokal, 1973) and leads to the
development of incomplete approaches that are not
based upon a structured theoretical framework. In this
context, most contributions that have analysed the phe-
nomena have complained about the lack of adequate
data and statistical information for the measurement
process. Nonetheless, it is necessary to remember that
this shortage may be explained by the delay that char-
acterizes the development of the underlying theoretical
models. For this reason, it is worth paying particular
attention to the proposal of new measurement frame-
works and not just to the improvement of the process
of data collection.
From a theoretical perspective it is possible to iden-
tify some common statements that are shared by
researchers who are interested in these issues (Martin
1995; Mansell and When, 1998; Ricci, 2000).
First, a key factor is represented by the role of infor-
mation as a critical resource because of the changes
in the way the information is organized, processed and
distributed. Second, the existing approaches aim at
integrating two major objectives. On the one hand they
focus on a quantitative dimension, i.e. the measure-
ment of the diffusion of digital phenomena. On the
other hand they concentrate on a qualitative dimen-
sion, i.e. the analysis of the impact of such phenomena
on social and economic systems.
Most programmes prepared by national governments
and by international organizations have dedicated a sub-
stantial amount of time and nancial resources to the
issue of the digital divide. However, quite surprisingly
there have not been so many efforts at developing and/or
improving the theoretical approaches and the statistical
Journal of Information Technology (2002) 17, 9–19
Measuring the digital divide: a framework
for the analysis of cross-country differences
NIC O L E T TA COR R O C HER
Research Centre on the Processes of Innovation and Internationalization (CESPRI), Bocconi University, Via S.
Mansueto, 5, 20122 – Milan, Italy
ANDR E A O R DAN IN I
Research Centre on the Digital Economy (I-LAB), Bocconi University, Viale Filippetti, 9, 20122 – Milan, Italy
This article proposes a new model for measuring the digital divide within a se t of countries or geograph-
ical areas. Starting from a series of elementary indicators the methodology groups these indicators into six
factors of digitalization and, subsequently, aggregates the factors in a synthetic index called the synthetic
index of digitalization. The dispersion in the distribution of the synthetic indexes of digitalization consti-
tutes the measure of the digital divide. This method is based upon a measurement approach, which is
different from the ones previously developed, since it uses principal components analysis for aggregating the
variables and avoids many of the problems and limits shown by existing mode ls. In the article an applica-
tion of the methodology is provided within a set of ten developed countries for 2000 and 2001. The measure-
ment framework for the digital divide presented here reveals new polic y implicatio ns for public institutions
and highlights opportunities and risks for managers wo rking in the ‘digital e conomy’ environment.
Journal of Information Technology
ISSN 0268–3962 print/ISSN 1466–4437 online © 2002 The Assoc iation for Information Technology Trust
http://www.tandf.co.uk/journals
DOI: 10.1080/0268396021 0132061
methodology directed at measuring this divide. A large
part of these proposals have indeed concentrated on the
de nition of policy issues related to the digital gap, more
than on the development of research projects for the
assessment of its actual magnitude and for the identi -
cation of appropriate evaluation techniques.
Furthermore, the digital divide has always been
analysed by comparing developed and developing coun-
tries: research works have often noted the existence of
relevant differences between these geographical areas,
but have not been able to explain them in terms
of different ‘speeds’ of digitalization (see ITU, 1999;
IDATE, 2000; Kenny, 2001). On the contrary, when
similar economic systems are taken into account, this
issue appears to be particularly interesting, both from
an informative point of view and from the perspective
of policy and managerial implications (as shown in the
nal section of this paper).
A brief review of the literature on
digitalization
In examining studies concerning digitalization it is pos-
sible to identify two broad lines of research. The rst
focuses on the patterns of diffusion of digital technolo-
gies at the macro-level of countries and at the micro-
level of sectors, consumers and public systems. The
second deals more speci cally with the issue of the dig-
ital gap and aims at quantifying and explaining the
differences in the development of digital technologies.
The present analysis follows the second strain of
research and, although a complete examination of the
studies on digitalization does not constitute the objec-
tive of this paper, it is believed that a synthetic review
of the most important contributions on the measure-
ment processes is quite useful in order to provide a
theoretical background to the model proposed in this
paper.
There are several international organizations within
the group of studies that have analysed the differences
in the level of digitalization that have dealt with the
digital divide for a long time.
The United States Department of Commerce (1999–
2000) developed one of the rst contributions in 2000.
This research measured the differences in access to
digital technologies within the US population, busi-
ness system and public administration. The Progressive
Policy Institute had already issued a report in 1999 on
the analysis of technological evolution employing a sim-
ilar methodology, but adopting a ‘regional’ perspective
analysing the US States. Using a comparative approach,
the study built a rank of the US States on the basis
of some indicators of digitalization (Progressive Policy
Institute, 1999).
In addition, the Organization for Economic Co-
operation and Development (OECD) has developed sev-
eral studies on the digital divide for a long time (OECD,
2000, 2001), whereby different countries are compared
on the basis of statistics related to the conditions of
access to ICTs and to the Internet. Similarly, in the con-
text of the ‘e-Europe’ initiative, the European Union
(EU) has developed a simple but extensive procedure for
benchmarking the diffusion of digital technologies within
its member countries (European Union, 2001).
