Developing successful public-private partnerships to foster investment in universal broadband networks

1 Introduction

ITU recognizes the importance of broadband services being accessible to consumers and businesses, irrespective of their location. The purpose of this report is to highlight the best practices used by publicprivate partnership (PPP) projects to successfully implement universal broadband projects, and thereby improve broadband access to unserved and underserved locations. A total of 13 PPP broadband projects have been researched as part of this project. This report takes a broad definition of what PPP means, i.e. any project where there is a mix of private- and public-sector involvement.

The projects were selected to provide insights and lessons learned from projects from emerging and developed markets across most world regions. To maximize the number and type and best practices and lessons learned, the projects were selected to include a wide variety of technologies, investment models and funding sources, as well as a variety of approaches by managing authorities to the projects. One of the projects – Lithuania's Rural Area IT Network (RAIN) project – has previously been described as part of the European Union Guide to broadband investment¹. The EU report has also been used as a model to provide a suitable structure for this report.

The ITU defines 'broadband' as a service with a minimum download speed of 256 kbit/s, but some of the projects referred to in this report define broadband as a service with a minimum download speed of 128 kbit/s. For the purposes of this report, 'next-generation broadband' refers to an evolution away from traditional, exchange- or central-office-based broadband technologies like Asymmetric Digital Subscriber Line (ADSL); however, it should be noted that there is no globally recognized definition for nextgeneration broadband.

1.1 Aims of this report

The aim of the report is to promote the sharing of best practices adopted by PPP broadband projects, whether they have been used to provide broadband access nationally, regionally, or in rural areas. The purpose of the document is not to focus on the approach followed by any single PPP broadband project, or to promote any one project, but to provide an overall view of the best practices implemented among all the projects. These best practices have been identified as beneficial for future PPP broadband projects, and which, if followed, should deliver successful outcomes.

The overall aim of this report is to help PPP broadband projects to achieve successful outcomes by:

• using public funds more effectively;

• monitoring outcomes, to ensure the project achieves the aim of providing broadband access or services to the locations/people it is supposed to target;

• managing risks, to ensure it is delivered on time and within budget;

• stimulating and creating demand to ensure the broadband infrastructure is used effectively.

The report has been prepared using secondary research based on publicly available information aimed at managing authorities with responsibility for preparing for, managing or monitoring PPP broadband projects.

1

http://ec.europa.eu/regional_policy/sources/docgener/presenta/broadband2011/broadband2011_en.pdf

2 The benefits of investing in broadband

As a first stage in the investment planning process, it is essential for a telecommunication regulator or other managing authority³ to define the aims of the broadband investment project – namely what the project needs to achieve, and why. Once a clear set of aims have been defined, these will guide the rest of the project and influence decisions throughout the planning process. These aims can also be used to determine certain milestones, for example the number of kilometres of core network fibre deployed, the number of homes covered by 3G (IMT-2000) or FTTH. These milestones may in turn be used to determine when the private partners in a PPP receive their subsidy payments. A managing authority should consider the appointment of a ‘champion’ to drive the project aims forward, and to monitor progress.

3

To avoid repetition, in the remainder of this report the term 'managing authority' is used to refer to whatever organization has responsibility for managing the PPP broadband project: this may be the national regulatory authority, another public authority such as a ministry, or an agency (e.g. an intermediate body such as a development agency) delegated to provide public support to these networks.

This section looks at a number of reasons why a managing authority may decide to make broadband investments; these are primarily associated with delivering socio-economic benefits, such as creating stronger community relationships, supporting regional development, promoting competition and attracting/retaining investment. A managing authority may also derive benefit from using the network for its own services (including playing the role of anchor tenant, which could help to support the business case).

Above all, it is important for a managing authority to be aware that access to affordable broadband has a positive effect in terms of meeting the most basic needs of the individuals, communities and businesses in a territory. The majority of projects highlighted in this document are intended to reduce the digital divide and improve access to broadband to unserved and underserved areas, whether these are urban or rural. The minimum access speeds to be provided by the various projects vary considerably: in some emerging markets a minimum download speed of 128 or 512 kbit/s is the target, while in other projects 10 Mbit/s is the minimum speed. It is important for a managing authority to keep these goals in mind, and prioritize the long-term benefit of individuals over the short-term gain of private entities.

Some of the example projects which were studied for this guide are in the process of being deployed, while others have already been completed and have been deemed to be a success by the managing authority, or by an external funder (for example, the ERDF).

3.1 Factors affecting the choice of infrastructure

3.1.1 Scope of the network

There are two main types of broadband infrastructure that a managing authority should consider investing in: access networks and backhaul/core networks. In addition, infrastructure to provide international connectivity may also be desirable.

An access network provides the connections between end users and the nearest network node (e.g. the local exchange or central office) at which the access network connects with the core network. Various options are available for providing broadband connections in the access network, depending on the requirements and available funding; these options include existing copper lines, new optical fibre cables and wireless networks (including satellite).

