The bad news is that the digital access divide is here to stay: Domestically installed bandwidths among 172 countries for 1986–2014

https://doi.org/10.1016/j.telpol.2016.01.006Get rights and content

Highlights

  • The digital divide measured in terms of bandwidth it not closing.

  • Inequality fluctuates up- and down with technological progress and diffusion.

  • Asia increased its global share in installed bandwidth from 23% to 51% in 30 years.

  • Bandwidth inequality is closely linked to income, which is notoriously unequal.

  • It is urgent to start measuring bandwidth, not merely counting subscriptions.

Abstract

In contrary to the common argument that the digital access divide is quickly closing and that the focus should shift to skills and usage, this article shows that access to digital communication is a moving target unlikely to ever be solved. While the number of subscriptions reaches population saturation levels, the bandwidth divide continuous to be dynamic. The article measures the nationally installed bandwidth potential of 172 countries from 1986 to 2014. The overarching finding is that the divide in terms of bandwidth does not show any clear monotonic pattern. It fluctuates up and down over the decades as the result of an intricate interplay between incessant technological progress and diffusion of technology. The bandwidth divide between high- and low income countries has first increased and only decreased below historic levels very recently during 2012–2014. In general it shows that the bandwidth divide is linked to the income divide, which is notoriously persistent. The bandwidth distribution among all countries is undergoing a new process of global concentration, during which North America and Europe is being replaced by Asia as the new global leader. In 2014 only 3 countries host 50% of the globally installed bandwidth potential (10 countries almost 75%). The U.S. lost its global leadership in 2011, being replaced by China, which contributes more than twice as much national bandwidth potential in 2014 (29% versus 13%). Despite this bad news about the continuous persistence of the digital access divide among countries, exploratory analysis from a global perspective brings the good news that many more individual people seem to enjoy more equal access to global bandwidth. All of this showcases the urgency to systematically develop indicators to track the digital divide in terms of bandwidth.

Introduction

The article takes inventory of the evolution of the international telecommunication infrastructure in terms of the installed telecommunication bandwidth capacity between 1986 and 2014. This matters both because of the importance of digital technologies throughout the world and the continuous evolution of telecommunication. Telecommunication access solution have evolved significantly during the past three decades, consisting exclusively of fixed-line telephony in the late 1980s, and a plethora of access solution with diverse performance levels. Over the last decade, the literature has increasingly pointed to the importance of bandwidth and especially broadband metrics, which have shown to have important socio-economic benefits (Dutton, Gillett, McKnight, & Peltu (2004), Koutroumpis (2009), Prieger (2013), Gruber, Hätönen, & Koutroumpis (2014), Lee, Park, & Hwang (2015)). Quantifying the digital divide in terms of subscriptions might not be sufficient anymore, as not all subscriptions are equal. In light of this, the key question of this article is if technological progress has rendered traditional metrics of the digital divide obsolete.

Traditionally the international digital divide is assessed in terms of telecommunication subscriptions (NTIA (1995), OECD (2001), ITU (International Telecommunication Union), 2015). On the international level, the most common go-to source are the statistics from United Nations’ International Telecommunication Union (ITU, 2014). ITU has undertaken a sustained effort over several decades to collect this data from administrative registries of national telecommunication authorities in a harmonized manner. These same databases have shown that the number of mobile and fixed telecom subscriptions per person are increasingly reaching a certain level of global saturation, including 6.8 billion mobile phone subscriptions worldwide for 7.0 billion people in 2014.

Since there seems to be a certain limit in how many technological devices a person handles (Hilbert, 2014a), any analysis that uses the number of subscriptions as a proxy for the digital divide must come to the conclusion that the divide is closing over time. As early as the year 2000, this perspective has led to the impression that “the gaps are rapidly closing” (Compaine, 2001). Over the years, this view has become as engrained into the way of looking at the digital divide that is has become natural to assume a national “carrying capacity of Internet users” (e.g. Neokosmidis, Avaritsiotis, Ventoura, & Varoutas, 2015). Once this carrying capacity is reached (once everybody has reached the limit of how many subscriptions can be handled), saturation sets in and the divide can only close. As a result, new technological solutions might create new divides, but in terms of their numbers the divide will always be closing over time, as it has happened with computer access, mobile phones, or broadband adoption (e.g. Vicente & López, 2011; Loo & Ngan, 2012; Prieger, 2013).