Among the more recent models, which are mostly
related to the Internet, it is worth mentioning the World
Times Information Society Index and McConnell Inter-
national’s ‘E-readiness Survey’ (IDC, 2000, 2001;
McConnell International 2000, 2001). These studies
implemented an approach that is similar to the one
adopted in the present research, which is based upon
the combination of elementary variables with the aim
of generating a synthetic index of Internet readiness.
The World Bank has also issued research on the mea-
surement of the digital divide (a section of the ‘World
Development Indicators’) that identi es different fac-
tors of digitalization and uses a statistical methodology
based upon the aggregation of elementary indicators.
Finally, there are several national and international
organizations’ World Wide Web (WWW) sites that
aim at analysing the digital divide (for example, www.
digitaldivide.org, www.digitaldividenetwork.org and
www.pbs.org/digitaldivide)
All of the works described above adopted an informa-
tive and descriptive approach, often with a strong policy
orientation, without concentrating too much on the
methodological aspects related to the measurement of
the phenomena. However, more recently some contri-
butions have examined the digital divide in a less des-
criptive way, focusing on the methodological dimension.
One of the rst attempts at rationalizing the pro-
cedures of measurement of the digital divide was
proposed by Ricci (2000), who illustrated and inter-
preted the development paths of European countries by
constructing an ‘adoption scale’ for digital technologies
that results from the aggregation of elementary indica-
tors. This approach is based upon the work of the task
force of the EU and, although it does not deal with the
most problematic methodological issues, it represents a
good starting point for the provision of a theoretical
framework.
Barbet and Coutinet (2001) recently proposed a
model for measuring the relationships between the
characteristics of the communications infrastructure
and the evolution of the digital economy and for
examining the impact of digitalization at the level of
countries, sectors and rms. This approach is quite
useful, since it tries to analyse the links between the
determinants and the effects of digitalization, but it
10 Corrocher and Ordanini
presents some methodological dif culties when posi-
tioning different indicators exclusively as inputs or
outputs of digitalization. These dif culties are high-
lighted in the debate on the measurement of infor-
mation technology (IT) productivity (for a review see
Brynjolfsson and Kahin (2000)).
Another interesting work is the one by Selhofer and
Mayringer (2001), who developed a methodology for
benchmarking the development of the information
society in European countries. This research provides
interesting insights into the identi cation of different
dimensions of digitalization and the development and
synthesis of composite indicators. However, the analy-
sis is characterized by some methodological limits in the
process of aggregation and synthesis of the indicators,
which is largely subjective and leads to results that are
quite similar to the ones that could be obtained if the
indicators were grouped using a simple mean instead of
a weighed mean.
Both the descriptive and the more scienti c contri-
butions provide interesting insights for the provision
of a measurement process related to the digital divide,
but raise some methodological questions that need to
be solved in order to align the research objectives and
the tools of policy intervention.
First, since the digital divide is a complex notion,
there are some dif culties related to the aggrega-
tion of the elementary indicators in one or in a few
synthetic indexes of digitalization. This process should
follow objective criteria of weighing and should not be
based upon subjective approaches for identi cation of
the levels of aggregation and the provision of weights.
Second, the digital divide represents dispersion in
the distribution of a speci c phenomenon (in this case
the diffusion of digital technologies) and is a systemic
measure that derives from the joint consideration of
all its elements. It therefore constitutes a more compre-
hensive result compared to the outcome of a simple
benchmarking process, through which it is possible to
derive the distances between single elements.
The following sections aim at elaborating a frame-
work for the measurement of the digital divide that
takes into account the main issues arising from the
literature, but which overcomes the above-mentioned
methodological limitations.
The objectives and characteristics
of the model
The measurement framework for the digital divide
elaborated on in this paper is the result of a 2 year
research project called Digital Italy that was developed
by I-LAB, which is a research centre on the digital
economy at Bocconi University in Milan.
This project aimed at developing a methodology for
measuring the level of digitalization within a set of
countries or geographical areas through a series of ele-
mentary indicators that are grouped in some digitaliza-
tion factors and are subsequently aggregated in a single
synthetic measure. The dispersion of the nal values,
which is calculated for each country/geographical area,
constitutes the measure of the digital divide within the
system.
The starting point for the development of the model
is the identi cation of the factors that explain the digi-
talization of a system, i.e. the dimensions that form the
theoretical background of the concept of a digital econ-
omy (or information society). The de nition of the
six factors, which are described in the next section,
accounts for the notion of digitalization that is adopted
in this study.
Second, the geographical levels of the analysis, i.e.
the context within which the digital divide is calculated
and the set of elementary indicators that are poten-
tially able to represent the factors of digitalization are
selected.
The nal clusters of indicators are synthesized
through a factor analysis in order to obtain a measure
of digitalization for each factor and, subsequently, a
synthetic index of digitalization. This last step, which
constitutes one of the most innovative aspects of the
present approach, de nes the value that is attributed
to the patterns of digitalization and will be described in
detail in the next section.
Once a synthetic index of digitalization is developed
for each country (or geographical area) it is possible
to evaluate the magnitude of the digital gap, i.e. the
distances in the levels of digitalization, by calculating
measures of dispersion. The general structure of the
model is illustrated in Figure 1.
