You are on page 1of 19

Telecommunications Policy 40 (2016) 755–773

Contents lists available at ScienceDirect

Telecommunications Policy
URL: www.elsevier.com/locate/telpol

A cost study of fixed broadband access networks for rural


areas$
Juan Rendon Schneir a,n, Yupeng Xiong b
a
Pompeu Fabra University, Department of Information and Communication Technologies, Tànger 122-140, 08018 Barcelona, Spain
b
Huawei Technologies, Western European Department, Hansaallee 205, 40549 Düsseldorf, Germany

a r t i c l e i n f o abstract

Article history: The deployment of high-capacity broadband access networks in rural areas in Europe lags
Received 24 November 2014 behind that in urban and suburban areas. This study assesses the cost implications for the
Received in revised form rollout of fixed access networks capable of providing citizens with downstream broad-
7 March 2016
band capacities of 30 Mbps or 100 Mbps, which have been defined in the European Digital
Accepted 12 April 2016
Available online 27 May 2016
Agenda as targets that should be met by 2020. A cost model was employed to determine
the cost of a home passed and the cost of a home connected for various fibre- and copper-
Keywords: based networks in rural areas. It was found that the cost of deploying a network outside a
Broadband town or village in a rural area is on average 80% higher than the cost of deploying the
Rural
network in the town or village. This situation may lead to a digital divide within the same
Cost
rural area. For all the geotypes analysed, the following order of costs (in descending order)
Fixed networks
was identified: FTTH, FTTdp-Building, FTTdp-Street, FTTRN, FTTC and CO-VDSL. Given the
long lengths of distribution, feeder and drop segments required, some network archi-
tectures will not be able to provide all households in some areas with the minimum
bandwidth of 30 Mbps as defined in the European Digital Agenda. Overall, it is possible
that operators will need to create a combination of various broadband access networks,
due to the significant cost differences between networks. Policymakers will need to ad-
dress several topics to promote the rollout of broadband networks in rural areas: how the
digital divide within a rural area can be avoided; a National Broadband Plan that clearly
addresses the provisioning of broadband in rural areas; elaboration of studies on broad-
band demand in rural areas; and the assessment of costs and technical capacity of wireless
networks in rural areas.
& 2016 Elsevier Ltd. All rights reserved.

1. Introduction

In several regions all over the world, policymakers have defined broadband targets that should be met within a certain
period of time in order to improve the national or regional telecommunications infrastructure. In Europe, the European
Commission defined the Digital Agenda, which states the following broadband targets: a) basic broadband, which is a
connection that enables at least 144 kbps or 1–2 Mbps for all citizens by 2013; b) download rates of 30 Mbps for all the


The article was prepared when the corresponding author was affiliated to Huawei Technologies. The views expressed in this article are those of the
authors and do not necessarily reflect the opinion of the authors' employers.
n
Corresponding author.
E-mail address: jrendons@gmail.com (J. Rendon Schneir).

http://dx.doi.org/10.1016/j.telpol.2016.04.002
0308-5961/& 2016 Elsevier Ltd. All rights reserved.
756 J. Rendon Schneir, Y. Xiong / Telecommunications Policy 40 (2016) 755–773

citizens by 2020; c) 50% of all the European households should have a broadband subscription of at least 100 Mbps by 2020
(European Commission, 2014a). Whereas basic broadband has been provided for all since 2013, the coverage of Next
Generation Access (NGA) networks, which can provide at least 30 Mbps downstream, was 68% in all of Europe in 2014,
compared to 62% in 2013. In rural areas, there was NGA coverage of 25% in 2014, compared to 18% in 2013 (European
Commission, 2015).
Overall, urban and suburban areas have better broadband coverage than rural areas (European Commission, 2015).
Operators are usually motivated to deploy broadband networks in urban and suburban areas rather than in rural areas,
because of the high density of users willing to pay for high-speed broadband services and the relatively low network rollout
costs in urban and suburban areas. Broadband diffusion begins in general in urban areas, before being expanded to suburban
areas and finally reaching rural areas.
One public policy tool that can be employed to facilitate the rollout of a high-speed broadband infrastructure in rural
areas is the use of state aid or subsidy (FSR, 2011). The European Commission has defined the conditions that should be met
in order to receive state aid for broadband rollout and has classified areas that could potentially receive it (European
Commission, 2013). According to the presence of operators that have or intend to have NGA infrastructures in a specific
region in Europe, the following three types of areas were defined: white areas are those where there is no NGA provider and
where there is unlikely to be any NGA provider in the near future; grey areas are those where there is only one NGA network
and where there will not be another in the near future; black areas are those that have at least two NGA networks. Under
certain conditions, white and grey areas may be eligible for state aid. The majority of rural areas can be classified as white
areas, due to the low NGA coverage in these regions. Different strategic, regulatory, economic, technical and cost aspects
need to be studied in order to understand the motivation of operators to invest in broadband networks in rural areas.
One research topic is therefore the analysis of suitable high-speed broadband access networks that can be deployed in rural
areas and the quantification of the costs of the corresponding network deployment. This is a subject that has drawn the
attention of operators, policymakers, researchers and analysts. A few studies have analysed the cost of broadband networks in
different regions. Elixmann, Ilic, Neumann, and Plückebaum (2008) described the cost of the deployment of fibre to the curb
(FTTC) and fibre to the home (FTTH) networks in various urban, suburban and rural areas for different countries in Europe.
Analysys Mason (2008) studied the cost of FTTC and FTTH networks in different geotypes in the United Kingdom (UK). A study
made by the European Investment Bank found that 41.1% of the total investment needed to provide high-speed broadband
services in Europe would be allocated to rural areas (EIB, 2011). The FTTH Council Europe (2012) analysed the total cost of
deploying FTTH networks in all of Europe. Point Topic (2013) showed that 63.4% of the entire investment required for de-
ploying NGA networks in Europe should be allocated to rural areas. Caio, Marcus and Pogorel (2014) analysed how FTTC/very
high speed digital subscriber line 2 (VDSL2) networks could be employed to meet the targets of the European Digital Agenda
in Italy. In the wireless arena, Frias, Gonzales-Valderrama and Perez Martinez (2015) compared the costs of providing 30 Mbps
downstream through FTTH and wireless Long Term Evolution (LTE) networks in rural Spain. Hallahan and Peha (2010) con-
ducted a cost assessment of a nationwide public safety wireless network in the United States. Prieger (2013) described the
differences in broadband availability between urban and rural areas in the United States.
The above mentioned studies did not, however, address in detail the issue pertaining to the cost of different fixed
networks that can be used to meet the targets of the European Digital Agenda in rural areas by 2020. It is likely that, when
some of the above-mentioned studies had been prepared, the only NGA fixed networks available were FTTC, FTTH and fibre
to the building (FTTB). The development of transmission speed over copper lines has since evolved and additional networks
that comprise hybrid fibre- and copper-based networks, such as fibre to the remote node (FTTRN) and fibre to the dis-
tribution point (FTTdp), can now be employed in rural areas. The objective of this article is to explain the costs of the
different fixed access networks that can be employed in rural areas to provide high-speed broadband services according to
the targets of the European Digital Agenda. The research questions that are addressed in this article are as follows:

1) What are the fibre- and copper-based access networks that can be used in rural areas to provide at least the 30 Mbps or
100 Mbps defined in the European Digital Agenda? 2) What is the cost of the rollout of such networks?

For the analysis of this matter, a cost model was employed to obtain the cost of the different networks. The following six
networks, which are being considered by different operators for a possible network rollout in different regions in Europe,
were analysed: central office VDSL (CO-VDSL), FTTC, FTTRN, fibre to the distribution point – Street (FTTdp-Street), fibre to
the distribution point – Building (FTTdp-Building) and FTTH. The input values employed for the cost model are based on
fibre and copper deployments in different regions in Europe. For the characterization of the rural area, we have employed
six geotypes that differ according to subscriber density and segment lengths. We have considered that there is only one
operator in charge of the rollout of the network, i.e., there is no competition between two or more operators that intend to
deploy the same type of network in the same area. The metrics derived were the cost of a home passed and the cost of a
home connected.
The rest of the article is structured as follows. A literature review of scholarly articles and studies about broadband in
rural areas is presented in Section 2. Section 3 describes the geotypes, network architectures and cost model employed for
the analysis. Section 4 presents the results obtained: the investment per home passed, the investment per home connected,
a sensitivity analysis and the total cost for all the geotypes considered. A discussion of the results and policy implications are
presented in Section 5. Finally, Section 6 presents a conclusion.
J. Rendon Schneir, Y. Xiong / Telecommunications Policy 40 (2016) 755–773 757

2. Literature review

A number of different studies have analysed various aspects of the problem of broadband provisioning and usage in rural
areas; further studies, while not addressing rural areas in particular, have drawn conclusions in their analyses about
broadband that are also relevant for rural areas. From an economic perspective, a useful classification can be made by
differentiating between supply- and demand-side studies. In this section, we will follow this classification in order to
present the studies. Moreover, we will present additional articles that are not exclusively or directly related to supply- or
demand-side aspects.