Based on this impression, scholars have long moved on to work on the digital divide in terms of differential usage patterns, caused by differences in skills, culture and other demographics and social variables (Mossberger et al., 2003, Warschauer, 2004, van Dijk, 2005, Vicente & López (2011), Deursen & Dijk (2014)). This sometimes referred to as the “second-level digital divide” (Hargittai (2002), Büchi, Just, & Latzer (2015)). In more advanced countries, the dimension of physical access has become a question of technology maintenance to sustain the level of subscriptions and devices (Gonzalez, 2015).

However, the fact that the number of telecom subscriptions per person seems limited and will eventually reach saturation levels does not automatically imply that inequality in terms of access to digitalized information is reducing as well. This is because bandwidth is not uniformly distributed among subscriptions. Fig. 1a shows this distinction was not relevant only a few decades ago. The late 1980s exhibited a linear one-to-one relationship between subscriptions and bandwidth, as there was only fixed line telephony around, all with the same bandwidth. Today’s digital infrastructure offers a myriad of different bandwidth options, which leads to an L-shaped non-linear relationship exhibit in Fig. 1b. The Figure shows that ICT diffusion seems to hit an invisible wall at around 2–3 subscriptions per person. However, the digital divide continues at this point, just along a new dimension: the bandwidth dimension.

“This implies that we have now moved into a second, more mature, and also more persistent stage of the digital divide” (Hilbert, 2014c). The first phase consisted of a universalization of the required technological infrastructure. The second stage consists of an endlessly evolving inequality of bandwidth. While today almost everybody in the world counts with a minimum level of connectivity potential through a digital mobile phone, some have access to significantly more bandwidth. Fig. 1b suggests that there might not be a clear “carrying capacity” in terms of the bandwidth per subscription and therefore also not automatically any level of saturation that would inevitably close the digital divide. The answer to the connectivity question moves from being a binary black-or-white choice (0–1) to a continuous and incessantly moving gray zone (1–∞).

This second stage of the digital access divide becomes extremely relevant in times where fast bandwidth solutions have become to take center stage in the discussion about social growth and progress (Dutton, Gillett, McKnight, & Peltu (2004), Koutroumpis (2009), Prieger (2013), Gruber, Hätönen, & Koutroumpis (2014), Lee, Park, & Hwang (2015)). Bandwidth matters especially for an age of big data, in which the quantity of data (not mere access) have become a driver of social-, cultural-, political-, and economic development (Manyika et al. (2011), Hilbert (2016)). Better understanding the bandwidth divide is also important in order to better frame the ongoing debate about net-neutrality, which centers on the question of creating a tiered internet (Krämer, Wiewiorra, & Weinhardt (2013), Hsu, 2014). As it stands, we do not have much insight into internet bandwidth distribution. With or without net-neutrality, it might well be that even without any kind of intervention into traffic patterns, plain differences in physical infrastructure availability around the world have already created a de-facto tiered structure among some with much and some with systematically little bandwidth: a digital bandwidth divide.

Despite these reasons, research continues to disregard this reality and still refers to the number of telecom subscription, especially for large-scale international statistical tests (for example Pick & Sarkar, 2015; Mardikyan, Yıldız, Ordu, & Şimşek, 2015; Mendonça, Crespo, & Simões, 2015). This seems to have less to do with the fact that scholars are not aware, but that subscriptions data is readily available, and bandwidth and traffic data is not. Reminiscent of the famous drunk who is looking for the lost keys under a well-lit lamppost far away from the dark site where the keys were dropped, international analysts continue to recur to the readily available and harmonized databases on subscriptions.