Measuring the digital divide 11
Dimensions of
digitalization
Context of
digitalization
Input of
digitalization
Value of
digitalization
Asymmetry of
digitalization
Definition of the factors
of digitalizatio n
Identification of
geographical ar eas
Collection of
elementary indicators
Synthesis of the
indicators
Computation of the
dispersion Digital divide
Figure 1 Logical steps for the development of the measure-
ment framework
The project aimed at developing an innovative model
for the evaluation of the digital divide and is character-
ized by some distinctive elements.
First, the model is composite since it takes into
consideration the existence of several ‘layers’ in the
digital economy and aggregates the elementary indi-
cators into six ‘factors of digitalization’ that represent
the outcome of the analysis concerning the complex
and multidimensional phenomena associated with the
diffusion of digital technologies.
The aim of this methodology is to avoid incomplete
measurement approaches in which the selected indi-
cators do not account for the interdependence between
the different variables.
At the same time, the measurement framework is syn-
thetic, i.e. it possesses an immediate explanatory power,
which derives from the use of synthetic indicators. In
this respect, the aggregation of the factors and the con-
struction of rankings for each factor represent funda-
mental steps in achieving synthetic and com-
parable cross-section and time-series data starting from
indicators that are not always homogeneous. The use
of the statistical procedure of multivariate analysis for
data aggregation allows some methodological problems
related to the establishment of a hierarchy of the indi-
cators and to the subjective attribution of weights in the
scaling process to be overcome (see below for details).
Third, the present methodology is transferable across
different contexts of application, i.e. it can be adopted
for the analysis of countries or geographical areas dif-
ferent from the ones initially considered. This is partic-
ularly relevant when comparing different economic
systems in order to provide important policy implica-
tions for promoting the growth and development of the
digital phenomena.
Finally, this model is exible since it is developed
through statistical procedures: this means that it can
be adapted and modi ed over time while maintaining
the original structure. This characteristic is particularly
useful for analysis of the performance of digitalization
over time since it allows the set of elementary vari-
ables to be updated, thereby increasing and improving
the number and type of indicators following the avail-
ability of better statistical sources or the dynamics of
technological development.
The factors of digitalization
As highlighted earlier the rst step of the analysis con-
cerns the identi cation of the factors of digitalization,
which represent the dimensions that characterize digital
development.
According to the conceptual framework of the pre-
sent study, the factors are meta-variables that not only
measure but also explain the differences between the
levels of digitalization of different countries: in this
way, they constitute explanatory factors of the digital
divide.
Several theoretical contributions have tried to distin-
guish the aspects of digitalization (Abramson, 2000;
Atrostic et al., 2000; Mesenbourg, 2000; University of
Texas, 1999, 2000). Here, this paper will refer to one
of the main research works (the model elaborated on
by the OECD Task Force on the digital economy
(Colecchia, 2000)) that analyses the strategic relevance
of the dimensions of digitalization in different phases of
the technological development of digital platforms. In
general, it is possible to identify a non-linear relation
between the time and diffusion of a speci c technolog-
ical phenomenon, which suggests the existence of and
need for considering different measurement priorities
in different stages of digital development. This appears
to be particularly important when considering the dig-
ital divide for policy purposes: indeed, different mea-
sures of the phenomenon can stem from different
interpretations of its determinants, which in turn can
highlight different strategic priorities and means of
intervention.
At the beginning of the use/application of the tech-
nology, differences between countries or regions are
explained by the speed of adoption. In this phase, the
factors that determine those differences are as follows.
(1) The communication infrastructures, which iden-
tify the availability of the physical resources that
allow access to the digital economy and stimu-
late its development. This factor includes aspects
related to the expansion of the Internet and of
WWW access devices as well as indicators con-
cerning the penetration and degree of techno-
logical advancement of other infrastructures that
account for the levels of connectivity in the sys-
tem, such as broadband cables and satellites.
(2) The human resources, which account for the
absorptive capacity of the system towards tech-
nological innovations on the basis of available
knowledge and education. In this context,
policies and programmes of formal education
and training play a central role, as well as the
employment conditions in the communications
sector.
(3) The competitiveness of the information and
communication providers and the degree of
competition among different operators, which
have a well-de ned role in fostering the provi-
sion of new services and in determining the pace
of adoption of new platforms and applications.
In the second stage, when the technology reaches a
critical mass of diffusion and is accepted as a common
12 Corrocher and Ordanini
standard, differences between countries or regions are
still in part explained by their speed of adoption, i.e.
by their basic infrastructure conditions, but the aspects
related to the intensity of adoption become more and
more important in the process of measurement. When
examining the digitalization in this phase, there is a
need for measuring the following variables.
(1) The diffusion of different devices for the use of
digital services and applications that can deter-
mine different patterns of digitalization in dif-
ferent systems.
(2) The size of the digital market, which identi es
the economic value of the technological appli-
cations de ning the ‘digital sector’.
In the third stage, when the technology becomes
mature, the measurement priorities become more and
more directed at qualitative aspects. In this respect, the
phenomena related to the impact of digitalization on
social and economic activities, on the structure of pro-
duction and consumption and on employment become
increasingly relevant.
Concluding, it is possible to identify three main
measurement priorities with reference to the diffusion
of the digital technologies that correspond to different
aspects of digitalization, as illustrated in Figure 2.