2.1. Supply-side aspects of broadband

Various articles have explored the supply-side aspects of broadband. Regarding the calculation of costs few studies have
analysed in detail the cost of broadband networks in rural areas. Frias et al. (2015) compare the costs of providing 30 Mbps
through FTTH and LTE networks to rural areas in Spain. The analysis made for different areas of Spain shows that in mu-
nicipalities with between 10,000 and 100,000 inhabitants, it is economically feasible to deploy FTTH networks, whereas in
municipalities with between 1000 and 10,000 inhabitants, it is more feasible in the majority of cases to deploy LTE net-
works. For the majority of municipalities with fewer than 1000 inhabitants, it is not feasible to deploy either FTTH or LTE
networks. Tahon et al. (2014) explore the cost reduction that can be achieved for the rollout of an FTTH network infra-
structure by employing a cooperation model between different utility operators. They show that a synergetic deployment of
new infrastructures can provide a cost reduction of up to 21%. However, the study was not conducted on rural areas. Rendon
Schneir and Xiong (2013) describe the cost reductions that one operator can achieve when employing a co-investment
scheme for FTTH deployments in urban, suburban and rural areas. It is shown that important cost reductions can be
achieved when a network sharing scheme is employed.
Various studies have analysed the regulatory and governmental strategies aimed at promoting broadband deployment.
Some of the different supply- and demand-side aspects that affect NGA deployment in the Netherlands, Sweden and the
United Kingdom are discussed by Ragoobar, Whalley and Harle (2011). The authors find that markets with geographical
constraints will need public support. In the Netherlands, there are no major differences in terms of broadband take up and
coverage between the urban, suburban and rural areas. In Sweden, the government established an information technology
(IT) infrastructure programme at an early stage (in 2001), to promote the deployment of broadband infrastructure in rural
areas, which has helped to mitigate the digital divide. In the United Kingdom, broadband availability in rural areas is quite
limited, due to geographical constraints among other issues. The authors argue that a universal service obligation (USO)
could boost NGA deployment in rural areas in the United Kingdom. The authors consider that demand-side public policies
could be effective after the NGA development has been made.
Briglauer and Gugler (2013) discuss whether the European Union (EU) regulatory framework provides enough incentives
for NGA deployment. By making an international cross-sectional comparison using data relating to FTTH/FTTB networks,
they find that the existing cost-based EU mandatory access regime will not help to achieve the targets of the Digital Agenda.
The authors identify three ways to motivate FTTH/FTTB rollout: a) market-based incentives, similar to the deregulation
strategies employed in the United States, together with an infrastructure competition approach; b) state subsidies, similar to
the ones used by East Asian countries and Australia and New Zealand, which can complement private investments and
would be beneficial for white areas (in many cases, rural areas); and c) favourable country-specific conditions that can
motivate fibre rollout, such as the massive public government initiatives and demand and supply characteristics of East
Asian nations, or the lower consumer migration costs in a few of the Eastern European economies.
Reggi and Scicchitano (2014) explore the European regional digital strategies by examining the allocation of the struc-
tural funds provided by the European Union. The results show that less-developed regions have a tendency to invest their
financial resources in those aspects in which they already demonstrate good relative performance. This unbalanced strategic
approach implies that regions try to improve their strengths, instead of addressing the weaknesses of their overall digital
strategy. Götz (2013) examines the impact of different regulatory regimes on telecommunications operators' incentives to
provide broadband access in areas with different population densities. It is concluded that for reasons of regional policy
supply-side subsidies can contribute to total welfare. Sundquist and Markendahl (2015) present the results of a cost model
of regulatory solutions aimed at addressing the lack of sufficient mobile broadband provisioning in rural areas. It is shown
that the build-out cost can be reduced by employing any or a combination of the following options: state subsidy of the
coverage licence, participation of the end-user in the cost, and network-sharing schemes between the operators.
Nucciarelli, Castaldo, Conte, and Sadowski (2013) examine three regional initiatives to promote broadband deployment
in Italy. It is mentioned that one of the regions studied is a rural area. They show that the main threats to local broadband
initiatives are the following: the geographical extension of projects and low incentives for private investors; an unclear
governance structure; a fast transition from copper to fibre, which does not consider different business strategies; and the
risk of a cannibalization of fixed-mobile services. Ruhle, Brusic, Kittl, and Ehrler (2011) explain various elements of the
nationwide broadband strategies that have been employed in different developed countries: tax reliefs, overall approach to
the ICT industry, public expenditure for network expansion, structures of functional separation, the regulation of network
access, and the introduction of universal service obligations for broadband. The article does not include a quantification of
the effectiveness of each policy instrument. Balmer (2015a) presents a review of the literature related to cooperative
758 J. Rendon Schneir, Y. Xiong / Telecommunications Policy 40 (2016) 755–773

investments in broadband networks. The author analyses different types of cooperative investment, such as joint ventures
and long-term access agreements, and proposes that regulators could employ regulated co-investment agreements to
complement current regulation. Balmer (2015b) presents a literature review that addresses the effect of the geographic
regulation of next-generation access networks on competition, investment and welfare in Europe. It concludes that, in many
cases, by the end of 2013, European regulators were not clear as to how to handle geographic regulation.
Amendola (2015) describes how the access obligations for FTTC networks that are imposed in subsidized areas, which in
many cases correspond to rural areas, could affect the state aid funding of ultra broadband. According to the author, the
wholesale access conditions and prices in subsidized areas may increase the necessary volume of state aid funding. Dauvin
and Grzybowski (2014) evaluate broadband diffusion in the European Union by employing data for 96 NUTS 1 regions, in
order to analyse the effects of regulation and competition on broadband penetration. Even though the data employed are
much more detailed than those employed in other country-based analyses, they still do not have the level of granularity that
would allow a detailed analysis of rural areas.
Prieger (2013) conducted an empirical study of the digital divide and broadband provision and usage in the United States.
Regarding fixed broadband, the author found that there is a higher availability and number of fast broadband providers in
urban areas than in rural areas, and there is a substantial difference in usage rates between urban and rural areas. With
respect to mobile broadband, it was found that rural areas have fewer broadband providers than urban areas, and mobile
broadband covers broadband gaps that are not covered by fixed broadband. For fixed and mobile broadband, the urban and
digital divide is greater among low-income households. Furthermore, the article includes a literature review on the topic of
broadband as a driver of rural economic development.
Very few governments have published a techno-economic analysis of different broadband networks that can be used to
meet the targets of a National Broadband Plan (NBP). In 2010, the Australian Government prepared a plan for the rollout of
its National Broadband Network. As part of this plan, an analysis of the economic resources needed to provide broadband to
all households was prepared. The results of the implementation study gave the following conclusions: 93% of the premises
were going to be served with FTTP (fibre to the premises) networks, the premises located in the range 93–97% were going to
be served with wireless networks, and the remaining 3% of the premises were going to be served with satellite networks
(Australian Government, 2010).

2.2. Demand-side aspects of broadband

In comparison with the studies relating to supply-side topics, there are not so many articles that have analysed the
demand-side aspects of broadband in rural areas in detail. Peronard and Flemming (2011) conducted an empirical study
concerning user motivation to adopt broadband in rural Denmark. The study is based on a set of interviews conducted in a
small village—with only 325 inhabitants—which has the typical features of a Danish parish village, in terms of population
type, employment type, presence of local institutions, and proximity to other nearby villages. Different motivational pat-
terns of rural users with regard to adopting broadband were found. Firstly, broadband helps to improve information ex-
change with other people. Secondly, a broadband connection leads to a better level of local activity—for example, more
people would attend local activities and meetings. Thirdly, a stable broadband connection gives the user a good level of
satisfaction—for example, users can enjoy the new applications and have the feeling of being at the forefront of innovation.
Fourthly, broadband helps to improve the level of transparency and convenience, and saves time—for instance, users can
buy products online. Finally, broadband can have various financial consequences—for example, broadband provides a
cheaper way of doing business. Based on these findings, the authors derive a number of policy implications and propose the
following four types of strategy, which can be employed to motivate the adoption of broadband in rural areas: informative
strategies, affective strategies, habitual strategies, and satisfaction strategies.
Belloc, Nicita and Rossi (2012) study the effect of different supply- and demand-side public policies on wireline
broadband penetration in 30 Organization for Economic Co-operation and Development (OECD) countries. The supply-side
policy variables considered in the empirical study are fiscal incentive programmes and subsidies, long-term loan pro-
grammes for broadband suppliers and national financing programmes, public-private partnership, territorial mapping
programmes, and administrative simplifications. The demand-side public policies analysed are public demand for specific
services, incentives for business demand, incentives for private demand, demand subsidies, and demand aggregation po-
licies. The results show that demand-side policies have a greater impact on broadband penetration than supply-side public
policies. Moreover, the study emphasized the importance of developing a broadband strategy that carefully combines de-
mand- and supply-side public policies for each individual country. However, the study does not make any particular dif-
ferentiation for rural areas; instead, it is based on country-based input data.
Ovando, Perez and Moral (2015) present a techno-economic study considering the feasibility of LTE providing 30 Mbps to
rural areas of Spain. The analysis was performed for areas that cover from 75.3% to 99.3% of the Spanish population, while
the last 0.7%, which corresponds to the inhabitants of extreme rural areas, was not considered. The authors show that
demand is sensitive to price, and that the ARPU should therefore be lowered in order to compete with other broadband
products. It is not clear in the study if the 30 Mbps provided are allocated to every user all of the time, or if there is some sort
of statistical multiplexing involved. Usually with statistical multiplexing, operators provide users with an average
throughput; however, if there are more users in the cell, the transmission speed is reduced. Puschita et al. (2014) describe
the proper end-user rural profile necessary to plan an appropriate broadband strategy from a marketing and technical
J. Rendon Schneir, Y. Xiong / Telecommunications Policy 40 (2016) 755–773 759

perspective in Romania. The empirical study is based on a survey conducted in rural areas in northwestern Romania. The
study shows that Internet access is based mainly on wired broadband networks. Furthermore, the majority of users are low-
income, medium- and low-educated adults. The analysis reveals low overall satisfaction with product service and main-
tenance. Moreover, the rural customer prefers data services to voice and video services.
The broadband demand situation in rural areas in South Korea is described in Park and Kim (2015). South Korea is a
leading country worldwide in terms of broadband penetration; over the last few decades, it has implemented a number of
policies that have helped narrow the broadband infrastructure difference between urban and rural areas. The study analyses
the broadband demand differences between urban and rural users. The data was obtained from a nationally representative
survey of South Koreans (N ¼3641). It was found that the frequency of online activity and the time spent online were similar
in both groups. Rural users take part more frequently in online participatory activities relating to political and social issues.
However, levels of trust and perceptions of the benefits of broadband connection are higher in the urban users' group. The
study suggests that a new type of digital divide exists, since rural users have a relatively low perception of the benefits of
broadband connection.