Backhaul and core networks provide links between network nodes to allow connectivity over large distances. Core networks link towns and cities across the country (also known as the backbones) and backhaul networks connect local exchanges to core networks. Because traffic from a large number of end users is aggregated as it passes through the backhaul/core network, optical fibre cable is often the technology of choice due to its high capacity. High-capacity wireless microwave links are also used, particularly in mountainous areas where digging trenches for fibre may be impractical.

The three network infrastructure components that may be considered as part of a broadband strategy are shown in Figure 1 below.

In addition to access and backhaul/core networks, investment in international fibre links (either landbased or submarine) is also sometimes necessary in emerging markets. In many cases the requirement arises because operators want more economical access to international connectivity via fibre, as opposed to low-bandwidth, high-priced satellite connectivity. Operators in emerging markets increasingly need to aggregate traffic from multiple core fibres and connect to international carrier networks which provide access to the Internet and support other data services. Land-based or submarine fibre provides links to international network hubs, to allow the transfer of Internet traffic and other data services. Optical fibre cable is usually used to provide the high-capacity links necessary to support the huge volumes of traffic to be distributed internationally.

3.1.2 Performance of the network

It is essential that a managing authority has at least a broad understanding of the technologies under consideration and the architectures that can be used to meet its requirements, so that it has an appreciation of the trade-off between cost and performance¹⁰.

It is also important that the managing authority does not specify what technology should be used: rather, it should specify its network requirements in a technology-neutral way. For example, the specification for a broadband access network could require that it should support a particular number of connections at a certain minimum speed. For a core network, the authority could specify that the infrastructure should be capable of supporting service providers requiring access to backhaul links, and also support those service providers' technologies.

It should be noted that a mix of technologies, rather than a single technology, may be appropriate in a particular region. While optical fibre cable usually delivers the highest connection speed, it is expensive to deploy over wide areas, and wireless and satellite technologies are likely to have a role to play in providing cost-effective wide-area coverage.

3.1.3 Ability of the network to support competition

Another important consideration is the impact that the new broadband network will have on competition in the market. In many countries a condition for granting government grants is that the recipient must provide open access to the infrastructure on a wholesale basis, regardless of whether or not the aid recipient has significant market power. It is generally accepted that if operators have access to the passive infrastructure (e.g. copper, dark fibre or underground ducts), they will have more freedom to develop innovative services and compete better with other operators, thereby delivering lower prices to consumers.

For the broadband projects in this report, the relevant guidelines regarding government grants included in the access obligations imposed on the infrastructure operator include providing access to both the passive and active levels of the infrastructure for up to 20 years or the lifetime of the network, without prejudice to any similar regulatory obligations that may be imposed by the regulator. The subsidized network has to be designed in a way that guarantees that alternative operators can have access at all levels: e.g. the infrastructure has to offer sufficient place in the ducts, space in street cabinets, and capacity on active equipment.

In the case of broadband networks, an argument may be put forward that in low-density areas access to the passive level alone will not result in additional competition since it may be not economically feasible to create an alternative network in such areas. Therefore the guidelines regarding government grants for broadband networks require that the new network should be opened at as many levels as possible – not just at the passive level – thus allowing market forces to decide which access products suit the market players best.

Access to the infrastructure (at which ever point it is provided) should not only be open, it should be offered on a transparent and non-discriminatory basis to allow for fair competition between retail service providers. To achieve this, the managing authority will need to design wholesale requirements which ensure that operators can compete effectively, regardless of who actually owns and operates the network. Wholesale obligations will need to be detailed in the requirements specification document used in the procurement. If the telecommunication regulator is not the managing authority, then consultation

10

Analysys Mason has conducted studies which are available in the public domain on the cost and capabilities of both

wireline and wireless technologies.

www.broadbanduk.org/component/option,com_docman/task,doc_details/gid,1036/

www.broadbanduk.org/component/option,com_docman/task,doc_view/gid,1246/Itemid,63/

with the regulator will be necessary: the wholesale requirements should be specified so they do not contradict the wholesale obligations stipulated by the regulator, and assurances should also be sought regarding future wholesale obligations that the regulator might advise.

Example broadband projects in which open-access principles are maintained by ensuring that other market players have non-discriminatory, open wholesale access can be found in Section 6 on monitoring and managing broadband projects.

When considering networks providing fibre to the home, competition considerations will also influence the choice of network architecture. There are two main options:

• A point-to-point (PTP) network provides a dedicated fibre connection to each home. This means that an operator can easily access any particular end user by connecting to the relevant fibre.

• An alternative architecture is a Gigabit Passive Optical Network (GPON) (ITU-T G.984) topology, in which each customer has their own connection into their home, although some of the access network is shared with other users (in a similar way to a cable network).

A GPON network may involve lower costs than a PTP network, but the options for competition are less straightforward as access to different customers must be managed electronically by the network operator. Most of the FTTH projects studied use a GPON architecture, although Qatar's Q.NBN also provides PTP specifically for enterprises, and Singapore's NGNBN supports both GPON and PTP.