If the current rate of diffusion of digital technolog-
ical platforms is looked at, the measurement priorities
need to focus on the phenomena associated with the
speed and intensity of digitalization. The present evolu-
tionary phase is such that most of the intercountry
differences in the digitalization process are explained
by the basic infrastructure conditions, which in turn
lead to different degrees of intensity of adoption in the
process of digitalization.
As far as the third aspect is concerned, i.e. the trans-
formations in economic and social systems, the effects
have just recently begun to appear, so that any analysis
in this respect would be misplaced because of both
the complexity of the phenomena, which requires
quantitative and qualitative appraisals and the lack of
signi cant statistical information.
On the basis of this line of reasoning the following
digitalization factors, the rst two related to the inten-
sity of adoption and the other four related to the speed
of adoption, are considered: markets, diffusion, infra-
structures, human resources, competitiveness and com-
petition. The list of elementary indicators for each
factor of digitalization is given in the Appendix.
The geographical context of the
digitalization and identi cation of
elementary variables
The digital divide is a ‘relative’ concept since it assumes
signi cance only if it is evaluated within the speci c
context of countries or geographical areas. The asym-
metries in the diffusion of digital technologies have a dif-
ferent meaning if, for example, they are examined within
a heterogeneous (as it is often the case) or within a
homogeneous group of countries in terms of economic
development.
In the rst case, comparative analysis of digitalization
strongly con rms the already existing differences at the
level of the economic system: this makes it more com-
plex to understand if the digitalization is a direct conse-
quence of economic development or can instead affect
its dynamics. On the contrary, this work investigates the
digital divide within a set of geographical areas that are
in large part homogeneous in terms of development, so
that the effects of the existing economic systems are
neglected. As will be seen in the nal section of the
paper, this approach of measurement of the digital
divide allows not only major policy issues to be high-
lighted, but also important economic and managerial
implications for the business arena to be illustrated.
Having said that, the following countries are included
in the analysis of the digital divide: Finland, France,
Germany, Italy, Japan, Norway, the UK, Spain,
the USA and Sweden. The rationale underpinning the
choice of these countries lies, on the one hand, in
the necessity of analysing the most representative
European countries in terms of economic performance
and, in terms of diffusion of digitalization (e.g. the
Scandinavian countries), on the other hand, in the
need for comparing the development of the digital econ-
omy in Europe with that of extra-European countries
that represent interesting systems from a comparative
perspective because of the early emergence of such
phenomena in these contexts (the USA and Japan).
After having selected the factors and the levels
of analysis, the development of the measurement
Measuring the digital divide 13
Digitalization
2001 t
Speed
Infrastructures
Human resources
Competition
Competitiveness
Inten sity
Markets
Diffusion Impact
Economic and
social changes
Figure 2 Measurement priorities for the digital divide and
digitalization factors
framework involves the choice of the basic indicators
that are suitable for the quantitative evaluation of the
phenomena.
First, a list of potentially useful indicators (approxi-
mately 100) based upon data deriving from of cial
statistics is prepared. Information coming from non-
of cial institutions is not taken into consideration in
order to avoid problems related to the reliability and
comparability of data sources as much as possible. The
initial data set was subsequently analysed and censored
according to the following rationales: the existence of
data for all of the countries, the homogeneity of the data
across different sources of information and the qual-
ity and reliability of the data sources and of the data
themselves.
The outcome of the selection process is a database
of 36 indicators for the geographical areas previously
identi ed. These indicators are subsequently classi ed
in the six factors of digitalization identi ed above (see
the Appendix for details).
The model previously shown in Figure 1 can now
be structured as follows (see Figure 3).
Methods for aggregating indicators
and calculating the digital divide
Following the choice of elementary indicators, the next
step entails the aggregation of such indicators rst into
the six factors of digitalization and then among the
factors themselves in order to obtain a synthetic index
of digitalization for each country.
This methodology represents one of the most valu-
able features of the present model, as previous studies
on the digital divide adopted methods of aggregation
based upon subjective approaches, particularly as far
as the process of weighing the elementary indicators is
concerned.
If there is consensus on the use of a weighting
approach for aggregating the indicators of digitalization
(Selhofer and Mayringer, 2001), then there is a need
for designing a model whereby loading factors do not
depend on the subjective choice of the researcher. This
is necessary in order to avoid different outcomes when
measuring the digital divide starting from the same set
of elementary indicators.
In the model used here principal components (fac-
tor) analysis is used, rst to aggregate the elementary
indicators into the six digitalization factors, i.e. market,
diffusion, infrastructure, human resources, competition
and competitiveness and then to aggregate such factors
in order to obtain a synthetic measure of digitaliza-
tion. Principal components analysis is a multivariate
statistics technique that allows the transformation of a
given set of variables into a group of new components
through linear combinations of the original variables.
The extraction method ranks the new components
according to decreasing shares of explained variance:
each of these components is the outcome of a linear
combination of the initial variables through different
factor scores (Rummel, 1970; Hair et al., 1998). In this
case, the share of explained variance (factor loading)
represents the weights assigned to each component in
explaining the phenomenon. (In this case the factor
analysis is used simply as a hierarchical procedure for
assigning weights to a set of indicators and not as a data
reduction technique for simplifying the analysis. With
this approach, the nal set of components has the same
informative power as the initial one (Gorsuch, 1990;
Dillon et al., 1989).)