2.3. Additional studies on broadband

A number of other articles have highlighted various aspects of broadband in rural areas that are not strictly related to
supply- or demand-side matters. Gruber, Hätönen and Koutroumpis (2014) address the economic benefits of implementing
broadband infrastructure that meets the targets of the European Digital Agenda in the European Union. Each EU member
state was split into three regions: urban, suburban and rural areas. For this study, the rural areas are those that have fewer
than 100 inhabitants per km2. The following four levels of service, which were defined in Hätönen (2011), were employed:
“minimum” (theoretical speed, with Internet centres serving rural areas); “base” (theoretical speed, coverage to household);
“advanced” (actual speed, coverage to the household); and “maximum” (actual symmetric speed, coverage to the house-
hold). For the Digital Agenda target of 30 Mbps, it was assumed that in rural areas the VDSL2 and Data over Cable Service
Interface Specification (DOCSIS) 3.0 technologies were able to provide advanced service, whereas FTTH was able to provide
maximum service. For the target of 100 Mbps, it was considered that FTTB and DOCSIS 3.0 were able to provide advanced
service, whereas maximum service could only be provided by FTTH. The cost results presented in Gruber et al. (2014) refer
to the costs of the four above-mentioned service scenarios in every country, i.e., the average values of the urban, suburban
and rural regions are presented, but the individual costs for the rural areas are not shown. A cost benefit analysis of the
Digital Agenda for all of Europe is provided, but there is no specific section that analyses the case of rural areas exclusively.
Kongaut and Bohlin (2014) discuss the following two strategies employed by policymakers in OECD countries to promote
NGA deployment: local loop unbundling (LLU) and infrastructure competition. Their analysis shows the conditions in which
the application of both strategies can be effective. However, as this is an inter-country study, the analysis does not differ-
entiate between urban, suburban and rural areas in a country. Bourreau, Cambini and Hoernig (2012) make a review of the
literature concerning differentiated wholesale remedies that could be implemented in different geographical areas in a
country, as part of an NGA network deployment. The topics that are addressed include the type of regulation that should be
employed in rural areas and what the corresponding access charges could be. The authors conclude that traditional cost-
based access methods might not be the best regulatory tool when promoting infrastructure investment in NGA networks.
Cave (2014) reviews the application of the ladder of investment in Europe as a means to promote the deployment of fixed-
access infrastructure. The interaction of copper pricings with measures to promote fibre deployment in different European
countries is described. In addition, different regulatory measures aimed to promote NGA deployment are criticized. The
article does not contain a specific analysis for rural areas; therefore, a matter for further study might be how the ladder of
investment could be effectively applied in rural areas.
The effect of broadband on productivity in Irish firms is studied by Haller and Lyons (2015). Two types of NUTS 2 region
were analysed: the Border, Midland and Western (BMW) region, and the more affluent South-Eastern (SE) region, which
includes Dublin. The authors found that, in both regions, broadband adoption does not increase firm productivity per se;
instead, industrial productivity is affected by specific applications that require access to the Internet. Pricewaterhouse-
Coopers (2015a) explores the benefits of a national broadband network (NBN) for areas in Ireland that are able to receive
public funding; many of these areas are rural areas. The following benefits are described: residential benefits, enterprise
benefits and eHealth benefits. Moreover, the study describes additional benefits in the following areas: jobs and en-
trepreneurship, education benefits, environmental benefits, social inclusion, balanced regional development, and public
service reform. Grzybowski (2014) considers fixed-to-mobile substitution (FMS) in 27 European Union countries. The author
concludes that the increase in quality and speed of mobile broadband could lead to a decrease in copper fixed-line con-
nections. However, by bundling fixed and mobile lines, this decline in copper-line connections could be slowed down.
Flacher and Jennequin (2014) have studied regulatory policies aimed at promoting investment in next-generation access
infrastructure. It was found that, if a regulator intends to increase the geographic coverage of Next Generation Networks
(NGNs) beyond a situation where there is no regulation, the regulator will need to use other regulatory tools than just
uniform access regulation.
Whitacre, Gallardo and Strover (2014) evaluate the contribution of broadband to economic growth in rural areas in the
United States. They find positive effects of broadband adoption in income growth and employment. However, as opposed to
adoption, broadband availability shows limited impact, which leads to the conclusion that broadband policies should put
760 J. Rendon Schneir, Y. Xiong / Telecommunications Policy 40 (2016) 755–773

more emphasis on demand stimulation. Brown, Browning and Clements (2015) examine the effect on economic perfor-
mance of the United States Department of Agriculture's programmes to provide funding for broadband infrastructure in
rural America. They find that there is a modest but significant statistical liaison between the funding programmes and
county employment and payroll. However, there is no relationship between the programmes and the number of total
business establishments in a community. Stenberg (2010) presents an analysis of the impact of broadband Internet on rural
America. The results show that investment in broadband Internet has a positive effect on the economy of rural communities.
Holt and Galligan (2013) review the United States' federal universal service programmes. Regarding the inclusion of
broadband service as part of the universal service provisioning, it is remarked that the Federal Communications Commission
(FCC) has not been able to address relevant questions relating to the rollout of broadband networks in rural areas.
A few authors have developed a few indexes related to broadband and telecommunications infrastructure. A broadband
achievement index for the United States was developed by Badasyan, Shideler and Silva (2011). The indicators employed for
this composite index are broadband availability, broadband adoption, broadband competition, broadband speed and digital
inclusion. All fixed broadband technologies that provide more than 200 kbps were included in the study, while mobile
wireless broadband was excluded from the paper. One outcome of the study is that the states with good positions in the
broadband index do well in all the indicators, while low-performing states have a below-average or low performance in the
indicators. Gerpott and Ahmadi (2015) have developed a nation's index, which assesses the availability, adoption and usage
of telecommunications networks and services. The index included 111 countries, and the main input data employed were
supply, adoption and usage. However, this study does not focus on rural areas.
The National Broadband Plans of Australia and New Zealand are analysed in Beltran (2014). As was explained before, the
aim of this publicly funded infrastructure project from the Australian Government was to deploy an FTTH network that
could cover 93% of Australian households; the rest of the households would be covered by wireless and satellite networks. In
New Zealand, the Ultra-Fast Broadband (UFB) network is a private-public partnership (PPP) initiative controlled by the
Crown Fibre Holdings (CFH), a state-owned company. The UFB network should be able to reach 75% of the population with
fibre-based networks, while the remaining population would have broadband access through wireless or satellite-based
networks, financially supported by the Rural Broadband Initiative (RBI). The analysis of the supply-side policies reveals that
Australia's NBN Co has had difficulties in meeting the construction targets, but it has a fibre uptake of around 16%. New
Zealand has exceeded its targets, but the uptake of the FTTH network has reached only 3%.
Clarke (2014) describes several methods that can be employed to increase the capability of 4G wireless technologies,
such as the allocation of additional spectrum and the usage of more spectrally efficient wireless technologies. The author
concludes that, in the United States, it is necessary to allocate additional spectrum to meet the demand of the wireless
market. Ling and Wu (2013) address broadband diffusion in OECD countries by identifying the five categories of broadband
adoption defined by Rogers (2003): innovators, early adopters, early majority, late majority, and laggards. The authors
conclude that, to promote broadband deployment, policymakers ought to implement appropriate strategies suited to the
main potential adopters in each broadband adoption stage, instead of applying one-size-fits-all strategies.

2.4. Summary of the literature review

In summary, the literature review shows that there is a lack of cost studies related to broadband networks in rural areas.
Several authors have addressed the topic of regulatory and governmental strategies for broadband rollout in rural areas.
Moreover, there are very few studies regarding broadband demand in rural areas. Different authors have worked on the
effect of broadband on economic growth.

3. Costing methodology

This section describes the procedure employed for the cost analysis, the features of the geotypes, the network archi-
tectures employed and the cost model.

3.1. Cost analysis procedure

The cost study was conducted by taking into account different sources of information. The general procedure employed
for the analysis is depicted in Fig. 1. The information about the geotypes is provided in Section 3.2 of the article. Several
features of the geotypes—i.e., the size of the geotype, the length of the different segments in the access network, the number
of subscribers, the number of street cabinets per central office, etc.—is derived from a study by Analysys Mason (2008),
concerning the deployment of fixed broadband access infrastructure in the United Kingdom. Based on the values of the
parameters of each geotype, it was possible to prepare a network design and define a network topology. Section 3.3 explains
the network architecture employed. With the market share value and the number of total subscribers per geotype, the
number of homes passed and homes connected could be derived. The calculation of the number of network elements was
performed using information about the network topology, and the number of homes passed and homes connected.
As is shown in Fig. 1, the cost per home passed is based on capital expenditures (CAPEX) values, whereas the cost per
home connected is based on CAPEX and operational expenditures (OPEX). Section 3.4 describes the cost model and different
J. Rendon Schneir, Y. Xiong / Telecommunications Policy 40 (2016) 755–773 761

Fig. 1. Flow chart with the procedure for the cost analysis.

input parameters employed, such as the unitary costs of network elements, as well as the financial and OPEX parameters.
The input values for the cost model were obtained over the period 2013–2014 by interviewing some of the manufacturers
and network deployment companies in charge of deploying and maintaining fixed access network infrastructure in Europe.

3.2. Geotype approach

For the definition of geotypes, we employed values taken from a study prepared for the case of the United Kingdom
(Analysys Mason, 2008). Compared with some of the other cost studies that present information about rural areas in Europe
(EIB, 2011; Elixmann et al., 2008; Point Topic, 2013), the information about the rural geotypes is much more exhaustive in
Analysys Mason's report (2008). However, since our study is not a study specifically aimed at the UK, for our analysis input,
we consider parameters that correspond to different countries in Europe (France, Germany and the UK). These values—i.e.,
the unitary costs of the network elements, the OPEX and financial parameters—were derived from sources different to the
study of Analysys Mason (2008). Section 3.4 provides further information on the input parameters employed.
According to Analysys Mason (2008), a rural area is characterized by having two geotypes: a and b (see Fig. 2). The
central office is located in geotype a. For both geotypes, the street cabinet is connected to the central office through the
feeder segment, whereas the premises are connected to the street cabinet through the distribution and drop segments.
Geotype a corresponds to the households located in a town or village. Geotype b refers to an area located outside the town
or village, and far away from the central office. Geotype b can also correspond to the households that are located along a
road that leads to the town or village. Both geotypes differ in terms of the lengths of the feeder, distribution and drop
segments, as well as in the number of street cabinets and premises per central office.
Analysys Mason (2008) defines 13 geotypes that characterize all of the UK. For our study, we have employed the last six
geotypes from the Analysys Mason study, which we have named Rural1a, Rural1b, Rural2a, Rural2b, Rural3a and Rural3b.
Table 1a and b shows the parameters employed for the definition of the six geotypes. For example, in Analysys Mason study,
it was defined that geotype Rural1a is a rural area, whereas geotype Rural1b is a remote area.
In general, different administrations do not share a single definition for what does and does not count as a rural area. The
OECD assumes that a rural area has a population density below 150 inhabitants per km2 (OECD, 2010). The Broadband
Report prepared by the European Commission defined “rural areas” as “those with less than 100 people per km2” (European
Commission, 2015). The Directorate-General for Regional and Urban Policy of the European Commission has prepared a
working paper with a harmonized definition of cities and rural areas (European Commission, 2014b). This paper defined
rural areas as those in which more than 50% of the population lives in rural grid cells. Rural grid cells exist outside urban
clusters, whereas urban clusters are “clusters of contiguous grid cells of 1 km2 with a density of at least 300 inhabitants per
km2 and a minimum population of 5000” (European Commission, 2014b). Moreover, the Government of the United King-
dom has made its own Rural Urban Classification; according to this classification, areas are rural “if they fall outside of
settlements with [a] more than 10,000 resident population” (Department for Environment, Food & Rural Affairs, 2013). The
following different types of rural areas were considered: town and fringe, town and fringe in a sparse setting, village, village
in a sparse setting, hamlets and isolated dwellings, and hamlets and isolated dwellings in a sparse setting.
Moreover, in some countries, there have been discussions about how broadband can reach the final 5% or 10% of the
households (Carnegie UK Trust, 2013). Table 1a shows that the sum of the premises of the six geotypes considered in our
study added up to 34.28% of households. However, if analysing only the four geotypes that correspond to Rural 2 and Rural
3, the percentage of households will only be 12.45%.