3.2 Access networks

3.2.1 Fibre to the home (FTTH)

FTTH involves laying an optical fibre cable to the home all the way from the central office, local exchange or other suitable local access node, such as a public-sector building. FTTH is the technology with the highest capacity, and therefore provides the highest degree of future-proofing. However, due to the long distances involved in deploying a connection all the way to the home, the deployment costs of FTTH can be very high. To date, commercial deployments of FTTH have been limited due to this high cost.

As discussed above, there are two main options for an FTTH architecture: GPON (ITU-T G.984) and PTP. GPON networks usually require less capex to deploy, particularly in less densely populated, rural areas. Studies have shown that the cost of deploying a PTP architecture is on average 10–20 per cent more than an equivalent GPON architecture. The cost difference is higher in rural (less dense) areas than in urban (more dense) areas. As mentioned above, most of the FTTH projects examined in this study use GPON.

However, PTP has the benefit of allowing all operators to have full use of a fibre between the local exchange and the end user (i.e. they allow full unbundling of the fibre local loop). For this reason, PTP networks can be more favourable from a competition point of view. The primary method of competition on GPON networks is via an electronic interface, which may restrict the level of control that an alternative service provider has over the services it can offer. The use of 'wavelength unbundling' technology on GPON networks may in the future offer a similar level of control as a dedicated fibre on a PTP network, but at the time of writing this technology is still being standardized.

PTP networks may also be better suited to providing symmetric services, i.e. connections with the same speed in both directions, as required by some business uses. They are also able to provide higher capacities to end users, and hence are considered to be more future-proof solutions – particularly in light of the prospect of both consumers and businesses moving gradually towards cloud-computing. Cloudcomputing provides users on-demand, hosted services – including software as a service (SaaS) and infrastructure as a service (IaaS). These entail the delivery of an application or an infrastructure-related service over the Internet. The solution is hosted by a vendor or service provider and accessible to users (or machines) over an Internet-capable device. The Internet-capable device does not need a special client to access the solution. The vendor or service provider usually charges the user on a per-month basis for the use of the solution.

Although the cost of deploying an FTTH network depends to some extent on whether a GPON (ITU-T G.984) or PTP topology is chosen, it should be strongly emphasized that for both architectures the cost is much more dependent on the ability to reuse existing infrastructure and the investment model adopted. Furthermore, the sustainability of the project (and therefore the ability to deliver long-term socioeconomic benefits) is more dependent on the business model and the expertise of the project partners, than the choice of technology.

In terms of the projects studied, those projects which have already deployed FTTH infrastructure have used GPON architecture. One of the projects is also using PTP. The following projects have deployed FTTH:

• Malaysia, NBI: the High Speed Broadband part of the project aims to provide downloads speeds of between 10 Mbit/s and 100 Mbit/s in major economic areas, using GPON.

• Qatar, Q.NBN: aims to accelerate the deployment of FTTH and deliver coverage of in excess of 95 per cent (at 100 Mbit/s). Q.NBN FTTH supports GPON for residential and business customers, and PTP for enterprises.

• Singapore, NGNBN: GPON FTTH network to deliver 1 Gbit/s download and 500 Mbit/s upload speeds and connect all homes, schools and businesses, reaching 95 per cent of population by mid-2012 and 100 per cent by 2015.

3.2.2 Fibre to the cabinet (FTTC)

FTTC involves laying fibre from the central office (or local exchange) to a street cabinet, or to the basement of an apartment block. Because the fibre does not go the whole distance to the home, significant cost savings can be realized relative to FTTH. However, as the existing copper network is used for the last part of the connection to the home, the speeds available on an FTTC network are significantly lower than with FTTH (the cost to deploy FTTC is around 80 per cent cheaper than the cost to connect a home). As with FTTH technologies, the cost is strongly affected by the ability to reuse existing infrastructure.

3.2.3 ADSL

In developed markets, basic broadband services are most often delivered over the existing copper network using ADSL technology. In contrast, in emerging markets low fixed-line availability may limit the ability to deploy ADSL, and it is often used in conjunction with wireless technologies. ADSL uses the existing access infrastructure and is therefore relatively cheap to deploy. However, the nature of the technology means that speeds are heavily affected by the distance between the local exchange (or central office) and the home; in many cases the speeds available are below 10 Mbit/s.

Four of the example projects involve investment in current-generation broadband (i.e. ADSL), with the intention of providing ADSL access and services to rural areas and unserved locations, delivering a minimum download speed of 128 kbit/s or 512 kbit/s; other technologies are also being used in most of these projects.

• Dominican Republic, Rural Broadband Connectivity Project: provision of broadband access, at speeds of at least 128 kbit/s, to 508 mostly rural communities using ADSL and UMTS.

• Malaysia, NBI: the BBGP (Broadband to the General Population) project will provide download speeds of between 256 kbit/s and 10 Mbit/s using ADSL and other access technologies (wireless HSPA and WiMAX).

• Pakistan, USF Broadband Programme: provision of broadband access with a minimum download speed of 128 kbit/s to unserved urban areas and rural communities, using ADSL and wireless HSPA and WiMAX.

• Paraguay, PNT: provision of broadband with minimum download speed of 512 kbit/s using ADSL and other access technologies (wireless HSPA, WiMAX or other wireless technologies).