The indicators are grouped using the factor analysis
according to an objective method of aggregation, which
allows going back to the initial set of variables at any
time by looking at the factor scores. Figure 4 shows
the main steps of the model, which is applied to the
six factors of digitalization.
Following this approach, it is possible to identify the
most important drivers of digitalization at the level of
each aggregate group of indicators. The rankings of
the countries for the six factors have been rescaled in
order to improve the comparison of the relative
distances: as the outcome of the factor analysis is
normalized (with mean equal to zero), it is possible to
transform the distribution so that its mean becomes
2.5. This value represents the ‘mean level’ of digital-
ization, i.e. the benchmarking value for the whole set
of countries.
14 Corrocher and Ordanini
Dimensions of
digitalization
Markets
Infrastructures
Competitiveness
Diffusion
Human resources
Competition
Definition of m factors
of digitalizatio n
Context of
digitalization
Finland
Japan
UK
France
Norway
USA
Germany
Spain
Italy
Sweden
Identification of
geographical areas
Input of
digitalization 36 elementary indicators
Collection of
elementary indicator s
Figure 3 The rst three phases of the measurement frame-
work
Figure 4 From elementary indi cators to a single measure
for each factor
Countries
Elementary indicators
Countries Weights
Ranking of countries
for the factor n
Final components
Final components
Factor n
Elementary indicators
Factor n
The next step of the measurement process is the
development of one single measure of digitalization
for each country. This index derives from a linear
combination of the six factors, once again using the
approach of principal components analysis (see Figure
5). This procedure allows ranking and comparing the
ten countries on the basis of a single measure of digi-
talization.
The proposed method permits the factors that
explain the differences in the levels of digitalization
among different countries to be highlighted, adopting
an objective approach for the measurement. In this
case, as the factor analysis is only used for de ning the
weights of the aggregations, the indicators are grouped
without any loss of information, contrary to what hap-
pens when the factor analysis is used for purposes of
data reduction. In addition, the principal components
technique is able to reveal the relative importance
of each indicator within each aggregate factor, which is
equal to its weight in the linear combination (Borgata
et al., 1986).
The nal step of the model refers to the measure-
ment of the digital divide (see Figure 1). After having
calculated the levels of digitalization for each country,
the meaning of the digital divide is that of a measure
of dispersion.
In this sense, the benchmarking value of 2.5 in the
ranking based upon the synthetic index of digitaliza-
tion, which is obtained from the translation of the
distribution of normalized factors, always represents
the mean value of digitalization in the set of countries
considered. This implies the following.
(1) The digital divide can be measured through the
standard deviation of the synthetic index of digi-
talization with respect to the mean of 2.5.
(2) The relative distance to this value always repre-
sents the advantage/disadvantage for each
country.
(3) The evolution of the digital divide can be
assessed over time, as represented by the value
of the mean for each year.
Figure 6 illustrates the meaning of the digital divide.
The outcome of the model: the digital divide
In this section, the model for measuring the digital
divide is applied to the set of countries indicated in
the previous section over two periods of time (2000
and 2001) and a comparison between the outcomes is
provided.
Figure 7 shows the value of the synthetic index of
digitalization, which is the synthetic indicator of digi-
talization, in the stated set of countries for 2001.
A rst comment on these data regards the substan-
tial differences in the level of digitalization in a set of
developed countries. The USA emerges as the point
of reference for the diffusion of digital technologies,
followed by the Scandinavian countries, the UK and
Japan.
Germany stands in the middle of the ranking, close
to the mean value of 2.5, while France, Italy and Spain
seem to be the slowest countries in terms of digital
diffusion.
The fact that a set of countries that are similar in
terms of economic development show important differ-
ences in terms of digitalization is an important outcome
of the analysis and implies a different meaning of the
Measuring the digital divide 15
Countries
Factors of digitalization
Countries Weights
Unique and final ranking
of digitalization
(synthetic index of digitalization)
Final factors
Final factors
Initial factors
Figure 5 From six factors to one measure of digitalization
for each country (the synthetic index of digitalization)
0 2.5
Standard deviation
Digital divide
5
Figure 6 A picture of the digital divide
0
Spain
Italy
France
Germany
Japan
Norway
UK
Sweden
Finland
USA
1 2 3 4 5
0.8
1.3
1.3
2.0
2.6
2.7
2.9
3.2
3.6
4.6
Figure 7 The synthetic index of digitalization for 2001
digital divide compared to the one usually employed
when comparing developed and developing countries.
Figure 8 compares the evolution of the levels of digi-
talization in each country between 2000 and 2001, 2.5
being the mean level of digitalization for both periods
of time.
In general, France, Norway and Sweden show a lag
in their adoption of digital technologies. If in the case
of the two Nordic countries the pattern is explained by
the early process of digitalization, which has already
reached a high level of diffusion in the system, in the
case of France the slow evolution is probably due to
lock-in phenomena related to the widespread diffusion
of Minitel, which is a standard that is an alternative to
the Internet.
More speci cally, it is possible to highlight the
following signi cant dynamics.
(1) The USA enlarges the distance with respect to
the rest of countries.
(2) The UK surpasses Norway and reaches fourth
position in the ranking.