Fig. 2. Main network elements of geotypes a and b (source: based on Analysys Mason, 2008).
762 J. Rendon Schneir, Y. Xiong / Telecommunications Policy 40 (2016) 755–773

Table 1
Parameters employed for the definition of geotypes (information taken from the UK): a) classification criteria and densities; b) number of components and
segment lengths (source: based on Analysys Mason, 2008). *Parameters added in order to consider the rollout of street cabinets.

a)

Geotype Classification criteria % of total number of % of total Premises den- Fraction of Average
(distances are straight premises (27.256.460 in area (UK) sity (per sq.km) homes in flats homes per
lines) the UK) block of flats

Rural 1 43000 lines a) 43000 lines o 1 km 10.12% 1.3% 876 8% 4.9


from Central Office
b) 43000 lines 41 km 11.70% 14.2% 93 4% 4.7
from Central Office
Rural 2 o3000 lines a) 41000 lines o 1 km 4.04% 1.6% 285 4% 4.6
and 41000 from Central Office
lines b) 41000 lines 41 km 4.21% 20.8% 23 2% 5.1
from Central Office
Rural 3 o1000 lines a) o 1000 lines o 1 km 1.60% 3.0% 61 2% 3.8
from Central Office
b) o 1000 lines 4 1 km 2.57% 48.9% 6 1% 4.4
from Central Office
Total – 34.28% 89.8% – – –
b)

Geotype Average lines per No. of street cabi- Average lines per Average feeder Average distribution Average drop Total average
Central Office nets per Central street cabinet segment (m) segment (m) segment (m) length (m)
Office

Rural1a 2751 13.4 205 477 246 9 732


Rural1b 3181 22.2 144 2083 705 37 2825
Rural2a 897 4.9 185 192 414 16 622
Rural2b 935 7.6 123 704 1316 70 2090
Rural3a 190 0.0 (2.0)* 0 (95)* 64 422 32 518
Rural3b 305 0.0 (3.0)* 0 (102)* 820 1318 126 2264

3.3. Network architectures

There are three types of fixed broadband access networks: copper, fibre and cable-based networks. For this study, we
considered copper- and fibre-based access networks. Fig. 3 shows the network elements of the different network
architectures.
The networks differ, among other things, in the extent that they reuse the existing copper network. CO-VDSL is a network
that reuses all existing copper-based access network located between the central office and the customer premises
equipment (CPE). With FTTC, the fibre should be deployed up to the street cabinet. With fibre to the remote node, a Mini-
digital subscriber line access multiplexer (DSLAM) is located in the distribution segment. There are typically two possibilities
for powering the Mini-DSLAM. In the first option, it can be powered from the street cabinet, which means that a remote
power unit (RPU) is deployed directly in the street cabinet. In the second option, the Mini-DSLAM is powered from the
central office, which means that there is a remote powering system in both the central office and the street cabinet. The
system in the central office transmits electricity to the street cabinet, which in turn transmits it to the Mini-DSLAM.
For this study, we considered local power from the street cabinet. For FTTdp-Street, a single-port distribution point unit
(DPU) cabinet is located in an area located between the distribution and drop segments. The DPU can use VDSL2/vectoring
over the copper line to provide 100 Mbps when the distance is at most 300 m, or G.fast to provide several hundred Mbps if
the distance is less than 100 m (ITU-T, 2014). For FTTdp-Building, a single-port DPU is located at the entrance of, or inside
the building. VDSL2/vectoring or G.fast can be employed over a copper line. FTTdp-Building has a similar network archi-
tecture as an FTTB network. For FTTdp-Street and FTTdp-Building, a reverse power feeding (RPF), located on the user's
premises, is used to provide energy to the DPU. A gigabit passive optical network (GPON) is used for the FTTH network. For
FTTRN, FTTdp-Street and FTTdp-Building, GPON is employed for transmission over the fibre line. For FTTC, it was assumed
that the existing main distribution frame (MDF), which is located in the street cabinet, could be reused. However, a new
cabinet with a DSLAM should be deployed. FTTC employs gigabit Ethernet (GE) in the feeder segment. In the central office,
the optical line terminal (OLT) contains the GPON and upstream Ethernet ports used for the FTTH, FTTRN and FTTdp net-
work architectures. The optical network terminal (ONT) is employed in the FTTH network architecture.
The design of the network architectures and network components was created so that every user could receive,
whenever possible, a transmission speed of at least 30 Mbps or 100 Mbps. For network architectures that have copper-based
segments, there is a broadband capacity limitation that depends on the length of the copper cable. It was assumed that for a
copper line length of at most 300 m, a downstream transmission capacity of 100 Mbps was possible, whereas for a copper
line length of at most 1 km, a downstream transmission capacity of 30 Mbps was possible. These broadband capacity values
J. Rendon Schneir, Y. Xiong / Telecommunications Policy 40 (2016) 755–773 763

Feeder Street Distribution Drop


a) Central Office
Segment Cabinet Segment Segment
In-building
Segment
DSLAM
MDF MDF

Feeder Street Distribution Drop In-building


b) Central Office
Segment Cabinet Segment Segment Segment
OLT ODF DSLAM MDF

Feeder Street Distribution Drop


c) Central Office
Segment Cabinet Segment Segment
In-building
Segment

Feeder Street Distribution Drop


d) Central Office
Segment Cabinet Segment Segment
In-building
Segment

Feeder Street Distribution Drop In-building


e) Central Office
Segment Cabinet Segment Segment Segment

Feeder Street Distribution Drop In-building


f) Central Office
Segment Cabinet Segment Segment Segment

Fig. 3. Network architectures: (a) CO-VDSL; (b) FTTC; (c) FTTRN; (d) FTTdp-Street; (e) FTTdp-Building; (f) FTTH.

can be reached when using VDSL2/vectoring. For this study, we assumed that VDSL2/vectoring was being used in the CO-
VDSL, FTTC and FTTRN architectures. VDSL2/vectoring or G.fast could be employed for FTTdp-Street and FTTdp-Building, We
focused on these two values (30 Mbps and 100 Mbps downstream), because they have been defined as broadband targets by
the European Commission, as was mentioned in Section 1 of this paper. Given the limitation that the maximum distance of
the copper line segment needed to be at most 1 km to provide all the households with 30 Mbps, in a few geotypes there are
a few networks that will not able to entirely meet the targets of the European Digital Agenda.
Table 2 shows the minimum capacity that can be provided by every type of network for the different geotypes. Table 2
provides the answer to the research question 1 posed in Section 1: what are the fibre- and copper-based access networks
764 J. Rendon Schneir, Y. Xiong / Telecommunications Policy 40 (2016) 755–773

Table 2
Minimum broadband capacity downstream that can be provided according to the broadband targets of the European Digital Agenda (30 Mbps or
100 Mbps). The term ‘distance’ refers to the copper line segment length.

Network type Geotype “a” Geotype “b”

CO-VDSL Rural 1a, 2a and 3a: 30 Mbps Rural 1b, 2b and 3b:
(300 mo distanceo 1 km) 30 Mbps not possible
for all households
(distance 41 km))
FTTC Rural 1a: 100 Mbps Rural 1b: 30 Mbps
(distance o 300 m) Rural 2a (distance o1 km) Rural
and 3a: 30 Mbps 2b and 3b: 30 Mbps
(distance o 1 km) not possible for all
households
(distance 41 km)
FTTRN 100 Mbps (distance o 300 m) 100 Mbps
(distance o300 m)
FTTdp-Street 100 Mbps (distance o 300 m) 100 Mbps
(distance o300 m)
FTTdp-Building 100 Mbps (distance o 300 m) 100 Mbps
(distance o300 m)
FTTH 100 Mbps 100 Mbps

that can be used in rural areas to provide at least the 30 Mbps or 100 Mbps defined in the European Digital Agenda? Given
the maximum total length of 1 km, it is possible for CO-VDSL to provide 30 Mbps in geotypes a, i.e., Rural1a, Rural2a and
Rural3a. Information regarding the total average segment lengths is shown in Table 1b. In the case of geotypes b, it was not
possible for CO-VDSL to provide all households with 30 Mbps; consequently, we did not consider CO-VDSL for geotypes b in
the cost analysis. FTTC could provide 100Mbps for geotype Rural1a, given the maximum distance of 300 m, but only
30 Mbps for geotypes Rural2a and Rural3a. For geotype Rural1b, FTTC could provide 30 Mbps, but could not provide all
households with 30 Mbps for geotypes Rural2b and Rural3b, because the total segment length was longer than 1 km.
Therefore, we did not consider FTTC for geotypes Rural2b and Rural3b in the cost analysis.
The network design for FTTRN was created so that the Mini-DSLAM could be placed at a point where the maximum
copper line length was 300 m from the subscriber's premises, which enabled a downstream transmission capacity of
100 Mbps. The Mini-DSLAM supports up to 48 VDSL2 ports; however, this maximum value of ports was not reached in our
network design, because the number of Mini-DSLAMs depends on the maximum distance up to the customer's premises, as
well as on the minimum bandwidth that should be provided. One splitting level of 1:4 was employed in the street cabinet
for the FTTRN network architecture, which provided on average a transmission speed of 625 Mbps per Mini-DSLAM. For
FTTdp-Street, FTTdp-Building and FTTH, a splitting ratio of 1:32 was employed in the street cabinet, which provided each
household on average a downstream transmission capacity of 78 Mbps.
For the design of the networks that could provide 100 Mbps downstream, we used a statistical multiplexing factor of
20%, a value that is similar to the ones employed in fibre-based networks in Europe. This implies that on average every user
should have at least 20 Mbps downstream at all times. Table 1a shows the percentage of households in flats; this value
ranged from 8% for Rural1a to 1% for Rural3b. Given the low number of households in multiple dwelling units (MDUs), we
have assumed for our cost calculation that all the households in all the geotypes were located in single dwelling units
(SDUs). For the feeder and distribution segments, the deployment was terrestrial, i.e., it was necessary to invest in digging in
order to lay out the fibre. For the drop segment, it was assumed that 45% of the drop segment would consist of aerial
deployment, which implies that the existing poles would be reused. It was assumed that two distribution segments were
connected to a street cabinet. In all cases, there was only one network operator in charge of deploying the network; for this
study, we did not analyse the case where different operators intended to reuse the same access network infrastructure.