3.2.4 Wireless and satellite

Terrestrial wireless technologies provide a link between the home and the nearest network node without the need for a physical wireline connection. Terrestrial wireless networks are complementary to fixed networks, and can be advantageous in areas where building a wireline network would be difficult and/or expensive (e.g. in mountainous terrain). However, because several users access the network via the same wireless link, the contention¹¹ for services can be much higher than on wireline networks, and the actual speed experienced may be much lower than the maximum speed quoted by the service provider. In order to ensure that an end user receives an assured level of service, more base stations may have to be added, depending on the minimum speed set by the telecommunication regulator or managing authority, which will increase costs. It should also be noted that if a large number of users on a wireless network demand high-speed rates, meeting this demand will often require additional investment in the fixed infrastructure (particularly the backhaul network) that supports the wireless network.

Satellite networks offer a useful solution primarily for areas that are not covered by terrestrial networks (either wireline or wireless), e.g. where the existing networks have ‘not spots’ (localized areas where there is no coverage). Despite their cost and capacity limitations, satellite technologies can therefore play a valuable role in broadband access. As with terrestrial wireless technologies, many users are accessing the same node (i.e. the satellite transponder) and so the effects of contention may have a greater impact than in wireline networks. In remote locations, there may be no choice but to use satellite links (e.g. Mongolia's ICT Infrastructure Development Programme).

Wireless technologies which could provide effective next-generation broadband services include terrestrial wireless broadband technologies such as LTE-Advanced,¹² WiMAX based on the 802.16m standard, and high-capacity satellites using Ka-band multi-spotbeam technology.

The costs of terrestrial wireless technologies vary according to a number of factors, including the terrain over which they are deployed, the data rate that must be delivered at the furthest point from the base station, and the overall traffic demand. Indeed, for both terrestrial wireless and satellite access technologies, the cost of deployment depends very heavily on the traffic demand to be supported. This is in contrast to fibre technologies, for which costs do not vary as strongly with traffic demand.

The example projects which feature wireless and satellite technologies are as follows:

• Dominican Republic, Rural Broadband Connectivity Project: provision of broadband access to 508 communities, mostly rural, via UMTS and fixed access ADSL.

• Malaysia, NBI: the BBGP (Broadband to the General Population) part of the project will deliver downloads speeds of between 256 kbit/s and 10 Mbit/s using ADSL and wireless HSPA and WiMAX.

• Mongolia, ICT Infrastructure Development Project: provision of broadband services using Wi-Fi to rural communities, hubbed to Ulaanbaatar via a VSAT satellite link.

• Pakistan, USF Broadband Programme: provision of broadband access to unserved urban areas and rural communities, using HSPA, WiMAX and ADSL.

• Paraguay, PNT: provision of broadband with minimum download speed of 512 kbit/s using wireless HSPA, WiMAX, other wireless technologies and ADSL.

11

Network contention is a measure of the number of users that are served by a single network node (e.g. a local

exchange or a terrestrial wireless base station). The capacity available to each end user at any point in time is

dependent on the capacity supplied to the node, divided by the number of users on that node that are currently using

services. 12

LTE is the abbreviation for Long Term Evolution; this wireless technology is sometimes referred to as 4G (IMT-Advanced).

• Saudi Arabia, Universal Service Project: an initiative for operators to provide voice and Internet access (minimum 512 kbit/s download speed) to underserved locations using 3G (IMT-2000) wireless access.

3.3 Backhaul/core networks

A large number of the example projects feature investment in fibre-based backhaul and core networks. A fit-for-purpose backhaul/core network is essential for providing effective broadband services. Since a backhaul/core network connects an extensive area, such a network may be a cost-effective way of providing coverage to a large number of end users, especially if there is an existing access network that is already sufficient to deliver broadband (e.g. basic broadband over copper lines). In other cases a new access network may need to be deployed through a separate investment – and sometimes the investment in backhaul/core infrastructure may provide a catalyst for investment to upgrade the access network. It is important for a managing authority to ensure that the new backhaul/core network is built using wellestablished technical standards, to allow effective competition in the access network.

Among the example projects, many of the backhaul/core initiatives are associated with supporting broadband access and services to specific regions, or to rural locations. Many also support the development of wireless access networks (WiMAX, UMTS) to provide last-mile access. The projects which include investment in a backhaul/core network are as follows:

• Argentina Connected: deploying a nationwide core network to double the amount of optical fibre infrastructure in Argentina from 2011 to 2015, in order to improve broadband access nationally.

• Latvia, Next-generation network for rural areas project: deploying core network to support improved broadband access in rural locations. The network will remain in public ownership.

• Lithuania, RAIN project: deployed a nationwide backhaul/core network to provide improved connectivity to existing access infrastructure.

• Pakistan, USF Broadband Programme: deploying a nationwide backhaul/core network to provide improved connectivity to all administrative districts.

• Paraguay, PNT: deploying a nationwide core network, with 1 000 kilometres rolled out per year over the period 2011–2015, to provide improved broadband access nationally.