(3) Japan gets close to the group of countries with
high levels of digitalization.
(4) Italy overtakes France in the ranking and this
is probably driven by the diffusion of mobile
communications technologies.
Considering that the synthetic index of digitalization
is the outcome of an aggregation process of six factors
of digitalization, i.e. markets, diffusion, infrastructure,
human resources, competitiveness and competition, it
is possible to analyse the rankings of the countries with
respect to each of the factors (see Table 1).
The positions reveal the role played by each factor of
digitalization and indicate the pattern of digitalization
followed by each country, which can be character-
ized by a symmetric or by an asymmetric evolution,
according to the value of the standard deviation (see
below). This analysis has several policy and managerial
implications, which are explained in the next section.
The digital divide may be represented by the level
dispersion in the distribution of the levels of the syn-
thetic index of digitalization of all the countries. It can
be observed that the standard deviation of the levels of
digitalization in 2001 is 1.17, which is smaller than that
calculated in 2000 (1.25) (see Figure 9): this means
that there has been a reduction in the digital divide
among the set of countries taken into consideration.
When excluding the USA, the relative distances
between countries seem to shrink: for instance, in 2001
the gap between the synthetic index of digitalization
for Germany and that for Finland is only 1.6 points.
On the contrary, in 2000 the Nordic countries were
closer to the USA and the other countries (Germany,
Japan, the UK and France) were behind in the ranking.
The application of the model reveals how, by using
an objective approach, cross-country comparisons con-
cerning the diffusion of digital technologies and the
drivers of the patterns of digitalization can be devel-
oped according to a dynamic perspective, which allows
the evolution of the digital divide over time to be
analysed.
Policy and managerial implications
The model for the measurement of the digital divide
proposed in this study allows not only distances
16 Corrocher and Ordanini
5.0
2.5
0.0
0.0 2.5
2001
5.0
2000
Spain
Italy
UK
Japan
USA
Finland
Germany
Sweden
Norway
France
Figure 8 A comparison between the levels of digitalization
(synthetic indexes of digitalization) of 2000 and 20 01
Table 1 The pattern s of digitalization: positions of the countries in the rankings of the digitalization factors
Finland France Germany Italy No rway UK Spain Sweden USA Japan
Markets 5 8 7 10 6 3 9 2 1 4
Diffusion 1 10 6 8 3 5 9 2 4 7
Infrastructure 5 7 4 9 1 6 10 3 2 8
Human resources 5 7 8 6 3 4 10 1 2 9
Competitiveness 2 6 7 8 10 4 9 5 1 3
Competition 2 10 8 6 4 5 9 7 1 3
SID 2 8 7 9 5 4 10 3 1 6
Mean ranking 3.1 7.9 6.7 8.1 4.6 4.4 9.4 3.3 1.7 5.7
Standard deviation 1.8 1.5 1.4 1.5 2.9 1.0 0.5 2.1 1.1 2.4
SID, synthetic index of digitalization.
between countries to be evaluated, but also the charac-
teristics of the pattern of digitalization in each country
to be analysed. This appears to be particularly relevant,
as it reveals many policy and managerial implications.
In this respect, most of the existing studies dealing
with issues related to the digital divide focus on the
comparison between developed and developing coun-
tries. According to such an approach, the digital divide
tends to be largely explained by different levels of
economic, technological and social development. This
type of analysis suggests that there is the need for poli-
cies directed at reducing these differences, but the
implementation of speci c policies in this context is
quite complex, since the digital gap may be a driver,
but also the result of differences in economic and social
development.
On the contrary, the model presented here measures
the digital divides between countries that are quite sim-
ilar in terms of their economic, technological and social
conditions. This implies that the emerging differences
are only marginally in uenced by variables other than
those strictly related to the diffusion of digital tech-
nologies and are directly affected by the pattern of dig-
italization in each country. This allows the real meaning
of the digital divide to be understood and important
policy and managerial implications to be derived.
Observing the features of the digitalization path that
characterizes each country (see Table 1), different
patterns seem to emerge.
Some countries, such as Japan, Sweden, Norway
and, partially, Finland, show asymmetric paths of digi-
talization, with high levels of standard deviation in the
position within the rankings referring to the digital-
ization factors: in these cases, situations of leadership
and backwardness co-exist. On the contrary, in other
countries, the levels of digitalization are more homoge-
neous across the six factors and their patterns of digi-
talization converge at high (the USA and UK) or low
(France, Italy and Spain) levels. This means that these
countries pursue an even digital development across
all its components.
Two considerations appear to be relevant in dis-
cussing policy and managerial implications. On the one
hand, the leadership in some factors of digitalization
may drive the development of the whole system. In this
respect, pursuing speci c objectives by exploiting the
technological and economic characteristics of the sys-
tems in terms of digitalization may be a winning strat-
egy, both for policy makers and for the rms involved
in the digital environment. On the other hand, a more
balanced pattern of digitalization can generate bene-
ts for the economy and society as a whole and this
implies that, according to a long-term perspective, pol-
icy interventions should aim at reaching similar levels
of development across all the factors of digitalization.
However, the heterogeneity of the digitalization levels
that emerges from this analysis, both in the group of
countries with an asymmetric pattern of digitalization
and in the group of countries with a symmetric pattern,
suggests that the two strategies may be considered as
complementary rather than as merely substitute.