3.4. Cost model

The cost per home passed and the cost per home connected were the two metrics used in this study. Whereas the cost of
a home passed did not involve all the network elements needed to connect a user, nor did it consider the effect of market
share value, the cost of a home connected considered all the network elements for an end-to-end connection in the access
network and the effect of market share value. For the cost of a home passed, a broadband coverage of 100% was assumed,
whereas for the cost of a home connected a broadband coverage of 100% and a broadband penetration equivalent to the
market share value was assumed. For FTTH, it was assumed that the cost of a home passed involved all the costs necessary to
deploy the fibre up to the basement of the building. For the calculation of the cost of an FTTH-connected home, it was
considered that a technician would need to go to the house to rollout the in-building fibre cable. Furthermore, the cost of the
CPE should be taken into account for the cost of a home connected for FTTH. For the FTTdp network architectures, the cost of
a home passed considered the rollout of the fibre up to the location of the DPU cabinet. The cost of a home connected
Table 3

J. Rendon Schneir, Y. Xiong / Telecommunications Policy 40 (2016) 755–773


Assumptions used to calculate the cost per home passed and the cost per home connected.

Network Cost per home passed Cost per home connected

All networks Network-specific All networks Network-specific

CO-VDSL – CAPEX – Deployment in the central office – CAPEX and OPEX – Maintenance of all copper infrastructures.
FTTC – 100% of homes are passed – Deployment of equipment in the central office and – 100% of homes are passed and the number – Maintenance of the copper infrastructure located
– Four years of initial invest- street cabinet up to the street cabinet. of homes connected depends on the mar- between the street cabinet and the customer's
ment are considered. ket share premises
FTTRN – No customer equipment – Deployment of Mini-DSLAM in the distribution – 15 years are taken into account – Maintenance of the copper line located between
segment. – CPE and customer equipment provided the street cabinet and the customer's premises
– Deployment of fibre up to the Mini-DSLAM
FTTdp-Street – Deployment of fibre up to the place where the DPU – Deployment of the DPU
will be located between the distribution and drop – Maintenance of the copper line located between
segments the DPU and the customer's premises
FTTdp- – Deployment of fibre up to the building – Deployment of the DPU
Building – Maintenance of the copper line located between
the DPU and the customer's premises
FTTH – Deployment of fibre up to the building – Deployment of in-building fibre

765
766 J. Rendon Schneir, Y. Xiong / Telecommunications Policy 40 (2016) 755–773

included the cost of the DPU cabinet, the cost of the installation of the DPU cabinet by a technician and the RPF and CPE. The
cost of a home passed for FTTRN included the cost of the Mini-DSLAM and the cost of the rollout of fibre up to the Mini-
DSLAM. For FTTC, the cost of the equipment in the street cabinet and the cost of the fibre rollout up to the street cabinet
were assigned to the cost per home passed. For the CO-VDSL network architecture, the cost of a home passed included the
cost of the VDSL2 equipment in the central office. Table 3 summarizes the assumptions employed when calculating the cost
per home passed and the cost per home connected. The cost per home connected included all the costs of the cost per home
passed, as well as additional costs.
A greenfield approach was considered for the deployment of the fibre-based infrastructure, i.e., it was assumed that there
was no existing fibre infrastructure available and that all ducts needed to be deployed. CAPEX and OPEX were considered in
the cost analysis. For the calculation of the majority of the CAPEX, a simplified bottom-up analysis was conducted. For the
calculation of the OPEX, a top-down analysis was conducted. To obtain the present value of the investments made, a cu-
mulative present value (CPV) with a discount rate (DR) of 10.5% was used. The discount rate consisted of a weighted average
cost of capital (WACC) of 8.5% and a risk premium of 2%. These numbers were similar to values employed by fixed operators
that deploy fibre-based access networks in Europe. For the calculation of the cost per home passed, the CPV of the CAPEX
employed for rolling out the network over the first four years was taken into account. To derive the cost per home con-
nected, the CPV of the CAPEX and the OPEX required to rollout the network and maintain it over 15 years was considered.
For this study, we assumed that the term cost of an individual network element referred to the purchase price. To derive the
annual CAPEX, all values of the cost of the individual network elements were added. Subsequently, the OPEX was derived.
Finally, the discount rate was employed to calculate the cumulative present value of the total investment.
The individual costs of the network elements were derived by obtaining information from eight of the companies in
charge of deploying fixed access infrastructure in France, Germany and the UK over the period 2013–2014. Thereafter,
average values were derived. These average values were in the order of magnitude for fibre- and copper-based access
networks in some regions in Europe. Some of these values had previously been used by the authors for a research study on
FTTdp and FTTH networks (Rendon Schneir and Xiong, 2015). The cost of trenching and duct deployment in the feeder and
distribution segments was €65 per metre; the cost of this item was €50 in the drop segment. The cost of the aerial de-
ployment of fibre in the drop segment was €12 per metre. The connection works for fibre deployment inside the building for
a new FTTH user was €220. The connection works for a new DPU cabinet for the FTTdp-Building and FTTdp-Street network
architectures was €90 and €140, respectively.
The CAPEX involves the investment needed to rollout the active and passive infrastructure. For fibre rollout in the feeder,
distribution and drop segments, the CAPEX consisted of the following items: the cost of digging, deployment of ducts and
the rollout of fibre and manholes. The OPEX included the cost of maintaining the active and passive infrastructure. The
following mark-up values were applied to the CAPEX in order to obtain the annual OPEX: 7.5% for the active network
elements and 1% for the passive infrastructure. In the central office, the OPEX also included the cost of floor space rental and
energy consumption. For the calculation of the energy consumption, a value of €0.16/kW h was employed. The OPEX in-
cluded the cost of the maintenance of the copper line. Several companies were asked to obtain costs for the maintenance of
the copper lines. The following average values were employed in the analysis for the monthly maintenance of a copper line:
€0.8, €0.3, €1.1 and €0.8 for the in-building, drop, distribution and feeder segments, respectively.
Regarding the asset lifetimes, the following values were employed: six years for the active network elements located
inside the customer's premises (CPE, ONT and RPF); eight years for the rest of the active network elements (DSLAM, DPU and
OLT equipment); 25 years for the cables and fibre; 45 years for the ducts. The renewal of the active equipment was taken
into account in this study, as it has a lifetime of less than 15 years, as considered in the cost analysis.
With regard to the time needed to deploy the network, it was considered that the network be deployed over the first four
years. The percentage of premises passed was 25%, 50%, 75% and 100% in years one, two, three and four, respectively. To
derive the cost of a home connected, the value of broadband penetration, which we labelled the target market share for our
analysis, was used. It was assumed than an operator would reach 22.5%, 45%, 67.5% and 90% of the target market share in
years one, two, three and four, respectively. Following on, the cost model considered an annual take-up rate of 0.9625%,
which would permit the operator to reach 100% of the target market share in year 15. Furthermore, the cost model used an
annual churn rate of 10% as of year two.
The following items were not considered in the cost analysis: the cost of aggregation, core and backhaul networks; the
cost of the provisioning of telephony, video or broadband services; marketing and sales costs; common costs, e.g., those
related to the management and administration tasks that are not allocated to individual services such as strategy, human
resources, research or regulatory departments (WIK-Consult, 2013); the cost of permits for deploying the needed infra-
structure; the cost of engineering drawings.

4. Results

This section provides the answer to the research question 2 posed in Section 1: What is the cost of fibre- and copper-
based access networks that can be employed in rural areas to provide at least the 30 Mbps or 100 Mbps defined in the
European Digital Agenda?
J. Rendon Schneir, Y. Xiong / Telecommunications Policy 40 (2016) 755–773 767

Table 4
Investment per home passed, CAPEX.

Rural1a Rural1b Rural2a Rural2b Rural3a Rural3b

CO-VDSL €17 – €19 – €52 –


FTTC €230 €697 €176 – €260 –
FTTRN €384 €1.131 €349 €1.241 €552 €1.257
FTTdp-Street €394 €1.195 €405 €1.329 €706 €1.302
FTTdp-Building €615 €2.104 €798 €3.049 €1.492 €4.397
FTTH €615 €2.104 €798 €3.049 €1.492 €4.397

4.1. Investment per home passed

Table 4 shows the results related to the investment per home passed, which were obtained by calculating CAPEX values.
The cost for FTTH and FTTdp-Building were the same, as it was assumed that the cost per home passed considered all the
network elements from the central office up to the basement of the building. As there was no drop segment when calcu-
lating the cost of a home passed for FTTdp-Street, the cost of FTTdp-Street was lower than that of FTTdp-Building and FTTH.
The cost of FTTRN was slightly lower than the cost of FTTdp-Street, because for FTTRN, the cost of the Mini-DSLAM was
included and the Mini-DSLAM was deployed in an area located no more than 300 m from the customers' premises. For every
rural scenario, the cost per home passed for geotype b was higher than for geotype a. This was the case because the feeder,
distribution and drop segments were longer in all cases for geotype b compared to geotype a. The cost for FTTC also included
the installation of the street cabinet with the DSLAM, whereas the cost for CO-VDSL included the cost of the VDSL2 ports.
The cost composition of the cost per home passed for geotypes Rural2a and Rural2b is shown in Table 5. The cost
percentage for the feeder segment ranged from 4% for FTTdp-Building and FTTH to 18% for FTTC for the Rural2a geotype. The
cost percentage of the distribution segment ranged from 36% for FTTdp-Building and FTTH in the case of geotype Rural2b to
84% for FTTRN in the case of geotype Rural2a. For FTTdp-Building and FTTH, the cost percentage of the drop segment was
48% and 56% for geotypes Rural2a and Rural2b, respectively.