3.4 International fibre networks

As discussed earlier, international fibre links – whether land-based or submarine – are necessary within a country to provide cost-effective international connectivity. Just one of the example projects features investment in an international fibre network. The government of Kenya entered a PPP with Etisalat (the main fixed operator in the United Arab Emirates) to deploy a submarine cable between Mombasa, Kenya and Fujairah (in the UAE). TEAMS Limited was created to construct and manage this cable. Originally the consortium was structured with the government having an 85 per cent ownership share, and Etisalat 15 per cent, but the government sold off part of its stake, resulting in ownership in TEAMS being split between private investors (83%) and the government (17%) – the Ministry of Finance holds the stake. Etisalat retains a 15 per cent stake in the submarine cable and TEAMS Kenya 85 per cent. All TEAMS consortium members are able to sell capacity, at a wholesale and retail level, equivalent to their share of the cable capacity.

3.5 Other infrastructure considerations

There are a number of other factors which a managing authority should take into consideration when choosing network architecture, including the use of appropriate expertise, technology obsolescence, and barriers to take-up. These are discussed briefly below.

3.5.1 Employing experts in the selection process

It is good practice for PPPs to use a range of appropriately qualified, independent experts to evaluate projects and assist in ensuring that all bids are evaluated in a fair and transparent manner against a range of clearly defined criteria. For example, in the Dominican Republic, the Rural Broadband Connectivity Project used engineers, economists and lawyers to evaluate projects and perform on-site visits to check project validity (e.g. engineers checked if a proposed wireless link could deliver line-of-sight connectivity).

3.5.2 Avoiding technologies that are nearing obsolescence

Technology obsolescence is a common issue in telecommunications, as technologies are constantly evolving. In large-scale projects that are intended to provide broadband services for many years, it is important that the technologies used are not nearing the end of their natural cycle. Older technologies will be more expensive to maintain (as fewer equipment vendors support them), and offer fewer opportunities for service innovation.

3.5.3 Using an appropriate technology mix

The use of a mix of technologies should be considered, irrespective of the aims of a project. In unserved and underserved locations, which are quite often rural or have difficult topographies, deployment should not be limited to one type of technology, and the fastest technology (in terms of broadband access speed) may not always be the most appropriate. For example, the national broadband project in Slovak Republic is clear that the core network should be fibre-based as this provides the best throughput, but for backhaul any appropriate technology may be considered – in particular wireless technologies to provide backhaul capabilities in mountainous areas.

3.5.4 Minimizing barriers to adoption

Public broadband investments should be designed to minimize the barriers to take-up by service providers and end users, in order to ensure that services are available and people actually use them. Earlier sections in this report have emphasized the importance of ensuring that service providers have access to active or passive infrastructure at a reasonable price and under conditions that are open, fair, transparent and non-discriminatory. For service providers, it is also important to define operational areas of an appropriate size: if there are too many small areas this may impose a burden on service providers that tender for all of the projects; conversely, defining a few large areas may make it difficult for an operator to create a viable business if those areas include large proportions of low-density users.

For end users, obstacles to take-up can include the cost and accessibility of services. Deploying the right infrastructure and promoting effective competition can help to minimize cost, while developing new services that (for example) do not require the use of a PC can ensure that services are accessible. An example could be to provide a healthcare monitoring service in someone's home to manage their glucose levels, without the need of deploying a PC. Some of the example projects have sought to increase adoption by providing free ICT equipment, or by implementing ICT literacy programmes:

• Argentina Connected: the project aims to provide 3 million netbooks to students. According to Conectar Igualdad, as of July 2012 netbooks had been provided to 1.9 million students since 2010. A Digital Literacy Program has also been used to provide PC and Internet training to communities.

• Malaysia, NBI: netbooks were provided at no cost to secondary school children whose annual household income was less than MYR 3 000 (USD 965), and computer and Internet training was provided to low-income and rural communities. Over 100 000 netbooks, bundled with access to Telecom Malaysia's High Speed Broadband (HSBB) network, were distributed in 2010. Broadband Experience Centres have also been developed throughout Malaysia, to increase ICT literacy and reduce the digital divide by providing training.

• Qatar, Q.NBN: the managing authority runs digital literacy programmes and provides free laptops or netbooks to those who cannot afford them.

3.5.5 Liaising with operators and industry stakeholders

Before developing a PPP, it is advisable to liaise with operators and other potential stakeholders, not just for their input on technology issues, but also to help understand the wide range of issues involved,

e.g. commercial, regulatory, deployment, demand-side and others. Three examples are provided below:

• Dominican Republic, Rural Broadband Connectivity Project: prior to launching its public tender, the managing authority, Indotel, conducted consultations with industry stakeholders and users, including operators, ISPs, the Ministry of Education, the Ministry of Health and local municipalities. The consultation was used to determine the need for Internet services and consumers' willingness to pay, assess the telecommunication infrastructure available, identify operators' plans, and discuss the challenges faced in deploying broadband infrastructure to rural locations. A technical evaluation team consisting of two engineers, an economist and lawyer supported the consultation process. It should be noted that it is difficult to determine end users' willingness to use and pay for Internet access and services, when end users have never used the Internet. This was a key finding of the project in the Dominican Republic.