Changing this current practice faces two challenges. One the one hand, important efforts are currently underway to explore new ways to measure relevant aspects of this second stage of the digital divide. Leading authorities like the U.S. Federal Communications Commission (FCC, 2014) or Ofcom (2014) in the UK have started to produce detailed annual reports at the national level. These efforts are struggling with finding adequate new metrics. For example the first generation of broadband reports of the FCC from the early 2000 distinguished between access solution below and above 200 kbps (FCC, 2000). A second generation of reports recognized that “existing definitions are not static” (FCC, 2004), and let to definitions which seem equally temporal in nature, such as distinctions between 200 kbps–2.5 mbps, 2.5–10 mbps, 10–25 mbps, 25–100 mbps (FCC, 2008); or more fine-tuned specifications, which seem as arbitrary as provisional, such as recommendations to measure download/upload ratios at 384 kbps/1.5 kbps and 3 mbps/768 kbps (FCC, 2012); or later assessments of ratios such as 10 mbps/1 mbps and 25 mbps /3 mbps (FCC, 2015).1 Given the lack of a universal metric, it is natural that researchers from academia have started to explore a combination of metrics from private and public sector sources to evaluate the implications of the diverse bandwidth landscape for growth and competition (e.g. Lehr, Heikkinen, Clark, & Bauer, 2011; Liebenau, Elaluf-Calderwood, & Kärrberg, 2013). These efforts usually go deep and explore different aspects, but are at the same time usually demographically limited in scope and in the analyzed timespan, since globally harmonized long term time series are not available for more detailed metrics.

On the other hand, the challenge consists in showing why we should care globally about this new dimension of the digital divide. Rather than exploring detailed metrics, these kind of studies ask why is it important to undertake a sustained effort to produce, harmonize and analyze the global evolution of bandwidth (e.g. Vicente & Gil-de-Bernabé, 2010; Hilbert & López, 2011; Riddlesden & Singleton, 2014; Gruber et al., 2014). This present article falls into this second group of research. As such, the article works with a rather rudimentary proxy for bandwidth, but is able to cast a wide global net to show a rough outline of ongoing dynamics for 172 countries2, corresponding to 96% of the world’s population and 99% of the world’s Gross National Income (GNI). The analyzed time series captures 29 years, which covers the entire transition from almost inexistent digitalization (less than 1% of the global information stockpile was digital in 1986), to the full-blown digital age with almost all of it in digital format (Hilbert & López, 2011).

Section snippets

Methodology: a statistical challenge

Much in line with Fig. 1, our two most important indicators refer to the number of subscriptions on the one hand, and bandwidth capacity on the other. The undertaking of creating global time series data faces two main statistical challenges, one related to the creation of national statistics among many countries (in space) and the other one to the creation of normalized time series (in time).

Results: bandwidth divides

Fig. 2 visualizes the source of the discrepancies between the tradition accounting of subscriptions, versus the accounting of bandwidth potential measured in optimally compressed kbps. Both, the accounting of subscriptions (Fig. 2a) and bandwidth capacity (2b) evidence the elimination of the dominance of fixed line telephony, which was the dominating form of distance communication in the late 1980s. In terms of the number of technological devices, this dominance was clearly replaced by mobile

Conclusions

During the discussions about the right choice of indicators to monitor the United Nations Sustainable Development Goals (SDGs) in 2015 (the “post-Millennium-Development-Goals”), two digital divide access indicators were discussed. The indicator “subscription to mobile cellular and/or fixed broad band internet” was evaluated as “feasible, suitable and very relevant (rating AAA)”, while the indicator “fixed and mobile broadband quality measured by mean download speed” was evaluated as “only

Acknowledgments

The author thanks SciDev.Net (Science and Development Network) and the DataPop Alliance for support in the preparation of partial aspects of the involved database. A previous and more limited version of this article was presented at the Hawaii International Conference on System Sciences. The author thanks the comments of reviewers and discussants from this conference.