4.2. Investment per home connected

To derive the cost of a home connected, the CAPEX and OPEX were considered over a period of 15 years with a market
share of 50%. Fig. 4 shows that in all cases, the cost per home connected indicated the following descending order: FTTH,
FTTdp-Building, FTTdp-Street, FTTRN, FTTC and CO-VDSL. For geotypes b, the cost of FTTdp-Building and FTTH was markedly
higher than the costs of the other network architectures. This was the case because of the investment needed to rollout the
fibre in the drop segment for both network architectures. For the three rural scenarios, Rural1, Rural2 and Rural3, the cost
per home connected for geotype b was higher than the cost for geotype a.
The cost composition for geotypes Rural2a and Rural2b is shown in Table 6. For CO-VDSL in geotype Rural2a, the cost
percentage of the central office was 62%, whereas for FTTC the cost percentage of the street cabinet was 54%. For geotype
Rural2b, the cost percentage of the distribution segment, which contains the Mini-DSLAMs and the fibre rollout up to the
Mini-DSLAMs, was 69% for geotype Rural2a and 75% for geotype Rural2b. For FTTdp-Street, the cost percentage of the
distribution segment was 39% and 62% for geotypes Rural2a and Rural2b, respectively. For FTTdp-Building and FTTH, the
major cost contribution was the distribution segment, with 34% and 43% for geotypes Rural2a and Rural2b, respectively.

4.3. Sensitivity analysis: effect of the market share on the cost

In order to understand the impact of the market share value on the cost per home connected, a sensitivity analysis
was conducted (see Fig. 5). As was explained in Section 2, there is a lack of sufficient studies concerning the demand
for broadband in rural areas of Europe. Therefore, for our study, we decided to assess the effect of different levels of
market share value on the cost. The market share values ranged from 20% to 80% and the geotype employed was Rural2a.

Table 5
Cost composition, investment per home passed.

Rural2a Rural2b

CO-VDSL FTTC FTTRN FTTdp-Street FTTdp-Building FTTH FTTRN FTTdp-Street FTTdp-Building FTTH

Central office 100% 8% 4% 8% 4% 4% 2% 3% 1% 1%


Feeder segment – 18% 9% 7% 4% 4% 15% 14% 6% 6%
Street cabinet – 74% 3% 5% 3% 3% 1% 2% 1% 1%
Distribution segment – – 84% 80% 41% 41% 82% 81% 36% 36%
Drop segment – – – – 48% 48% – – 56% 56%
768 J. Rendon Schneir, Y. Xiong / Telecommunications Policy 40 (2016) 755–773

Fig. 4. Investment per home connected, CAPEX and OPEX, 50% market share.

Table 6
Cost composition, investment per home connected, CAPEX and OPEX, 50% market share.

Rural2a Rural2b

CO-VDSL FTTC FTTRN FTTdp-Street FTTdp-Building FTTH FTTRN FTTdp-Street FTTdp-Building FTTH

Central office 62% 12% 9% 11% 9% 9% 4% 5% 4% 4%


Feeder segment 6% 7% 5% 4% 3% 3% 12% 11% 7% 7%
Street cabinet 1% 54% 3% 3% 2% 2% 2% 2% 1% 1%
Distribution segment 8% 7% 69% 39% 34% 34% 75% 62% 43% 43%
Drop segment 2% 2% 1% 1% 20% 19% 1% 1% 33% 33%
In-building segment 6% 5% 4% 3% 3% 21% 2% 1% 1% 8%
DPU elements – – – 29% 20% – – 13% 8% –
(ONT/RPF)þ CPE 15% 13% 9% 10% 9% 12% 4% 5% 3% 4%

In Section 3.2, the baseline value of the market share employed for the analysis was 50%. Fig. 5 shows that for all market
share values, the cost values had the following descending order: FTTH, FTTdp-Building, FTTdp-Street, FTTRN, FTTC and CO-
VDSL. This order was also observed for the rest of the geotypes. For the lowest values of the market share, the cost of the
FTTRN, FTTdp-Street, FTTdp-Building and FTTH was much higher than those of CO-VDSL and FTTC, because of the invest-
ment needed to deploy the fibre in the distribution segment.

Fig. 5. Effect of the market share on the investment per home connected, CAPEX and OPEX, 50% market share, geotype Rural2a.
J. Rendon Schneir, Y. Xiong / Telecommunications Policy 40 (2016) 755–773 769

Fig. 6. Cumulative value of investment per home connected, CAPEX and OPEX, 50% market share. Previous cost values up to 66% of the population were not
considered.

4.4. Total costs

The costs necessary for rolling out a network at the national level have been calculated. This was done by multiplying the
number of households located in a geotype by the cost per home connected in the geotype. Fig. 6 shows the total costs for
the six geotypes analysed in the study. The previous cost values up to 66% of the population were not considered because
this study focused exclusively on the six geotypes that covered around 34% of all the households (see Table 1a). If all the six
geotypes were to be covered with FTTH, the total cost would be €33.6 billion, whereas the total cost would be €25.7 billion if
the six geotypes were to be covered with FTTdp-Street. For the initial three geotypes, Rural1a, Rural1b and Rural2a, the FTTC
cost will be €10.3 billion. The cost of CO-VDSL for the Rural1a geotype was €2.2 billion.

5. Assessment of the results and policy implications

5.1. Assessment of the results

Various characteristics were detected from the results presented in the article. Firstly, the cost of rolling out a network in
geotype b was in all cases higher than in geotype a. The average cost of all the networks in geotype b was 97%, 98% and 44%
higher than the average cost of all the networks in geotype a for the Rural1, Rural2 and Rural3 areas, respectively. These
values lead to a total average cost increase of 80%. Analysys Mason (2008) found that the costs per home connected through
an FTTH/GPON deployment for networks in b geotypes are around 240% higher than the equivalent costs in a geotypes. This
is basically due to the longer lengths of the distribution and drop segments required in b geotypes.
Secondly, in terms of the cost of a home connected, the cost of FTTdp-Building was on average 1.4% lower than the cost of
FTTH. In both cases, the majority of the costs—at least 68%, as shown in Table 6—were allocated to network elements located
between the central office and the basement of the building. Given this relatively low difference between the costs of FTTH
and FTTdp-Building, it is understandable that in some cases operators prefer to continue reusing the copper line inside the
house, in order to avoid the management overheads and efforts incurred by in-house fibre deployment.
Thirdly, four network architectures were able to provide 100 Mbps downstream for all geotypes: FTTRN, FTTdp-Building,
FTTdp-Street and FTTH. The mean costs of FTTdp-Building and FTTH were on average 30% and 58% higher than the mean
costs of FTTRN and FTTdp-Street in geotypes a and b, respectively. Therefore, if the aim is to provide 100 Mbps downstream,
FTTRN and FTTdp-Street would be able to achieve this at a lower cost. The cost of FTTdp-Street was on average 18% higher
than the cost of FTTRN. Some of the different studies mentioned in Sections 1 and 2 of the present article—such as Analysys
Mason (2008), EIB (2011), Elixmann et al. (2008), FTTH Council Europe (2012), Frias et al. (2015) and Tahon et al. (2014)—are
based on cost calculations made for FTTH and/or FTTC networks. In the future, other types of networks, such as FTTRN and
FTTdp-Street, could also be employed to calculate the deployment costs of high-speed access networks in rural areas.
Fourthly, CO-VDSL had the lowest cost and was able to provide 30 Mbps in geotype a. FTTC had the lowest cost and was
able to provide 30 Mbps in geotype Rural1b. For geotype Rural1b, the cost of a home connected with FTTC was 36% lower
than the cost of FTTRN. Therefore, from the cost perspective, the CO-VDSL solution is the most economic one. However, it
would only be effective in a geotypes, leaving many of the premises located in b geotypes without a proper NGA connection.
Finally, it was found that for all the market share values, the cost values had the following descending order: FTTH,
FTTdp-Building, FTTdp-Street, FTTRN, FTTC and CO-VDSL. It is not surprising to find that the FTTH network is the most
expensive one. Moreover, given the limited amount of investment needed to deploy the CO-VDSL infrastructure, it was
expected to derive low values for the cost of CO-VDSL networks. However, the descending order of costs, based on CAPEX
and OPEX over a period of 15 years, leads to the following conclusion: the more fibre is deployed, the more expensive the
770 J. Rendon Schneir, Y. Xiong / Telecommunications Policy 40 (2016) 755–773

network will be.