For instance, if the cases of Spain and the USA are
considered, it can be seen that both of them show
balanced patterns of digitalization, even if they are
placed at the opposite extremes in the ranking of the
synthetic index of digitalization. Similarly, Japan and
Sweden exhibit asymmetric patterns of digitalization,
but their levels of digitalization diverge from each
other.
These situations suggest that, within each country,
policy makers and rms involved in the digital envi-
ronment have to exploit the technological competitive
advantages in order to sustain the process of digital-
ization (see the growth of mobile telecommunications
in Italy and in the Nordic countries). At the same time,
they have to implement policies and strategies directed
at reducing their technological weaknesses in order to
converge to a more even pattern of digitalization.
Furthermore, when dealing with policy implications
it is worth bearing in mind that priorities may vary
considerably between national governments and supra-
national organizations. At a national level, for instance,
policy makers may prefer a focus on speci c issues
related to digital development, i.e. improvement of
the infrastructure, while at an international level (for
instance at the level of the EU) institutions are gener-
ally interested in pursuing an even pattern of digital-
ization across different geographical areas. Sometimes
these two approaches may collide, with mixed impli-
cations for the expected trajectories of digitalization.
One nal consideration refers to the importance of
different factors of digitalization, in particular for man-
agers of rms involved in the digital economy. As
already stated in the section on the factors of digitaliza-
tion, measuring the digital gap in a set of countries that
are similar in terms of economic and social background
implies that the digital divide can be explained by the
strategic patterns followed by rms that operate in dif-
ferent national (or international) digital environments.
Measuring the digital divide 17
0 2.5
2000
Standard devia tion = 1.25
5 0 2.5
2001
Standard devia tion = 1.17
5
Figure 9 The digital divide: a 2000–2001 comparison
In this context, analysis of the digital divide can reveal
opportunities and risks for managers operating in busi-
nesses that are linked to the digital economy and, there-
fore, represents a powerful analytical tool for driving
the investment strategies and competitive behaviour of
rms. Given the pervasiveness of digital technologies, a
growing number of businesses are in uenced by the
evolution of digitalization (Sampler, 1998).
The technological opportunities related to the digital
economy emerge more clearly and can be exploited and
sustained over time if the measurement of the digital
divide is based upon a well-structured analytical frame-
work. On the one hand, rms may develop advanced
competencies for building a competitive advantage in
relevant technological areas. On the other hand, policy
makers may bene t from a clear explanation of the
drivers of digitalization at a national and international
level, so that they can support the system, thereby fos-
tering the competition between rms and implement-
ing ad hoc policies for stimulating digital development
in society as a whole.
References
Abramson, B.D. (2000) Internet globalisation indicators.
Telecommunications Policy, 24, 69–74.
Atrostic, B.K., Gates, J. and Jarmin, R. (2000) Measuring
the Electronic Economy: Current Status and Next Steps (US
Census Bureau, Washington DC).
Barbet, P. and Coutinet, N. (2001) Measuring the digital
economy: state-of-the-art developments and future pros-
pects. Communications and Strategies, 42, 153–84.
Borgata, E.F., Kercher, K. and Stull, D.E. (1986) A caution-
ary note on the use of principal components analysis.
Sociological M ethods and Research, 15, 160–8.
Brynjolfsson, E. and Kahin, B. (2000) Understanding the
Digital Economy (MIT Press, Cambridge, MA).
Colecchia, A. (2000) De ning and Measuring Electronic Com-
merce – Towards the Development of an OECD Methodology
(OECD Press, Paris).
Dillon, W.R.N., Mulani, N. and Frederick, D.G. (1989) On
the use of component scores in the presence of group
structure. Journal of Consumer Research, 16, 106–12.
Dordick, H. and Wang, G. (1993) The Information Society:
A Retrospective View (New Delhi Sage Publication,
Newbury Park, CA).
Gorsuch, R.L. (1990) Common factor analysis versus com-
ponent analysis: some well and little known facts. Multi-
variate Behavioural Research, 25, 33–9.
Hair, J.F., Anderson, R.E., Tatham, R.L. and Black, W.C.
(1998) Multivariate Data Analysis, 5th edn (Prentice-
Hall International, Upper Saddle River, NJ).
Hill, M.W. (1999) The Impact of Information on Society
(Gower, Aldershot).
Kenny, C. (2001) Prioritising countrie s for assistance to over-
come the digital divide. Communications and Strategies,
41, 17–36.
Lyon, D. (1988) The Information Society: Issues and Illusions
(Basil Blackwell, Oxford) (Italian edition: La Società
dell’Informazione (Il Mulino, Bologna)).
Mansell, R. and When, U. (1998) Knowledge Societies Infor-
mation Technology for Sustainable Development (Oxford
University Press, Oxford and New York).
Martin, W.J. (1995) The Global Information Society (Bowker
Saur, London).
Mesenbourg, T.L. (2000) Measuring Electronic Business:
De nitions, Underlying Concepts and Measurement Plans
(US Census Bureau, Washington DC).
Ricci, A. (2000) Measuring information society dynamics of
European data on usage of information and communi-
cation technologies in Europe since 1995. Telematics and
Informatics, 17, 141–67.
Rummel, R.J. (1970) Applied Factor Analysis (North-western
University Press, Evanston, IL).