• Qatar, Q.NBN: the project was planned in consultation with network operators, such as Qtel and Vodafone, to determine the extent to which existing infrastructure could be used and how the deployment of FTTH could be used to support the development of broadband services.

• Saudi Arabia, Universal Service Project: prior to the project, the Communications and Information Technology Commission launched a consultation regarding the creation of a USF to support its universal access and universal service policies, originally conceived in 2006. In April 2010, the universal service policy was enacted (Decision No. 257/1431). The consultation involved operators, government departments and other organizations, and aimed to determine the effectiveness of the USF in supporting the deployment of voice and broadband services to all locations with 100 or more people.

4.1 Bottom-up

The bottom-up or local community model involves a group of end users (typically comprising local residents and/or businesses) organizing themselves into a jointly owned and democratically controlled organizational group (frequently a cooperative) capable of overseeing the contract to build their own local network. In this model it is likely that the public sector has no role in owning or running the project, but rather passes the funding to the group itself to oversee the project. Given that the local group is likely to have little experience of telecommunications, it is likely that the day-to-day running of the network will be outsourced to an operator with the necessary expertise.

None of the projects studied have implemented bottom-up models. However, the following examples have occurred in developed markets in Europe:

• Finland, eRegio project in North Karelia¹³: end users provided the investment, and demand often did not materialize until the network deployment reached their premises.

• Netherlands, the OnsNet¹⁴ project in Nuenen: this included a six-week demand aggregation scheme and free services for the first year.

• Sweden, Rural Development Programme¹⁵: end users provided investment (either financial or ‘in kind’) to attract additional bottom-up investment.

Advantages of the bottom-up model

• As the investment is generally undertaken by non-profit organizations made up of end users, it is usually considered on a long-term basis. For this reason, infrastructures such as FTTH can be deployed, which provides the highest level of future-proofing.

• Cooperative organizations have the effect of generating and aggregating demand in an area, which ensures that maximum social benefit is derived from the investment, even if only a small amount of funding is available.

13

DG REGIO, available at

http://ec.europa.eu/regional_policy/sources/docoffic/official/communic/smart_growth/comm2010_553_en.pdf

14

Municipal broadband access net works in the Netherlands

http://w3.ele.tue.nl/fileadmin/ele/TTE/ECO/Files/Pubs_2006/Kramer_AccessNets_06_presentation.pdf

15

State aid to broadband within the framework of the rural development programme

http://ec.europa.eu/eu_law/state_aids/comp-2010/n030-10-en.pdf

Disadvantages of the bottom-up model

• This approach may not be suited to providing widespread coverage, as individual projects can be very localized. This may mean that some areas are missed out and those networks that are built have differing technical standards, which may mean that competition from other operators is limited.

• The cooperative organizations are unlikely to have specific expertise in telecommunication networks, and higher-cost ‘turnkey’ solutions may be required.

• If the funding is to come from the end users themselves, then the need to produce this funding upfront may create a barrier. In this case, the public sector can help by guaranteeing or underwriting loans.

• It is highly unlikely that end users in unserved or underserved locations in emerging markets would be able to finance any such project without substantial support from the public sector (central or local government). In these situations the public DBO model is more appropriate.

Overall, the bottom-up model is most appropriate for targeting localized areas in developed markets and for gaining the most benefit from small amounts of funding.

4.2 Private DBO

The private DBO model involves a private-sector organization receiving some level of public funding (often a grant) to assist in its deployment of a new network offering open wholesale access. Critically, in this model the public sector has no specific role in the ownership or running of the network, but it may impose obligations relating to either of these in return for the funding.

Advantages of the private DBO model

• The cooperative or partnership organizations are unlikely to have specific expertise in telecommunication networks, and so high-cost turnkey solutions may be required.

• This approach imposes only a limited burden on the public sector, which is not involved in running the network. This in turn can lead to faster deployments than other investment models.

• It has a number of advantages for the private operator, particularly because ownership of the network assets is likely to prove valuable in the long term.

Disadvantages of the private DBO model

• It is essential that sufficient funding is available to attract interest from private operators, as significant investment may be required to make a viable business case, especially in rural areas.

• As the managing authority has limited ongoing control, the social benefit that the public sector is looking to create may be restricted if the private operator has little interest in delivering this and instead focuses on generating a financial return. (This issue can be addressed to some extent by the terms of the agreement).

The private operator is exposed to more risk in this model; in other models the private entity continues to share some portion of financial exposure with the public sector throughout the project. Because of this, an additional risk premium will be included by potential private partners when they specify the funding requirements for the project.

Overall, the private DBO model is appropriate for larger-scale investments than the bottom-up model where sufficient funding is available to attract interest from operators to work in rural areas, and where the operations (and risk) of the network can be effectively transferred to an operator with little ongoing control from the managing authority.