5.2. Policy implications

The analysis undertaken in the previous sections highlights a series of issues that need to be addressed by policymakers.
The topics mentioned in the following deserve further research and consideration; the first of these is the risk of a digital
divide within a single rural community. From a cost perspective, this study has shown that, in rural areas, the cost of pro-
viding a high-speed broadband service is on average 80% higher for geotype b than for geotype a. This could lead to a
situation in which operators are more interested in rolling out high-speed broadband networks in geotype a than in geotype
b. Some authors have already identified this situation in other regions; for example, Gijon, Whalley and Anderson (2015)
have identified broadband speed differences between different areas in the city of Glasgow. Grubesic (2010) describes a
process for collecting data concerning broadband provisioning in the United States, and identifies “issues of equity and [the]
availability of broadband for all residents within a community”.
Policymakers should be aware of the differences in transmission speeds that can be produced within a rural community
when a particular type of technical solution is chosen. For example, for the case of rural Ireland, it was found that the
technical solution chosen can result in an internal digital divide. One report on the broadband strategy for Ireland states that
a telecom network consists of two parts: a) a core and backhaul network, where the fibre is deployed up to the street
cabinet or central office, and which would therefore imply an FTTC or a CO-VDSL solution, respectively; and b) an access
network—the section known as “the last mile”—which connects the premises to the point where the backhaul network is
available (PricewaterhouseCoopers, 2015b). Given the current technical limitations, a downstream capacity of 30 Mbps can
only be provided if the premises are located less than 1 km away from the street cabinet or central office. Five villages in
rural Ireland were analysed as case studies and it was found that, on average, 45% of the premises were located at a distance
of less than 1 km from a street cabinet or central office. As a consequence, they would be able to receive a minimum
transmission speed of 30 Mbps by using FTTC or CO-VDSL. The remaining 55% of the premises fell outside of the 1 km range
and would therefore need a different type of technical solution.
The second topic concerns a proper National Broadband Plan. Several jurisdictions have already defined (or are in the
process of preparing) the National Broadband Plan that will be implemented. OECD (2011) compares the National Broad-
band Plans of different developed countries and concludes that, with the exception of heavily urbanised countries such as
Singapore and Luxembourg, in the majority of cases NBPs have not made commitments to provide every home with a very
high-speed broadband service. A National Broadband Plan should specify, among other things, what type of broadband
transmission capacity will be provided in rural areas, who is going to bear the rollout costs of different broadband networks
so that the targets of the European Digital Agenda are met, and when these targets should be achieved. However, what is
currently not clear enough is the overall strategy of each country or region in Europe for meeting the broadband targets in
rural areas, in terms of NGA provisioning, in 2020. With an NGA coverage of only 25% in rural areas in Europe (European
Commission, 2015), it would be desirable for the National Broadband Plans to make a much more detailed analysis of the
expected broadband diffusion in rural areas over the next years.
The third topic for policymakers is the question of funding and the allocation of subsidies. The cost of network rollout in
rural areas is higher than in urban or suburban areas (Analysys Mason, 2008; Elixmann et al., 2008), which implies that
some sort of state aid might be needed to provide broadband services to the grey and white areas. Policymakers should
address the following question: what type of broadband transmission capacity should be provided by networks that receive
state aid—30 Mbps, 100 Mbps, or a different value? As we have seen in this study, there are important cost differences
between the different types of network. Briglauer and Gugler (2013) explain that state subsidies can be beneficial for white
areas (i.e., rural areas). Moreover, the principle of universal service has been extensively applied to telephony services and
the provisioning of basic broadband in several regions across the world. With regard to high-speed broadband transmission
capacity, it should be considered whether this principle should be applied, and if so, how (Alleman, Rappoport and Banerjee,
2010; Bohlin and Teppayayon, 2010; OECD, 2012). Ragoobar et al. (2011) have stated that a universal service obligation
would help to promote NGA rollout.
Fourthly is the issue of technology neutrality and the combination of different networks. In Europe, the principle of tech-
nology neutrality is applied, which ensures that no particular type of technology or network is promoted or discriminated.
The application of this principle, together with the fact that access networks have different costs, might lead to a situation
where different types of access networks might be used in the same EU Member State. For example, for broadband pro-
visioning of 30Mbp in geotype Rural1a, the least expensive technology type to use is CO-VDSL. In geotype Rural1b, FTTC has
the lowest cost. Meanwhile, for broadband provisioning of 100 Mbps in geotype Rural2b, FTTRN would be the least ex-
pensive network. The evolution of copper-based technologies could provide an extension to the lifespan of copper lines and
complement the deployment of fibre-based networks to provide NGA rollouts in some rural environments.
The fifth area on which policymakers should focus is studies on broadband demand in rural areas. The present article has
focused on the cost of the rollout of broadband networks. In order to construct a business case, detailed studies of the
demand for high-speed broadband services in rural areas should be conducted. Furthermore, an affordability analysis of
broadband services in rural areas should also be made; Analysys Mason and Tech4i2 (2013) includes a 30 Mbps broadband
demand study for Europe, but they do not distinguish between rural and other areas. Elsewhere, Peronard and Flemming
(2011) explain broadband demand in rural Denmark, and Park and Kim (2015) explain the situation of broadband demand in
J. Rendon Schneir, Y. Xiong / Telecommunications Policy 40 (2016) 755–773 771

rural South Korea. Updated studies about broadband demand in rural regions should be conducted in different European
countries.
The sixth topic concerns the costs when a co-investment or network sharing scheme is employed in rural areas. The analysis
conducted for this article considers only one operator in charge of deploying the infrastructure. It is possible that innovative
business models may emerge that change the underlying assumptions regarding the provision of infrastructure in rural
areas. Rendon Schneir and Xiong (2013) show how the costs are allocated when a co-investment scheme is employed for
FTTH deployments. Further studies are needed in order to understand better the cost implications of network sharing
schemes for broadband rollout in rural areas.
Finally, the cost of other broadband networks should be studied. The current study has presented the results for the rollout
of fibre- and copper-based access networks in rural areas. In general, it is believed that wireless networks are capable of
providing an appropriate broadband service in rural areas. Certain features of wireless networks such as long-term evo-
lution (LTE), LTE-Advanced and small cells should be addressed. For example, what is the cost of a wireless network if it has
to provide 30 Mbps or 100 Mbps downstream for all households in a rural area? In this sense, Frias et al. (2015) describe the
cost of an LTE network that provides 30 Mbps in rural Spain. The cost results for wireless networks should be compared with
the results for the rollout of a fixed access network, which would help us to better understand the cost advantages and
drawbacks of fixed and wireless networks. Moreover, the cost of cable-based access networks should also be analysed. It is
not clear whether the cost of satellite-based networks should be considered as part of an NGA network rollout study, since
the downstream transmission speed of these networks nowadays is lower than 30 Mbps.

6. Conclusions

The rollout of high-speed broadband access networks in rural areas in Europe lags behind the rollout of broadband
networks in urban and suburban areas. In order to understand the cost implications of the rollout of broadband networks
that meet the broadband targets of 30 Mbps or 100 Mbps as defined in the European Digital Agenda, this article addresses
the following questions: 1) What are the fixed access networks that can meet the targets of the European Digital Agenda in
rural areas? And 2) What is the cost of these networks? Different fibre- and copper-based networks were analysed.
In the cost model employed, we differentiated between a geotype labelled a, which included households located in a
town or village close to the central office and a geotype labelled b, which included households located outside the town or
village. It was found that the cost of broadband deployment for geotype b was on average 80% higher than for geotype a.
This situation can lead to the creation of a digital divide within a rural area. For all the geotypes analysed, the following
order of costs (in descending order) was identified: FTTH, FTTdp-Building, FTTdp-Street, FTTRN, FTTC and CO-VDSL. Given
the extended lengths of the distribution, feeder and drop segments, some network architectures will not able to provide all
households with the minimum bandwidth of 30 Mbps as defined in the European Digital Agenda. Overall, the study showed
that operators will likely use a combination of different broadband access networks due to the significant differences in
costs for implementing different networks in different regions.
The article has highlighted the following topics that should be addressed in detail by policymakers in order to provide a
solution to the low deployment of high-speed broadband networks in rural areas: the risk of a digital divide within a rural
area; the definition of how broadband deployment in rural areas will be addressed as part of a National Broadband Plan; and
the allocation of subsidies to support the rollout of networks capable of providing 30 Mbps or 100 Mbps. Further research is
also needed to understand the demand of broadband networks in rural areas, a cost-benefit analysis for rural areas, the
network costs when a co-investment scheme is employed, and the broadband capacity and deployment costs of wireless
networks such as LTE, LTE-Advanced and small cells.

Acknowledgements

The authors would like to express their gratitude to the reviewers and editors of the Telecommunications Policy journal
as well as the participants of the European Regional ITS Conference 2014 held in Brussels, Belgium, for the useful remarks
provided.

References

Alleman, J., Rappoport, P., & Banerjee, A. (2010). Universal service: a new definition? Telecommunications Policy, 34(1–2), 86–91.
Amendola, G. B. (2015). Ultra-Broadband for all in Europe: can access regulation hinder innovation and welfare maximisation? In Proceedings of the 26th
European regional ITS conference. Madrid, Spain, June.
Analysys Mason (2008). The Costs of Deploying Fibre-based Next-generation Broadband Infrastructure. Analysys Mason report for the Broadband Stakeholder
Group. Available from: 〈http://www.analysysmason.com/PageFiles/5766/Analysys-Mason-final-report-for-BSG-(Sept2008).pdf〉.
Analysys Mason & Tech4i2 (2013). The Socio-economic Impact of Bandwidth. Study prepared for the European Commission by Analysys Mason & Tech4i2.
Available from: 〈http://ec.europa.eu/digital-agenda/en/news/study-socio-economic-impact-bandwidth-smart-20100033〉.
Australian Government (2010). Implementation Study, National Broadband Network. May 6. Available from: 〈http://www.aph.gov.au/About_Parliament/Par
liamentary_Departments/Parliamentary_Library/pubs/BN/1011/NBN#_Toc268871752〉.
772 J. Rendon Schneir, Y. Xiong / Telecommunications Policy 40 (2016) 755–773