Sampler, J.L. (1998) Rede ning industry structure for
the information age. Strategic Management Journal, 19,
343–55.
Selhofer, H. and Mayringer, H. (2001) Benchmarking the
information society. Development in European coun-
tries. Communications and Strategies, 43, 17–55.
Sneath, P.H.A. and Sokal, R.R. (1973) Numerical Taxonomy
(W. H. Freeman, San Francisco).
Webster, F. (1995) Theories of the Information Society (Rout-
ledge, Lond on).
Statistical sources
European Union (2001) Information Society Statistics EU,
Brussels.
IDATE (2000) Prioritising Countries for Assistance to Overcome
the Digital Divide, IDATE, Montepellier.
IDC (2000–2001) World Times Information Society Index
International Data Corporation, Framlingham, MA.
ITU (1999 ) Challenges to the Network: Internet for Development
International Telecommunications Unit, Geneva.
ITU (2000–2001) World Telecommunication Indicators Inter-
national Telecommunications Unit, Geneva.
McConnell International (2000–2001) Ready? Net. Go!
Partnerships Leading the Global Economy, Washington DC.
Organization for Economic Co-operation and Development
(1998–2000) Information Technology Outlook OECD,
Paris.
Organization for Economic Co-operation and Development
(1999–2001) Communications Outlook OECD, Paris.
Organization for Economic Co-operation and Development
(2000) Measuring the ICT Sector OECD, Paris.
Organization for Economic Co-operation and Development
(2001) Understanding the Digital Divide OECD, Paris.
Progressive Policy Institute (1999) The State New Economy
Index PPI, Washington DC.
University of Texas (1999–2000) The Internet Economy
Indicators University of Texas, Austin.
US Department of Commerce (1999–2000) Falling Through
the Net: Toward D igital Inclusion US Census Bureau,
Washington DC.
18 Corrocher and Ordanini
Biographical notes
Nicoletta Corrocher Degree in Economics (Bocconi
University, Milan, Italy), MSc in Science and Tech-
nology Policy (SPRU – Science and Technology Policy
Research, University of Sussex, Falmer, Brighton, UK),
PhD in Economics and Management of Innovations
(Sant’ Anna School of Advanced Studies, Pisa, Italy).
Current research areas include economics of ICT (and
the Internet), diffusion of ICT in the nancial sector,
the evolution of industrial districts, the emergence of
new industries and of new technologies in high tech
elds.
Andrea Ordanini Degree in Management (Bocconi
University, Milan, Italy). PhD in Business Adminis-
tration. Visiting researcher at the London School of
Economics and Political Sciences (London, UK).
Current research topics are economics and manage-
ment in the ICT sectors, e-business strategies, compet-
itive advantage of high-tech rms.
Address for correspondence: Andrea Ordanini,
Research Centre on the Digital Economy (I-LAB),
Bocconi University, Viale Filippetti, 9, 20122 – Milan,
Italy.
Appendix: the elementary indicators
Markets
(1) Share of the IT hardware market over gross
domestic product (GDP).
(2) Share of the IT software and services market
over GDP.
(3) Share of the telecommunications equipment
market over GDP.
(4) Share of the telecommunications services mar-
ket over GDP.
(5) On-line consumer expenditure over GDP.
Diffusion
(1) Computer hosts per 1000 inhabitants.
(2) Internet users per 100 inhabitants.
(3) Percentage of WWW buyers over Internet users.
(4) Number of secure WWW servers for electronic
commerce per 1 000 000 people (a proxy for
the diffusion of the electronic commerce).
(5) Number of mobile subscribers per 100 inhabi-
tants.
(6) Number of cable television subscribers.
(7) Diffusion of national domain names per inhab-
itants.
(8) Digital television subscribers among households.
Infrastructures
(1) Number of personal computers (PCs) per 100
inhabitants.
(2) Number of devices connected to the Internet
per 1000 inhabitants.
(3) Number of WWW servers per 100 inhabitants.
(4) Number of access lines per 100 inhabitants.
(5) Number of ISDN subscribers per 1000 inhab-
itants.
(6) Penetration of broadband Internet access.
(7) Number of PCs per employee (as an indicator
of the penetration of the IT infrastructure within
rms).
Human resources
(1) Share of employment in the ICT hardware
sector over total employment.
(2) Share of employment in the telecommunications
services sector over total employment.
(3) Share of employment in other ICT services over
total employment.
(4) Share of expenditure on education over GDP.
(5) Percentage of the population with a university
degree.
(6) Percentage of schools with an Internet connec-
tion.
(7) Number of PCs per student.
Competitiveness
(1) Number of leaders in the telecommunications
sector (according to the ranking of the top 100
companies classi ed by revenues).
(2) Mean ranking of the above-mentioned leaders
(calculated as the relative size of the com-
panies on the basis of the above described
ranking).
(3) Share of research and development investments
in the ICT sector.
(4) Share of ICT patents over total ICT research
and development.
Competition
(1) Costs of telecommunications services.
(2) Internet access costs.
(3) Concentration ratio in the IT sector (market
share of the top ten operators).
(4) Market share of the new entrants in the xed
telecommunications segment.
(5) Market share of the incumbent in the mobile
telecommunications segment.
Measuring the digital divide 19