The private DBO model has been used by six of the example projects:

• Dominican Republic, Rural Broadband Connectivity Project: the project was won by Codetel, which did not to use the funding available, but instead used 2 × 15 MHz of spectrum in the

• 3.5 GHz band which was available for no fee. The unused spectrum was offered by the managing authority to another operator as an alternative means to make broadband access available to rural communities.
• Malaysia, NBI: for this project Telekom Malaysia entered into a PPP with the government. The main fixed operator provides open access to its High Speed Broadband (HSBB) network but with no regulation of pricing. At the time of writing there is an ongoing consultation, and service providers are pushing for regulated open access. In March 2012, REDtone signed an agreement to access the HSBB network on a wholesale basis in order to provide services to business customers.

• Mongolia, ICT Infrastructure Development Project: a grant was provided by the government of Mongolia, the government of Japan and the World Bank to provide broadband services in rural communities using Wi-Fi.

• Pakistan, USF Broadband Programme: grants were provided from the USF to operators to deploy broadband to unserved urban areas and rural communities, and nationwide core network. The grants have attracted many operators, and broadband access projects have been awarded to the main fixed operator (PTCL), Wateen and Worldcall, while fibre projects have been awarded to PTCL and Wateen.

• Saudi Arabia, Universal Service Project: grants were used to enable operators to provide voice and broadband access to underserved locations. Four projects were supported between 2010 and 2011, including one pilot, and all have used 3G (IMT-2000) wireless access.

• Singapore, NGNBN: the network is being built and operated by OpenNet, a consortium of the main fixed operator SingTel (30%), Axia NetMedia (30%), Singapore Press Holdings (25%) and Singapore Power Telecommunications (15%). OpenNet makes use of SingTel’s existing passive infrastructure assets, such as ducts, manholes and exchanges – SingTel has transferred these assets to a neutral party (the Asset Company or AssetCo), an independent and separately managed company owned by a registered business trust. By 2014, SingTel is anticipated to reduce its stake in AssetCo to a level approved by the Infocomm Development Authority. From 2013, OpenNet will be subject to a universal service obligation to install fibre to end-customer points. The operating company is called Nucleus Connect; it is responsible for operating the active Layer 2 and administering open access to Layer 3 by retail service providers. It began offering wholesale services on 31 August 2010. Retail services providers are responsible for selling services to end users and businesses. As of June 2012, 12 service providers were delivering NGNBN-based services.

4.3 Public outsourcing

Under a public outsourcing model a single contract is awarded to a private-sector organization, covering all aspects of the design or construction of the network. The major characteristic of this model is that the network is built and operated by the private sector, but the public sector retains ownership and some control of the network.

4.4 Joint venture (partnering)

A joint venture is an agreement where ownership of the network is split between the public and private sectors. Construction and operational functions are likely to be undertaken by a private-sector organization.

Advantages of the joint venture model

• The joint venture model has a number of advantages over the public outsourcing model described above, as both parties can maintain a long-term financial stake in the network. This is attractive to managing authorities which are reluctant to relinquish full ownership of the network as they see long-term strategic value in owning the assets.

• The joint venture approach has the ability to broadly balance the interests of the public and private sectors; it can also balance the sharing of risk – some forms of joint venture have required the private partner to increase its stake in the project when certain key performance indicators (such as take-up) are achieved, which represents a form of risk-sharing arrangement.

• This model often features the creation of special-purpose vehicles (SPVs). These SPVs can be of almost any size, which makes the model very scalable (i.e. able to address from local

communities to subnational regions). The SPV mechanism also allows investment to be gathered from comparatively innovative sources, such as institutional investors.

Disadvantages of the joint venture model

• With two stakeholders in the network each with different interests, it may be difficult to align those interests and set up the joint venture, or to continue it over a long period.

Overall, the joint venture model should be used only where the interests of the public and private sectors can be closely aligned. Indeed, only one of the example projects – the TEAMS submarine cable project in Kenya – follows this investment model. The lack of project examples suggests that the joint venture model is likely to be unattractive to both the public and the private sectors.

4.5 Public DBO

A public DBO model involves the managing authority operating without any private-sector intervention, except at a service provider level (involving either wholesale or retail service providers). All aspects of network deployment and operation are managed by the public sector. A network company is formed by the managing authority and typically offers wholesale services, with the potential to offer retail services (although this is not common).

Advantages of the public DBO model

• This model allows the managing authority to retain control of the network, and may have benefits such as: ensuring that social capital targets are given a high priority, ensuring that there are no conflicts of interest in achieving effective competition, and enforcing common technical standards.

• This approach is also suitable when the managing authority does not have confidence that the available legal mechanisms (e.g. competition regulations) will be sufficient to ensure effective competition.

Disadvantages of the public DBO model

• Sole ownership of the network by the managing authority increases its exposure to the risk of a failed venture. Organizations set up in this manner may struggle to meet targets (e.g. for coverage and take-up), often because of their lack of relevant commercial and technical expertise (which few public-sector organizations possess). The public will be aware that its money is being spent, and if any failings receive substantial publicity this could result in a loss of confidence in the project.

• Networks deployed under this model may be limited in size and scope due to the finite amount of expertise held within the managing authority. Therefore the ability for investment under this model to provide widespread network coverage may be reliant upon the network providing a catalyst for other investments.