Badasyan, N., Shideler, D., & Silva, S. (2011). Broadband achievement index: moving beyond availability. Telecommunications Policy, 35(11), 933–950.
Balmer, R. (2015a). Cooperative investments in next generation broadband networks: a review on recent practical cases and literature. In Proceedings of the
26th European regional ITS conference. Madrid, Spain, June.
Balmer, R. (2015b). Geographic regulation of next generation access networks: a review on recent practical cases and literature. In Proceedings of the 26th
European regional ITS conference . Madrid, Spain, June.
Belloc, F., Nicita, A., & Rossi, M. A. (2012). Whiter policy design for broadband penetration? Evidence from 30 OECD countries. Telecommunications Policy, 36
(5), 382–398.
Beltran, F. (2014). Fibre-to-the-home, high-speed and national broadband plans: tales from down under. Telecommunications Policy, 38(8–9), 715–729.
Bohlin, E., & Teppayayon, O. (2010). Broadband universal service in Europe: review of policy consultations 2005–2010. Communications and Strategies,
21–42.
Bourreau, M., Cambini, C., & Hoernig, S. (2012). Ex-ante regulation and co-investment in the transition to next generation access. Telecommunications Policy,
36(5), 399–406.
Briglauer, W., & Gugler, K. (2013). The deployment and penetration of high-speed fiber networks and services: why are EU member states lagging behind?
Telecommunications Policy, 37(10), 819–835.
Brown, S., Browning, K. & Clements, M. (2015). Rural utilities service broadband loans and economic performance in Rural America. In Proceedings of the
TPRC 43: the 43rd research conference on communications, information and research policy . Arlington, USA, September.
Caio, F., Marcus, S. & Pogorel, G. (2014). Achieving the Objectives of the Digital Agenda for Europe (DAE) in Italy: Prospects and Challenges. Report of the expert
advisory team appointed by President Letta. Available from: 〈http://www.governo.it/backoffice/allegati/74621-9208.pdf〉.
Carnegie UK Trust (2013). Going the Last Mile: How Can Broadband Reach the Final 10%? Available from: 〈http://www.carnegieuktrust.org.uk/getattachment/
a88d402f-26c9-4f74-9f53-24e10c8e74d3/Going-the-Last-Mile—How-can-broadband-reach-the-.aspx〉.
Cave, M. (2014). The ladder of investment in Europe, in retrospect and prospect. Telecommunications Policy, 38(8–9), 674–683.
Clarke, R. N. (2014). Expanding mobile wireless capacity: the challenges presented by technology and economics. Telecommunications Policy, 38(8–9),
693–708.
Dauvin, M., & Grzybowski, L. (2014). Estimating broadband diffusion in the EU using NUTS 1 regional data. Telecommunications Policy, 38(1), 96–104.
Department for Environment, Food & Rural Affairs. (2013). Rural Urban Classification. United Kingdom. Available from: 〈https://www.gov.uk/government/
collections/rural-urban-definition〉.
EIB (2011). An Assessment on the Total Investment Requirement to Reach the Digital Agenda broadband targets. Study prepared for the European Investment
Bank (EIB) by Pantelis Koutroumpis.
Elixmann, D., Ilic, D., Neumann, K.H. & Plückebaum, T. (2008). The Economics of Next Generation Access – Final Report. WIK-Consult report for ECTA. Available
from: 〈〈http://wik.org/uploads/media/ECTA_NGA_masterfile_2008_09_15_V1.pdf〉〉.
European Commission (2013). The Broadband State Aid Rules explained; an eGuide for Decision Makers. WIK-Consult report for the European Commission.
Available from: 〈http://ec.europa.eu/digital-agenda/en/news/handbook-decision-makers-broadband-state-aid-rules-explained〉.
European Commission (2014a). Broadband Strategy and Policy. Available from: 〈http://ec.europa.eu/digital-agenda/en/broadband-strategy-policy〉.
European Commission (2014b). A Harmonised Definition of Cities and Rural Areas: the New Degree of Urbanisation. Regional Working Paper 2014, WP 01/2014.
Available from: 〈http://ec.europa.eu/regional_policy/sources/docgener/work/2014_01_new_urban.pdf〉.
European Commission (2015). Broadband Coverage in Europe 2014, Mapping Progress Towards the Coverage Objectives of the Digital Agenda. Study prepared
for the European Commission by HIS and VVA. Available from: 〈https://ec.europa.eu/digital-agenda/en/news/study-broadband-coverage-europe-2014〉.
Flacher, D., & Jennequin, H. (2014). Access regulation and geographic deployment of a new generation infrastructure. Telecommunications Policy, 38(8–9),
741–759.
Frias, Z., Gonzales-Valderrama, C. & Perez Martinez, J. (2015). Keys and challenges to close the broadband rural gap: the role of LTE networks in Spain. In
Proceedings of the 26th European regional ITS conference. Madrid, Spain, June.
FSR (2011). Broadband Diffusion: Drivers and Policies. Study for the Independent Regulators Group (IRG) prepared by the Florence School of Regulation (FSR).
Available from: 〈http://www.irg.eu/streaming/CN%20(11)%2081_FSR_Study_on_BB_Promotion_FINAL.pdf?contentId ¼ 547201&field ¼ATTACHED_FILE〉.
FTTH Council Europe (2012). The Cost of Meeting Europe's Network Needs. Report prepared by the FTTH Council Europe. Available from: 〈http://www.
ftthcouncil.eu/resources?category_id¼ 6&location ¼&topic¼ Business þ andþ Financing#〉.
Gerpott, T. J., & Ahmadi, N. (2015). Advancement of indices assessing a nation's telecommunications development status: a PLS structural equation analysis
of over 100 countries. Telecommunications Policy, 39(2), 93–111.
Gijon, C., Whalley, J., & Anderson, G. (2015). Exploring the differences in broadband access speeds across Glasgow. Telematics and Informatics
Götz, G. (2013). Competition, regulation and broadband access to the Internet. Telecommunications Policy, 37(11), 1095–1109.
Gruber, H., Hätönen, J., & Koutroumpis, P. (2014). Broadband access in the EU: an assessment of future economic benefits. Telecommunications Policy, 38(11),
1046–1058.
Grubesic, T. H. (2010). Efficiency in broadband service provision: a spatial analysis. Telecommunications Policy, 34(3), 117–131.
Grzybowski, L. (2014). Fixed-to-mobile substitution in the European Union. Telecommunications Policy, 38(7), 601–612.
Hallahan, R., & Peha, J. M. (2010). Quantifying the costs of a nationwide public safety wireless network. Telecommunications Policy, 34(4), 200–220.
Haller, S. A., & Lyons, S. (2015). Broadband adoption and firm productivity: evidence from Irish manufacturing firms. Telecommunications Policy, 39(1), 1–13.
Hätönen, J. (2011). The economic impact of fixed and mobile high-speed networks. EIB Papers, 16(2), 30–59.
Holt, L., & Galligan, M. (2013). Mapping the field: retrospective of the federal universal service programs. Telecommunications Policy, 37(9), 773–793.
ITU-T (2014). Recommendation ITU-T G.9700. Fast Access to Subscriber Terminals (FAST)-Power Spectral Density Specification. Available from: 〈https://www.itu.
int/ITU-T/workprog/wp_item.aspx?isn ¼9064〉.
Kongaut, C., & Bohlin, E. (2014). Unbundling and infrastructure competition for broadband adoption: implications for NGA regulation. Telecommunications
Policy, 38(8–9), 760–770.
Ling, M.-S., & Wu, F. S. (2013). Identifying the determinants of broadband adoption by diffusion stage in OECD countries. Telecommunications Policy, 37(4–5),
241–251.
Nucciarelli, A., Castaldo, A., Conte, E., & Sadowski, B. (2013). Unlocking the potential of Italian broadband: Case studies and policy lessons. Tele-
communications Policy, 37(10), 955–969.
OECD (2010). OECD Regional Typology . Directorate for Public Governance and Territorial Development. Available from: 〈http://www.oecd.org/gov/regional-
policy/42392595.pdf〉.
OECD (2011). National Broadband Plans. OECD Digital Economy Papers, No. 181. Available from: 〈http://www.oecd-ilibrary.org/docserver/download/
5kg9sr5fmqwd.pdf?expires¼ 1457337149&id¼ id&accname ¼guest&checksum ¼F9963C61B6A9BD2A9886FE468DFBE780〉.
OECD (2012). Universal Service Policies in the Context of National Broadband Plans. OECD Digital Economy Papers, No. 203. Available from: 〈http://www.oecd.
org/officialdocuments/publicdisplaydocumentpdf/?cote ¼DSTI/ICCP/CISP (2011)10/FINAL&docLanguage ¼En〉.
Ovando, C., Perez, J., & Moral, A. (2015). LTE techno-economic assessment: the case of rural areas in Spain. Telecommunications Policy, 39(3–4), 269–283.
Park, S. & Kim, G. (2015). Same access, different uses, and the persistent digital divide between urban and rural users. In Proceedings of the TPRC 43: the 43rd
research conference on communications, information and research policy. Arlington, USA, September.
Peronard, J.-P., & Flemming, J. (2011). User motivation for broadband: a rural Danish study. Telecommunications Policy, 35(8), 691–701.
Point Topic (2013). Europe's Broadband Investment Needs: Quantifying the Investment Needed to Deliver Superfast Broadband to Europe. Available from: 〈http://
point-topic.com/wp-content/uploads/2013/05/Point-Topic-Europes-superfast-broadband-investment-needs-20130520-1.2.pdf〉.
PricewaterhouseCoopers (2015a). National Broadband Plan: Benefits of High Speed Broadband. Draft Report prepared for the Department of Communications,
Energy and Natural Resources of Ireland, July. Available from: 〈http://www.dcenr.gov.ie/communications/en-ie/Broadband/Pages/Strategy-Interven
J. Rendon Schneir, Y. Xiong / Telecommunications Policy 40 (2016) 755–773 773

tion—Expert-Reports.aspx〉.
PricewaterhouseCoopers (2015b). Broadband Strategy for Ireland. Report prepared for the Department of Communications, Energy and Natural Resources of
Ireland, July. Available from: 〈http://www.dcenr.gov.ie/communications/en-ie/Broadband/Pages/Strategy-Intervention—Expert-Reports.aspx〉.
Prieger, J. E. (2013). The broadband digital divide and the economic benefits of mobile broadband for rural areas. Telecommunications Policy, 37(6–7),
483–502.
Puschita, E., Constantinescu-Dobra, A., Colda, R., Vermesan, I., Moldovan, A., & Palade, T. (2014). Challenges for a broadband service strategy in rural areas: a
Romanian case study. Telecommunications Policy, 38(2), 147–156.
Ragoobar, T., Whalley, J., & Harle, D. (2011). Public and private intervention for next-generation access deployment: possibilities for three European
countries. Telecommunications Policy, 35(9–10), 827–841.
Reggi, L., & Scicchitano, S. (2014). Are EU regional digital strategies evidence-based? An analysis of the allocation of 2007–13 structural funds. Tele-
communications Policy, 38(5–6), 530–538.
Rendon Schneir, J., & Xiong, Y. (2013). Economic implications of a co-investment scheme for FTTH/PON architectures. Telecommunications Policy, 37(10),
849–860.
Rendon Schneir, J., & Xiong, Y. (2015). Economic aspects of fibre to the distribution point with G.fast. Telecommunications Policy, 39(6), 450–462.
Rogers, E. M. (2003). The diffusion of innovations ((5th ed.). New York: The Free Press.
Ruhle, E.-O., Brusic, I., Kittl, J., & Ehrler, M. (2011). Next generation access (NGA) supply side interventions – an international comparison. Tele-
communications Policy, 35(9–10), 794–803.
Stenberg, P. (2010). The rural effect of broadband internet service. In Proceedings of the TPRC 38: the 38th research conference on communications, information
and research policy. Arlington, USA, October.
Sundquist, M., & Markendahl, J. (2015). A case study cost modelling of regulatory alternatives to mitigate the mobile network coverage and capacity
problems in rural areas. In Proceedings of the 26th European regional ITS conference. Madrid, Spain, June.
Tahon, M., Van Ooteghem, J., Caiser, K., Verbrugge, S., Colle, D., Pickavet, M., & Demeester, P. (2014). Improving the FTTH business case – a joint telco-utility
network rollout model. Telecommunications Policy, 38(5–6), 426–437.
Whitacre, B., Gallardo, R., & Strover, S. (2014). Broadband's contribution to economic growth in rural areas: moving towards a causal relationship. Tele-
communications Policy, 38(11), 1011–1023.
WIK-Consult (2013). Estimating the Cost of GEA. WIK-Consult Report for TalkTalk. Available from: 〈http://stakeholders.ofcom.org.uk/binaries/consultations/
fixed-access-markets/responses/TalkTalk_Group_second_addit1.pdf〉.

You might also like