References (65)

  • J.E. Prieger

    The broadband digital divide and the economic benefits of mobile broadband for rural areas

    Telecommunications Policy

    (2013)
  • D. Riddlesden et al.

    Broadband speed equity: a new digital divide?

    Applied Geography

    (2014)
  • M.R. Vicente et al.

    Assessing the broadband gap: from the penetration divide to the quality divide

    Technological Forecasting and Social Change

    (2010)
  • M.R. Vicente et al.

    Assessing the regional digital divide across the European Union−27

    Telecommunications Policy

    (2011)
  • Akamai. (2014). The State of the Internet (Quarterly Report). Retrieved from...
  • Bauer, S, Clark, D, & Lehr, W. (2010). Understanding broadband speed measurements. In Proceedings of the 38th research...
  • G.E.P. Box et al.

    Empirical model-building and response surfaces

    (1987)
  • M. Büchi et al.

    Modeling the second-level digital divide: a five-country study of social differences in Internet use

    New Media Society

    (2015)
  • Cisco Systems. (2015). Cisco Visual Networking Index: Global Mobile Data Traffic Forecast Update 2014–2019 White Paper....
  • B.M. Compaine

    The digital divide: facing a crisis or creating a myth?

    (2001)
  • A.J. van Deursen et al.

    The digital divide shifts to differences in usage

    New Media Society

    (2014)
  • W.H. Dutton et al.

    Bridging broadband Internet divides: reconfiguring access to enhance communicative power

    Journal of Information Technology

    (2004)
  • Ericsson. (2015). Ericsson Mobility Report (OpenArticle). Retrieved from...
  • European Commission

    Quality of Broadband Services in the EU. DG Communications Networks

    Content Technology

    (2014)
  • FCC (Federal Communications Commission). (2000). Second Broadband Progress Report. Washington D.C. Retrieved from...
  • FCC (Federal Communications Commission). (2004). Fourth Broadband Progress Report: Availability of Advanced...
  • FCC (Federal Communications Commission). (2008). Fifth Broadband Progress Report. Washington D.C. Retrieved from...
  • FCC (Federal Communications Commission). (2012). Eighth Broadband Progress Report. Washington D.C. Retrieved...
  • FCC (Federal Communications Commission). (2014). Measuring Broadband America − 2014. FCC’s Office of Engineering and...
  • FCC (Federal Communications Commission). (2015). Tenth Broadband Progress Report 2015. Washington D.C. Retrieved from...
  • C. Gini

    Measurement of Inequality of Incomes

    The Economic Journal

    (1921)
  • A. Gonzales

    The contemporary US digital divide: from initial access to technology maintenance

    Information, Communication Society

    (2015)
  • Cited by (103)

    • Urbanization and sustainable development for inclusiveness using ICTs

      2022, Telecommunications Policy
      Citation Excerpt :

      Fundamentally, ICT are substantially changing the nature of socio-technical, ecological, and socio-cultural relationships around the world, and acts as a vector of socio-economic, socio-cultural, and environmental changes. In this respect, infrastructural investments in ICT have a remarkable impact on productivity rates of the economy, its competitiveness and economic complexity (Gouvea et al., 2018; Hilbert, 2016). To achieve the Sustainable Development Goals (SDGs) and implement the Paris Agreement, developing countries need to change their energy systems, ecosystem management, agriculture and land use, urban management, material use, gender outcomes, health, education, governance, and other areas.

    • Are we ready to assess digital readiness? Exploring digital implications for social progress from the Network Readiness Index

      2022, Technology in Society
      Citation Excerpt :

      Likewise [82], revealed that usage of mobile data might complement fixed broadband access, i.e., consumers may use mobile data to sample online content, but they complete their online activity using fixed access at home. Finally, quantifying subscriptions is not sufficient [15]. reveals that whereas the digital divide in terms of subscriptions is closing, it is evolving regarding bandwidth.

    View all citing articles on Scopus
    View full text