• The model may potentially exclude certain aspects of private-sector expertise, which could be valuable in ensuring the efficient deployment and operation of the network. More broadly, this model does not exploit the economies of scale and scope that private-sector operators can bring.

Overall, the public DBO model is appropriate when a managing authority needs to have absolute control over the operations of the network (perhaps to ensure competition), or when it is confident that a targeted investment will inspire investment from other sources.

The following example projects feature the public DBO model:

• Argentina Connected: this project has followed a mixed model consisting of public DBO and public outsourcing. AR-SAT, the managing authority, has responsibility for deploying and operating the national core fibre network; this arrangement follows a public DBO model. To

• complement this, in large cities and in certain regions where AR-SAT does not have the capability to deploy fibre, it subcontracts deployment via public outsourcing.
• Lithuania, RAIN project: this has deployed a new national backhaul/core network using the public DBO model in order to ensure absolute control of the network, and therefore promote effective competition.

• Qatar, Q.NBN: to accelerate the deployment of FTTH, and deliver coverage of in excess of 95 per cent by 2015 (100 Mbit/s). Q.NBN is 100 per cent owned by the Qatari government provides equal, non-discriminatory access to the FTTH network, enabling any operator to use the infrastructure to deliver services.

5.1 Universal service funding

USFs have been used by managing authorities to fund broadband projects, particularly in unserved and underserved areas, and in rural areas. Five of the example projects have used USFs:

• Dominican Republic, Rural Broadband Connectivity Project: the USF is available to support this project, although the winning bidder, Codetel, chose to use some unassigned spectrum that was available for no fee instead of taking the available funding. Operators contribute 2 per cent of their gross income to the USF.

• Saudi Arabia, Universal Service Project: a universal access/universal service policy was created by the telecommunication regulator CITC in 2006, and a USF was created to fund these policies following consultation with the industry. A decision to levy one per cent of operators' revenues was agreed. The USF is used to assist operators to provide mobile voice and broadband Internet access to unserved or underserved communities.

• Mongolia, ICT Infrastructure Development Project: the government introduced a USF in July 2006 and began collecting the levy in December 2006, amounting to 2 per cent of operators' annual taxable income. Between end-2006 and end-2010, the fund accumulated USD 7.5 million.

• Pakistan, USF Broadband Programme: grants have been has provided from the USF to operators to deploy core broadband access to unserved urban areas and rural communities. The USF was created in 2006 and has been collected since 2007 at a rate of 1.5 per cent of operators' adjusted revenues. No other government funds have been used to fund the Broadband Programme.

• Malaysia, NBI: the USF was created in 2002, and operators¹⁶ contribute 6 per cent of their service revenues to the fund if their weighted average net revenue exceeds MYR 2 million. The fund has been used to provide voice telephony, wireless broadband coverage and to provide communities with access to computers, broadband services and IT training.

Argentina created a USF in 2007, and operators contribute one per cent of their revenues. The fund was set up with the intention of funding telecommunication projects in underserved and unserved areas.

16

Excluding content application service providers.

However, as yet the Argentina Connected project has not drawn on the USF, but instead has been funded by other government grants.

5.2 Government grants

Government grants have been used to support around half of the broadband projects studied. These projects have followed a variety of investment models: public outsourcing, public DBO, joint venture, and private DBO.

• Argentina Connected: this project is using government grants to fund a national core fibre and backhaul network, using a mix of public outsourcing and public DBO.

• Lithuania, RAIN: this used government grants as well as external funding from the ERDF to fund deployment of a nationwide backhaul/core network, using a public DBO model.

• Slovak Republic, national broadband project: a backhaul/core network is being jointly funded by the Government of Slovakia (EUR 11.32 million (USD 14.3 million)), the ERDF (EUR 96.22 million (USD 121.1 million)) and operators (EUR 5.66 million (USD 7.1 million)). A public outsourcing model is being followed.

• Singapore, NGNBN: this project is using USD 2 billion of government grants to fund the roll-out of FTTH, in line with a private DBO model.

• Malaysia, NBI: a government-funded project in conjunction with Telekom Malaysia to provide an open-access broadband network, on a commercially negotiated wholesale basis. A private DBO model is followed

• Kenya, TEAMS: the government has a 17 per cent share in this joint venture to build a submarine cable linking Mombasa with the UAE.

5.3 External funds

External funding provided by organizations such as the World Bank and the ERDF, or by foreign governments, has been used to finance a small number of the broadband projects identified in this report.

• Latvia, next-generation network for rural areas: this backhaul/core project is being funded entirely by the ERDF, which will provide a total of EUR 119 million (USD 149.9 million) by the end of 2018.

• Lithuania, RAIN: this backhaul/core project has received significant support from the ERDF, and has also benefited from government grants.

• Mongolia, ICT Infrastructure Development Project: used a mix of external financing from the government of Japan and the World Bank, plus USF funding.

• Slovak Republic, national broadband project: a backhaul/core network is being jointly funded by the ERDF (EUR 96.22 million (USD 121.2 million)), the Government of Slovakia (EUR 11.32 million (USD 14.3 million)) and operators (EUR 5.66 million (USD 7.1 million)).